The 13th German-Japanese Bridge Symposium
Program and Abstracts
Osaka Metropolitan University, Osaka, Japan
29th – 31st August, 2023
Organized by
Osaka Metropolitan University
Co-hosted by
Committee on Steel Structures, Japan Society of Civil Engineers
Committee on Hybrid Structures, Japan Society of Civil Engineers
Foreword
First of all, we are very pleased to have the 13th German-Japanese Bridge Symposium in OSAKA, Japan.
In 1982, Professor Kurita at Osaka Institute of Technology stayed in Germany for one year to enrich
collaborative research between the German team led by Professor Roik at Ruhr University Bochum and
the Japanese team led by Professors Maeda and Naka. This collaboration has led to be held the First
German-Japanese Joint Bridge Colloquium in 1994 at the Technical University of Munich chaired by
Professor Albrecht. Since then, four colloquia and eight symposia have taken place between Germany and
Japan every 2years. After the 12th Symposium was held in Germany in 2018, the 13th Symposium which
is initially scheduled for 2020 at Osaka City University, was postponed due to the Covid-19 pandemic. And
then, the Covid-19 pandemic has been subsided in the beginning of 2023, the 13th Symposium has been
rescheduled in 29th of August 2023. This symposium will provide the chance to communicate between
German and Japanese bridge researchers/engineers.
This symposium has 66 papers from both countries. Additionally, 4 keynote lectures will be presented,
which provide very interesting and useful/practical information for both sides' delegates.
I hope that this symposium gives all the participants many chances to exchange academic and practical
engineering information, expertise on bridge design, construction and maintenance. I also wish the
friendship between German and Japanese delegates extend firmly to the future.
Finally, I will appreciate the participation of all and deeply express the gratitude to those supported and
contributed to the organization of this symposium.
We are looking forward to seeing you on the symposium.
Takashi Yamaguchi
Chairman of 13th GJBS Symposium
Prof. of Osaka Metropolitan University
Table of Contents
Organizing Committee Members ....................... 2
Exhibition .......................................................... 3
Schedule ............................................................ 5
Symposium Information .................................... 6
Symposium Program ........................................13
Abstracts ...........................................................22
1
Organizing Committee Members
Advisory Committee
Prof. I. Mangerig
University of the Bundeswehr Munich, Germany
Prof. S. Inoue
Osaka Institute of Technology, Japan
Prof. K. Sugiura
Kyoto University, Japan
Prof. M. Keuser
University of the Bundeswehr Munich, Germany
Chairman
Prof. T. Yamaguchi
Osaka Metropolitan University, Japan
Members
Prof. T. Yamaguchi
Osaka Metropolitan University, Japan
Prof. H. Onishi
Iwate University, Japan
Prof. O. Ohyama
Osaka Institute of Technology, Japan
Prof. Y. Kitane
Kyoto University, Japan
Prof. T. Kitahara
Kanto Gakuin University, Japan
Prof. T. Tominaga
Shizuoka Institute of Science and
Technology, Japan
Prof. H. Higashiyama
Kindai University, Japan
Prof. T. Matsumoto
Hokkaido University, Japan
Prof. M. Matsumura
Kumamoto University, Japan
Prof. Y. Mikata
Osaka Institute of Technology, Japan
Prof. G. Siebert
University of the Bundeswehr Munich, Germany
Prof. M. Mensinger
Technical University Munich, Germany
Prof. O. Fischer
Technical University Munich, Germany
Prof. T. Braml
University of the Bundeswehr Munich, Germany
Prof. M. Spannaus
University of the Bundeswehr Munich, Germany
Dr. C. Braun
Maurer SE, Germany
Assoc. Prof. Y. Imagawa
Osaka Institute of Technology, Japan
Assoc. Prof. K. Nagata
Nagoya Institute of Technology, Japan
Assoc. Prof. K. Hashimoto
Kobe University, Japan
Assoc. Prof. M. Hirohata
Osaka University, Japan
Asst. Prof. G. Hayashi
Osaka Metropolitan University, Japan
Secretary General
Asst. Prof. G. Hayashi
Osaka Metropolitan University, Japan
Secretary:
Mr. R. Sakura
Osaka Metropolitan University, Japan
Mr. Y. Chen
Osaka Metropolitan University, Japan
2
Exhibition
Company exhibition booths
Booth F
Booth D
Booth G
Booth I
Entrance
column
column
Booth C
Booth J
Booth B
Booth A
Entrance (closing)
Entrance
(Symposium venue)
Booth No.
A. Hanshin Expressway Co., Ltd.
Hanshin Expressway Research Institute for Advanced Technology
B. Honshu-Shikoku Bridge Expressway
Co., Ltd.
G. SHO-BOND Corporation
C. Japan Bridge Association
H. West Japan Railway Company
D. The Japan Iron and Steel Federation
I. Sumitomo Mitsui Construction Co.,
Ltd.
E. Kawakin Core-Tech Co., Ltd.
J. West Nippon Expressway Co., Ltd
F. MAURER SE
3
A.
F.
B.
G.
C.
H.
D.
I.
E.
J.
4
Schedule
9:00 9:20
9:40
10:30
10:40
11:30
11:55
12:45
-
9:40
10:30
10:40
11:30
11:55
12:45
13:15
13:15 - 15:00
August 29, Tuesday
Registration
Room-A
Opening Ceremony
Keynote Lecture 1
Coffee Break
Keynote Lecture 2
Group Photo
Lunch
See Exhibition
Room-A
Session 1-A:
Room-B
Session 1-B:
Design Codes and Bridge
Engineering 1
Steel Structures 1
Coffee Break
15:00 - 15:15
15:15 - 17:00
18:15 -
Session 2-A:
Session 2-B:
Design Codes and Bridge Engineering 2
Steel Structures 2
Reception
@ AOI NAPOLI in TENNOJI
August 30, Wednesday
Room-A
Keynote Lecture 3
Keynote Lecture 4
9:00 - 9:50
9:50 - 10:40
10:40 - 10:55
10:55 - 12:10
Coffee Break
Room-A
Session 3-A:
Room-B
Session 3-B:
Composite Structures 1
Steel Structures 3
Lunch
See Exhibition
12:10 - 13:00
13:00 - 13:30
13:30 - 14:45
Session 4-A:
Session 4-B:
Composite Structures 2
Vibration and Monitoring
Short Break
14:45 - 14:50
Session 5-A:
Session 5-B:
14:50 - 16:05 Advances in Bridge Engineering and
Fatigue 1
Technologies 1
Coffee Break
16:05 - 16:20
Session 6-A:
Session 6-B:
16:20 - 17:35 Advances in Bridge Engineering and
Technologies 2
17:40 - 18:00
19:30 -
Fatigue 2
(Composite & Steel Structures)
Room-A
Closing Ceremony
Banquet
@ MIYAKO CITY OSAKA TENNOJI
5
Symposium Information
Venue
Osaka Metropolitan University, Sugimoto Campus
-Conference hall, 10F at Media Center
How to get to the Sugimoto Campus
Access by Public Transport
⚫ 5 min. walk from Sugimoto-cho Station JR Hanwa Line
⚫
Access by Public Transport
Access from the Main Terminals
⚫
From Kansai International Airport (KIX) :
Take the Kansai-Airport Rapid Service, change at Sakai-shi to a local train for Tennoji and get off
at Sugimoto-cho Station.
⚫
⚫
From Shin-Osaka Station :
Take the subway Midosuji Line and get off at Abiko Station.
From Osaka International Airport (Itami) :
Take an Airport Limousine Bus to Abenobashi Station, then take a local train on the JR Hanwa
Line from JR Tennoji Station and get off at Sugimoto-cho Station.
Secretariat Office & Information Desk
Secretariat Office of GJBS2023 will open at #C309, Faculty of Engineering Building from 9am
to 5pm.
The information desk is also located at the registration desk.
On-Site Registration
The on-site registration can proceed at the registration desk in the 10F at Media Center.
Registration fee
General
40,000 JPY
Student
20,000 JPY
6
Symposium Venue
Osaka Sta.
Tennoji Sta.
Sugimoto Sta.
Osaka Metropolitan Univ.
7
ACCESS
View-1
Ticket
Gate
★
View-2
★
View-3 ★
JR Subway Station
(Sugimotocho)
OMU
Bridge Eng., Lab.
Conference hall
10F
Meeting Place
(Media Center 1F)
Nonohana House
1F
View-4 ★
Metasequoia
View-4
View-1
View-2
8
Entrance
Conference hall
10th Floor
Study Room for Researching Staff
Study Room for Researching Staff
Meeting
Room
Meeting Room for
EV Researching Staff
Conference
Room
EV
EV
Exhibition
Coffee Break
Registration Desk
WOMEN
MEN
Room A
Opening Ceremony
Keynote Lecture
Closing Ceremony
Parallel Session A
9
Room B
Parallel
Session B
Lunch Venue
Symposium Lunch will be served in Nonohana Hause and Metasequoia.
※Nonohana Hause is for German delegates and Organizing committee members only.
※Metasequoia is for Japanese general participants.
※1
Nonohana House
※2
Metasequoia
※1：Nonohana Hause is on the 1st floor at Media Center.
※2：Metasequoia is on the southwest side of Osaka Metropolitan University.
10
Reception Venue
Reception will be held in AOI NAPOLI (Italian cuisine) on Tuesday, 29th August.
Reception will start at 18:15.
We recommend that you take the train that leaves at 17:29 / 17:44
from Sugimotocho station.
AOI NAPOLI
Google Map
Tennoji sta.
Banquet Venue
Banquet will be held in MIYAKO CITY OSAKA TENNOJI (French cuisine) on Wednesday,
31st August.
Banquet party will start at 19:30.
We recommend that you take the train that leaves at 18:44 / 18:57
from Sugimotocho station.
Tennoji sta.
Google Map
MIYAKO CITY
OSAKA TENNOJI
11
Technical Site Visit for German Delegates
Excursion will be held on Thursday, 31st August.
The below technical site visit timetable summaries are tentative.
Please ensure that return times may vary due to traffic and travel estimations.
9:30
10:30
12:00
14:10
18:00
-
11:30
13:30
15:10
16:30
21:00
August 31, Thursday
Technical Site Visit
Bus to depart from Tennoji
Shin-Meishin Expressway, Joyo Junction
Lunch
Yodogawa Bridge
Return to Osaka sta.
Farewell Party at Umeda
Dismissed at JR Osaka Sta.
Farewell Party will be held in Umenohana at Umedaon on Thursday, 1st
September. Please scan QR code for restrant information.
Technical Tour to Akashi-Kaikyo Bridge
& Hanshin Expressway Earthquake Museum
for German Young Engineer Delegates
Technical site visit will be held on Friday, 1st September.
The below site visit timetable summaries are tentative.
Please ensure that return times may vary due to traffic and travel estimations.
9:45
10:00
10:30
12:00
13:00
14:30
16:30
-
10:30
12:00
13:00
14:30
16:00
1st September, Friday
Technical Site Visit
The meeting place: JR Maiko Station
Tour of Akashi Kaikyo BRIDGE EXHIBITION CENTER
Tour of the main tower of the Akashi-Kaikyo Bridge
Lunch
JR Maiko Station to JR Konan Yamate
Tour of Hanshin Expressway Earthquake Museum
Dismissed at JR Konan Yamate Sta.
12
Symposium Program
Day 1
Keynote Session
Keynote Lecture 1
Chair: Takashi Yamaguchi & Geralt Siebert
9:40-10:30
Mr. Yasutomo Komatsu
(Osaka City Government, Japan)
Bridge Projects in Osaka City
Abstract:
The bridges located in Osaka City are often referred to as the ”808 bridges of Naniwa”. They
play a vital role in supporting the transportation needs of Osaka City, also known as the City of
Water. These bridges are not only functional but are also recognized by many citizens for their aesthetic contribution to the urban landscape. Simultaneously, since Osaka City was at the forefront
of early urbanization in Japan, some of its bridges are showing signs of aging and deterioration.
To counteract these issues, Osaka City is actively engaged in the maintenance and renovation of
these bridges. In this keynote, we will explain the bridge projects that Osaka City is currently
undertaking.
Keynote Lecture 2
10:40-11:30
Prof. Martin Mensinger
(Technical University Munich, Germany)
Advancements in Robust, Material-Efficient, and Sustainable Steel and Composite
Bridges at TUM
Abstract:
The keynote addresses current challenges in bridge construction in Germany and presents three
projects related to the key themes of material efficiency, durable and long-lasting structures and
preservation of heritage-listed bridges. These projects have been worked on at the Chair of Metal
Construction at TUM in recent times.
13
Session 1-A
13:15-15:00
Design Codes and Bridge Engineering 1
Chair: Hiroshi Higashiyama & Kazutoshi Nagata
Consideration of Imperfections with Temperature Differences Measurement of
Sections at the Thulba Viaduct
. . . . . . 23
Nadine Thomas, Agnes Weinhuber, Martin Mensinger, Joseph Ndogmo, Cristoph Holst
Fundamental Study on the Behavior of Curved Box Girder Bridge Subjected to
Temperature Change
. . . . . . 25
Kyogo Nakayama, Masahide Matsumura
Reliability Assessment of Existing Concrete Bridges with Geometrical NDT Results Case Studies
. . . . . . 27
Stefan Küttenbaum, Christian Kainz, Thomas Braml
Development of Fast-Setting UHPFRC for Bridge Deck Overlay
. . . . . . 29
Shunji Aoki, Takayoshi Tomii, Jun Homma, Koji Tamataki
Isogeometric Analysis of Bridge Structures: State of the Art and Potential Advantages
. . . . . . 31
Florian Zimmert, Leo Lapidus, Josef Kiendl, Thomas Braml
Verification of Soundness Judgement of RC Slabs by FWD
. . . . . . 33
Hiroki Akamatsu, Masaya Tsukamoto, Satoshi Tada, Hiroshi Higashiyama
Numerical Study on the Influence of Corrosion Damage on Reaction Force at the
Support of a Steel Box Girder Bridge
. . . . . . 35
Yuga Shutoku, Yasuo Kitane, Kunitomo Sugiura, Yoshinao Goi, Iksong Kim, Yasuo Hanaoka, Nobuhito Okubo
Session 1-B
13:15-15:00
Steel Structures 1
Chair: Takeshi Kitahara & Johann Kollegger & Yu Chen
Analytical Study on the Influence of the Misalignment of the Bottom Flange Joint
of a Box Girder on the Tightening Axial Force and Slip Strength
. . . . . . 39
Jianpeng Lai, Takashi Yamaguchi
Epoxy Coated Strands for International Stay Cable Applications
. . . . . . 41
Jannik Gawlista, Werner Brand
Experimental Study on Slip Load and Clamping Force Relaxation of Frictional
High Strength Bolt Connection with Entire Corrosion
. . . . . . 43
Lingbo Yao, Ming Li, Takashi Yamaguchi
Slip Tests of Double-lap Joints Consisting of Non-projected and Sandglass-shaped
Bolts with High Strength and Durability
. . . . . . 45
Masashi Takayama, Hitoshi Moriyama, Masayori Yoshimi, Takashi Yamaguchi, Gen Hayashi
Noise Reduction of Modular Expansion Joints on the Example of the New Pattullo
Bridge(CAN)
. . . . . . 47
Torsten Ebert, Christian Braun, Toshihisa Mano
Study on Load-carrying Capacity of Built-up Column Lost Lacing Bars Focused
on Buckling Mode
. . . . . . 49
Kenta Morimoto, Toshikazu Takai, Takao Miyoshi, Kaname Iwatsubo, Kazuya Tamada
Numerical Study of Stiffened Plates Joined by Thermal Spraying
Eitaro Horisawa, Kunitomo Sugiura, Yasuo Kitane, Yoshinao Goi
14
. . . . . . 51
Session 2-A
15:15-17:00
Design Codes and Bridge Engineering 2
Chair: Kazutoshi Nagata & Hiroshi Higashiyama
LT Bridge - A New and Fast Construction Method for Cost-efficient Bridge Structures
. . . . . . 55
Franz Untermarzoner, Michael Rath, Johann Kollegger
Fast Erection of Deck Slabs for Steel-concrete-composite Bridges
. . . . . . 57
Johann Kollegger, Franz Untermarzoner, Michael Rath
Shin-Meishin Expressway where Bridge Construction Progresses
. . . . . . 59
Yoshinori Wada, Joon-Ho Choi, Takafumi Omura, Shinya Hiraoka, Shinya Maehara, Moeka Tokutsu, Masafumi Udo
The Use of Glass for Bridges - Basics, Special Questions, Codes and Application
Examples
. . . . . . 61
Geralt Siebert
Structural Design of Glass-Elements in Bridge-constructions
. . . . . . 63
Alexander Pauli, Geralt Siebert
The Arnulfpark Bridge - Glass as Contact Protection and Design Element
. . . . . . 65
Barbara Siebert, Tobias Herrmann
Fiber Distribution Pattern Recognition in UHPFRC Based on Deep Learning
Technology
. . . . . . 67
Xin Luo, Takashi Matsumoto
Session 2-B
15:15-17:00
Steel Structures 2
Chair: Johann Kollegger & Takeshi Kitahara & Yu Chen
A Fundamental Study on Application of Two-dimensional Hermitian Elements to
In-plane Bending Deformation Problems of Plates
. . . . . . 71
Tsukushi Okabe, Masaki Sakai, Naoki Kaneko, Kyosuke Yamamoto
Experimental Investigation on Corrosion Deterioration in Defective Areas of
Paint-coated Steel
. . . . . . 73
Feng Jiang, Kazuki Ojima, Mikihito Hirohata
Analytical Study on Transferred Load of Friction-bearing Hybrid Joints with
Mechanical Bearing Blind Rivet Bolts
. . . . . . 75
Masataka Komura, Takashi Yamaguchi
Energy Absorption of Bolted Patch Plate Repaired Member in Ultimate Behavior
. . . . . . 77
Souta Masudome, Toshikazu Takai
Corrosion Assessement of Weathering Steel Bridges in Osaka and Wakayama
Prefectures (Japan)
. . . . . . 79
Wint Thandar, Shen Hui, Yasuo Hanaoka, Nobuto Okubo, Testuya Iida, Tomonori Tomiyama, Kunitomo Sugiura
Evaluation of Debonding of CFRP Bonded onto Steel Plate by AE Method
. . . . . . 81
Morimune Mizutani, Toshiyuki Ishikawa, Yoshimichi Fujii
Study on Relationship between Whole Displacement and Bearing Deformation of
Bolt Holes in High-strength Frictional Bolted Joints
Zice Qin, Hitoshi Moriyama, Takashi Yamaguchi
15
. . . . . . 83
Day 2
Keynote Session
Keynote Lecture 3
Chair: Osamu Ohyama & Oliver Fischer
9:00-9:50
Prof. Yasuo Kitane
(Kyoto University, Japan)
Structural response and remaining capacity of steel plate girder subjected to fire
conditions
Abstract:
Bridge fires may not occur as often as the other disasters such as earthquakes and floods, and
most bridges are not designed against fires. However, once it happens, a bridge may sustain a
major damage, which will greatly affect economic activities. This presentation will cover a series
of studies on structural response and remaining capacity of a steel plate girder subjected to elevated temperatures simulating fire conditions and a repair method for its deformed web panel. In
addition, an estimation method for the maximum temperature of steel bridge subjected to a fire
is introduced.
Keynote Lecture 4
9:50-10:40
Prof. Thomas Braml
(University of the Bundeswehr Munich, Germany)
Digital twins for the infrastructure
Abstract:
Bridge maintenance is an essential task for the infrastructure in the coming years. Especially
in the case of bridges with small and medium spans in the municipal road network, in many cases
there are no personal available for structural inspection and maintenance. Therefore, it will be
necessary in the future to develop concepts for a partially automated bridge inspection. If information from loads and resistances of structures is available, the measured values contain important
information on the condition of the structure. One concept for this can be digital twins. The presentation shows the implementation of the concept on a 1-span prestressed concrete frame bridge.
The installation of the sensors, the digital twin in the management shell and the evaluation of the
data are presented.
16
Session 3-A
10:55-12:10
Composite Structures 1
Chair: Yousuke Imagawa & Max Spannaus
Building Bridges with Thin Walled Semi-precast Concrete Elements-experimental
Torsional Investigations
. . . . . . 87
Michael Rath, Franz Untermarzoner, Johann Kollegger
Proposal on Rigid Connection between Steel Deck Plate Girder and RC Abutment in Replacement Project
. . . . . . 89
Yasuo Tawaratani, Naomitsu Akashi, Mikinao Goto, Osamu Ohyama, Yusuke Imagawa
Innovative Developments of Composite Columns with High-strength Steels
. . . . . . 91
Michael Schäfers, Rudolf Röss, Martin Mensinger
Precast Modular Bridge Structures - Current Developments, Pilot Projects and
Experimental Investigations
. . . . . . 93
Oliver Fischer, Nicholas Schramm
Analytical Study on the Mechanical Behavior of the Intermediate Support in the
Composite Structure Using Bearing Plates
. . . . . . 95
Kenta Nakaoka, Takashi Yamaguchi, Satoshi Kimura, Taro Tonegawa
Session 3-B
10:55-12:10
Steel Structures 3
Chair: Toshikazu Takai & Thandar Wint & Ryo Sakura
Slip Behavior between Cast Iron Deck Module and Steel Main Girder Using High
Strength Bolted Frictional Joints with Slotted Hole
. . . . . . 99
Yugo Shirai, Takashi Yamaguchi, Ryo Yamashita, Hironobu Tobinaga
Sufficient Choice of Steel Material for Bridge Bearings to Avoid Brittle Fracture
. . . . . . 101
Natalie Hoyer, Bertram Kühn
MAURER Uplift Spherical Bearing
. . . . . . 103
Toshihisa Mano, Christian Braun, Torsten Ebert
Significance of Treating Initial Imperfection in FE Simulation for Compressive
Behavior of Welded Steel Structural Members
. . . . . . 105
Yuxuan Cheng, Shuhei Nozawa, Mikihito Hirohata
Evaluation of Load Capacity of Temporary Bridges Using End-plate Connections
under Pure Bending Moments: A Proposal for a Simplified Calculation
Ruoxi Li, Yu Chen, Isao Matsuda, Hirotoshi Azuma, Takashi Yamaguchi
17
. . . . . . 107
Session 4-A
13:30-14:45
Composite Structures 2
Chair: Max Spannaus & Yousuke Imagawa
Bond Behavior of CFRP Plates with Tapered Ends for Steel Structure Reinforcement
. . . . . . 111
Shunta Sakurai, Yuya Hidekuma, Kazuo Ohgaki, Yoshiaki Okui
The Effect of Prestressing on the Shear Capacity of Post-tensioned Concrete
Beams
. . . . . . 113
Sebastian Lamatsch, Oliver Fischer
Crack Prevention Methods of Pre-flexed Beam Prefabricated by Segmental
Method
. . . . . . 115
Hiroaki Fujibayashi, Naoki Noro, Shota Tsuzi, Osamu Ohyama, Shigeyuki Matsui
Imaging of Ultrasonic Echo Measurements for Reconstruction of Technical Data
of Bridges – Possibilities, Limitations and Outlook
. . . . . . 117
Stefan Maack, Ernst Niederleithinger
Effect of Fire Damage on Residual Prestress and Load Carrying Capacity of
Pretensioned Prestressed Concrete
. . . . . . 119
Dennise, Yasuhiro Mikata, Susumu Inoue
Session 4-B
13:30-14:45
Vibration and Monitoring
Chair: Thandar Wint & Toshikazu Takai & Ryo Sakura
Study of a Monitoring Plan and Behavior Analysis to Verify the Performance of
an Integrated Column by Multiple Steel Pipes
. . . . . . 123
Shinsuke Akamatsu, Masahiro Hattori, Yasumoto Aoki, Yoshiki Taniguchi, Kunitomo Sugiura
Application of Bridge Weigh-in-Motion on a Bridge with Prestressed Concrete
Girders
. . . . . . 125
Marcel Nowak, Oliver Fischer
Natural Frequency of Lightweight Foamed Concrete Composite Slabs (LFCCS)
. . . . . . 127
Zainorizuan Mohd Jaini, Kunitomo Sugiura, Sakhiah Abdul Kudus
Geo-referenced Localisation of SHM Sensors on New Bridge Construction Based
on the Example of the Digital Bridge Schwindegg (Germany)
. . . . . . 129
Johannes Wimmer, Thomas Braml
Standardisation in Structural Health Monitoring (SHM) - A Concept Proposal
Thomas Braml, Johannes Wimmer, Fabian Seitz, Max Spannaus
18
. . . . . . 131
Session 5-A
14:50-16:05
Advances in Bridge Engineering and Technologies 1
Chair: Yoshinao Goi & Kyosuke Yamamoto
Improving Damage Prediction by Assessing Structural Damage Through Sensor
Measurements in Combination with Virtual Building Models
. . . . . . 135
Nathalie Nießer, Geralt Siebert
“SmART Strand” Prestressing Steel Strand with Optical Fiber Sensor for Tension
Monitoring
. . . . . . 137
Masashi Oikawa, Shinji Nakaue, Naoki Sogabe, Michio Imai
Use of Data from BIM Method for New and Existing Concrete Bridges - Practical
Report and Possible Improvements
. . . . . . 139
Christian Kainz, Gertraud Wolf
Study on Damage Detection of Simply-supported Bridges Using Structural Responses of Girder Ends
. . . . . . 141
Phyoe W. Hein, Yoshinao Goi, Yasuo Kitane, Kunitomo Sugiura
Reuse of Structural Steel Products
. . . . . . 143
Christoph Ehrenlechner, Christina Radlbeck, Martin Mensinger, Matthias Müller, Thomas Ummenhofer
Session 5-B
Fatigue 1
14:50-16:05
Chair: Risa Matsumoto & Osamu Ohyama
Analytical Study on the Reinforcement of Intersections Structure in Orthotropic
Steel Deck by U-rib Cutting Method
. . . . . . 147
Qihang Shen, Takashi Yamaguchi
Steel Castings in Infrastructure Projects
. . . . . . 149
Sven Nagel, Max Spannaus
An Investigation on Prevention of Weld Root Fatigue Crack by Assistance with
Adhesive Bonding
. . . . . . 151
Yifei Xu, Mikihito Hirohata, Jiahao Mao
Crack Propagation Calculations with Scattering Material Parameters for the Assessment of Welded Bridges
. . . . . . 153
Dorina Siebert, Christina Radlbeck, Martin Mensinger
Experimental Investigation of the Ultra-Low-Cycle-Fatigue (ULCF) Behaviour of
Full-scale Steel components
Sergey Chernyshov, Andreas Taras
19
. . . . . . 155
Session 6-A
16:20-17:35
Advances in Bridge Engineering and Technologies 2
Chair: Kyosuke Yamamoto & Shinya Watanabe
Real-time Damage Assessment of Bridge Structures Based on Reduction of Natural Frequency under Ambient Vibration Measurement
. . . . . . 159
Khuyen Trong Hoang, Hiroyuki Uchibori, Naoki Nagamoto
Influence of the Longitudinal Reinforcement Ratio of Prestressed Beam Elements
on the Development of Strain and Compression Softening in the Cracked Web
. . . . . . 161
Sebastian Thoma, Oliver Fischer
A Simulation Model for Heating Correction on I-Shaped Welded Steel Bridge
Members
. . . . . . 163
Xiaoyu Guan, Mikihito Hirohata, Satoshi Mukawa, Seiji Okada
Development of a Temperature Model for Small-sized Box Girders
. . . . . . 165
Malik Ltaief, Martin Mensinger
Experimental Study of Sound-based Hammer Test on Composite Structure
. . . . . . 167
Yiran Yu, Yoshinao Goi, Kunitomo Sugiura
Session 6-B
16:20-17:35
Fatigue 2 (Composite & Steel Structures)
Chair: Osamu Ohyama & Risa Matsumoto
Fatigue Analysis of RC Slab Repaired with Early-Age Ultra-High Performance
Fiber Reinforced Concrete
. . . . . . 171
Amatulhay Pribadi, Takashi Matsumoto
Wheel Running Fatigue Test for Steel Plate-concrete Composite Deck Using Peculiar Shape Ribs with Multi-functional Projections
. . . . . . 173
Kozo Iwata, Risa Katsuki, Shota Nakagawa, Shigeyuki Matsui, Hiroshi Higashiyama
Experimental Fatigue Test on Historic Railroad Bridge
. . . . . . 175
Fabian Seitz, Max Spannaus
Lifetime Fatigue Reliability Analysis Considering Different Distribution Types
Mohamed Zied Mili, Kunitomo Sugiura, Yasuo Kitane
20
. . . . . . 177
SESSION 1-A
Design Codes and Bridge Engineering 1
13th - Japanese-German Bridge Symposium, Osaka, Japan
Consideration of imperfections with temperature differences
Measurement of sections at the Thulba viaduct
Dr.-Ing. Nadine Thomas*
Agnes Weinhuber M. Sc.**
Univ. Prof. Dr.-Ing. Martin Mensinger; Dr.-Ing. Joseph Ndogmo; Univ. Prof. Dr.-Ing. Christoph Holst
* Technical University Munich, Chair of Metal Structures, Germany, n.thomas@tum.de
** Technical University Munich, Chair of Engineering Geodesy, Germany, a.weinhuber@tum.de
Abstract:
The Material Testing Institute (MPA) of the Technical University of Munich was commissioned by the
Autobahndirektion Nordbayern to measure sections of the Thulba viaduct various stages of construction. The bridge is
being built using the incremental launching method. The measurements will be taken place at different positions on the
bridge and at different temperatures. The measurements are carried out by the Chair of Engineering Geodesy. The outof-plane imperfections of six webs are measured using 3D laser scanning. This paper present some of the results of the
day and night measurements without launching in the meantime. It was investigated whether the imperfections depend
on the temperature. The results of the measurements show a change in component deformation at the different times.
However, these changes are very small. In addition, there is no general reduction in imperfections due to cooling of the
components during the night. Considering the results at defined points and comparing them with the manufacturing
tolerances according to EN 1090-2:2018 [1] and the geometric equivalent imperfections according to EN1993-1-5
Annex C [2], the values are not exceeded in any case.
Keywords: imperfection, buckling behaviour, longitudinal stiffened panels, incremental launching, steel bridges
1 Introduction
During the launching of the box girder, complex loads occur. The decisive factor here is the biaxial stress state in the
bottom plates and webs. Due to the compressive stresses, these components are particularly sensitive to imperfections
and are at risk of buckling. The real imperfections out of the plane of the longitudinally stiffened webs and bottom
plates can therefore have a significant effect on the buckling behavior. Imperfections for the buckling check using FEM
are specified in EN 1993-1-5:2019 [2] Annex C. Regulations for the manufacture tolerances are given in EN 10902:2018 [1]. This paper compares the measurement results to these two standards and shows the influence of temperature
on the imperfections. Therefore, a day and a night measurement of the same bridge section (39-41; see Figure 1) was
carried out in September, as this is the period when the largest possible temperature difference is expected. For this
study, the web curvature is considered locally between the stiffener sections. Imperfections in the longitudinal stiffeners
are not considered in this report. Since imperfections perpendicular to the web are critical for the buckling analysis, case
No. 7 from Table B.1 of EN 1090-2 [1] is used. This case governs the curvature of the web out of its plane. For the
longitudinally stiffened webs, it is assumed that instead of the given flange in case No. 7 [1], the longitudinal stiffeners
are welded to the web. There is no regulation for longitudinally stiffened webs in EN 1090-2:2018 [1]. Annex C of EN
1993-1-5:2019 [2] provides equivalent geometric imperfections for the buckling verification using FEM. These
equivalent imperfections include both geometric and structural imperfections. For a local equivalent imperfection for
subpanel and single plates, see row 3 of Table C.2 and the corresponding Figure C.1 shall be used. To ensure maximum
coverage of all webs and redundant observation of neighbouring segments, eight viewpoints are used.
Figure 1: Representation of the evaluation points of each panel and overview bridge section
The definition of the coordinate system is shown in Figure 1. To define the reference plane, the positions of the targets
are recorded, and a reference plane is estimated through them. There are four targets on each web, placed in each edge.
Due to the definition of the reference plane, the deviation of the neighboring field is included in the measurement
results in this consideration. This can lead to higher deviations, as the reference plane would have to be defined
separately for each subpanel to compare the single panels with the two specified cases.
23
2 Results
In the plot of the measured values, results that deviate in the positive y-direction are shown in red and those that deviate
in the negative y-direction are shown in blue. Figure 1 shows the position (P1 to P4) of the evaluation points of each
panel and the bridge sections (39-40; 40-41; 41-42). The surface temperature was measured at these points after each
measurement. Both measurements are plotted on a graph to show the change in imperfection due to cooling. For the
segments, the temperature difference is determined by the horizontal offset of the measured points of the segments. No
relationship can be detected for the imperfections (see Figure 2). The largest differences in surface temperature can be
seen in components 41-42. Components depending on orientation (East = O/ West = W) and position (39-40; 40-41)
have a similar surface temperature over the height of the web. The imperfections are different. The measured values of
the day and night measurements are always lower than the permissible values according to [1] or the values given as
geometric equivalent imperfection in [2]. This can be seen in Figure 2 on the right. In this diagram, the regression line
has been created from the values according to [1] and [2]. This is shown in grey. All measured values (day and night
measurements) are below the straight line. From this plot it is easy to see that all results are within the acceptable range.
The manufacturing tolerances for web curvature according to EN 1090-2:2018 [1] are thus fulfilled. Geometric
equivalent imperfections for the buckling verification according to EN 1993-1-5:2019 Annex C [2] are defined to be
large enough for comparison in the center of the panel with b/200 for local buckling.
Figure 2: Comparison between Day- and Night-measuring and comparison of the results with [1] and [2]
3 Summary and Discussion
During the launch of the Thulba viaduct, three segments of the cross-section were measured at different times of day
and launching positions. This paper presents the analysis of the day and night measurements. The daytime
measurements show that there are large temperature differences on the component surfaces and in the air temperature.
This is due to the direction from which the sun shines on the component and the rapid drop in temperature in the
evening. A time difference of just one hour results in a temperature difference of approximately 5°C. A correlation
between temperature and out-of-plane component deformation cannot yet be established from the following night
measurement with the evaluation of the available measurement. The changes are generally very small. The maximum
measured difference occurs in section 40-41 East subpanel P3 with -0.62 mm at a temperature difference of 10.8°C.
Comparing the defined positions from the measurements with [1] and [2], all values are smaller than those given there.
To compare with [1] and [2], the reference area would have to be determined separately for each subpanel. In the
present investigation, each subpanel is already deformed, which means that the values in the center of the subfields are
also influenced by them. The values evaluated in this way and compared with the values in the center of each subpanel
given in [1] and [2] are always smaller. There are measured maximum values that would exceed these values. As these
points are at the edges of the plate, the question arises as to whether or how they should be included in the evaluation.
The values given in [1] and [2] also refer to the position in the center of each plate or the hole stiffened plate, but they
assume that all other positions are smaller than this value. On the one hand, the imperfection in the load introduction
zone is critical for plate buckling. During launching, the load introduction zone passes through all sections of the bridge.
On the other hand, critical areas are unsupported slender sections of the bridge. These components are the most
sensitive to geometric imperfections. These are deviations in the central region of the subpanel, or globally of the whole
plate. It is therefore unlikely that the assessment can be made by comparing a single value. Rather, it would be
necessary to investigate which areas of the plate’s maxima could become decisive for buckling. This range should then
be compared with [1] and [2]. In [1] there is no regulation for longitudinally stiffened webs. If for the web, the
subpanels can be compared with the web curvature according to No. 7 of Table B1 [1] using the position of the
longitudinal stiffeners instead of the flange can be discussed.
4 References
[1] EN 1090-2:2018; Execution of steel structuresand aluminium structures – Part 2: Technical requirements for
steel structures; German version
[2] EN 1993-1-5:2019; Eurocode 3 - Design of steel structures - Part 1-5: Plated structural elements; German version
24
13th - Japanese-German Bridge Symposium, Osaka, Japan
Fundamental Study on the Behavior of Curved Box Girder Bridge
Subjected to Temperature Change
Master course Kyogo Nakayama *
Univ. Prof. Dr.-Eng. Masahide Matsumura **
* Kumamoto University, Department of Civil and Environmental Engineering and Architecture, Japan,
235d8310@st.kumamoto-u.ac.jp
** Kumamoto University, Center for Water Cycle, Marine Environment and Disaster Management, Japan,
matsumura-m@kumamoto-u.ac.jp
Keywords: Curved box girder bridge, Temperature change, FEM analysis, Radius of curvature
1 Introduction
In August 2021, a noise was reported at the Ushibuka-Haya Bridge in Amakusa City, Kumamoto Prefecture, and a visual
inspection was conducted. Figure 1 shows the damage to the bearings. As a result of the investigation, cover plates were
found to have fallen off, rollers were ruptured, and pressure plates were damaged at three pivot roller bearings that were
movable in the longitudinal direction. After two months, an internal inspection of all bearings revealed damage to several
other bearings as well. The complex behavior of the girder due to temperature changes may have repeatedly transmitted
loads from directions other than the direction of roller movement to the pivot roller type bearing, causing a phenomenon
similar to ultra-low cycle fatigue failure in the bearing, which led to the damage [1]. So in this study, a straight bridge
model with steel bearings and a curved bridge model with four different radii of curvature were created to estimate the
reaction force and displacement of the bearings when temperature variations were considered.
2 FEM Analysis
The analytical model is a three-span curved box girder with a constant radius of curvature and a span of 133.3 m with a
total length of 399.9 m. The model is based on the north side (P3-P6) of the main line of the Ushibuka-Haiya Bridge, and
is created using temperature-displacement coupled shell elements with box-shaped cross sections as shown in Figure 2.
The analytical models were curved bridge models (D15, D30, and D45) and straight bridge model (S) with different radii
of curvature and girder end crossing angles of 15°, 30°, and 45°, as shown in Figure 3. Piers and longitudinal slopes were
not considered. Table 1 [2, 3] shows the material properties. The entire girder was subjected to a temperature change of
+30°C after the model was loaded with dead load. Model S with different radii of curvature and bearing conditions were
compared to study the deformation behavior and the effect of the temperature change on the bearing. The analysis software
used was Abaqus 2017 [4].
Figure 1: Ground plan of the superstructure of the main line section
of the Ushibuka-Haiya Bridge and the damaged parts of the bearings [1]
Figure 3: Alignment of each model
Table 1: Material properties [2, 3]
Model D15-XY
Model name Bearing movable direction
X: Transvers direction
Y: Longitudinal direction
Z: Vertical direction
Figure 2: Cross-sectional shape
25
Figure 4: Analysis pattern name
3 Analysis result and consideration
Figure 5 shows the bearing displacements in the coupled temperature-displacement analysis. For bearing Type-XY, the
displacement of the bearing located outside of P6 is the largest, and tends to be larger for models with a smaller radius of
curvature. The longitudinal displacement has different characteristics depending on the type of bearing. For bearing TypeXY, the displacement of the outer bearing of the P6 pier of Model S was 143.8 mm, the largest among all bearings, and
the displacement tended to be smaller for models with a smaller radius of curvature. On the other hand, for the Type-Y
bearing that restrains transvers movement, the difference in longitudinal displacement between the straight and curved
bridges is smaller. Figure 6 shows the direction of girder movement when Model D45 is subjected to temperature changes.
It can be read that the girder moves in the longitudinal direction for bearing Type-Y, whereas the direction of the girder
movement does not coincide with the longitudinal direction for bearing Type-XY. Figure 7 shows the transvers and
vertical reaction forces; it can be read that for Type-Y, the bearing reaction force values of P4 and P5 are larger than those
of P3 and P6. The effect of the radius of curvature on the transvers reaction force is small, but the constraining conditions
of the bearings generate large reaction forces; for the vertical reaction force of Type-Y, negative values were read for the
bearings located inside P3 and P6. Therefore, when a curved bridge is thermally deformed by external temperatures, the
smaller the radius of curvature, the greater the transvers movement. Restraining a girder that is also trying to move
transvers direction, as in the case of bearing Type-Y, will generate large transvers reaction forces on the bearing, which
will also affect the vertical reaction forces.
4 Conclusion
Curved box girder bridges that are not constrained in the transversal direction move in the longitudinal and transversal
directions due to girder expansion and contraction caused by temperature changes, and the amount of movement strongly
affects the bridge length and radius of curvature. On the other hand, a curved box girder bridge constrained in the
transversal direction expands and contracts along the longitudinal direction, and a reaction force is generated in the
constrained transversal direction. In addition, there is a large difference in the vertical reaction force values of bearings
installed on the same piers. This effect is thought to be related to the radius of curvature. Further research is needed to
determine the extent to which this effect acts on the bearings and to clarify the process of bearing damage.
5 References
[1] Kumamoto Prefecture: The 1st Technical Investigation Committee for Permanent Countermeasures of UshibukaHaiya Bridge due to Bearing Damage. Kumamoto Prefecture, JAPAN, 2021.
[2] Japan Highway Association: Road Bridge Specifications and Commentary II: Steel Bridges and Steel Members,
pp. 61, JAPAN, 2017.
[3] Yusuke Imagawa, Kazuo Takehara, et al: Analytical Study on Load Carrying Capacity Evaluation of SteelConcrete Simple Composite Girder Bridges, Journal of Structural Engineering, JSCE, Vol.53A, pp.1107-1116,
JAPAN, 2013.
[4] Dassault Systems Simulia: ABAQUS User's Manual, Ver 6.8, 2008.
(c) Longitudinal direction Type-Y
(a) Transvers direction Type-XY
(b) Longitudinal direction Type-XY
Figure 5: Transvers and Longitudinal displacement in Coupled Temperature-Displacement Analysis [mm]
Figure 6: Synthetic displacement in
Model D45 [mm]
(b) Vertical direction Type-Y
(a) Transvers direction Type-Y
Figure 7: Transvers and Vertical reaction force in Coupled TemperatureDisplacement Analysis [N]
26
13th - Japanese-German Bridge Symposium, Osaka, Japan
Reliability assessment of existing concrete bridges
with geometrical NDT results – case studies
Dr.-Ing. Stefan Küttenbaum *
Dipl.-Ing. Christian Kainz **
Univ.-Prof. Dr.-Ing. Thomas Braml **
* BAM – Federal Institute for Materials Research and Testing, Div. 8.2, Berlin, Germany, stefan.kuettenbaum@bam.de
** University of the Bundeswehr Munich, Inst. of Struct. Eng., Neubiberg, Germany, thomas.braml@unibw.de
Extended abstract:
The reliability analysis of existing structures employing measured data is particularly useful when required information
is not available or incomplete, when justified doubts about the condition of available information have arisen, or when
the available information is outdated. The non-destructive impulse echo methods, i.e., ultrasonics and ground penetrating
radar, are capable to detect and measure, for example, component thicknesses and the position of tendons and steel rebars
(see Figure 1). The significance of precise knowledge about these internal and external dimensions of structures became
evident, e.g., in an analysis of 723 damage events that occurred primarily in European countries, some of which had
considerable consequences [1]. In this study, 109 (15 %) of these damages were attributed to wrong dimensions or to the
incorrect placement of the reinforcement. An example of deviations determined with the aforementioned volume methods
can be found in [2]. If no verified construction documents are available in the course of a structural reassessment, the asbuilt condition must be surveyed and compared with any existing plans [3]. A regional investigation of 157 bridges, which
are public easements of German municipalities, has e.g. shown, that the as-built drawings of 42 bridges are missing [4].
Against this background, this paper addresses the clarification of the inner structure of concrete bridges (position of
tendons and steel rebars) with the aim to explicitly incorporate quality-assured, measured information into the reliability
assessment of existing structures. The presented case studies are part of the national, German pre-standardization project
ZfPStatik, which aims at the structure-specific and measurement data-based modification of partial safety factors [5].
Figure 1: A) Principle of impulse echo methods [6]; B) One of the investigated bridges with GPR results [7]
The procedure for the reliability assessment of existing road bridges in Germany is specified in the recalculation guideline
[3,8]. Once the recalculation premises (e.g., the target load level) have been defined and the available documents on the
structure reviewed, structural analyses are performed, the evaluated and verified results of which are provided to the
owner of the structure who determines further actions to be taken. On-site inspections are not mandatory when evaluating
structural reliability. The respective observations can, however, increase the level of approximation of the computation
models. One information source are non-destructive testing methods. NDT-supported reliability assessment can be guided
by the following reassessment procedure. The three case studies excerpted in section 3 each address one of these steps:
1.
2.
3.
Reliability assessment without measured data
Aim: Targeted definition of quantities to be measured and specification of needed testing accuracies.
Non-destructive bridge inspections
Aim: Quality-evaluated NDT- and measurement results characterizing properties of the investigated structure.
Reliability assessment using measured data
Aim: Refined reassessment results based on the quality-evaluated on-site testing results.
The overall project goal is the integration of most modern inspection methods into the assessment process of existing
structures by supplementing the current rules with a practicable and standard-compliant procedure for NDT data-based,
structure-specific modifications of partial safety factors. This will facilitate more realistic structural assessments,
optimized maintenance strategies, higher infrastructure availability and, in general, the targeted allocation (and thus
conservation) of resources. The project will be completed in summer 2024.
27
2
Bridge profile of the case study in section 3.1, photos taken by BAM
Existing bridge carrying a four-lane federal highway in Schleswig-Holstein, Germany (built 1980)
Cross-section: Slab- &-beam with two haunched main girders
System:
Longitudinally and transversely prestressed,
four span continuous beam
Dimensions
Length
Width Beam height Slab height
/m
95,80
> 23
1,2…1,6
< 0,50
Investigated limit
ULS/SLS/Fatigue:
states acc. to German
1. Proof of the stirrup reinforcerecalculation guideline
ment (torsion, shear)
(stage 2*) [3,8]:
2. Decompression
3. Fatigue of shear reinforcement
* cf. full paper
Target load level: LM 1
Identified weaknesses:
1.
2.
Performed inspections:
‐
Based on a girder grillage model, the decompression proof, fatigue proof of the shear
reinforcement, and torsion proof of the stirrups could not be successfully performed.
The bridge could be successfully assessed using a FE shell model; however, the plans
contained inconsistent information about the transverse tendon curves.
Vertical position of the transverse tendon ducts with ultrasound-echo and ground
penetrating radar (GPR) in decisive cross-sections
Bridge profile of the case study in section 3.3, photos provided by HFR Ingenieure GmbH and BAM
Former existing bridge carrying a two-lane federal highway in Bavaria, Germany (built 1965)
Cross-section: Single-cell hollow box with variable height
System:
Longitudinally and transversely pre-stressed,
three span continuous beam
Dimensions
Total length
Width
Construction height
/m
133
12
1,20 … 1,45
Investigated limit
ULS/SLS in longitudinal direction:
states acc. to German
1. Robustness reinforcement
recalculation guideline
2. Bending and axial force
(stage 1 & 2) [3,8]:
3. Shear and torsion
4. Decompression
Target load level:
5. Concrete, reinforcement and
LM 1
prestressing steel stresses
6. Crack width limitation
Identified weaknesses:
Performed inspections:
1.
2.
3.
4.
‐
‐
‐
insufficient amounts of reinforcement to avoid failure without prior notice
serious deficits in the tension strut and deficits in the compression strut proof
insufficient amount of reinforcement to connect the slabs to the webs
minor deficits in decompression proof and torsion analysis
structural clarification: detection and localization of the longitudinal tendons in center
span and the shear reinforcement (pier area) using ultrasound and GPR,
half-cell potential (corrosion activity), concrete cover, air permeability [9],
concrete strength using drill cores, and further monitoring activities incl. PLT.
References
[1] Matousek M., Schneider J.: Untersuchungen zur Struktur des Sicherheitsproblems bei Bauwerken. ETH Zürich, 59 1976.
[2] Taffe A. et al.: Bauwerkscanner zur automatisierten und kombinierten Anwendung zerstörungsfreier Prüfverfahren im
Bauwesen. Beton- und Stahlbetonbau, 106 (4) 2011, p. 267–276. doi:10.1002/best.201100004.
[3] BMVBS: Richtlinie zur Nachrechnung von Straßenbrücken im Bestand (Nachrechnungsrichtlinie). 05/2011.
[4] Rechnungshof Rheinland-Pfalz: Bericht nach § 111 Abs. 1 LHO über die Erhaltung und den Zustand von Brücken in
kommunaler Baulast: Az.: 2-P-0057-39-1/2011. Speyer, 10.10.2013.
[5] BAM: Bridge safety – (…): https://www.bam.de/Content/EN/Press-Releases/2022/Infrastructure/2022-12-01-bridgesaftey-testing-methods.html (retrieved 2023-06-23). Press release. Berlin, 2022.
[6] Maack S. et al.: Die Ultraschall ‐ Echomethode – von der Messung zur bautechnischen Kenngröße. Beton- und
Stahlbetonbau, 116 (3) 2021, p. 200–211. doi:10.1002/best.202000091.
[7] Küttenbaum S. et al.: Ways to unlock the potential of non-destructive concrete testing for the reliability assessment of our
built environment. In: 8th International Workshop on Reliability of NDT/NDE @ SPIE Smart Structures + Nondestructive
Evaluation 2023. SPIE, 12.03.2023 - 17.03.2023, p. 1249107-1…11. doi: 10.1117/12.2658736.
[8] BMVI: Richtlinie zur Nachrechnung von Straßenbrücken im Bestand (Nachrechnungsrichtlinie). 04/2015.
[9] Maack S. et al.: Testing to Reassess – (…). In: Proc. of the 1st Conference of the European Association on Quality Control
of Bridges and Structures (Editors: C. Pellegrino et al.). Cham: Springer Int. Publ., 2022, pp 678–686.
28
13th - Japanese-German Bridge Symposium, Osaka, Japan
Development of Fast-Setting UHPFRC for Bridge Deck Overlay
Shunji Aoki *
Takayoshi Tomii *
Jun Homma **
Koji Tamataki ***
* Civil Engineering Renewal Department, OBAYASHI Corporation, Tokyo, Japan,
aoki.shunji@obayashi.co.jp, tomii.takayoshi@obayashi.co.jp
** Engineering Department, OBAYASHI Road Corporation, Japan, jun-homma@obayashi-road.co.jp
*** Research Institute Concrete Laboratory, Mitsubishi UBE Cement Corporation, Japan, koji.tamataki@mu-cc.com
1 Introduction:
There are approximately 730,000 road bridges in Japan. About half of them will be 50 years old by 2030. Many of these
bridges are suffering from fatigue deterioration of the slabs due to the heavy repeated wheel loads and number of
vehicles, chloride attack caused by the application of de-icing agents and damage caused by the freeze-thaw effect.
Deteriorated slabs have been repaired by the bridge deck overlay method, which involves removing and replacing the
top surface. However, the conventional material used for the repair, which is Steel Fiber Reinforced Concrete (SFRC),
has poor workability, integrity, and degradation has recured.
Therefore, as a material for the bridge deck overlay method, we developed a fast-setting Ultra-High Performance FiberReinforced Cement-based Composite (UHPFRC). UHPFRC is a suitable material for repairing and reinforcing slabs
because it has high crack resistance, density, and excellent durability. Since the compressive strength of the fast-setting
UHPFRC is quickly developed, the traffic can be opened in three hours after pouring the UHPFRC. In addition, its
fluidity can be adjusted to suit the construction area with cross-slopes up to 10%. With developing the fast-setting
UHPFRC, the dedicated batch plant, transporter, and concrete paving machine have been also deployed. By means of
the above-mentioned developments construction can be carried out in only one lane at night. This results in reducing the
impact on road users.
Keywords: bridge deck overlay, UHPFRC, fast setting
Figure 1: Large power screed
Figure 2: UHPFRC
2 Outline of fast-setting UHPFRC
Table 1 shows the UHPFRC target performance and measured values.
2.1 Fluidity
The flow rate of UHPFRC was set in a range of 150 to 280 mm to suit the slopes of the construction area over. The flow
rate is set so that the higher the slope, the lower the flow rate, and has been confirmed to be used up to 5% on bridges
and up to 10% in laboratory tests. The flow rate of the mortar was adjusted by setting the amount of high performance
water reducer and retarder to be added for each ambient temperature between 5 and 40 degrees Celsius.
2.2 Compressive strength
The compressive strength of SFRC used for the bridge deck overlay method has a management standard value of 24
N/mm2 or more at a given age (usually 3 hours) according to the Structural Construction Management Guidelines [3]
(Management Guidelines) used by Japanese expressway companies, and this material also met this performance. In
addition, the target value of 120 N/mm2 or more was met for compressive strength at 28 days.
29
2.3 Adhesive strength
UHPFRC showed high adhesive strength of 2.7 N/mm2 at 28 days of age. In the bridge deck overlay method, the
integrity of the existing and overlay sections is important in order to restore and improve the load-bearing capacity of
the slab. The adhesive strength between the concrete and the repair material was found to be sufficient compared to the
1.5 N/mm2 or more adhesive strength specified in the Management Guidelines.
2.4 Crack initiation strength
The average crack initiation strength at 28 days was 9.25 N/mm2, which was lower than the UFC Guidelines average of
11.7 N/mm2. The compressive strength of this material was approximately 130-150 N/mm2, which was lower than the
UFC Guidelines average of 194 N/mm2. Therefore, the crack initiation strength and compressive strength are
considered to be low.
2.5 Percentage change in length and mass
Figure 3 shows the measured rates of change in length and mass. The rate of length change was 211 μ expansion strain
at 28 days. According to the Management Guidelines, the dimensional stability of the bridge deck overlay method was
defined as a shrinkage rate of 250 μ or less, UHPFRC satisfied this requirement.
2.6 Accelerated neutralization depth
The depth of neutralization after 52 weeks of acceleration test in a 5% concentration environment was 0 mm for all
sample dimensions and no neutralization was observed.
2.7 Apparent diffusion coefficient of chloride ion
The apparent diffusion coefficient of chloride ions after 52 weeks of immersion in a 10% NaCl solution was
0.032cm2/year, a larger value compared to the apparent diffusion coefficient of UFC (0.0019cm2/year), but smaller than
that of ordinary concrete (0.14 to 0.9 cm2/year).
Table 1: Target performance of UHPFRC
UHPFRC (Material age)
Target value
Measured value
Flow (mm)
JIS R 5201(Static)
150 to 280
185
Air volume (%)
JIS A 1128
4.0 or less
3.0
Compressive strength
(N/mm2)
JIS A 1108
24.0 (3 hours)
34.8 (3 hours)
120(28 days)
150(28 days)
Static modulus of elasticity
(kN/mm2)
JIS A 1149
-
44.3(28 days)
Adhesive strength
(N/mm2)
JIS A 1171
1.0 (3 hours)
1.5 (3 hours)
1.5(28 days)
2.7(28 days)
Crack Initiation Strength
(N/mm2)
JIS A 1113
6.0(28 days)
9.25(28 days)
Length change
Test Method 439
Less than 250µ(28 days)
211µ expansion(28 days)
Neutralization(mm)
Test Method 439
-
0
Apparent diffusion coefficient of
chloride ion (cm2/year)
JCSE-G572 and
JCSE-G574
-
0.032
3 Summary
It was confirmed that the fast-setting UHPFRC satisfied the Management Guidelines standard values for adhesive
strength and dimensional stability of cross-sectional repairs used for the bridge deck overlay, and that the depth of
neutralization and the apparent diffusion coefficient of chloride ions were less and more durable than those of ordinary
concrete or SFRC.
The field application of the fast-setting UHPFRC confirms that construction can be completed within one-lane
restrictions on an in-service expressway, and that fluidity suitable for construction can be managed by adjusting the mix
to match the ambient temperature.
Based on the above, we believe that the fast-setting UHPFRC can be effectively used as a replacement for the SFRC
used in the bridge deck overlay method.
30
13th - Japanese-German Bridge Symposium, Osaka, Japan
Isogeometric Analysis of bridge structures:
State of the art and potential advantages
Florian Zimmert, M.Sc. *
Leo Lapidus, M.Sc. **
Univ.-Prof. Dr.-Ing. Josef Kiendl ***
Univ.-Prof. Dr.-Ing. Thomas Braml ****
* Universität der Bundeswehr München, Institut für Konstruktiven Ingenieurbau, Germany, florian.zimmert@unibw.de
** Universität der Bundeswehr München, Institut für Konstruktiven Ingenieurbau, Germany, leo.lapidus@unibw.de
*** Universität der Bundeswehr München, Institut für Mechanik und Statik, Germany, josef.kiendl@unibw.de
**** Universität der Bundeswehr München, Institut für Konstruktiven Ingenieurbau, Germany,
thomas.braml@unibw.de
Abstract
The Isogeometric Analysis is a novel method for the numerical solution of boundary value problems of different types.
Since its introduction in 2005, it has successfully been applied to different problems of structural mechanics, among
others. It offers significant advantages compared to the classical Finite Element Method. Nevertheless, this method is
barely used in practice nowadays. In this contribution we first summarise the fundamentals of Isogeometric Analysis
using Non-uniform rational B-Spline basis functions for both the geometric description of a structure and the numerical
approximation of a boundary value problem solution field (e.g. deformations). We then offer an overview of recent
applications of Isogeometric Analysis in the context of structural engineering with special focus on bridge constructions.
Finally, we highlight potential advantages of applying Isogeometric Analysis in future bridge design. In this contribution,
we focus on the benefits of a CAD-integrated, parametric design and analysis process, the advantages of geometric
reduction of three-dimensional systems as well as the consistent data exchange in a digital workflow. These advantages
are demonstrated using the example of a bridge superstructure.
Keywords: IGA; NURBS; bridges; design; digitalisation
1 Introduction
The Isogeometric Analysis (IGA) has first been introduced by Hughes et al. in 2005 [1]. Since then IGA has been
successfully applied in different fields of structural mechanics, fluid dynamics, vibration and wave analysis, and multiphysics, among others [2]. In IGA, Non-uniform rational B-Spline (NURBS) basis functions are used to describe both
the geometry of the structure to be analysed and the boundary value problem (BVP) solution field of interest (e.g.
deformations). Different elements for the Isogeometric Analysis of structures, like the Euler-Bernoulli beam or the
Kirchhoff-Love shell, have since been developed and can be applied to analyse bridge structures [3, 4]. Despite obvious
advantages, IGA is barely used for the design and calculation of bridge structures nowadays. Available research articles,
providing insight to possible applications and advantages in this field, are summarised in this contribution. By exploiting
the unique properties of NURBS basis functions and IGA elements, different advantages for the design and calculation
of bridge structures can be shown. In the following sections, three of them are briefly introduced.
2 CAD-integrated, parametric design and analysis
NURBS-based IGA allows for the calculation of deformations, internal forces, and stresses of a structure, directly using
a predefined geometric model. A discretisation and simplification of the geometric model and consequently the generation
of an additional numerical model is not necessary [1]. Especially in early planning phases of a bridge, when changes in
geometry are performed regularly, this offers considerable procedural advantages. Using NURBS, even complex systems
can be described and analysed without a discretisation of the geometry. Due to the high continuity of NURBS basis
functions and extended methods of refinement, the numerical analysis using IGA is performed with a lower computational
effort [1]. This allows even demanding nonlinear calculations, e.g. for the analysis of solid composite structures, to be
performed within an acceptable time-range. An efficient CAD-integrated design and analysis process also provides a
good basis for algorithmic optimisation processes. Furthermore, by exploiting the convex-hull property of NURBS curves,
surfaces, and solids, bridge structures can be integrated into the landscape automatically, respecting predefined geometric
boundary conditions [5].
3 Geometric reduction
Due to the tensor-product structure of NURBS solids and surfaces and the possibility to easily calculate local derivatives
and gradients, it is possible to automatically derive geometrically reduced numerical models from three-dimensional
geometric representations. This allows the advantages of a geometrically reduced numerical model to be exploited, for
example for the design of reinforced concrete bridge components, while at the same time retaining the digital
representation of the three-dimensional model [6]. This procedure is shown in Figure 1.
31
4 Consistent data exchange in a digital workflow
When using a single model for the design and numerical calculation of a bridge superstructure, all digital data of the
construction are updated in each design and calculation step and provided for further purposes. They can be used e.g. for
the evaluation of a design step in terms of cost, construction material consumption, or CO2 equivalent. Furthermore, the
raw data of the geometric description of structural components may be used for digital fabrication and automated
manufacturing, see Figure 1 [6].
Figure 1: Geometric reduction of a bridge superstructure and consistent data exchange in the framework of IGA [6]
5 Conclusions
In this contribution, we summarise the fundamentals of IGA and the state of science and technology for the application
of IGA in bridge design. The following areas, offering potential advantages in the application of IGA for bridge design
and calculation, are discussed:
▪
CAD-integrated, parametric design and analysis
▪
Geometric reduction
▪
Consistent data exchange in a digital workflow
6 Acknowledgements
This research work has been carried out within the project DEFINE and is funded by dtec.bw – Digitalization and
Technology Research Center of the Bundeswehr, which we gratefully acknowledge. dtec.bw is funded by the European
Union – NextGenerationEU. Within the project DEFINE, methods of free-form modelling as well as structuralmechanical and building-physical optimisation are applied to a novel AC/DC converter station for medium voltage
networks.
7 References
[1] Hughes, T. J. R.; Cottrell, J. A.; Bazilevs, Y.: Isogeometric analysis: CAD, finite elements, NURBS, exact geometry
and mesh refinement. Computer Methods in Applied Mechanics and Engineering, vol. 194, 39-41, pp. 4135–4195,
2005, doi: 10.1016/j.cma.2004.10.008.
[2] Cottrell, J. A.; Hughes, T. J. R.; Bazilevs, Y.: Isogeometric Analysis: Toward Integration of CAD and FEA. John
Wiley & Sons. Chichester, UK, 2009.
[3] Bauer, A. M.; Breitenberger, M.; Philipp, B.; Wüchner, R.; Bletzinger, K.-U.: Nonlinear isogeometric spatial
Bernoulli beam. Computer Methods in Applied Mechanics and Engineering, vol. 303, pp. 101–127, 2016, doi:
10.1016/j.cma.2015.12.027.
[4] Kiendl, J.; Bletzinger, K.-U.; Linhard, J.; Wüchner, R.: Isogeometric shell analysis with Kirchhoff–Love elements.
Computer Methods in Applied Mechanics and Engineering, vol. 198, 49-52, pp. 3902–3914, 2009, doi:
10.1016/j.cma.2009.08.013.
[5] Piegl, L.; Tiller, W.: The NURBS Book, 2nd ed. Springer Verlag Berlin, Heidelberg, New York, 1997.
[6] Zimmert, F.; Braml, T.: Freiformbauteile im Stahlbeton-, Spannbeton- und Verbundbau: Berechnung von
Querschnittswerten. Beton und Stahlbetonbau, vol. 118, no. 5, pp. 341–352, 2023, doi: 10.1002/best.202200110.
32
Verification of Soundness Judgement of RC Slabs by FWD
Hiroki Akamatsu *
Masaya Tsukamoto **
Satoshi Tada ***
Hiroshi Higashiyama ****
* Toa Road Corporation, Tsukuba, Japan, h_akamatu@toadoro.co.jp
** Toa Road Corporation, Minato-ku, Japan, m_tukamoto@toadoro.co.jp
*** Toa Road Corporation, Tsukuba, Japan, s_tada@toadoro.co.jp
**** Kindai University, Higashiosaka, Japan, h-hirosi@civileng.kindai.ac.jp
1 Introduction
In the methods for evaluating the soundness of bridge RC slabs, there are some indices such as crack density, width, and
interval from the visual inspection of cracks on the lower surface of slabs, and measured deflection values compared to
elastic theory or numerical analysis [1]. On the other hand, FWD (Falling Weight Deflectometer) tests, which have been
used to evaluate the soundness of pavement structures so far, are conducted as one of the soundness evaluation methods
of bridge slabs. This method is effective in measuring the deflection of slabs on the bridge surface. The authors focused
on the deflection area proposed by Abe et al. [2] as a soundness evaluation index for RC slabs using FWD and have
confirmed the linear relationship between the deflection area and the deflection at the loading point from the analysis and
measurement results on an actual bridge slab [3]. Since the deflection measured with FWD on the bridge surface having
asphalt pavement is affected by the pavement temperature, Higashiyama et al. proposed the correction method based on
the results of 3D elastic FEM analysis [4]. Moreover, Higashiyama et al. proposed the classification of RC slabs in four
stages to judge the soundness [5].
In this study, we evaluated the reliability of the soundness evaluation of RC slabs by FWD after the deflection correction
on an existing actual bridge before and after re-pavement construction.
2 Measurement results
2.1 Measurement overview
The target bridge in this study was a 4-span continuous steel non-composite plate girder bridge with a total bridge length
of 157.6m, a design slab thickness of 210mm, and an asphalt pavement thickness of 55mm. The measurement points were
hereinafter referred to as P1 to P28 as shown in Fig. 1. The measurement was performed on the asphalt pavement before
removing the existing asphalt pavement and on the exposed RC slab after removing the existing asphalt pavement. Fig. 2
shows each measurement condition. Since each RC slab panel was diamond-shaped, the deflection sensors in the
transverse direction to the bridge axis were set parallel to the crossbeams as shown in Fig. 3.
Fig. 1 Plan view of bridge and FWD measurement point
Fig. 2 FWD measurement conditions (left: on asphalt pavement , right: on RC slab)
33
Fig. 3 Installation locations of deflection sensors on each slab panel
2.2 Soundness evaluation results
The soundness evaluation results of each slab panel are shown in Table 1. The soundness of each slab panel obtained
from the data measured on the asphalt pavement and on the RC slab was consistent at 19 panels out of 28 panels in total,
and a high hit rate of about 70% was obtained. From the above, it can be said that the soundness evaluation of RC slabs
using FWD can be performed with high accuracy even from the asphalt pavement of the bridge surface.
Table 1. Soundness evaluation results of each panel on the measured bridge
Panel No.
P1
P2
P3
P4
P5
P6
P7
P8
P9
P10
Measurement
Asphalt pavement
Ⅲ
Ⅱ
Ⅱ
Ⅰ
Ⅱ
Ⅲ
Ⅰ
Ⅰ
Ⅰ
Ⅰ
location
RC slab
Ⅳ
Ⅱ
Ⅱ
Ⅰ
Ⅱ
Ⅱ
Ⅰ
Ⅱ
Ⅰ
Ⅰ
Match
×
○
○
○
○
×
○
×
○
○
Panel No.
P11
P12
P13
P14
P15
P16
P17
P18
P19
P20
Measurement
Asphalt pavement
Ⅰ
Ⅰ
Ⅰ
Ⅰ
Ⅲ
Ⅰ
Ⅱ
Ⅰ
Ⅳ
Ⅳ
location
RC slab
Ⅲ
Ⅰ
Ⅰ
Ⅰ
Ⅱ
Ⅰ
Ⅰ
Ⅰ
Ⅲ
Ⅲ
Match
×
○
○
○
×
○
×
○
×
×
Panel No.
P21
P22
P23
P24
P25
P26
P27
P28
Measurement
Asphalt pavement
Ⅱ
Ⅲ
Ⅱ
Ⅰ
Ⅰ
Ⅰ
Ⅱ
Ⅲ
Match
location
RC slab
Ⅱ
Ⅲ
Ⅱ
Ⅰ
Ⅰ
Ⅰ
Ⅱ
Ⅱ
19/28
○
○
○
○
○
○
○
×
Match
3 References
[1] Matsui, S., Maeda, Y.: A study on degradation judgment method for RC slabs of road bridges, Journal of JSCE, No.
374/I-6, pp.419-426, 1986.
[2] Abe, N., Sekiguchi, M.: Examination of applicability to soundness evaluation of reinforced concrete slabs of road
bridges using portable FWD, Proceedings of the 59th annual conference of the JSCE, V-620, 2004.
[3] Higashiyama, H., Mashito, H., Tsukamoto, M., Abe, N., Sekiguchi, M., Nagami, T.: Study on soundness evaluation
of bridge slabs by falling weight deflectometer, International Journal of GEOMATE, Vol.15, Issue 51, pp.106-112,
2018.
[4] Higashiyama, H., Mashito, H., Tsukamoto, M., Abe, N. Sekiguchi, M.: Analytical study on temperature correction
of asphalt concrete on RC slabs for soundness evaluation using FWD, Proceeding of 10th symposium on decks of
highway bridge, pp. 249-254, 2018.
[5] Higashiyama, H., Tsukamoto, M. and Mashito, H.: A proposal on soundness evaluation method of slabs using FWD,
Proceedings of the 11th symposium on decks of highway bridges, pp. 35-40, 2020 (in Japanese).
34
13th - Japanese-German Bridge Symposium, Osaka, Japan
Numerical Study on the Influence of Corrosion Damage
on Reaction Force at the Support of a Steel Box Girder Bridge
Yuga Shutoku1, Prof. Ph.D. Yasuo Kitane2, Prof. Ph.D. Kunitomo Sugiura3,
Asst. Prof. Dr. of Eng. Yoshinao Goi4, Iksong Kim5, Yasuo Hanaoka6, and Dr. of Eng. Nobuhito Okubo7
1
Kyoto University, Kyoto, Japan, shutoku.yuga.24w@st.kyoto-u.ac.jp
2
Kyoto University, Kyoto, Japan, kitane.yasuo.2x@kyoto-u.ac.jp
3
Kyoto University, Kyoto, Japan, sugiura.kunitomo.4n@kyoto-u.ac.jp
4
Kyoto University, Kyoto, Japan, goi.yoshinao.2r@kyoto-u.ac.jp
5
Kyoto University, Kyoto, Japan, kim.iksong.65x@st.kyoto-u.ac.jp
6
TAKADAKIKO Co., Ltd, Wakayama, Japan, n_ookubo@takadakiko.co.jp
7
TAKADAKIKO Co., Ltd, Wakayama, Japan, y_hanaoka@takadakiko.co.jp
1 Introduction
In steel bridges, when load-carrying members are damaged due to corrosion, the stiffness of the damaged member is
reduced, resulting in the load redistribution, which may cause a change in reaction force at the support from the healthy
state. This study focuses on the bearing on the abutment, which is relatively easy to access among bridges, and examines
the possibility of evaluating the integrity of steel bridges and estimating the damage level and location based on the
reaction force at the support. However, the relationship between corrosion damage and reaction force change at the
support in steel bridges has not been clarified. Therefore, this study
investigates the effect of different corrosion damages on the reaction
force by using a full bridge FE model of a non-composite steel box girder
bridge.
2 Target bridge and analysis method
2.1 Target bridge
Figure 1 shows a general view of the subject bridge. The bridge , which
was temporarily assembled at the Wakayama Works of TAKADAKIKO
Co., Ltd, is a simple non-composite narrow steel box girder bridge with
a length of 70.5 m, a width of 10 m, and a longitudinal slope of 4.300%.
2.2 Analysis model
The general-purpose finite element analysis software Abaqus2020, was
used. Figure 2 shows the analytical model. The analytical model was
created based on the drawings, using solid elements for the slab, concrete
on the end crossbeams, and sole plate, and shell elements for the steel
girder. The material properties are given in Table 1, and the dead load is
given by the acceleration of gravity (9.8 m/s2). Boundary conditions of
bearings are given on the line on the underside of the sole plate, as shown
in Figure 3, and SH1G1 and SH1G2 are fixed bearings and SH2G1 and
SH2G2 are movable bearings. The analysis is performed without
considering material and geometrical nonlinearities.
(a) Elevation view
(b) Cross-section view
Figure 1: General view
(a) Boundary condition setting location
Figure 2: Analysis model
Table 1: Material properties
E(N/mm2)
ν
Steel girder
Concrete on the
end crossbeams
2.00×10
Concrete slab
2.80×10
2.50×10
5
4
4
ρ(ton/mm2)
0.3
7.80×10
0.2
2.50×10
0.2
2.50×10
35
-9
-9
-9
(b) Schematic diagram
Figure 3: Bearing condition
3 Validation of the analytical model through temporary assembly test
Figure 4 shows a temporary assembly test. In the temporary assembly test, load
cells were installed at supports, and main girders were supported by four load
cells to measure reaction forces at the supports as shown in Figure 5. The loading
test was conducted by placing various weights on the bridge by using a crane,
and reaction forces were measured at the four support locations for each loading
case. Table 2 shows the loading cases. As shown in Figure 6, these loading cases
were reproduced in the analytical model by applying equally distributed loads at
weight locations in the test to validate the FE model. Figure 7 shows the change
in the reaction force between two loading cases at SH1G1. The graph shows that
the analytical and experimental values are close, so the analytical model is
considered to accurately reproduce the distribution of the reaction force at the
support on the steel girders of the target bridge.
Figure 5: Support point
40
20
0
-20
-40
Amount of
change (kN)
1
2
3
4
5
Table 2: Loading cases
Loading conditions
No loading
3t weight on SH1G1
5t weight on center span of G1
5t weight on center span of G1 and G2
5t weight on center span of G2
Figure 4: Temporary assembly test
Analysis
1→2
Figure 6: Loading point
2→3
3→4
Test
4→5
Transition in loading cases
Figure 7: Amount of change in reaction
force at SH1G1
4 Relationship between corrosion damage and reaction force at the support
4.1 Analysis case
To check whether the composite action of the slab causes a
change in the reaction force, a model with a slab modelled as
solid elements rigidly connected to the main girder (Figure 8
(a)) and a model without slab elements but with the slab load
placed on the upper flange of the main girder as an equally
distributed load (Figure 8 (b)) were created. Corroded regions
considered in the model are shown as blacked-out areas in
Figure 9. Corrosion in box girder bridges tends to occur on the
upper surface of the lower flange due to water retention inside
the box girder. Here, the amount of thickness reduction was
set to 11 mm, which is half the thickness of the smallest value
in the lower flange of the girder end portion. The weight loss
due to this thickness reduction is 109 kN
(a) Case M
(b) Case W
With modeled slab
With weight of slab
Figure 8: Analysis model
Amount of
change (kN)
4.2 Analysis result
Figure9: Corrosion part
Figure 10 shows amount of change in reaction force at each
100
M
W
support from the intact case. The result of W case shows that
50
most of the reduction in the reaction force occurs at the support
0
of the corroded girder(G1). This indicates that the load
-50
distribution effect of the crossbeams is small when the two
-100
main girders are connected only by crossbeams.
-150
Result of M case shows different changes in reaction forces
SH1G1 SH1G2 SH2G1 SH2G2
when compared to the results of W case. In M case, the
Figure10: Amount of change in reaction force at the
composite slab has a significant effect on reaction forces when
support from intact condition
thickness reduction due to corrosion is introduced. Reaction
force increases at SH1G1 support which is one of the supports
of the corroded girder, and decreases at SH1G2 which is the support of the intact girder. Therefore, it is necessary to
consider the deck slab when considering the effect of corrosion damage on reaction force at the support.
5 Summary
1) Comparison of the results of the temporary assembly tests and the analysis showed that the analytical model accurately
reproduced the load distribution of the main girders.
2) The analysis result for the corrosion-damaged model without slab elements showed that crossbeams had almost no
effect on the load sharing between two girders.
3) The change in reaction force due to corrosion damage from the healthy girder differs depending on whether a composite
slab is present or not.
36
SESSION 1-B
Steel Structures 1
13th - Japanese-German Bridge Symposium, Osaka, Japan
Analytical Study on the Influence of the Misalignment of the Bottom Flange Joint of a Box
Girder on the Tightening Axial Force and Slip Strength
PhD Student Jianpeng LAI*
Professor Takashi YAMAGUCHI**
* Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, su23503i@st.omu.ac.jp
** Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, yamaguti-t@omu.ac.jp
Keywords: High-Strength Bolted Friction Joints, Misalignment, FEM analysis
1 Introduction
High-strength bolted friction joints [1]-[3] (hereafter referred to as 'joints'), one of the in-situ connection methods for steel
bridges, have been widely used since the 1980s due to their advantages such as construction convenience and fatigue
resistance. Due to the long-term service after construction and the rapid increase in traffic, many bridge replacement
projects [4] are currently being carried out. In urban construction, the construction time and worksites are limited, and
there is a high possibility of construction errors such as misalignments between girders due to construction restrictions.
In a previous study [5], a strip girder joint was considered as an object of study by taking out a part of the girder joint and
setting 0, 1, 2 and 5 mm difference between the base plates of the joint, and an analytical study was carried out. As a
result, it was reported that the total axial force and slip strength at the completion of fastening were 9.3% and 20.2%
lower, respectively, for a joint with a 1 mm misalignment in comparison with a joint with a 0 mm misalignment. On the
other hand, in the joint of an actual steel box girder, the deformation condition of the plate is complicated by the increase
in the number of bolts and the increase in the size of the joint members because of the multiple matrix arrangement of the
bolts when the joint has a misalignment.
In this study, taking into account the tightening in-situ construction sequence, the influence of the joints on the bolt axial
forces and slip strength is evaluated from FEM analysis for steel box girder lower flange joints with 0, 1, 2 and 3 mm
joints between the base plates.
2 Analysis
2.1 Analysis model and method
The FE Model which focused on the lower flange joints and the boundary conditions of the analysis are shown in Figure
1. The analysis cases in Table 1. As shown in Figure 2, referring to the bolt location and tightening sequence in execution.
Considering the implementation procedure, the bolts at the joints of the upper flange were tightened before the joints of
the lower flange were tightened, as the road surface steps had to be secured. Based on the above, boundary conditions
were given to prevent vertical displacement of the central plane of the base plate at the joint. In order to prevent rigid
body movement of the connecting plate, restraint is given to the end of the connecting plate before the 60% tightening of
the bolts, and the restraint is released from the main tightening.
2.2 Slip definition in the analysis
The slip definition was determined as the maximum load up to the point where a load drop occurred.
3 Analysis results and discussion
3.1 Variations in bolt axial force
In order to quantitatively evaluate the influence of the misalignment, the rate of reduction of the bolt axial force is used.
The axial force reduction rate of the row1 bolt (J1-P1, J2-P1, J3-P1) fastened in each joint ①, ② and ③ is the highest.
Focusing on the rate of axial force reduction for each case, in the cases of 1, 2 and 3 mm, the axial force reduction rates
for the row1 tightened bolts (J1-P9, J2-P1 and J3-P1) were 11.5%, 21.6% and 30.0% respectively, with the axial force
reduction rate also increasing with the amount of the difference. Meanwhile, the average percent reduction in axial force
for each case in the J1-J4 joint was 1.9%, 3.3% and 5.6% for 1, 2 and 3 mm, respectively.
The non-dimensionalised axial force reduction rate, vertical displacement and the amount of misalignment at the first
tightened bolt (2mm-LF, J1-P1) are shown in Figure 3. The presence of a clearance between the base plate and the
splice plate at the J1-P1 bolt position leads to a higher vertical deformation at the J1-P1 bolt position due to the
tightening of the J1-P5 bolt.
3.2 Slip strength and slip coefficients
The slip strength and slip coefficients for each case are listed in Table 4. The reduction in slip strength was 5.2%, 7.1%
and 10.9% for the case of 1, 2 and 3 mm joints, respectively. The decrease in slip capacity and slip coefficient increased
with the amount of the misalignment. The reduction in the total axial force at the completion of fastening was 0.4%, 1.7%
39
Table 1 Analysis case
UZ＝URX=0
Axial force
up to 60%
UX＝10
UY＝UZ
＝０
Slip
stage
Plate width
direction
Axial
direction
Bolt axial force reduction rate
60
𝜂1 = 1 −
40
0mm-base
1mm-LF
2mm-LF
3mm-LF
0
1
2
3
Detail
Base case
Misalignment
parameter
A5 J1-A1 J1-P1 P5
P9
P13
A14 A10 A6
A2
P2
P6
P10 P14
A15 A11 A7
A3
P3
P7
P11 P15
A16 A12 A8
A4
P4
P8
P12 P16
J1
(Short for
Joint 1)
Figure 1 boundary conditions of the analysis
50
Misalignment
(mm)
A13 A9
UZ=0
Axial force
up to 100%
Vertical
direction
Case
As an example for
the location of bolts
Figure 2 Bolts location and tightening sequence
𝜂1 : Bolt axial force reduction rate
𝑁1
× 100%
𝑁𝑑 : design bolt axial force(=205kN)
𝑁𝑑
𝑁1 : Axial force of each bolt at completion of tightening
30
0mm-LF
1mm-LF
2mm-LF
3mm-LF
20
10
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
-P 1-P 1-P 1-P 1-P 1-P 1-P 1-P 1-P -P1 -P1 -P1 -P1 -P1 -P1 -P1
J
J
J
J
J
J
J
J
J1
J1
J1
J1
J1
J1
J1
J1
Bolt locations
Figure 3 The rate of decrease in axial force after completion of tightening
Figure 4 Timing of vertical
displacement variations
Table 4 The slip strength and slip coefficients
Case
0mm-LF
1mm-LF
2mm-LF
3mm-LF
Number Design slip
of bolts coefficient
52
0.5
Total design bolt
Total axial force on
Slip
At slip Total
slip
slip
axial forces
completion of tightening Strength axial force coefficient coefficient
/kN
/kN
/kN
/kN
μ1
μ2
10561
9563
9641
0.45
0.45
10521
9074
9220
0.43
0.43
10660
10385
8882
9063
0.42
0.43
10103
8528
8653
0.40
0.42
and 4.3% for the cases with 1, 2 and 3 mm of misalignment, respectively, compared to the total axial force at the
completion of fastening for the 0 mm-LF case. Therefore, the cause of the difference between the rate of decrease of the
total axial force and the rate of decrease of the slip strength at the completion of tightening is considered to be the
imbalanced contact between the base plate and the splice plate due to the influence of the misalignment.
4 Summary
This study focused on the total axial force and slip strength of the box girder lower flange joint with joints at the
completion of fastening, using a model with joints and analytical investigations. The following conclusions were obtained.
1) The axial force of the first bolt fastened tends to decrease in accordance with the fastening sequence in situ. Compared
to the case with no joint gap, the total axial force at the completion of tightening in the cases of 1, 2 and 3 mm was 1.9%,
3.3% and 5.6%.
2) The slip strength was reduced due to the effect of the misalignment, with 5.2%, 7.1% and 10.9% of the total slip
strength for the 1, 2 and 3 mm cases compared to the no misalignment case.
5 References
[1] AASHTO : AASHTO LRFD Bridge Design Specifications,2010.
[2] Japan Road Association (JRA). Specifications for highway bridges – Part 2 Steel Bridges and Members. Tokyo,
Japan: Maruzen, 2017. (in Japanese)
[3] EN 1993-1-8Eurocode 3: Design of steel structures–Part 1–8: Design of joints, 2005
[4] Ministry of Land, Infrastructure and Transport: Repair and Renewal of Road Structures, 2017 (in Japanese)
[5]Toshikazu Takai Analytical study on influence of irregularity on slip strength of high strength bolted friction type
joint. The journal of Structural Engineering, 61A, 605-613, 2015. (in Japanese)
40
13th - Japanese-German Bridge Symposium, Osaka, Japan
Epoxy coated strands for international stay cable applications
M.Eng. Jannik Gawlista *
Dipl.-Ing. Werner Brand **
* DYWIDAG-Systems International GmbH, Unterschleissheim, Germany, jannik.gawlista@dywidag.com
** DYWIDAG-Systems International GmbH, Unterschleissheim, Germany, werner.brand@dywidag.com
Keywords: Epoxy coated strand; Cable-stayed bridges; durability; corrosion protection
Extended Abstract:
Cable-stayed bridges have gained significant popularity in recent years, particularly for medium to long-spans due to
their cost efficiency, enhanced stiffness compared to suspension bridges, and typically simpler construction. With this
increased popularity stay cable technology has also evolved, and upper design boundaries and records have
continuously progressed. Besides record spans, cable-stayed bridges have also developed in other areas. Extradosed
bridges for instance have become a popular construction type worldwide for shorter to medium spans, with the first
application in Japan already in 1994 at the famous Odawara Blueway Bridge.
At this bridge epoxy coated strands have been introduced to address concerns regarding durability and corrosion
protection. Epoxy coated strands, either with or without outer PE sheathing, provide an alternative to the typical bare or
galvanized, waxed and PE-sheathed stay cable strand types and are regulated in standard stay cable recommendations,
such as PTI DC45 and fib Bulletin 89. The robust and thick coating film of epoxy coated strands provides a very
durable, continuous, and reliable corrosion protection system without the need to peel off the coating in the anchorage
zone.
This paper presents the latest international applications of DYWIDAG’s epoxy-coated strand systems in large-scale
cable-stayed and extradosed-bridges. The growing demand for superior corrosion resistance has led to the increased
adoption of epoxy coated strands in bridge projects beyond Japan, where they have been used for many years.
Sumitomo Electric Industry (SEI) offers various epoxy coated strand variants with different sizes, mechanical
properties, and corrosion inhibiting layers. The anticorrosive properties of epoxy coated strands are caused by the high
adhesiveness and durability of the epoxy resin. Its anticorrosive attributes are further achieved by the complete filling of
voids between the wires with the epoxy resin. For an additional redundancy in corrosion protection also an additional
PE-sheathing can be added.
This superior corrosion protection is not just limited to cable-stayed or extradosed bridges but is also very advantageous
for pre-tensioned strands, ground anchors, internal bonded tendons as well as external unbonded tendons. Negative
experiences with cement grouting and flexible filler have led to an increased application of this technology also outside
of Japan, where it has already been frequently adapted for many years. To apply epoxy-coated strands at stay cable
projects some enhanced requirements on single strand testing and full-scale fatigue and tensile testing need to be
considered. The strand developed by SEI which is used at worldwide stay cable applications exhibits remarkable
relaxation and creep values and surpasses even galvanized strands in standard salt spray tests.
Axial fatigue and tensile full-size testing as well as leak tightness tests and fatigue bending tests of the DYWIDAG
DYNA Grip® anchorage system with epoxy coated strands yielded numerous excellent results, as presented in detail.
The performance of the combination with multitube stainless steel tubes in a frictional saddle configuration, known as
cradle, has been successfully tested, confirming adherence to frictional, fatigue, and strength requirements. Moreover,
the system has successfully passed fire resistance tests, making it applicable for projects with special requirements.
The paper highlights completed international bridge applications that have successfully employed the prescribed
DYWIDAG epoxy-coated stay cable system, some of which have been in use for several decades. Furthermore,
impressive ongoing stay cable bridge projects employing these advanced systems are also presented.
41
42
13th - Japanese-German Bridge Symposium, Osaka, Japan
Experimental Study on Slip Load and Clamping Force Relaxation of Frictional High Strength
Bolt Connection with Entire Corrosion
PhD Student Lingbo Yao ∗
Associate professor Ming Li ∗∗
Professor Takashi Yamaguchi ∗∗∗
∗
∗∗
Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, sw22522k@st.omu.ac.jp
Department of Civil Engineering, Suzhou University of Science and Technology, China, seuliming@foxmail.com
∗∗∗
Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, yamaguti-t@omu.ac.jp
Keywords: High strength bolt connection, Corrosion, Bolt clamping force, Shear resistance
1 Introduction
Frictional high strength bolt connections are widely used in steel structures such as steel bridges, buildings, and transmission tower for connecting components. The slip load of the frictional high strength bolt connection is positively related to
the bolt clamping forces and the slip coefficient.
With the service of structure in various environments, many severely corroded high strength bolts connections were
observed in steel structures with long term service like figure 1. Based on previous studies, the slip load of the frictional
high strength bolt connection is positively related to the bolt clamping forces and the slip coefficient. The corrosion can
introduce the damages to the bolt, which has the potential to alter clamping force, what’s more, the corrosion products
generated on the surface of the plate, introduce influences on the slip coefficient. consequentially, the steel corrosion of
the high strength bolt connection can have impacts on the slip load of the frictional high strength bolt connection. What’s
more, the variation in the clamping force of the bolts has rarely been precisely monitored in the steel corrosion process.
The influence of the decreased bolt clamping force, which may be induced by steel corrosion, on the frictional shear
strength of the high-strength bolt has not been fully investigated.
2 Experimental procedure and results
In this study, for the purpose of studying the influence of corrosion on the bearing capacity of the high strength bolt connection like figure 2, a set of experiments were conducted. Firstly, due to its fine core diameter and strong electromagnetic
interference capacity, the fiber bragg grating (FBG) sensor was used for measuring the clamping force. The calibration
experiment was conducted for each FBG sensor to ensure the reliability of the measured data of the high strength bolts.
Secondly, the clamping force decrease of each high strength bolt after finally screwed was monitored for quantitative
analysis of the clamping force decrease influence on the bearing capacity of the high strength connection. It is conducted
that the clamping force decreases rapidly in the first 2 hours, decreases slowly after 5 hours and the clamping force is
almost no change from the 7th day to the 21st day. The average and standard deviation of the final clamping forces of
bolts are 20.7 % and 11.4% respectively like figure 3.
Furthermore, experiments of accelerate steel corrosion and the tensile tests were conducted to obtain the slip load and slip
coefficient to study the influence of corrosion on the high strength bolt connection. From the experiments results, those
were shown the clamping force reduced by 1.69% to 4.55 % result from the entire corrosion in this research. However, the
slip loads increased up to 258 % and slip coefficient of the corroded specimen increased up to 184%, both were generally
lager than that of the uncorroded specimens like figure 4.
Figure 1: Corroded bolt connections of onsite steel
structures
Figure 2:
43
The corrosion states of specimen
Figure 3: Time history of bolt axial strain after final tight
Figure 4: The slip coefficients and slip loads of high strength bolt connection
3 Conclusions and discusses
Based on the experimental results and analysis, the conclusions can be drawn as following:
1. The average of the clamping force relaxation after final tightened is 20.7 %, exceeding which required by the specification, there has the possibility that when assemble the connection, the tiny gaps exist between the main plate and splice
plate, result in the excessive shrink of the bolt shank, lead to the much relaxation of high strength bolt connection.
2. With the limited corrosion of high strength bolt connection, in this study, the maximum corrosion is corresponding
to the 1.02 mm of thickness reduction of splice plate result from the corrosion, the clamping force reduced by 1.69% to
4.55 % result from the entire corrosion of the connections, however, both slip coefficient and slip load were significantly
enhanced because of corrosion. There is the possibility that the slip coefficient was enhanced by the mechanical interlocking force caused by the red rust. The red rust was stripped after the specimen was damaged which can be considered
as evidence for the red rust contribution to the friction coefficient.
3. Based on the experiment results, it can be conducted that the limited corrosion degree can reduce the clamping force
but improve the slip load and the slip coefficient, and making the slip load meets the specification required. However, in
the condition that with the higher corrosion degree of the frictional high strength bolt connection, the development of the
slip load and slip coefficient influenced by the corrosion is still unclear, what is more, the damage of the plate result from
the corrosion may have impact on the bearing strength which the net section of the plate have the potential to yield. Those
situations of the frictional high strength bolt connection should be investigated in further study.
4 References
[1] M. Tendo, K. Yamada, and Y. Shimura, “Stress Relaxation Behavior at High-Tension Bolted Connections of StainlessSteel Plates ,” Journal of Engineering Materials and Technology, vol. 123, pp. 198–202, 09 2000.
[2] M. Xia, Y. Wang, and S. Xu, “Study on surface characteristics and stochastic model of corroded steel in neutral salt
spray environment,” vol. 272, p. 121915, 02 2021.
[3] J. H. Ahn, J. M. Lee, J.-H. Cheung, and I.-T. Kim, “Clamping force loss of high-strength bolts as a result of bolt head
corrosion damage: Experimental research a,” Engineering Failure Analysis, vol. 59, pp. 509–525, 2016.
44
13th - Japanese-German Bridge Symposium, Osaka, Japan
SLIP TESTS OF DOUBLE-LAP JOINTS CONSISTING OF NON-PROJECTED AND
SANDGLASS-SHAPED BOLTS WITH HIGH STRENGTH AND DURABILITY
Graduate Student Masashi Takayama *
Senior Lecturer Hitoshi Moriyama **
Masayori Yoshimi ***
Prof. Takashi Yamaguchi ****
Assistant Prof. Gen Hayashi *****
* Tokushima University, Tokushima, Japan, c612331016@tokushima-u.ac.jp
** Tokushima University, Tokushima, Japan, moriyama.hitoshi@tokushima-u.ac.jp
*** Nippon Steel Bolten Corporation, Osaka, Japan, m.yoshimi@bolten.co.jp
**** Osaka Metropolitan University, Osaka, Japan, yamaguti-t@omu.ac.jp
***** Osaka Metropolitan University, Osaka, Japan, hayashi-g@omu.ac.jp
Abstract: Our research group has been developing a non-projected and sandglass-shaped bolt with high strength and
durability, known as the Double Spindle Fastener (DSF) comprising a countersunk head bolt and countersunk sleeve nut,
to enhance the anticorrosive performance of bolted in-situ connections. This study conducted slip tests of double-lap
joints with DSFs or hexagonal head bolts to confirm the difference in slip behaviour by the two bolt types. The after-slip
behaviour, like ultimate resistance and failure modes, was also confirmed. Obtained results indicated that the average slip
coefficient of five specimens was almost the same in two cases and more than design value 0.45 specified in the Japanese
code for bridge codes. As for the after-slip behaviour, the failure mode of the joint with DSF was bolt shear failure at the
unengaged threads or the first engaged thread from the countersunk head. Conversely, the countersunk sleeve nut was not
deformed by the nut-side splice’s bearing force. Consequently, the maximum load can be moderately estimated with the
effective shear resistance of the countersunk head bolt’s thread, not including the sleeve nut’s shear-resisting area.
Keywords: Slip tests, Friction-type bolted joints, High-strength countersunk head bolts, Countersunk sleeve nut
1 Slip tests of friction-type bolted joints with DSFs and hexagonal head bolts
Figure 1 shows a prototype of Double Spindle Fastener (hereafter called DSF) [1]. Two cases focus on bolt types: frictiontype double-lap joints with DSFs or high-strength hexagonal head bolts (hereafter called HEX). Figure 2 illustrates the
geometrical dimensions of specimens and measuring locations of the relative displacement between the connected plate
and splices. Parameters are bolt types and DSF’s tension. DSF’s designed tension is 150kN, corresponding to 74 per cent
of the yield resistance. Since DSF’s tightening capacity is required to be that of 8.8 Class-M22 bolts, HEX’s designed
tension is set to 165kN, the designed tension of 8.8 Class-M22 bolts. However, in this test, 10.9 Class-M22 bolts were
used instead of 8.8 Class bolts by adjusting the tension value because of material procurement. The introduced tension of
both types of bolts is 1.1 times the designed tension. Table 1 summarizes mechanical properties written in inspection
certificates, designed bolt tensions, the tension before test and obtained the number of specimens in all cases is five. In
the case of “DSF-135-F”, the after-slip behaviour was investigated. The faying surface treatment for every plate like
splice is inorganic zinc-rich paint after blast cleaning. The slip coefficient was evaluated with the tension before the test
considering influences of tension reduction behaviour by
applied load, and calculated by eq. (1).
𝜇1 =
𝑃𝑠𝑙𝑖𝑝
𝑛𝑚𝑁1
(1)
Here, Pslip is slip load, n is the number of bolts, m is the
number of faying surfaces, N1 is the average bolt tension
before the test.
Figure 1: Double Spindle Fastener (DSF)
(b) Case: HEX
(a) Case: DSF
Figure 2: Geometrical dimensions and relative displacement’s measuring points of specimens (unit: mm)
45
500
400
160
300
0
0.00
0.05
0.10
0.15
δr (mm)
200
100
1
3
5
0
0.00
0.50
1.00
240
400
160
1.50
δr10 (mm)
0.20
300
0.00
0.05
0.10
0.15
δr (mm)
200
0
2.50 0.00
0.20
300
0
0.25
0.00
0.05
0.10
0.15
δr (mm)
0.20
0.25
200
100
100
2
4
δr=0.20
2.00
240
80
0
0.25
320
160
80
P (kN)
P (kN)
80
400
Avg. kr10
= 9,278 (kN/mm)
320
P (kN)
P (kN)
400
240
500
400
Avg. kr10
= 8,776 (kN/mm)
320
P (kN)
400
Avg. kr10
= 8,410 (kN/mm)
P (kN)
500
0.50
1.00
1.50
2.00
δr10 (mm)
0
2.50 0.00
0.50
1.00
1.50
2.00
δr10 (mm)
2.50
(b) DSF-135
(c) DSF-150
(a) HEX-165
Figure 3:Applied load P versus relative displacement at 10mm location from the connected plate’s end δr10 curves
2 Results and discussion
Table 1: Bolt’s mechanical properties
Figure 3 shows the relationship between the applied load P and the relative
and slip test results
Bolt's Bolt's Desiged Avg. bolt
Slip
Slip
Slope
Max.
displacement at 10mm from the connected plate’s end δr10. The P - δr10
yield tensile
bolt
tension
load coefficient
of
load
Experistrength strength tension brfore
evaluated P vs δ
curves demonstrate a linear relationship and included an obvious slippage
mental No.
the test
by N
curve
case
σ
σ
P
N
N
regardless of bolt type used. The slope before the slippage in cases of
k
P
μ
(MPa) (MPa) (kN)
(kN)
(kN)
(kN/mm) (kN)
“DSF-135” and “DSF-150” was found to be larger than that of “HEX-165”,
1
165.4
299.7
0.453
8,978
2
168.5
326.2
0.484
8,579
although HEX’s designed tension is the highest. The average μ1 of “DSFHEX3
1,035
1,078
165
171.6
291.1
0.424
8,944 N/A*
165
150” was 0.459, the same as that of “HEX-165” (= 0.488). “DSF-150” can
4
168.4
367.6
0.546
8,049
5
166.9
355.5
0.533
7,503
also secure the specified value of 0.45 in the Japanese design code for
1
147.0
258.1
0.439
9,447
highway bridges [2].
2
145.3
244.8
0.421
8,610
r 10
1
y
Figure 4 indicates the relationship between the applied load P and the test
machine’s displacement δ in the case of “DSF-135-F”. The failure mode
of the joint with DSF was bolt shear failure at the unengaged threads or the
first engaged thread from the countersunk head, as shown in Figure 5. In
contrast, the countersunk sleeve nut was not deformed by the nut-side
splice’s bearing force. The average maximum load of five specimens was
548kN. Therefore, the maximum load can be moderately estimated with
the effective shear-resisting area in two shear planes considered only for
the countersunk head bolt’s male thread part, not including the sleeve nut’s
shear-resisting area. Eq. (2) expresses the above assumption and expects
the ultimate resistance to be 542.4kN, almost the same as the average
maximum load confirmed in the test.
(2a)
𝜏𝑡 =
DSF135-F
d
3
150
1
slip
r 10
1
144.0
275.9
0.479
9,187
4
147.4
318.5
0.540
10,155
5
144.8
241.9
0.418
8,991
1
134.0
256.7
0.479
8,635
2
132.8
218.5
0.411
9,067
134.9
207.5
0.385
8,360
4
134.8
222.2
0.412
9,020
5
3
1,342
1,443
135
3 Conclusions
This study conducted slip tests of double-lap joints with DSFs or hexagonal
bolts to confirm their slip behaviours and DSF’s after-slip behaviour.
Obtained results are as follows;
(1) P - δr10 curves before the slippage became a linear relationship. The
slope in cases of DSF was also larger than that of HEX, although
HEX’s designed tension is the highest. The average slip coefficient μ1
of “DSF-150” was the same as that of “HEX-165”
(2) The failure mode of the joint with DSF was bolt shear failure. The
maximum load can be moderately estimated with the effective shearresisting area in two shear planes considered only for the countersunk
head bolt’s male thread part.
Acknowledgement
max
N/A*
N/A*
134.8
200.4
0.372
8,799
1
131.0
195.5
0.373
9,097
536.7
2
138.5
224.0
0.404
9,452
551.0
131.9
190.0
0.360
9,014
512.4
4
135.9
205.0
0.377
7,910
573.5
5
133.1
210.6
0.396
9,795
518.5
3
135
[NOTE] N/A*: Not applicable
1000
1
3
5
Pbod_2
800
2
4
Pbod_1
(2b)
√3
Here, σt is DSF’s tensile strength, τt is shear strength, Ae-m (= 162.8mm2) is
the effective area of the countersunk head bolt’s male thread part.
4
DSF135
P (kN)
𝑃𝑏𝑜𝑑_1 = 𝑛𝑚𝐴𝑒−𝑚 𝜏𝑡
𝜎𝑡
DSF150
t
Failure
points
600
Slippage at
fixed side
400
Slipping at
DSF side
200
0
0.0
4.0
8.0
12.0
δ (mm)
16.0
20.0
Figure 4: P versus δ curves
Shear plane
Failure
position
Shear plane
The JSPS KAKENHI, Grant Number JP20H02235, supported this work.
5
inner
References
[1]
T. Hashimoto et al. “Numerical exploration of a high strength and durability non-projected and
sandglass-shaped bolt for steel structures’ connection”, J. Steel Const. Eng. (JSSC), Vol.30,
No.118, pp.45-56, 2023.
[2]
Japan Road Association, “Specifications for highway bridges Part Ⅱ Steel Bridges”, 2017.
46
Bearing side
(toward joint center)
outer
Tensile side
Figure 5: DSF’s appearance after tests
13th - Japanese-German Bridge Symposium, Osaka, Japan
Noise reduction of modular expansion joints
on the example of the new Pattullo Bridge (CAN)
Dr.-Ing. Torsten Ebert *
Dr.-Ing. Christian Braun **
Dr.-Ing. Toshihisa Mano ***
*MAURER SE, Frankfurter Ring 193, 80807 München, Germany, t.ebert@maurer.eu
**MAURER SE, Frankfurter Ring 193, 80807 München, Germany, c.braun@maurer.eu
***MAURER SE, Frankfurter Ring 193, 80807 München, Germany, t.mano@maurer.eu
Abstract
MAURER MSM® Swivel Joist Expansion Joints with the capacity to carry out large movements in longitudinal and
transverse direction will be used at the new Pattullo Bridge in Canada. The expansion joints are in the main deck as well
as at the access ramps and have different characteristics due to the expansion length to be covered. Since the new Pattullo
Bridge is located in an inner-city area, the client required a noise reduction system which should reduce the over-rolling
traffic noise. A very effective system for reducing over-rolling traffic noise consists of welded rhombic elements on the
centre beams and sinusoidal plates on the edge beams. However, those noise-reducing elements severely restrict the
transverse movement that can be absorbed when the modular expansion joint is closing. The challenge was to develop a
new element geometry that makes this transverse movements in close and open situations of the expansion joints possible.
Keywords
modular expansion joint, joint, noise reduction, rhombic element
1 Introduction
The paper deals with the development process of the new rhombic element for noise reduction on lamella expansion
joints. In the second chapter the geographical location of the new Pattullo Bridge is discussed and the types of expansion
joints – located in the main deck and at the access ramps – are presented. The movements to be absorbed by the expansion
joints, which significantly influence their size, are presented in the third chapter. In chapter four, the normatively defined
penetration bodies are presented, which must be used to demonstrate the safety for vehicles and cyclists to pass over as
well as the walkable for pedestrians. Taking these specifications into account, a standard rhombus as well as a rhombus
for use in seismic areas was developed. In chapter five, the geometry of the rhombic elements is discussed, and the
movement capacity tests required as part of the approval procedure are presented.
2 Construction site of Pattullo Bridge
The Pattullo Bridge is one of the main bridges to connect the communities of Surrey and New Westminster near
Vancouver. The bridge gets the name from Thomas Dufferin Pattullo. He was a famous Politician and journalist in British
Columbia during the 30th and 40th years of the last century. The old bridge construction should be replaced through a new
four-lane toll-free bridge for vehicles as well as for cyclists and pedestrians. Both are separated from the traffic lane to
achieve a high safety standard. The new bridge is scheduled to open in 2024 and will be built to allow for potential future
expansion to six lanes.
3 MAURER modular expansion joint for Pattullo Bridge
MAURER modular expansion joints will be located on the ramps and the main bridge. The required movement capacity
is composed of the thermal movements (ULS) and the 2475-year seismic movement. In case of largest seismic opening
movements, the gaps between the centre beams may open more than 150 mm and the support bar lengths are designed to
handle these movements. For large seismic closing movements which results in the joint fully closing the longitudinal
fuse will be released to prevent the expansion joint from being crushed.
4 European and German standards for noise reduction elements
In Germany exists different standards, which regulated the use of noise reduction elements. ZTV-ING [4] includes the
material specification and contains the reference to TL/TP FÜ [3] for construction rules. Steel support components should
have the grade J2 and 3.1 certificate acc. EN 10204 [2] for transability. There is no limit to the steel strength, with S235
or S355 usually being used. The cleaning and maintenance of sealing profiles must be carried out without removing
structural parts, such as elements for noise reduction.
The gaps and voids surfaces check should be carried out for the ultimate limit displacements. Depending on the road user
(car, bike, pedestrian) there are different penetrators whose dimensions are defined in [1] and [3]. With the help of these
penetration bodies should be carry out an over-rolling safety check. Local deformations within the deformations from
47
wheel loads within the roadway transition do not have to be considered. In case that the noise reduction elements are
located on the top of the lamella of a modular expansion joints the distance between the adjacent elements may not exceed
a maximum of 100 mm and there must not be a continuous gap line parallel to the lamella axis.
5 MAURER noise reduction elements
5.1 MAURER Standard-Rhombus for noise reduction
Expansion joints represent a discontinuity in the roadway, which leads to driving noise when driving over it. It can be
reduced by sinuous elements on the top surface. For this purpose, MAURER uses rhombuses welded onto the centre
profiles at intervals of 150 mm with the ground plan dimensions 190 × 131 mm. The elements can be used for gap widths
from 0 to 100 mm in longitudinal direction, which corresponds to the maximum permissible gap in the Ultimate Limited
State (ULS) load case acc. EAD [1] and TL/TP FÜ [3].
5.2 MAURER Pattullo-Rhombus for noise reduction in earthquake zones
The new Pattullo Bridge is located in an earthquake zone. During an earthquake event the expansion joints must absorb
movements in all directions. The Standard-Rhombus can hardly absorb transversal movements when the joint gap is
closed. For this reason, a new geometry of the noise reduction elements had to be developed to allow greater movements
in the Y-direction without damaging the joint construction.
The approval procedure was carried out on a model on 1:1 scale with three strips in a width of 90 mm and a height of 33
mm to represent the adjacent centre beams too. Sheet metal strips of 3 mm thickness and different widths were placed
between them to reproduce the respective gap width. The two outer centre beams were provided with two rhombuses, the
centre beam, fixed on a clamping plate, with three rhombuses.
Compared to the standard version, the new rhombus allows an increase of the transverse movement capacity of ±30 mm
in the closed state of the expansion joint and can be used up to a maximum opening gap of 80 mm. The transverse
movement capacities of the expansion joint per individual gap can be determined from the specifications for traffic safety
as well as the limitation by contact of the components. The rhombuses are welded onto the centre profiles at intervals of
185 mm with the ground plan dimensions 190 × 169 mm. In addition to the new rhombus geometry the sinuous plates for
the edge profiles must be adaptive. For this challenge the distance between the trident would be done larger.
6 Conclusion
The new Pattullo Bridge is in an inner-city area, which is why the client had invited expansion joints with noise reduction
elements. As the bridge is furthermore located in an earthquake zone and in the event of an earthquake large transverse
movements must be absorbed by the expansion joint without collision risk. That’s why a modification of the geometry of
the MAURER standard rhombus elements had become necessary. Despite the welded-on elements for noise reduction,
the joint had to allow collision-free transverse displacement even when closed. For this purpose, the distance between the
rhombuses was increased compared to the standard design and, in addition, the base area was changed by laterally
protruding wings in the plane of the centre beam. This ensures that the increased gap width could be compensate and
guarantees a safety over-rolling for all bridge users.
7 References
[1]
EAD 120113-00-0107 (2019-08). Modular expansion joints for road bridges, EOAT.
[2]
EN 10204 (2005-01) metallic products – types of inspection documents.
[3]
TL/TP-ING (2021-03) Technische Lieferbedingungen und Technische Prüfvorschriften für Ingenieurbauten,
Teil 8 Abs. 1 Technische Lieferbedingungen und Technische Prüfvorschriften für Fahrbahnübergänge TL/TP FÜ,
Bundesanstalt für Straßenwesen.
[4]
ZTV-ING (2022-01) Zusätzliche Technische Vertragsbedingungen und Richtlinien für Ingenieurbauwerke,
Bundesministerium für Digitalisierung und Verkehr.
48
13th - Japanese-German Bridge Symposium, Osaka, Japan
Study on Load-Carrying Capacity of Built-up Column Lost Lacing Bars Focused on
Buckling Mode
Kenta Morimoto *
Toshikazu Takai **
Takao Miyoshi ***
Kaname Iwatsubo ****
Kazuya Tamada *****
* Kyushu Institute of Technology, Kitakyushu, Japan
** Kyushu Institute of Technology, Kitakyushu, Japan, takai@civil.kyutech.ac.jp
*** National Institute of Technology (KOSEN), Akashi College, Akashi, Japan
**** National Institute of Technology (KOSEN), Kumamoto College, Yatsushiro, Japan
***** National Institute of Technology (KOSEN), Maizuru College, Maizuru, Japan
A built-up column was often used in old bridges. The column consists of shaped steels connected by lacing bars and tie
plates. And these components are usually tightened by rivets. Recently, the column lost its lacing bars due to corrosion is
found in some cases. The detail of the mechanical behavior of such built-up columns lost some lacing bars is uncleared.
The findings and knowledge of the behavior of the columns are useful for maintaining the old bridges. In this study, finite
element analysis was conducted to investigate the mechanical behavior and load-carrying capacity of the built-up columns
that lost their lacing bars.
First, buckling eigenvalue analysis was conducted to clear the buckling mode and direction of the built-up column in
some cases where the columns lost lacing bars. And next, load-bearing analysis was carried out to investigate the
mechanical behavior of the column in compression. In the latter analysis, the components included rivets meshed exactly
by solid elements to obtain detailed behavior.
The loss of the lacing bar reduced the maximum load. In case of the buckling direction is normal to the plane given by
lacing bars, the loss of lacing bars on the compression side reduced the load more than that on the tensile side. The
evaluation in case the lacing bars are lost on the compression side gives more safety results. The maximum loads in cases
where the initial deflection in the buckling direction or both directions were the same. The giving initial deflection in both
directions is useful in case the buckling direction is unknown. Although the case where few lacing bars were lost reduced
the maximum load slightly, the case where lacing bars and rivets were lost obviously reduced the load. Whether the rivets
are lost is important to estimate load-carrying capacity.
Figure 1: von Mises stress distribution at the maximum load (Deformation scaling: 40)
49
50
13th - Japanese-German Bridge Symposium, Osaka, Japan
Numerical Study of Stiffened Plates Joined by Thermal Spraying
M.Eng. Eitaro Horisawa *
Ph.D. Kunitomo Sugiura **
Ph.D. Yasuo Kitane ***
Ph.D. Yoshinao Goi ****
* Kyoto University, Kyoto, Japan, horisawa.eitaro.44w@st.kyoto-u.ac.jp
** Kyoto University, Kyoto, Japan, sugiura.kunitomo.4n@kyoto-u.ac.jp
*** Kyoto University, Kyoto, Japan, kitane.yasuo.2x@kyoto-u.ac.jp
**** Kyoto University, Kyoto, Japan, goi.yoshinao.2r@kyoto-u.ac.jp
Abstract:
Thermal spraying is one method to form coatings on the surface of machine parts and steel structures. This method is
expected to be applied to repairing, reinforcing, and joining materials such as stainless steel because the heat input to the
base material can be kept below 100°C. As a fundamental study, this paper shows the compressive behaviour of plates
with a stiffener joined by thermal spraying, which was investigated by finite element analysis. In the analytical model,
debonding between the sprayed metal and the plates was considered. Parametric study was performed by varying the
bond strength, dimensions of the stiffener, and the magnitude of geometric imperfection. The maximum compressive load
of the stiffened plates bonded by thermal spraying was more than 95% of the welded stiffened plates when the initial
deflection was sufficiently small, and the plates were relatively stocky. On the other hand, as the initial deflection and the
stiffness ratio of the stiffener increased, the maximum compressive load of the stiffened plates decreased.
Keywords: thermal spraying, stainless steel, stiffened plates, finite element method
1 Introduction
Thermal spraying is a technology to form ceramic and metal coatings on a substrate by spraying and depositing the melted
materials (Figure 1). There are few restrictions on the spray and base material, and the technology has been used for
various applications such as corrosion prevention and improvement of fatigue strength of welds in bridge engineering.
One of the advantages of thermal spraying is small heat input to the substrate during the formation of coatings. When the
metal particles are sprayed, the thermally sprayed metal is heated to its melting point, but the substrate material does not
need to be heated. Therefore, joining stainless steel which is difficult to weld due to its large thermal deformation, becomes
easier with thermal spraying, that requires little heat input to the base material.
To extend the service life of structures by promoting the use of stainless steel, a joining method using thermal spraying
was proposed by the authors. The method is to form metal deposits by thermal spraying at the corner when attaching a
plate at right angles to the base plate, thereby joining the two together. In this study, prior to the experimental investigation
of members fabricated by thermal spraying, the mechanical behaviour of the stiffened plates joined by thermal spraying
was investigated by numerical calculation.
2 Methodology
The geometry and boundary conditions of the finite element model of the stiffened plate are shown in Figure 2. The
numerical model reproduces half length of the stiffened plate and assumes a centrally plane-symmetric boundary
condition. The panel plate and the stiffener are made of stainless steel SUS304 and are joined by thermally sprayed metal
of stainless steel SUS420J2. The mechanical properties of these materials obtained from experiments are listed in Table
1 [1, 2]. The dimensions of the stiffned plates were determined by slenderness parameters R and stiffness ratio of stiffeners
described in JSHB [3]. Other than these parameters, bond strength and the magnitude of geometric initial imperfection
Figure 1 Thermal spraying
Figure 2 Symmetric model of stiffened plate
51
SUS304
SUS420J2
Plates
Sprayed metal
Table 1 Mechanical properties [1, 2]
Elasitc limit
Young’s modulus
(GPa)
(MPa)
196
190
65
311
Yeild stress
(MPa)
308
587
Poisson‘s ratio
0.28
0.15
C = 10
0.8
C=5
0.6
0.4
R = 0.5
S=1
D = 1.0
0.2
0
0
1
2
3
Compressive displacement u/uy
4
Figure 3 Load-displacement curves
1.2
S1-C5
S2-C5
1
S1-C10
S2-C10
0.8
0.6
0.4
0.2
0
0.1
0.5
1
Scale of initial deflection D
Figure 4 Ultimate load (R=0.5)
1.2
S1-C5
S2-C5
1
S1-C10
S2-C10
0.8
0.6
No convergence
Tie
Max. compressive load Pmax/Pmax-tie
1
Max. compressive load P max/Pmax-tie
Compressive load P/Py
1.2
0.4
0.2
0
0.1
0.5
1
Scale of initial deflection D
Figure 5 Ultimate load (R=1.0)
were varied as numerical parameters. Cohesive behavior was introduced on boundary surfaces between plates and
deposits to represent the debonding of sprayed metal. The secondary stress criterion and a linear damage evolution law
were used in the cohesive behaviour, and an effective separation amount of 0.001 mm after damage occurrence and 0.001
of viscosity coefficient were introduced. In this study, normal bond strength of 5 and 10 MPa (shear bond strength is five
times higher) were used to investigate the effect of bond strength on the compressive behavour. The above numerical
calculations were performed using the commercial finite element analysis software Abaqus/standard 2020.
3 Result and Discussion
The relationships between compressive load and compressive displacement of stiffened plates joined by thermal spraying
are shown in Figure 3. The dotted line with a note Tie in the figure shows the result of the model with tied joint surface
between the panels and the sprayed metal. It can be confirmed that the stiffened plates exhibited the same curves as Tie
model until they reached the maximum load. Subsequently, the smaller the bond strength, the smaller compressive
displacement is required for the drop of load. After the load was decreased, the convergence of the numerical calculations
deteriorated, and almost no subsequent curves were obtained after those described. The above trends were also observed
for varying slenderness parameters and the magnitude of initial imperfection.
Figures 4 and 5 summarize the maximum load on the stiffened plates Pmax normalised by the maximum load on the Tie
model Pmax-tie. The load carrying capacity of all stiffened plates is less than Tie model, and the decrease in bond strength
causes a decrease in the capacity. Also, the maximum load tends to decrease as the stiffness ratio S, the magnitude of
initial imperfection D, and slenderness parameter R increase. On the other hand, all stiffened plates with R=0.5, D=0.1
have a maximum load greater than 95% of the one of Tie model. It is expected that the out-plane deformation of the
stiffened plates accelerates debonding of sprayed metal and reduces the maximum load, especially when the relative
deformation between the panel plate and the stiffener increases due to the increase in stiffness ratio.
4 Conclusion
A new joining method by thermal spraying was considered as a mean to avoid the welding difficulties of stainless steel
in this study. Numerical simulations revealed how stiffened plates joined by thermal spraying behave when subjected to
compressive load. The results showed that the out-plane deformation of the stiffened plates caused debonding of the
sprayed metal, which led to local buckling of the stiffener and the decrease in load-carrying capacity.
References
[1] Horisawa, E., Sugiura, K., Yasuo, K.: Study on Fatigue Strength of Lean Duplex Stainless Steel Base Metal, Steel
Construction Engineering, Vol.28, Issue 111, 107-117, 2021.
[2] Horisawa, E., Sugiura, K., Kitane, et al.: Mechanical Properties of Stainless Steel Coatings Formed by Build-up
Spraying, Proceedings of the Eighth International Conference on Structural Engineering, Mechanics and
Computation, pp.1385-1390, 2022.
[3] Japan Road Association: Specification for highway bridges: Part-II steel bridges., Maruzen Publishing, 2017.
52
SESSION 2-A
Design Codes and Bridge Engineering 2
13th - Japanese-German Bridge Symposium, Osaka, Japan
LT-Bridge – A new and fast construction method for cost-efficient bridge structures
Dipl.-Ing. Franz Untermarzoner ∗
Dipl.-Ing. Michael Rath ∗
Univ.-Prof. Dr.-Ing. Johann Kollegger ∗
∗
Technische Universität Wien, Institute of Structural Engineering, Vienna, Austria
franz.untermarzoner@tuwien.ac.at; michael.rath@tuwien.ac.at; johann.kollegger@tuwien.ac.at
Keywords: LT-bridge; Post-tensioned bridge; Thin-walled; Precast deck slab element; Balanced lowering method;
Precast girder
1 The concept of the LT-Bridge construction method
1.1 Idea
Since good experience has been gained with thin-walled precast elements, a new construction method for plate-girder
bridges was developed at TU Wien. This construction method unites the advantages of a fast construction progress for
the whole bridge structure with the quality of a bridge, which is cast using in-situ concrete.
Two types of precast elements have been designed. First, a precast girder corresponding to the span length is required.
This girder spans in the longitudinal direction of the bridge (L). The other type of precast element is a deck element placed
on the already installed longitudinal girder. This element spans in the transverse direction of the bridge as built (T). The
bridge construction method was named after the span directions of the elements – LT Bridge Construction Method.
The new construction method includes conceptual considerations of the precast segmental span-by-span erection method
(with transverse joints between the segments) and the method for precast girder erection by launching gantry (with girders
oriented in the longitudinal direction). With this new method, a weekly cycle for producing a construction section with
a length corresponding to the span can be achieved. This fast construction progress is possible because only limited
reinforcement laying work must be done at the installation site. For the placing of the precast girders, a launching gantry
was developed. It can move the precast girders to their final position. Moreover, the launching gantry is designed in such
a way that it can also move the precast elements for the deck slab to their final position.
1.2 Description of the precast elements
1.2.1 Precast girder
The precast girder consists of several hollow box segments connected on the finished deck slab or at an assembly area.
The number of segments depends on the span length of the bridge, the maximum possible transportation length, and the
allowable transport weight in the respective country. For spans up to 30 m, the longitudinal girder can be cast from one
piece. In general, the precast element sizes depend on the lifting capacities available in the precast plants, which are
limited to 80–100 t in Austria, for example. An example of a precast girder is shown in Fig. 1. In this case, the girder
consists of three hollow box segments. The segments are connected with an additional in-situ concrete layer on the bottom
plate and bonded tendons. Before the segments are clamped together, the joints must be grouted.
1.2.2 Precast element for the deck slab
The precast elements for the deck slab consist of plates, which are connected by cross-beams. Fig. 1 shows a precast
element for the deck slab, which already has the entire width of the bridge in the final state. Three plates are connected by
two reinforced cross-beams. The element has upstands at the cantilever ends to create a vertical formwork for the in-situ
concrete layer. To create a sufficient bond between the in-situ and the precast concrete, the surface of the plates must have
a certain roughness. The deck slab element is dimensioned so that it can support its dead load, an additional concrete
Figure 1: Precast elements for the LT-Bridge construction method; left: Precast girder; right: Precast deck slab element
55
layer, and a construction crew. Therefore, the element acts as a formwork that can be fully included in the structural
analysis. The connection between the precast girder and the precast elements for the deck slab is achieved with the shear
reinforcement, which protrudes from the top surface of the hollow box segments, and the reinforcement necessary for
shear between the web and flanges. Almost all the reinforcement required for the final state is contained in the precast
elements. Only the splice reinforcement, the upper longitudinal, and the upper transverse reinforcement must be installed
at the construction site.
1.3 Construction possibilities for the assembly of the precast components
For the erection of an LT Bridge, the following construction methods can be used:
• Span-by-span erection with a launching gantry, or
• Span-by-span erection with cranes and an auxiliary steel truss.
Depending on the given boundary conditions, the erection with a launching gantry or the erection with cranes can be
more efficient and economical. Below is a non-exhaustive list of decision parameters for which construction method it is
sensible to choose:
• Length and number of the spans,
• total length of the bridge,
• topographical conditions, and
• weight and dimensions of the precast elements to be lifted.
2 Conclusion
The LT bridge construction method can be regarded as a combination of the construction method with precast girders
laid side by side (”Precast girder erection by launching gantry”) and the span-by-span construction method (”Precast
segmental span-by-span erection by launching gantry”). In the LT bridge construction method, two longitudinal girders
with hollow box-shaped cross-sections are arranged in the longitudinal direction (L) to construct a structural section of
a multi-span prestressed concrete bridge. Deck slab elements are placed on the longitudinal girders in the transverse
direction (T). The longitudinal girders and the deck slab elements are connected by shear reinforcement and a top layer
of concrete. The range of application of the LT bridge construction method is for spans between 30 m and 60 m. With the
construction method presented here, building one span of a multi-span pre-stressed concrete bridge will be possible per
week.
Fewer building materials are consumed since a large part of the webs remains hollow. Because material is saved, the
new construction method is superior to segmental construction concerning sustainability. During the development of the
construction method, the material was removed in those places where it was not statically necessary. In the example of the
Pinkabach bridge, the longitudinal girder was designed as a hollow box girder, which led to material savings compared to
the initial design.
56
13th - Japanese-German Bridge Symposium, Osaka, Japan
Fast erection of deck slabs for steel-concrete-composite bridges
Prof. Dr.-Ing. Johann KOLLEGGER *
Dipl.-Ing. Franz UNTERMARZONER **
Dipl.-Ing. Michael RATH ***
* Institute of Structural Engineering, TU Wien, Vienna, Austria, johann.kollegger@tuwien.ac.at
** Institute of Structural Engineering, TU Wien, Vienna, Austria, franz.untermarzoner@tuwien.ac.at
*** Institute of Structural Engineering, TU Wien, Vienna, Austria, michael.rath@tuwien.ac.at
Abstract
Different methods have been developed to produce deck slabs of large steel-concrete composite bridges. Often
such deck slabs are cast with in-situ concrete with the aid of a formwork carriage. The formwork of the deck slab
is supported by the formwork carriage, which can be moved longitudinally along the bridge. Placing of the
reinforcement takes place at the installation site resulting in a construction progress of 15 m to 25 m of the deck
slab per week. To speed up the construction, construction methods with precast partial- and full-depth elements
have been employed. Usually, a crane is needed for the placement of the prefabricated elements. A significant
disadvantage of the construction with precast elements is the fact that a supporting steel structure has to be
provided. Cantilevering transverse steel girders have to be installed in the bridge deck, whose sole purpose is to
support the precast elements.
A new construction method was developed at TU Wien, which enables a very fast erection of the concrete deck
slab and which avoids the installation of additional transverse steel girders. In this method, precast deck slab
elements with reinforced concrete cross-beams are employed. The precast deck slab elements are placed on top of
the steel girders using cranes or a novel transportation carriage. The transportation carriage consists of two parts,
which are connected by two longitudinal trusses, and can be moved longitudinally along the bridge. To finish the
deck slab, a layer of in-situ concrete is applied on top of the precast deck slab elements at the construction site.
The new construction method will make the production of 15 m to 30 m of the deck slab section per day possible.
Keywords: Steel-concrete-composite bridge, deck slab, precast element, transportation carriage
1 Precast deck slab element
In order to speed up the construction process for the erection of the deck slab of steel-concrete-composite bridge
a deck slab element was invented. The precast element is shown in Fig. 1. For clarity, the reinforcement already
included is not drawn. The elements consist of thin plates stiffened by one to two cross beams. The load capacity
of the deck slab element is dimensioned so it can carry its dead load, the cast-in-place concrete, and a construction
crew, which can use this element as a working platform. A comparison of the erection of a deck slab with a
formwork carriage and a precast deck slab element is shown in Fig. 1.
Figure 1 – Erection of a deck slab with a formwork carriage (top) and deck slab element for the erection of the
deck slab of a steel-concrete-composite bridge (bottom)
57
The formwork carriage hast two functions:
(i) It provides the formwork for a section of the deck slab.
(ii) It supports the formwork and carries the dead loads of the formwork and the fresh concrete as well as the
weight of the workers who are positioned on the formwork.
Both of these functions can be fulfilled by the precast deck slab element. The shape of the deck slab element
matches the underside of the deck slab of the bridge, therefore no formwork in required at the construction site.
The cross beams of the deck slab element are dimensioned in such a way, that no further support structure is needed
during the pouring of the in-situ concrete layer. The deck slab elements can be placed on top of the steel structure
with the aid of a crane or with a transportation carriage.
2 Concept of the new construction technique
At an assembly area next to the construction site, several precast deck slab elements are connected. After that, this
construction section can be moved to the installation site with a transportation carriage. Subsequently, the
transportation carriage can be moved back to the assembly area and can bring a new construction section to the
installation site. At the same time, the reinforcement laying works can start. Only the top longitudinal
reinforcement and the splice reinforcement for the bottom longitudinal reinforcement must be placed on site. The
top and bottom transverse reinforcement, the bottom longitudinal reinforcement, and the shear reinforcement are
already installed in the precast deck slab elements. After the reinforcement is installed, a first concrete layer can
be poured to create a partial bond between the steel structure and the thin precast elements. This step creates a
cross-section in which the thin plates already participate in the load-bearing behavior for the heavy in-situ concrete
layer.
The transportation carriage consists of a front part and a rear part, which are connected by a longitudinal truss. For
the longitudinal movement of the transportation along the bridge roller supports are used. The same roller supports
as for the movement of a formwork carriage are used (see Fig.1).
3 Concluding remarks
This novel method combines fast a construction progress with an excellent quality of the deck slab, with no
continuous joints over the height of the cross-section. Using this method, up to eight deck slab elements can be
installed daily. This corresponds to a construction progress of approximately 30 m per day. The placing of
additional reinforcement at the construction site and casting of the top concrete layer is independent of the
installation of the deck slab elements. Furthermore, no formwork is needed on site. Thus no formwork stripping
time must be observed, which is usually on the critical path.
The placement of a deck slab element for a small, single track railway bridge in Austria is show in in Fig. 2.
Figure 2 – Placement of a deck slab element for a railway bridge in Austria
58
13th - Japanese-German Bridge Symposium, Osaka, Japan
Shin-Meishin Expressway where Bridge construction progresses
Master of Civil Engineering. Yoshinori Wada*
Ph.D of Civil Engineering. Joon-Ho Choi**
Master of Civil Engineering. Takafumi Omura***
Master of Civil Engineering. Shinya Hiraoka****
Master of Civil Engineering. Shinya Maehara*****
Master of Civil Engineering. Moeka Tokutsu******
Bachelor of Civil Engineering. Masafumi Udo*******
* West Nippon Expressway Co., Ltd., Manager, japan, y.wada.ab@w-nexco.co.jp
** West Nippon Expressway Co., Ltd., Deputy Manager, japan, j.choi.aa@w-nexco.co.jp
*** West Nippon Expressway Co., Ltd., Chief, japan, t.omura.ab@w-nexco.co.jp
**** West Nippon Expressway Co., Ltd., Chief, japan, s.hiraoka.aa@w-nexco.co.jp
***** West Nippon Expressway Co., Ltd., Chief, japan, s.maehara.aa@w-nexco.co.jp
****** West Nippon Expressway Co., Ltd., Chief, japan, m.tokutsu.aa@w-nexco.co.jp
******* West Nippon Expressway Co., Ltd., Chief, japan, m.udo.aa@w-nexco.co.jp
1 Introduction
The E1A Shin-Meishin Expressway (hereinafter referred to as Shin-Meishin) is 174km expressway connecting Nagoya
to Kobe.
NEXCO West has jurisdiction over the area west of the Koka Tuchiyama IC. The following section was opened: Kusatsu
JCT - Kusatsu Tanakami IC in March 2005 (initially as the E1 Meishin Expressway (hereinafter referred to as Meishin)
IC), Kusatsu Tanakami IC - Kameyama JCT in February 2008, Joyo JCT/IC - Yawata Kyotanabe JCT/IC in April 2009,
Takatsuki JCT - Kusatsu Tanakami IC - Kameyama JCT in February 2008, Joyo JCT/IC - Yawata Kyotanabe JCT/IC in
April 2009, Takatsuki JCT/IC - Kawanishi IC in December 2009, and Kawanishi IC - Kobe JCT in March 2008.
Currently, construction is underway on approximately 36 km of Otsu - Takatsuki JCT (Figure 1) and a six-lane project
on approximately 29 km between Koka Tsuchiyama IC and Otsu JCT (tentative name).
The 25.1 km section between Otsu JCT (tentative name) and Joyo JCT, which connects to the Keinawa Expressway
(hereafter, Otsu-Joyo), runs from Otsu City, Shiga Prefecture to Ujitawara Town and Joyo City, Kyoto Prefecture. The
expressway passes through steep mountainous areas, most of which are designated for erosion control, and crosses the
center of Joyo City via continuous viaducts, and also crosses the JR Nara Line and the Kintetsu Kyoto Line via bridges.
The expressway consists of 14.2 km of earthworks, 8.2 km of bridges, and 2.7 km of tunnels.
Otsu JCT
Takatsuki
JCT/IC
Kobe JCT
Yawata-Kyotanabe
JCT/IC
Joyo JCT/IC
Figure 1: Location map of Shin-Meishin (between Otsu JCT (tentative name) and Takatsuki JCT)
59
The 10.7 km section between Yawata Kyotanabe JCT and Takatsuki JCT, which connects to the Second Keihan Road,
runs from Yawata City, Kyoto Prefecture to Hirakata City and Takatsuki City, Osaka Prefecture. After passing a series
of factories and residences through a tunnel, the expressway crosses over the Keihan Railway main line, the Yodo River,
the National Route 171, the JR Tokaido Shinkansen, the Hankyu Kyoto Line, the JR Tokaido Line, the Meishin Route on
both sides, and then goes through a tunnel in the mountains. The expressway has a high ratio of structures.
This paper reports on the characteristics of four representative bridges between Otsu and Takatsuki area, including the
effects of the Shin-Meishin maintenance.
2 Integration of ancient civil engineering technology with modern highway technology
The Tenjin River Bridge (tentative name) is a double layers RC arch bridge with a full-length arch spanning the Tenjin
River, a first-class river in the Shin-Meishin (Otsu - Joyo).
Based on the historical background of the bridge site and the concept of fusing ancient civil engineering technology with
modern highway technology, a two-tiered arch structure was adopted, with a full arch on the upper level and an open arch
on the lower level. Figure 2 shows the image of the completed bridge.
The bridge length is 552 m. The lower arch spanning is a 7 spans continuous RC open arch with side spans of 72 m and
center span of 84 m. The upper arch is a 46 spans continuous RC filled arch that allows the same pavement structure as a
general earthwork section by filling with soil.
Construction of the foundation, substructure and lower arch is underway at the site. Photo 1 shows the site conditions as
of the end of June 2023.
Figure 2: Image of the completed Tenjin River Bridge (tentative name)
Photo 1: Tenjin River Bridge (tentative name) site conditions (view toward Joyo JCT, taken on 2023.6)
3 Conclusion
This paper introduces four representative concrete bridges between the Otsu JCT (tentative name) to Takatsuki JCT, but
many other concrete and steel bridges are under construction in addition to those mentioned above. We hope that this
paper will be useful for the development of bridge planning adapted to a wide variety of local conditions.
4 References
[1] https://corp.w-nexco.co.jp/activity/branch/kansai/shinmeishin/outline/outline01/01/index.html
[2] Yoshinori. W., Masato. F.: Shin-Meishin Expressway where bridge construction progresses. CEMENT &
CONCRETE. No.899, pp18-25, 2022.1.
60
13th - Japanese-German Bridge Symposium, Osaka, Japan
The use of glass for bridges – basics, special questions, codes and application examples
Univ.-Prof. Dr.-Ing. Geralt Siebert*
* University of the Bundeswehr Munich, Faculty for civil engineering and environmental sciences, Institute and
laboratory for construction engineering, Chair for structural design and building physics, Neubiberg, Germany,
Geralt.Siebert@UniBw.de
Abstract:
The growing use of glass in structural engineering can not only be seen in the wide field of buildings and related
applications but also in the field of bridges. In addition to a possible use as walking surface for pedestrian bridges the
application of architectural glass is common for railings and transparent sound screens or tunnels for noise protection,
even for highway-bridges. In a first part, a short overview is given about the basics of glass design including special
questions related to brittleness of glass and measurements to nevertheless guarantee sufficient safety. A second part with
information about the present situation of regulations in Germany and Europe (Eurocode) follows; here special focus is
upon the mentioned applications in terms of bridge building. In a last part especially of the full paper and presentation,
several examples of projects with the use of glass including interesting questions are presented.
Keywords: load carrying glass elements, fracture consequences, residual strength, redundancy, safety concept
1 Introduction
Glass as a perfect example for a brittle material shows sudden failure without any yielding (like steel or aluminium) or
visible slow growing cracks (like timber or reinforced/prestressed concrete). To overcome the related challenges for
building a safe structure several strategies do exist on material side as well as on design side. Having these in mind, long
lasting, safe and redundant structures can be built. The following chapters have a short view on aspects of material as
well as codification. As bridges are – at least subconsciously – regarded to have higher safety requirements than e.g.,
residential houses, the use of glass in bridge building is still extraordinary to most engineers.
2 Glass – characteristics and basic design principles
Due to production procedure maximum sizes of glass panes have to be kept in mind. Usually, the production steps on
single glass panes following the production of raw material always mean a reduction in size, additive manufacturing or
welding single parts together is – at the moment – not possible. An additive production process (in thickness) is lamination
of single panes to a “package” of laminated glass, consisting of several glass panes and interlayer. In addition, combination
of several glass panes with edge spacer and seal form insulating glass units – which are usually not used in bridge building
and therefore not considered within this paper.
The behaviour of glass in terms of crack growth of always existing, mostly with naked eye not visible cracks until fracture
can be described using simple linear fracture mechanics. To improve overall strength, thermal prestressing with
compressive stress on outer surfaces and tensile stress in intact inner volume is done. The (thermally) added energy has
the effect of smaller glass fragments in case of fracture, up to less harmful small pieces of safety glass, see Fig. 1.
Figure 1: Crack growth due to fracture mechanics and principle of thermally prestressed glass (left),
example of broken safety glass: size approximately 350mm² and detail view with mm-scale (right)
After the stress due to loading (and loading can be “classical loads” like wind, snow and traffic – or impact, local
temperature, climatic loads) exceeds the resistance (i.e., maximum value of strength, which can be determined by fracture
61
mechanics) a breakage occurs, the glass is no longer intact. Due to the sudden failure of brittle material without any
warning, regarding safety of people different aspects or consequence scenarios have to be considered:
-
(direct) injury of person in direct contact with glass element e.g., falling against it
-
(indirect) injury of person hit by falling glass sherds
-
Residual strength or load carrying capacity of broken glass element as well as alternative load path (redundancy)
Tempered glass breaks into a huge number of small pieces in case of fracture, so it might be sufficient for the first aspect,
but for the second aspect only in few cases (and not e.g., if broken glass elements is inclined or installed at bigger height)
and for third aspect in even less installation situations (e.g. in vertical installation with all sided linear bearing and no
perpendicular loading). Wired glass is, by the way, not considered as material with a safe-breakage behaviour.
So often laminated safety glass is used. Here bigger sized glass fragments after breakage are superior, so annealed or heat
strengthened glass are preferred; for illustration of different behaviour see Fig. 2. The adhesion of glass sherds to interlayer
(foil, sheet, hardening liquid material…) and the deformation behaviour of the interlayer itself are the crucial points for
determining the fulfilment of the above-mentioned requirements. Quantification and numerical modelling of these
interlayer properties and behaviour of broken laminated glass under consideration of different influence factors like stress
level, time or temperature and even moisture is field of actual research, also at chair and laboratory of the author.
Figure 2: Breakage behaviour of steel ball tested canopy of different glass types (from left to right): (fully) tempered
(safety) glass only, laminated safety glass of (fully) tempered (safety) glass, laminated safety glass of heat
strengthened glass, laminated safety glass of (fully) tempered (safety) glass using a special (stiff) interlayer
3 Codes for glass and glass design
When considering standardization, a distinction must be made between standards for construction products and standards
for their application. The former is a European competence, uniform standards are defined against the background of a
free European internal market. Design (in sense of application of products) is within national competence, the planned
Eurocode is to define uniform procedures, the level of verification is defined individually by each country.
As production level is defined by European product standards, one has to deal with the existing product range – or define
special conditions if needed. The upper mentioned design-scenarios regarding consequences of possible breakage may be
part of considerations to fulfil the requirements of the design situations FLS (Fracture Limit State) or PFLS (Post Fracture
Limit State) implemented in CEN/TS 19100, the basis document for the future Eurocode for Design of Glass structures.
At the moment there is no code or standard specific for design of glass elements for bridges; so usually the glass design
standards made for buildings are applied, sometimes with additional or further considerations.
4 Application example
Due to restricted space, in the abstract only pictures of one bridge are presented: a bridge crossing the 34 railway tracks
at Munich main railway station, see Fig. 3. Almost vertical glass elements serve as transparent protection elements.
Figure 3: Arnulfparksteg with point fixed (almost) vertical laminated safety glass elements made of tempered glass
62
13th - Japanese-German Bridge Symposium, Osaka, Japan
Structural Design of Glass-Elements in Bridge-Constructions
M.Sc. Alexander Pauli *
Univ. Prof. Dr.-Ing. Geralt Siebert **
* University of the Bundeswehr Munich, Institute for Structural Design, Germany, alexander.pauli@unibw.de
** University of the Bundeswehr Munich, Institute for Structural Design, Germany, geralt.siebert@unibw.de
Abstract:
Aesthetics and transparency are becoming increasingly important in building construction, structural engineering, and
especially in bridge design. In order to meet these requirements, the use of structural glazing is becoming more and
more common. Whether as noise protection elements, fall-proof infill of bridge railings, glass floors, or glazing in the
area of tunnel portals, structural glazing is increasingly finding its way into structural bridge engineering. For all these
applications, increased requirements must be taken into account. Due to the brittle failure of glass, the use of laminated
safety glass (LSG) is indispensable. LSG consists of at least two panes of glass, bonded with a polymeric interlayer.
The interlayer provides a shear transfer between the panes when the glasses are in the intact state and limits the damage
when the glasses are broken. It ensures a certain residual load-bearing capacity and protects against dropping glass
fragments. Furthermore, LSG provides redundancy in comparison to monolithic glazing. However, polymers often
behave strongly nonlinear with respect to time and temperature. To represent this complicated behaviour, material
models of viscoelasticity are required. This makes the simulation of LSG a difficult task and still an open research topic.
In this work, LSG elements with two different interlayer materials, commonly used for glazing in bridge design, are
investigated by means of numerical simulations. Therefore, different loading durations and temperatures are considered.
The calculated glass stresses of the two different elements are then compared to each other in order to show the decisive
influence of the interlayer. Overall, this work gives an insight into the design of glass elements in bridge structures,
considering the time and temperature-dependent behaviour of laminated glass.
Keywords: Glass Design, Laminated Glass, Numerical Simulation
1 Introduction
Glass has become an indispensable transparent building material in building construction. It is used for transparent
façades, roofing and fall protection balustrades for instance. In addition, the use of glass in bridge construction, especially
for pedestrian bridges, is becoming increasingly interesting. For this application, it can be used, for example, as balustrade
or as a glass floor. Furthermore, the use as a bracing element is conceivable. For example, the Bach Long Bridge in
Vietnam's Moc Chau distric, which is 632 m long and runs across a 150-meter-deep valley, has a walkable glass floor.
The respective glass elements are constructed from 3 layers of 40 mm thick toughened SGG Diamond glass, made by the
company St.Gobain, and intermediate layers of SentryGlas® (SG), made by the company Kuraray. Another example is
the “Arnulfsteg” between Munich Pasing and Munich main station. It has a balustrade with infill fall protecting glass
elements. These consist of two layers of 10 mm thermally toughened safety glass and a 1.52 mm interlayer of
Polyvinylbutyral (PVB).
Figure 1: Arnulf-Steg (© Dr.-Ing. Tobias Herrmann)
Within this work, numerically simulations on two four-side supported laminated glass elements are carried out for
different loading scenarios in the intact state. Both models consist of two 6 mm thick layers of glass connected by a 1
mm thick interlayer. However, for the first model a PVB interlayer is considered and for the second model a SG
interlayer is considered. For both models the glass stresses with respect to loading duration and temperature are
evaluated, compared to each other, and discussed.
63
2 Numerical Model
The numerical simulations are carried out with the commercial software Ansys (version 2023 R1). The laminates are line
supported along the bottom edges, subjected to a surface load perpendicular to the plane. The glass layers as well as the
interlayers are constructed with volume elements of the type “Solid186” [1], fully bonded at the interface between glass
and interlayer (cf. エラー! 参照元が見つかりません。). To save computation time only a fourth of the pane is modeled
by means of symmetry considerations.
Figure 2: Numerical Model of Laminated Glass
The glasses are modeled linear elastic, the interlayers linear viscoelastic by means of a Prony series with respect to a
reference temperature (Eq. (1)). However, to account for different temperatures, a Williams-Landel-Ferry (WLF) TimeSuperposition-Principle (TTSP) is added to the Prony series (Eq. (2)).
𝜎 = ∫ 2𝐺(𝑡 − 𝑠)
σ
ε
𝑑𝑠 + 𝐼 ∫ 𝐾(𝑡 − 𝑠)
𝐺(𝑡 − 𝑠) = 𝐺 𝑔 + ∑
𝑔 𝑒𝑥𝑝 −
𝐾(𝑡 − 𝑠) = 𝐾 𝑔 + ∑
𝑔 𝑒𝑥𝑝 −
𝑙𝑜𝑔 (𝛼 ) =
(
(
)
)
𝑑𝑠
(1)
(2)
Figure 3: Material Model of Viscoelasticity, [2]
Based on dynamic, mechanical, thermal analyses (DMTA), the respective parameters (depicted in Figure 3) for the
material models of the interlayers are evaluated by the use of an algorithm, developed at the Institute of Structural
Design of the University of the Bundeswehr Munich [3]. The parameters for the glass are taken from literature.
3 Results
Within the simulations, the maximum values of the glass stress in the first principle direction and the deflections
perpendicular to the plane are evaluated separately for the different loading durations and temperatures. The results for
both laminates are then compared. As expected, the glass stresses in the LSG with SG interlayer are much lower than
the ones of the LSG with PVB interlayer. However, for low temperatures and short durations, the difference is
considerably smaller than for high temperatures and long loading durations.
4 Conclusion, Outlook and Further Research
It could be shown how the glass stresses are directly influenced by the properties of the interlayer along with the loadduration and temperature. It is therefore a crucial step in the structural design of glass elements in bridge design to
evaluate the interlayer type in combination with different load scenarios. In addition to the investigation of the intact
state, the description of laminated safety glass in the broken state is also subject of current research. This topic is dealt
with by the authors as well.
5 References
[1] Ansys, Inc.: Element Reference, Release 2021 R1
[2] Ansys, Inc.: Material Reference, Release 2021 R1
[3] Kraus, M. A., and M. Niederwald: Generalized collocation method using Stiffness matrices in the context of the
Theory of Linear viscoelasticity (GUSTL). Technische Mechanik-European Journal of Engineering Mechanics
37.1, 2017
64
13th - Japanese-German Bridge Symposium, Osaka, Japan
The Arnulfpark Bridge – Glass as Contact Protection and Design Element
Dr.-Ing. Barbara Siebert *
Dr.-Ing. Tobias Herrmann **
* Dr. Siebert + Partner Consulting Engineers PartGmbB, Munich, Germany, bsi@ing-siebert.de
** Dr. Siebert + Partner Consulting Engineers PartGmbB, Munich, Germany, the@ing-siebert.de
Abstract:
The foot and bicycle bridge “Arnulfsteg” crosses all 37 railroad tracks between Hacker- and Donnersberger-Bridge in
Munich. The superstructure with its varying cross section made of steel was built by means of incremental launching
method in 2018 and 2019. The installation of the contact protection glazing was carried out subsequently, but before
destacking the superstructure to its final position. The strip-shaped printed glazing fills the rectangular openings between
top and bottom girder. It is linear and point supported and inclined towards the bridge axis. Therefore, in addition to the
requirements for impact resistance according to DIN 18008-4, the verification of the residual load-bearing capacity must
also be fulfilled. Due to the deviations from the technical rules, a special building permit was requested, which was
granted on the basis of pendelum impact tests and an expert's opinion. Along the connecting stairs to two S-Bahn platforms
and on an adjacent noise barrier, further large-format, point supported glazing was installed. In the static calculation, the
load case train pass-by had to be considered among other things.
Keywords: Footbridge, Glass, Point-Fittings, Contact Protection
1 The bridge
1.1 Location and function
The barrier-free new pedestrian and cycle bridge connects two new Munich city quarters. Over a length of 240 m, the
bridge crosses a total of 37 tracks, including the train-cleaning-system. Convenient usability was just as much a
requirement for the designers as the consideration of railroad concerns during construction and later during maintenance
and repair. Regular rail operations had to be ensured, and any necessary shutdowns had to be scheduled early enough.
1.2 Design and execution
The office Lang Hugger Rampp Architekten developed together with the structural engineer office SSF Engineers AG
the concept up to the execution design. At the same time the pivoting of the upper and lower girders leads to a constantly
changing shading of the surfaces. A consortium of the companies Mayerhofer and Stahlbau Plauen is responsible for the
execution. The construction of the bridge structure was subject to difficult boundary conditions: Work in the track area
only during the track closures, keeping the clearance with the overhead cables, limited space between the tracks, pile
foundations, complex geometries of the reinforced concrete ramps, etc. Together with the need for spans of 60 to 90 m,
the choice fell on the incremental launching method. Subsequently the glazing was installed and coatings repaired. After
that the superstructure was stacked to its final position.
Figure 1: Top view and elevation (tender documents)
2 The contact protection on the superstructure
2.1 Requirements
Since full-surface contact protection is required above the overhead cables, but at the same time a view of the tracks and
neighbouring buildings was desired, the transparent and at the same time safe laminated safety glass was chosen. In
addition, the glazing must be impact-resistant and resistant to residual stress.
2.2 Constructionand analysis
The glass-panes are held by continuous linear supports at the lower edges and by four hinged point fittings (diameter 80
mm) in the upper area. The structural safety and serviceability of the glazing were verified by finite element analysis in
accordance with DIN 18008-3. For the smaller pane sizes, laminated safety glass made of 2 × 8 mm fully tempered glass
was selected; for the larger panes, the static calculation resulted in laminated safety glass made of 2 × 10 mm fully
tempered glass. Due to the stress peaks at the glass holes, the screen printing was excluded here in order to be able to
apply the full strength of the fully tempered glass. The screen printing was provided for protection against bird strike. The
overhead cables under the bridge made it necessary to design the glazing as effective protection against contact in order
to protect people from electric shock. This means that it was necessary to ensure in the design that it was not possible to
65
reach through between the glass panes. This was achieved by positioning the glass-panes on the outside of the
superstructure cross-section, which thus covers the vertical pane joints and upper edges. It is not possible to reach through
between the steel structure and the lower edge of the glass, as the glass edge is directly framed in a line on the steel
structure. In this U-shaped welded steel profile, the glass-panes are placed on heavy-duty bearing blocks. EPDM seals
close the joints between the steel and the glass. Any water that nevertheless penetrates is drained away below the bottom
edge of the glass-panes through appropriate openings. The workshop design for the glazing is based on that of the steel
structure. However, the necessary superelevations of the steel structure had to be removed again for the glazing design.
Possibilities for tolerance compensation had to be provided. In this way, it was possible to avoid breakage of the panes
during assembly and stacking and to ensure a smooth continuation of the construction process.
2.3 Special permit
Towards the middle of the bridge, the openings between the upper and lower girder become larger, so that the lower girder
can no longer be regarded as a railing with sufficient height. Therefore, the glass-panes in these areas had to be designed
to be drop-resistant in accordance with Category A of DIN 18008-4. Since at the same time some of the panes slope more
than 10° inwards from the vertical, they are classified as horizontal (overhead)glazing according to the definition of DIN
18008-2. DIN 18008-3 on the use of point-fixed glazing requires the use of laminated safety glass with a large breakage
pattern for this application. For this reason and due to the fact that the top edge of the pane overhangs the point-fixing by
more than 30 cm (limit value according to DIN 18008-3 for horizontal glazing), a special approval of the responsible
building supervisory authority was required. A certified testing laboratory carried out pendulum impact tests on the
largest glass pane. For this purpose, the test specimen was mounted on a purpose-built, inclinable test frame. The
pendulum was dropped onto the glazing in a vertical orientation. All impacts were recorded without damage or permanent
deformation. Subsequently, the glass pane was inclined according to the later installation situation and both layers of the
laminated safety glass were hit with hammer and center punch. The residual load-bearing capacity was evaluated as
sufficient after a standing time of the construction of 24 hours.
2.4 Assembly
The glazing had to be installed before the superstructure was stacked in its final position on the bridge bearings, as there
would then not have been sufficient space above the overhead cables for the installation scaffold. The working scaffold,
which could be moved along the superstructure, together with a compact crane placed on the bridge, allowed convenient
and safe installation of the panes from outside during normal rail operations below. In order to be able to exclude any
forced stresses caused by the lowering of the steel structure for the glass panes, the bolts on the connecting plates of the
point supports were only tightened after stacking.
3 Conclusion
Glazing along pedestrian bridges must satisfy both design and safety aspects. Of course, this also applies to glazing in
building construction. However, the boundary conditions for an engineering structure - even more so in such a prominent
location - are considerably more demanding. In particular, the operational constraints require a high degree of design
accuracy, technical competence, coordination between the disciplines, and rapid response to unexpected situations during
execution.
Figure 2: the verification of the residual load capacity
Figure 3: Finished superstructure
Figure 4: Finished superstructure
4 Stakeholders and References see full paper
Figure 5: Finished superstructure
66
13th - Japanese-German Bridge Symposium, Osaka, Japan
Fiber Distribution Pattern Recognition in UHPFRC
Based on Deep Learning Technology
Doctoral Student Xin LUO*
Univ. Prof. Dr. Takashi Matsumoto**
*Graduate School of Engineering, Hokkaido University, Japan, xin.luo.a1@elms.hokudai.ac.jp
** Faculty of Public Policy, Hokkaido University, Japan, takashim@eng.hokudai.ac.jp
1 Introduction
UHPFRC's remarkable mechanical properties have led to its use in various structures. Key to its performance is fiber
distribution, enhancing behavior by reinforcing the matrix and improving strength and ductility. Fiber arrangement
influences overall performance, particularly bending resistance and load capacity. Proper distribution controls cracks
and stress. Established patterns ensure load transfer, reduce stress concentration, and minimize cracks, enhancing
strength, durability, and performance. To optimize, advanced techniques like image analysis and AI algorithms,
including YOLO series, address fiber distribution challenges. YOLO divides images into a grid, aiding detection based
on patterns. YOLO tech, especially YOLOv8, provides advantages in UHPFRC's fiber pattern recognition, removing
manual inspection, enabling rapid analysis of large datasets, and offering objective analysis. This study explores
YOLOv8 for recognizing UHPFRC fiber distribution. By training on extensive UHPFRC image dataset, it aims to
create an automated analysis tool. Results will advance UHPFRC understanding, optimizing fiber mixtures and
placements. This study provides insight for UHPFRC design and application, contributing to materials science through
deep learning application.
2 Materials and Methods
This research involved the use of nine beam shaped UHPFRC specimens, sourced from two batches. Each specimen
measured 240mm in length, 60mm in width, and 25mm in thickness. The specimens underwent a dual process of fourpoint bending tests and X-ray scanning. The purpose of the four-point bending tests was to ascertain the flexural loadcarrying capacity of the specimens, while the X-ray scanning was conducted to generate CT scan images containing
valuable fiber distribution data.
To create a foundational image dataset, a sequence of preprocessing steps was applied to the CT scan images. The fiber
distribution information extracted from these images was categorized into three distinct classes: random distribution,
defect distribution, and ideal distribution. The resulting raw image dataset was subsequently annotated with two types:
defect distribution and ideal distribution to establish the training dataset. Following this, a deep learning algorithm
YOLOv8 was employed to train the dataset, ultimately leading to the development of a deep learning model.
(a)
(b)
Figure 1. Classes of fiber distributions (a)Three types of fiber distribution (b) Two types of labels.
3 Results
The trained model demonstrates strong performance, effectively detecting the remaining specimens' CT images. The
detection outcomes are subsequently subjected to statistical categorization. Through this statistical analysis, the fiber
distribution patterns of UHPFRC specimens are classified into three categories: "transverse distribution dominant,"
"random distribution dominant," and "axial distribution dominant."
Upon investigating the relationship between these three distribution patterns and their flexural load-carrying capacities,
notable insights emerge. The "axial distribution dominant" pattern exhibits the highest average flexural load-carrying
capacity, followed by the "random distribution dominant" pattern, while the "transverse distribution dominant" pattern
exhibits the lowest load-carrying capacity.
67
Figure 2. Three types of the fiber distribution pattern.
4 Conclusion
The CT scan images, containing crucial fiber distribution information from UHPFRC specimens, served as the
fundamental dataset for this study. By meticulously labelling defect distribution and ideal distribution, a cutting-edge
object detection algorithm called YOLOv8 was employed to train a highly efficient deep learning model with
impressive mean average precision (mAP). This model exhibited remarkable proficiency in accurately identifying
diverse fiber distribution patterns within the CT images.
With the trained model in hand, we proceeded to apply it for recognizing fiber distribution in CT scan images of
UHPFRC specimens. The results obtained were subjected to rigorous statistical analysis, enabling the classification of
UHPFRC's internal fiber distribution into three distinctive patterns. Subsequently, these patterns were juxtaposed
against the outcomes of the four-point bending test. Fascinatingly, the flexural load-carrying capacity of UHPFRC
specimens featuring an axial distribution dominant pattern surpassed those with a random distribution dominant pattern,
and significantly outperformed specimens showcasing a transverse distribution dominant.
Through our study, we have effectively provided substantial data evidence, facilitating the prediction of UHPFRC
specimen flexural strength based on fiber distribution patterns. The comprehensive analysis of fiber distribution patterns
within the realm of UHPFRC materials holds immense practical significance and augments our understanding of
UHPFRC behavior. This research offers a promising pathway for the application of deep learning techniques in the field
of material science.
5 References
Song, Q., et al., Optimization of fibre orientation and distribution for a sustainable Ultra-High Performance Fibre
Reinforced Concrete (UHPFRC): Experiments and mechanism analysis. Construction and Building Materials, 2018.
169: p. 8-19.
Song, Q., et al., Key parameters in optimizing fibres orientation and distribution for Ultra-High Performance Fibre
Reinforced Concrete (UHPFRC). Construction and Building Materials, 2018. 188: p. 17-27
Liu, K., L. Peng, and S. Tang, Underwater Object Detection Using TC-YOLO with Attention Mechanisms. Sensors
(Basel), 2023. 23(5).
68
SESSION 2-B
Steel Structures 2
13th - Japanese-German Bridge Symposium, Osaka, Japan
A Fundamental Study on Application of Two-Dimensional Hermitian Elements
to In-Plane Bending Deformation Problems of Plates
Tsukushi Okabe*
Masaki Sakai*
Naoki Kaneko*
Ph.D. (Eng.), Kyosuke Yamamoto**
* Graduate Student, University of Tsukuba, Japan, s2220830@u.tsukuba.ac.jp
* Graduate Student, University of Tsukuba, Japan, s2220848@u.tsukuba.ac.jp
* Graduate Student, University of Tsukuba, Japan, s2220837@u.tsukuba.ac.jp
** Assistant Professor, University of Tsukuba, Japan, yamamoto_k@kz.tsukuba.ac.jp
Abstract:
This research presents the application of two-dimensional first-order Hermitian elements in solving in-plane bending
deformation problems in plates. Conventionally, in-plane deformation problems are frequently solved using twodimensional first-order Lagrangian elements. However, this scheme primarily assumes expansion and contraction
deformations, making it challenging to solve for bending. Specifically, when the mesh division is coarse, Lagrangian
elements can exhibit shear locking, leading to a marked decline in accuracy. Common countermeasures include
increasing the order of the base function or applying the reduced integration method, but both come with their pros and
cons. Conversely, first-order Hermitian elements, using higher-order polynomials as basis functions, are less susceptible
to shear locking, enabling more accurate solutions for bending problems. Thus, the application of Hermitian elements
can solve deformation problems using fewer elements without inducing shear locking. Nevertheless, reducing the
number of mesh divisions can result in mismatches between load points and nodes. This study validates through
numerical simulations that accurate in-plane bending deformations can still be achieved using equivalent nodal forces,
even when load points and nodes do not align. These simulations confirms that the multivariate two-dimensional firstorder Hermitian elements can accurately solve the in-plane bending problems of plates, even in the models using
equivalent nodal forces.
Keywords: Hermitian element, Finite Element Method, In-plane bending problem
1 Introduction:
The Finite Element Method (FEM) is among the most prevalent numerical solution method for partial differential
equations (PDEs). FEM defines numerical solutions at nodes within the domain where the PDEs are defined, and
constructs approximate solutions using established basis functions. By substituting those approximate solutions into the
weighted residual equations based on the original PDEs, numerical solutions are obtained. Finite elements, defined by
the division of the domain, are characterized by their basis functions, most commonly Lagrange polynomials. However,
first-order Lagrangian elements often encounter issues with shear-locking. While the application of higher-order
Lagrangian elements mitigates this issue, it requires a laborious process of element subdivision. In contrast, EulerBernoulli beams and Kirchhoff-Love bending plates often employ Hermitian elements[1]. Since Hermitian elements are
C1 elements, they can prevent the shear locking. Recent advancements in the development of bending plate elements
using Hermitian basis functions [2][3] suggest potential applications within bridge engineering. A significant advantage
of Hermitian elements is that their constituent nodes do not change even for higher orders, which paves the way for
high-precision computations in the future. Implementing high-precision Hermitian elements can result in models with
fewer elements. Even when load points and nodes do not coincide due to a rough mesh, the application scope broadens
if equivalent nodal forces can maintain adequate precision. Specifically, considering the application to girder bridges
needs the verification of the computational accuracy of the in-plane bending deformation of equivalent nodal forces of
Hermitian elements.
2 Basic Theory:
The displacement fields 𝑢 and 𝑣 in a first-order Hermitian element can be approximated by the following equations:
𝑢(𝑥, 𝑦) = 𝑵 ⋅ 𝒖 = 𝐀𝑯(𝜉, 𝜂) ⋅ 𝒖
{
𝑣(𝑥, 𝑦) = 𝑵 ⋅ 𝒗 = 𝐀𝑯(𝜉, 𝜂) ⋅ 𝒗
(1)
In these equations, 𝒖 and 𝒗 denote the nodal displacement vectors in the 𝑥 and 𝑦 direction, respectively. 𝜉 and 𝜂 are
normalized coordinate. Assuming that the one-dimensional Hermitian basis functions can be expressed as:
𝜓1 (𝜉) = (𝜉 − 1)2 (𝜉 + 2)⁄4
𝜓2 (𝜉) = (𝜉 − 1)2 (𝜉 + 1)⁄4
𝜓3 (𝜉) = − (𝜉 + 1)2 (𝜉 − 2)⁄4
𝜓4 (𝜉) = (𝜉 + 1)2 (𝜉 − 1)⁄4
The correction matrix 𝐀 and basis function vector 𝑯(𝜉, 𝜂) can be expressed as:
71
(2)
1 0
0
⎡
𝜕𝑥 𝜕𝑥⎤
⎢0 𝜕𝜉 𝜕𝜂 ⎥
𝐀=⎢
⎥ (3),
⎢
𝜕𝑦 𝜕𝑦 ⎥
0
𝜕𝜉 𝜕𝜂⎦
⎣
𝐻 𝜉, 𝜂
𝜓 𝜉 𝜓 𝜂 ,
𝐻 𝜉, 𝜂
𝜓 𝜉 𝜓 𝜂 ,
𝐻 𝜉, 𝜂
𝜓 𝜉 𝜓 𝜂 ,
𝐻 𝜉, 𝜂
𝜓 𝜉 𝜓 𝜂 ,
𝐻 𝜉, 𝜂
𝜓 𝜉 𝜓 𝜂 ,
𝐻 𝜉, 𝜂
𝜓 𝜉 𝜓 𝜂 ,
𝐻 𝜉, 𝜂
𝜓 𝜉 𝜓 𝜂 ,
𝐻 𝜉, 𝜂
𝜓 𝜉 𝜓 𝜂 ,
𝐻 𝜉, 𝜂
𝜓 𝜉 𝜓 𝜂 ,
𝐻
𝜓 𝜉 𝜓 𝜂 ,
𝐻
𝜓 𝜉 𝜓 𝜂 ,
𝐻
𝜓 𝜉 𝜓 𝜂 ,
𝜉, 𝜂
𝜉, 𝜂
𝜉, 𝜂
(4)
𝐁 matrix of the plane stress problem becomes:
𝜕𝑵⁄𝜕𝑥
0
𝐁
0
𝜕𝑵⁄𝜕𝑦
𝜕𝑵⁄𝜕𝑦
𝜕𝑵⁄𝜕𝑥
(5)
The finite element equation to be solved is given by the following equation:
𝒖
𝒗
𝐁𝐃𝐁 d𝑉
𝒇
𝒈
(6)
3 Numerical Simulation:
This study employs a numerical simulation using a two-dimensional first-order Hermitian elements to simulate the inplane bending deformation of a plate. The plate model used in this study is shown in Fig. 1. A four-point bending test is
performed on this plate model. A pin supports and a pin roller support are set at the lower end of the plate. Two
concentrated loads of 100 N/m are applied at two points, 1.0 m and 4.0 m from the left end of the upper side. By
comparing three patterns with mesh divisions of 250, 10, and 4, respectively. In the third model, the load points and
nodes do not match.
100 N/m
1.0 m
100 N/m
3.0 m
= 20000 N/m,
0.5 m
1.0 m
= 0.25
A
B
Fig. 1 The beam model for numerical verification
4 Results and Discussion:
From Fig.2 and Fig.3, it can be observed that the Hermitian element allows for stable calculations even with rough
mesh divisions. From Fig.3 and Fig.4 indicate that the accuracy is not compromised when the load points do not
coincide with the nodes. These results suggest that the equivalent nodal forces with Hermitian elements do not affect the
accuracy of in-plane bending deformation of plates.
1
1
1
0
0
0
-1
-1
-1
-2
-2
-2
0
1
2
3
4
5
Fig.2 The result of dense split model
0
1
2
3
4
5
Fig.3 The result of rough split model
0
1
2
3
4
5
Fig.4 The result in mismatches
between load points and nodes
5 Conclusion:
The performance of the equivalent nodal forces with Hermitian elements is verified, in the numerical simulation.
6 References:
[1] Bogner, F. K., Fox, R. L., Schmit, L. A.: The Generation of inter-element-compatible stiffness and mass matrices
by the use of interpolation formulas, Proc. of the Conference on Matrix Methods in Structural Mechanics,
pp.397-444, 1965.
[2] Beheshti, A.: Novel quadrilateral elements based on explicit Hermite polynomials for bending of Kirchhoff–Love
plates, Computational Mechanics, 62, pp.1199-1211, 2018.
[3] M. Bacciocchi, N. Fantuzzi, A.J.M. Ferreira: Conforming and nonconforming laminated finite element Kirchhoff
nanoplates in bending using strain gradient theory, Computers and Structures, 239, 106322, 2020.
72
13th - Japanese-German Bridge Symposium, Osaka, Japan
Experimental Investigation on Corrosion Deterioration in
Defective Areas of Paint-coated Steel
Jiang Feng *
Ojima Kazuki *
Hirohata Mikihito *
* Department of Civil Engineering, Osaka University, Japan, f-jiang@civil.eng.osaka-u.ac.jp
* Department of Civil Engineering, Osaka University, Japan, k-ojima@civil.eng.osaka-u.ac.jp
* Department of Civil Engineering, Osaka University, Japan, hirohata@civil.eng.osaka-u.ac.jp
Abstract:
Anti-corrosion measures, particularly paint-coatings, are crucial for protecting structures like bridges. Regular inspections
are needed due to their deterioration over time. This study investigates two types of paint-coated steel samples from real
bridges, introducing artificial defects to simulate corrosion progression. The samples underwent atmospheric and
accelerated corrosion tests, with surface measurements taken at different stages. Blistering data was used to assess
corrosion degradation. The study aims to validate the practical significance of blistering-related indicators in evaluating
corrosion degradation at paint-coated steel defects.
1 Introduction
In Japan, infrastructure deterioration due to corrosion, particularly in steel bridges, is a significant issue. Various corrosion
protection measures, including anti-corrosion paint-coatings, are employed to maintain these bridges. These coatings,
which are simpler and more commonly used, need replacement several times during bridges' lifespan due to their shorter
service life.
Regular inspections are crucial for effective maintenance of these coatings. However, visual inspections may not
accurately reflect the underlying corrosion condition. Therefore, methods that evaluate corrosion without removing the
coating are beneficial, especially in areas with initial defects that allow corrosive factors to penetrate.
This study conducted experimental investigations to verify the validity and engineering significance of metrics used to
assess corrosion degradation in areas of coated steel defects. Both atmospheric exposure tests, which provide reliable
data, and accelerated corrosion tests, which yield significant data in a short time, were used on specimens to investigate
degradation characteristics.
2 Experiments
This study examined and analyzed the deterioration of coatings and corrosion of two types of paint-coated steels. Both
steel materials were SS400. One of the paint-coated steel specimens, referred to as H-steel in this study, was cut from a
steel component of a highway bridge. The other paint-coated steel specimen, referred to as R-steel, was cut from a steel
component of a railway bridge. Both steels underwent corrosion tests in four different corrosive environments. These
included two atmospheric exposure tests conducted in Choshi and Miyakojima, Japan, and two accelerated corrosion
tests, namely ISO 16539 Method B and CCT Method A. For each corrosion environment, three specimens were used for
both H steel and R steel, resulting in a total of 12 specimens for each type of steel undergoing corrosion tests. Figure 1(a)
presents the appearance of the specimens, while Figure 1(b) provides specific dimensions. All specimens were rectangular
plates measuring 150 mm × 70 mm with a thickness of 8 mm. Artificial defects of 2 mm × 50 mm were machined into
the steel substrate of each specimen. At the end of the corrosion test, the corroded surface within the 50 mm × 70 mm
area (indicated by the red box) was observed and measured.
Figure 1: Specimen setup: (a) Appearance of the specimens; (b) Dimensions of the specimens
73
3 Experimental results
Coating deterioration can generally be assessed on the basis of rusting, cracking, and spalling. Other factors like blistering,
discoloration, and staining are also considered during visual inspections. However, these factors are often subjectively
judged. This study focuses on blistering, aiming to quantitatively assess coating deterioration.
Blistering occurs when water and oxygen permeate defects in the paint-coating, causing rust formation on the exposed
steel substrate. The rust expands, reducing the steel plate's thickness and causing the paint-coating to bulge and form
blisters.
Figure 2 presents the images of the paint-coating surface appearances of the specimens subjected to the atmospheric
exposure test and the accelerated corrosion test. Visual observations confirmed blistering near the defects. Blistering was
more pronounced in accelerated corrosion tests than in atmospheric exposure tests, presumably due to the slower corrosion
rate in the latter. Even after a longer corrosion time, the atmospheric exposure test did not achieve the same level of
corrosion progression as the accelerated corrosion test.
4 Quantitative analysis of blistering
To assess blistering quantitatively, a threshold for blistering height is required. This threshold should exclude surface
unevenness due to paint inhomogeneity but include paint-coating pushed up by rust expansion. This study compares
surface roughness before and after paint-coating removal to understand corrosion thickness reduction under deteriorated
paint-coating.
Blistering was defined as an area greater than 50 µm from the intact part of the coating, and corrosion as an area less than
-30 µm from the intact part of the steel substrate. Five indicators were used to measure blistering and corrosion: blistering
height, blistering area, blistering volume, corrosion depth, and corrosion extension distance.
Correlation coefficients were calculated between the indicators related to corrosion depth and blistering. The results show
that the correlation coefficient with blistering height is the largest. Scatter plots of the correspondence between the
corrosion extension distance and blistering metrics show high correlation coefficients with blistering area and blistering
volume.
5 Conclusions
This study examined two types of paint-coated steel through atmospheric exposure and accelerated corrosion tests. Visual
inspections showed that coatings in accelerated tests deteriorated more than those in atmospheric tests. Rust and blistering
were observed near defects in all specimens, with blistering more pronounced in accelerated tests.
Metrics were established to quantify blistering and corrosion, revealing correlations between "blistering height and
corrosion depth", "blistering area and corrosion extension distance", and "blistering volume and corrosion extension
distance". The study validates the use of these indicators for assessing corrosion degradation in defective areas of paintcoated steel without removing the paint-coating.
Figure 2: The surface appearance of specimens at different stages after corrosion tests: (a) Atmospheric exposure test;
(b) Accelerated corrosion test.
74
75
76
13th - Japanese-German Bridge Symposium, Osaka, Japan
Energy absorption of bolted patch plate repaired member in ultimate behaviour
Souta Masudome*
Toshikazu Takai **
* Kyushu Institute of Technology, Kitakyushu, Japan, masudome.souta523@mail.kyutech.jp
** Kyushu Institute of Technology, Kitakyushu, Japan, takai@civil.kyutech.ac.jp
Abstract
One of the causes of deterioration of steel bridges is corrosion damage. The damage reduces the thickness of a steel
plate and load-carrying capacity. To recover the structure from the damage, bolted patch plate repair is adopted in some
cases. The patch plate is attached around the damaged area and assembled by tightening high-strength bolts. The
attached patch plate appends the cross-sectional area lost by the damage. Not only load resistance but also deformability
is an important factor in realizing ductile structure considering seismic performance. In this study, we conducted finite
element analysis to investigate the characteristics of energy absorption in the ultimate behavior of the patch plate
repaired member.
The investigated member was like a small bolted connection as a standard specimen of slip test of bolted connection.
Although the main plates of a bolted connection are divided into two sides, the main plate of the member is continuous.
In advance, the reproduced analysis of an experiment of plate repaired structure reported in a previous study was carried
out to check the validity of the analysis method. After that, parametric analyses were conducted. The parameters
focused on in the analyses were the depth of corrosion, the thickness of repaired plate, and the bolt arrangement. The
load resistance and deformability were evaluated considering the energy absorption.
As a result, the failure modes of almost all cases were the tensile fracture in the net cross-section of the main plate, as
shown in Fig. 1. In some cases where the depth of corrosion was large, and the thickness of the repaired plates was
thinner than the depth, the maximum load was reduced. However, the maximum load and elongation at the time were
almost the same except the case, even the depth of corrosion and the thickness of the repaired plate was different.
Therefore, in the case where the thickness of the patch plate is enough to fill the corroded depth, the performance of
energy absorption hardly changes.
Figure 1: Mises stress distribution at the maximum load.
77
78
13th - Japanese-German Bridge Symposium, Osaka, Japan
Corrosion Assessment of Weathering Steel Bridges in Osaka and
Wakayama Prefectures (Japan)
Wint Thandar*, Shen Hui **, Yasuo Hanaoka***, Nobuto Okubo****,
Tetsuya Iida*****, Tomonori Tomiyama******, Kunitomo Sugiura*******
*, ******Advanced Material Research Centre, PWRI, Japan., thandar-w177cn@pwri.go.jp, tomiyama@pwri.go.jp
** JIP Techno Science Co., Ltd., Japan., shenhui_jts@jp.nttdata.com
***, **** Takadakiko Co., Ltd., Japan., y_hanaoka@takadakiko.co.jp, n_ookubo@takadakiko.co.jp
*****Takigami Co., Ltd., Japan., t.iida@takigami-grp.jp
*******Department of Urban Management, Kyoto University, Japan., sugiura.kunitomo.4n@kyoto-u.ac.jp
1 Introduction:
Weathering steel, well known for its outstanding corrosion resistance, is made by adding a small amount of chemical
such as Cr, Cu, P and Ni to resist weathering effect to the ordinary carbon steel. The application of weathering steel in
construction of infrastructure benefits reducing life cycle costs by eliminating initial and repainting processes. However,
the ability to corrosion resistance of weathering steel depends on the quality of protective rust layer formed on the steel
substrate after exposure in the environments. The states of rust formation change with period of exposure and relative
exposure environment, such as the amount of airborne salt and exposure posture. In Japan, the construction of weathering
steel bridges has considerably increased within a year of 1995 to 2005 since the first building of weathering steel bridge
was initiated in 1978 [1]. Therefore, the maintenance of weathering steel bridges is an issue in Japan these days. In order
to achieve an efficient maintenance of weathering steel bridges, appropriate inspeciton technology must be adopted, just
as conventional bridge. In this study, the observation of corrosion stages and protective qualities of four weathering steel
bridges located in Osaka and Wakayama prefectures are discussed according to evaluation by the visual observation and
microstructural study of rust products by the XRD analysis. This study reports the evaluation results of characterization
and compositional analysis of rust layer of weathering steel bridges as a reference for future weathering steel bridge
maintenance.
2 Experimental details
N
Bridge D
Konohana, Osaka
↑
The four weathering steel bridges located in Osaka and
Pref.
Wakayama Prefectures as shown in figure 1 are investigated for rust
Bridge C
Hirano Higashi,
layer’s composition and protectiveness. The rust thickness, rust layer
Osaka Pref.
hardness and physical appearance observation are checked to
Bridge A
evaluate the rust status. The powder rust sampling for
Hidakagawa,
Wakayama Pref.
microstructural analysis has done by scraping of rust layer on the
bridge’s surface using razor blade. The visual observation and
Bridge B
adhesive tape sampling are also conducted on bridges. The
Nishikatsuura,
protective ability index of rust layer is calculated according to the
Wakayama Pref.
result of percentage of phase composition. The state of the rust layer
Fig. 1 Location of inspected bridges
is specified as active, inactive and protective rust layers correspond
to categories Ⅰ, Ⅱ and Ⅲ, which are characterized by the β-FeOOH
+ spinel-type iron oxide rich, γ-FeOOH rich and α-FeOOH rich domains, respectively. The state of the rust layer of
bridges in this study is characterized by using the ternary diagram of composition of rust layer.
3 Results
The maximum rust thickness is measrured in the Bridge D with the amount of thickness is nearly 200 µm. As the
resutls of measurements on different memebers of brdige, it found out that horizontally located memebers have higher
rust thickness than other surfaces. The rust apperance rating is observed within 3 and 5. The phase analysis from the XRD
analysis shows that γ-FeOOH and α-FeOOH are major components of rust in the studied bridges. The formation of γFeOOH is not observed in rust with highly chloride concentrated bridges in Osaka city, however, an amorphous form rust
δ-FeOOH is found very rich. The other forms of amorphous rust such as hematite and magnetite or mahemite contain in
rust with high chloride content. In a mean time, geothite is found at very low peak in the rust from the Bridge D located
in a severe seaside region. Due to the chloride concentration in rust, the rust stabilization process and compositions of
rust are varied. By seeing comparison with the results of rust thickness measurement results, it is undestood that a poor
quality rust with high rust thickness surface condition is observed in Bridge D. But during this time survey, the rust
thickness is still under the limit of good condition and the difference does not necessary to be concerned. The state of the
rust layer is determined by plotting rust composition in simulated tenary plot. The ternary plot of rust compositon and
calculated PAI of each bridge are presented in Fig. 2. Similary to the rust thickness and surface attached salt, the calculated
PAI vlaue of bridges are arranged as Bridge A > Bridge B > Bridge C > Bridge D. The conditions of rut states in three
inspected bridges are existed in the inactive rust state in the ternary plot. According to the XRD result of phase
composition in these rust, the initiation of protective rust is confirmed in the bridge A, B and C, and it is also expected to
79
NaCl content in rust (mg)/ Rust (g)
transform into fully protective state after the time being. The active rust layer is explored in the bridge near the severe
seaside region with high chloride content and it turns out that the higher chloride content in rust layer causes retard in
producing of protective corrosion phase goethite (α-FeOOH). As shown
0
100
in Fig. 3, the amount of chloride concentration in rust of Bridge D is higher
than others and it is assumed that high concentration of chloride in Bridge
10
90
D inhibited the formation of α-FeOOH as in a form of stabilized rust layer.
20
80
It is noticeable that the chloride concentration in rust would affect not only
Protective
30
70
rust: III
on the corrosion kinetics but also on the morphology and protective
α/γ*=1
40
60
characteristics of corrosion protective films. Transparent adhesive tape test
50
50
was used for the size of rust particles, density and quality of rust layer.
From the results of tape adhesion, the properties of rust layers are different
Inactive
rust:
II
60
40
within the inspection location in a same bridge. This type of variation
A
70
30
pointed out an importance of necessary for inspection of weathering steel
Active rust: I
B
80
20
bridge at the several different locations, especially near the drainage or
C
D
90
10
horizontal surface of the members, even the inspection has taken only on
100
0
one bridge. Rust flakes are measured to the neasrest 5 mm at all measured
0
10
20
30
40
50
60
70
80
90 100
bridges. Although the tape-adhesion test shows the adhesive quality of top
loosen-rust, the visual observation of adhesive test needs profession to
β-FeOOH+ spinel
(β+S)/γ*=0.5
decide properly the protective state of stable rust layer. However, the
Fig. 2 Ternary plot of rust composition
physical appearance of rust and thickness measurement results of bridges
had a good agreement with the PAI index calculation. From this point, it is noted that the ratio of α/γ* are possible to
evaluate the protective ability of weathering steel rust layer in real bridges located in seaside and severe seaside regions
with a certain amount of chloride concentration.
45
The results of this study confirm that corrosion products vary
Bridge D
40
in each bridge related to local environmental condition. The ability
35
of rust protectiveness in the bridges are related to the composition
Bridge A
30
of rust layer. Hardness of rust layer, as an indicator of mechanical
25
properties of protective rust is measured on different measurement
20
points of bridge’s members. Multiple measurements were
Active rust
15
conducted on different bridge’s members. According to the
10
measurement conducted bridges, the results of both surface
Inactive rust
5
salinity and hardness of rust layer are different. Results of several
0
measurement points at inspected bridges are summarized in Fig. 4
Bridge A
Bridge B
Bridge C
Bridge D
where it is representing that the hardness tends to decrease as the
PAI:
0.408
0.312
0.134
0.113
salt concnetration increases. This scenario explains that excessive
amount of deposited salt erodes the rust layer leading to a
Fig. 3 Chloride concentration and PAIs
formation of unstable rust layer in which δ-FeOOH is a major
components of rust.
5 mm
5 mm
Rebound hardess (HV)
300
4 Discussion
Bridge C
Bridge D
Bridge A
Bridge B
250
High contents of amorphous phases such as α-FeOOH and
Bridge E
δ-FeOOH are mainly observed in the rust simulated in severe
200
seashore environment. The applicability of rust appearance index
150
as weathering steel management and maintenance score in
bridges are confirmed by comparing the results of rust
100
appearance evalaution and pahase compositions of rust. The
50
appearance index and thickness of rust layer are relevant to the
calculated protective indices. On the other side, the result of tape0
adhesion test can be used with professional skill to judge the rust
0
200
400
600
800
Surface salinity (mg/m²)
properties. At the moment, the PAIs of all inspected bridges are
lower than 0.5, and it is assumed that the rust layer is still
Fig. 4 Hardness varation with surface salinity
developing to achieve fully protective rust on a base steel
material. During the detection, significant oxide film degradation has not been identified and ranked the state of rust layer
as acceptable conditions. In terms of mechanical properites of rut layer, it is confirmed that the hardness of rust layer has
a negative linear relationship with an amount of salt. The physical appearance and thickness of base rust layer can be
useful to represent the state of protective rust layer, and chloride contamination in rust palys a major role in the rust
stabilization process of uncoated weathering steel briges.
5 References
[1] Japan Bridge Construction Association: Weather-resistant Steel Bridge Performance Data Collection, 24 th Edition,
Japan, March 2019.
80
13th - Japanese-German Bridge Symposium, Osaka, Japan
Evaluation of Debonding of CFRP bonded onto Steel Plate by AE Method
Morimune Mizutani *
Toshiyuki Ishikawa **
Yoshimichi Fujii ***
* Kansai University, Faculty of Environmental and Urban Engineering, Japan, mizutani@kansai-u.ac.jp
** Kansai University, Faculty of Environmental and Urban Engineering, Japan, t-ishi@kansai-u.ac.jp
*** Kanazawa Institute of Technology, Graduate School of Engineering, Japan
Abstract
In the CFRP bonded repair of steel members, the stress in the steel members is transmitted to the CFRP by the adhesive
layer. The debonding of the CFRP should be prevented to sustain the repair effect. Therefore, early detection of CFRP
debonding is necessary for the CFRP bonded repair of steel members. This study conducted debonding tests on CFRP
bonded steel plates using strain gauges and installed AE sensors. Consequently, elastic waves were detected by the AE
sensor at an earlier stage than debonding detection by strain gauges. The position of elastic wave generation during
debonding was determined by the AE, and the debonding propagation behavior was compared with that given by the
energy release rate. The results indicated that the two trends were almost identical.
Keywords: Debonding, Acoustic emission, bonded joint, CFRP
1 Introduction
Carbon fiber reinforced polymer (CFRP) bonded repair is a repair method for corrosion-damaged steel members. CFRP
is a composite material with excellent material properties such as light weight, high strength, and corrosion resistance. In
CFRP bonded repairs, adhesive stress is concentrated on the CFRP edge, which may cause debonding. If the CFRP
debonds from the edge, the repair effectiveness is lost because the required bond length to transmit the force is insufficient.
Therefore, techniques for the early detection of CFRP debonding are necessary to ensure the safety of steel bridges. In
this study, the authors focused on the AE method as a technique for monitoring the debonding of CFRP bonding repairs
and conducted debonding tests using strain gauges and AE sensors installed on the specimens.
2 Debonding Test of CFRP Bonded Steel Plates
In this study, the cantilever plate bending test shown in Figure 1 was conducted to debonding of CFRP bonded steel
plates. The CFRP plates are preformed with the fiber direction oriented in the longitudinal direction. The CFRP was
bonded along the fiber direction of the CFRP, corresponding to the axial direction of the specimen. A two-component
epoxy resin was used as an adhesive. After CFRP bonding, the specimens were cured in a room at 20°C for at least 24 h
before loading tests were conducted.
Strain gauges and AE sensors were placed on the specimen at the positions shown in Figure 1. Two AE sensors were
mounted on the top surface of the specimen, 200 mm apart. The AE sensor was set to a threshold value of 40 dB and
sampling frequency of 5 MHz.
3 Test Results
Figure 2 shows the applied load–strain relationship obtained from the loading tests. The filled circles within the
experimental values are measured per second. The theory shown in Figure 2 is based on shear lag theory derived for
Figure 1: Dimension of specimen and test setup
(unit: mm)
Figure 2: Relationship between the applied load and the
strain for specimen-2
81
Figure 3: Cumulative counts with respect to the elapsed
time and applied load for s pecimen-2
Figure 4: The applied load and position of elastic wave
generation for s pecimen-2
conditions similar to cantilevered beam bending tests [1]. As shown in Figure 2, the experimental and theoretical values
were in good agreement with each other in the lower applied load range, where the experimental results showed linearity.
As shown in Figure 2, as the applied load increased, the values of the strain gauges SG-1 through SG-3 on the CFRP side
reached zero. In the case of SG-3, where the measured strain gradually decreased after reaching a maximum value, the
lag in force transmission owing to the movement of the bonded edge caused by the debonding process might be affected.
Figure 3 shows the cumulative counts with respect to the elapsed time and applied load. The drawn AE result is the only
event used in the analysis of the source of the elastic wave with the two AE sensors described below. The time of
debonding at zero strain at each location is indicated by dashed lines in Figure 3. Notably, the cumulative count of AE
increased with the time of debonding at SG-1 to debonding at SG-3.
Figure 4 shows the results of the applied load and position of elastic wave generation. The position of the generated elastic
wave was linearly evaluated using the time difference between the events when the generated elastic wave arrived at the
two AE sensors. The velocity of sound propagating through the steel was assumed to be 5,900 m/s. To reduce the noise
in location determination, a filter was applied to modify the threshold value to 50 dB. The solid line in Figure 4 shows
the relationship between the debonding load and the debonding length from the CFRP end which is determined by the
energy release rate [2]. In this study, the energy release rate is calculated using the debonding load of SG-2, since signs
of brittle debonding were obtained for SG-2 of specimens-1 and 2. Using the energy release rate due to the debonding
load of SG-2 as the critical value, the Relationship between theoretical debonding load and remaining bond length was
obtained.
As shown in Figure 4, as the load increased, the debonding tip location moved from the fixed side to the center of the
specimen. Figure 4 indicates that the evaluation of debonding propagation by the energy release rate has a tendency
similar to the results of debonding propagation given by the AE. Therefore, in conclusion, the debonding propagation
behavior was captured by the AE sensors. However, elastic waves were also observed at points where the CFRP was not
adhered (x = –25 to 0 mm). Therefore, the measurement method must be improved to capture only the elastic waves
caused by debonding and evaluate the AE parameters. In addition, it should be noted that only two AE sensors were used
in this study; therefore, evaluation in the width direction was not possible.
4 CONCLUSIONS
In this study, to evaluate the debonding of CFRP, debonding tests were conducted on CFRP bonded steel members using
strain gauges and installed AE sensors. AE events occurred faster than the strain gages responded, suggesting that AE
sensors can detect minute delamination that cannot be detected by strain gages. The source of the elastic waves could be
estimated using two AE sensors. As the applied load increased, the location of the elastic waves moved from the CFRP
end on the fixed side to the center of the specimen. The energy release rate was used to evaluate the debonding
propagation, and the results were almost identical to the AE results.
5 References
[1] Ishikawa, T., Shimizu, M., Hattori, A. & Kawano, H. 2012. Effect of loading conditions on adhesive stresses of
steel members strengthened by bonding CFRP plates, Journal of Japan Society of Civil Engineers, Ser. A2
(Applied Mechanics (AM)), Volume 68, Issue 2, I_715-I_726. (In Japanese)
[2] Mizutani, M., Ishikawa, T. & Fujii, Y. 2022. Evaluation of debonding damage in CFRP bonded steel plates by
acoustic emission method, Proceedings of Constructional Steel, Volume 30, 159-164. (In Japanese)
82
13th - Japanese-German Bridge Symposium, Osaka, Japan
Study on Relationship between Whole Displacement and Bearing Deformation of
Bolt Holes in High-Strength Frictional Bolted Joints
PhD Student Zice QIN *
Senior Lecturer Hitoshi MORIYAMA **
Professor Takashi YAMAGUCHI *
* Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, su23537p@st.omu.ac.jp
** Graduate School of Technology, Industrial and Social Sciences, Tokushima University, Japan,
moriyama.hitoshi@tokushima-u.ac.jp.
* Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, yamaguti-t@omu.ac.jp
Keywords: High-strength frictional bolted joints, whole displacement of joints, bearing deformation, residual
deformation
1 Introduction:
The deformation capacity of high-strength frictional bolted joint is related to a summation and ratio of tensile deformation
and bearing deformation. However, since this relationship has not been investigated, the deformation capacity is not
quantitatively evaluated. Therefore, tensile tests of joints composed of one bolt or three bolts were conducted to
investigate the relationship among failure modes, whole displacement of the joint, and the bearing deformation. It would
be desirable to induce the split failure mode for developing the deformation capacity from obtained results. The whole
displacement is about the same as residual bearing deformation, although depending on failure modes.
This paper conducted tensile tests of frictional bolted joints, which vary geometrical configurations and grades of plate
and bolts, as a fundamental study focused on the ultimate limit state of the high-strength frictional bolted joints. The
relationship among the failure mode, whole displacement and bearing deformation of joints was investigate to search the
failure mode which has a high ductility.
2 Tensile Test
The considered parameters of tensile tests are number of bolts n (=1, 2, 3), steel grades, bolt grades, end distance e1, pitch
p and width w. To evaluate the entire behavior of the joint, its displacement δ was measured. The strain of the side surface
of the connected plate was measured to investigate the strain distributions after a major slip. In plate failure modes cases,
the applied load was removed at 95% of the maximum load after the peak to observe the bearing deformation of the bolt
hole δres. In bolt shear failure mode cases, loading was continued until bolt breakage occurred due to brittle failure.
3 Test Results
3.1 Failure modes
As shown in Fig.4, the failure modes confirmed in the test were shear failure (SH), split failure (SP), net cross-section
failure (N), net cross-section failure (N), net cross-section failure and plate shear yielding (N(SH)), bolt shear failure (BO),
(a)
SH
(b)
SP
(c)
N
(d)
N(SH)
(e) BO
(f) BO(SH)
(g) BO(N)
(h) BO(N+SH)
N(SH): Net cross-section failure occurs in conjunction with plate shear yielding.
BO(SH): Bolt shear failure occurs in conjunction with plate shear yielding.
BO(N): Bolt shear failure occurs in conjunction with net cross-section yielding.
BO(N+SH): Bolt shear failure occurs in conjunction with plate shear yielding and net cross-section yielding.
Figure 1: Failure modes confirmed in the test
83
6.0
5.0
n=2(d=16mm)
n=3(d=16mm)
4.0
S
Ptnd/Pbod
N
BO
Pesd / P bod
Pesd / P bod
SH
SP
N
N(SH)
BO
BO(SH)
BO(N)
BO(N+SH)
3.0
2.0
4.0
3.0
24
18
2.0
12
1.0
6
δ = 0.8262 δres
R² = 0.7531
Ptnd/Pbod
1.0
0.0
0.0
0.0
SH
SP
N
N(SH)
BO
BO(SH)
BO(N)
BO(N+SH)
y=x
L.R.Eq.
30
δ (mm)
Pesd / P bod Pesd / P bod
n=1(d=22mm)
5.0
36
6.0
n=1(d=16mm)
1.0
2.0
3.0
Ptnd/Pbod
4.0
5.0
6.0
0
0.0
1.0
2.0
3.0
4.0
5.0
Ptnd/Pbod
(a) Expected failure modes
(b) Actual failure modes
Figure 2: Classification of the failure modes by Ptnd/Pbod and Pesd/Pbod
6.0
0
6
12
18
δres (mm)
24
30
36
Figure 3: Relationship of δ - δres
bolt shear failure and plate shear yielding (BO(SH)), bolt shear failure and net cross-section yielding (BO(N)), bolt shear
failure and net cross-section yielding and shear yielding (BO(N+SH)). These modes are the same as those of mild steel
joints [1, 2, 3], as well as HSS joints in other countries [4, 5, 6].
The expected and actual failure modes classified by Ptnd /Pbod and Pesd / Pbod, which are the ratios of design ultimate
resistances for the three major failure modes in joints as known (shear failure S, net cross-section failure N, and bolt shear
failure BO), are shown in Fig.6. Mode SH, N, and BO can be almost classified using the aforementioned conventional
equations developed for mild steel joints. Coupled modes such as SP, N(SH), BO(SH), BO (N), and BO(N+SH) occurred
as Ptnd/Pbod and Pesd/Pbod decreased. Especially in the case of SP, Pesd/Pbod and Ptnd/Pbod were both approximately 1.0.
3.2 Relationship between Whole Displacement and Bearing Deformation
Figure 8 shows the relationship between the whole displacement of the joint δ at the maximum load Pmax and the residual
deformation of the bolt hole δres. As shown in Figure 8, δ is generally in a linear relationship with δres. The relatively small
difference between δ at Pmax and δres is due to the fact that δ includes the bolt hole deformation, the elongation of the
general area of the connected plate, and the shear and bending deformations of the bolt. In plate failure modes, the
proportion of elastic deformation of the connected plate and shear/bending deformation of the bolt in the whole
displacement δ is smaller.
In bolt shear failure modes (BO, BO(SH), BO(N), BO(N+SH)), the reason why δ is greater than δres is that the plastic
deformation around the bolt hole is relatively small. As a result, the proportion of the elastic deformation of the connected
plate and the shear/bending deformation of the bolt increases in the whole displacement δ.
4 Conclusion
In this study, tensile tests of high-strength frictional bolted joints with HSS were conducted to investigate the relationship
among the failure mode, whole displacement and bearing deformation of joints, to search the failure mode which has a
high ductility. The following conclusions can be drawn.
(1) It is confirmed that failure modes of HSS joints are the same as those of joints made of mild steel.
(2) These failure modes can be classified by Ptnd /Pbod and Pesd / Pbod, which are the ratios of design ultimate resistances
as known.
(3) The whole displacement of the joint δ at the maximum load Pmax shows a generally linear relationship with the
residual deformation of the bolt hole δres, with some slight variations depending on the failure mode. This is
considered because the plastic deformation around the bolt hole is dominant in δ.
(4) According to (3), if the deformation of the bolt hole and the shear/bending deformations of the bolt included in the
whole displacement δ are quantified, it is possible to estimate δ from δres.
5 References
[1] Eurocode 3: Design of Steel Structures - Part 1-1: General Rules and Rules for Buildings, EN1993-1-1, European
Committee for Standardization (CEN), 2005.
[2] Eurocode 3: Design of Steel Structures - Part 1-8: Design of Joints, EN1993-1-8, European Committee for
Standardization (CEN), 2005.
[3] Eurocode 3: Design of Steel Structures - Part 1-12: Additional Rules for the Extension of EN 1993 up to Steel
Grades S700, EN1993-1-12, European Committee for Standardization (CEN), 2007.
[4] Može P., Beg D. and Lopaticˇ J., “Net cross-section design resistance and local ductility of elements made of high
strength steel”, Journal of Constructional Steel Research, 63(11), 1431-1441, 2007.
[5] Može P. and Beg D., “High strength steel tension splices with one or two bolts”, Journal of Constructional Steel
Research, 66(8-9), 1000-1010, 2010.
[6] Wang Y.B., Lyu Y.F., Li G.Q. and Liew J.Y.R., “Behavior of single bolt bearing on high strength steel plate”,
Journal of Constructional Steel Research, 137, 19-30, 2017.
84
SESSION 3-A
Composite Structures 1
13th - Japanese-German Bridge Symposium, Osaka, Japan
Building bridges with thin-walled semi-precast concrete elements – experimental torsional
investigations
Dipl.-Ing. Michael RATH *
Dipl.-Ing. Franz UNTERMARZONER **
Prof. Dr.-Ing. Johann KOLLEGGER ***
* Institute of Structural Engineering, TU Wien, Vienna, Austria, michael.rath@tuwien.ac.at
** Institute of Structural Engineering, TU Wien, Vienna, Austria, franz.untermarzoner@tuwien.ac.at
*** Institute of Structural Engineering, TU Wien, Vienna, Austria, johann.kollegger@tuwien.ac.at
Abstract
A new bridge construction technique, called LT-bridge construction method [1], is currently being developed at TU
Wien. This method utilizes thin-walled precast concrete elements for the construction of concrete bridges. During this
new bridge construction method, thin-walled precast concrete elements are joined together to form hollow box girders,
spanning in the longitudinal direction (L). On top of the longitudinal girders, specially designed precast deck slab
elements are placed, spanning in the transverse direction (T). For multi-span bridges, the longitudinal hollow box girder
is connected to a previously prepared pier segment by the means of post-tensioning. To form a continuous deck and
bottom slab, in-situ concrete is poured on the deck slab elements as well as the thin bottom slab, allowing to place
reinforcement over the joints between the hollow box girder and the pier segment. While this process prevents
unreinforced joints in the deck and bottom slab, the joints in the webs remain unreinforced. This design aspect needed
to be considered in more detail regarding the load carrying mechanisms.
Keywords: LT-bridge construction method, thin walled bridges, torsional testing, joints, hollow box girder
1 Experimental investigations
Experimental tests on the influence of such a joint in the webs regarding torsional loads were carried out at the
laboratory of the Institute of Structural Engineering of TU Wien. Two large-sized test specimens in the form of thinwalled box girder sections, one representing a regular hollow box girder without joints and one with joints in the webs,
but both with continuous deck and bottom slabs, were subjected to torsional loads. The test specimens were 1.00 m
wide, 1.30 m high, 0.10 m thick and 9.20 m long. In the middle of the girders, a diaphragm was provided, against which
the webs of one of the girders were connected with a shear joint. Both test specimens were reinforced with a
longitudinal torsion reinforcement of Ø8/15 and a stirrup reinforcement of Ø10/15 placed all around the cross-section.
In the case of the girder with joints, the longitudinal reinforcement of the webs had to end in the joint and was therefore
no longer be effective. Since it is equivalent to use a concentrated corner reinforcement with the same cross-sectional
area instead of a longitudinal reinforcement placed all around, two additional continuous Ø12 bars were placed in the
bottom and top slab in each corner of this girder. Their cross-sectional area corresponds almost exactly to the ineffective
web reinforcement. Figure 1 shows a test specimen already installed in the test setup. The hydraulic presses used to
generate the torsional moment are visible in the foreground.
Figure 1: Finished test specimen in the test setup [2]
87
The test specimens were prestressed against the floor of the laboratory in the centre of the girder, at the point where the
diaphragm was concreted, by means of hydraulic presses and threaded rods. Thus, a restraint was created that could absorb
the torsional moments that were induced at the end of the girder at the location of another diaphragm with hydraulic
presses. Due to the created restraint, the second half of the test body remained unloaded and could be subjected to another
torsion test after a rotation of the test body. Thus, a total of four torsion tests could be carried out with the two specimens
produced. Simplifications were made in comparison to a cross-section of the LT bridge construction method. For example,
it was decided not to apply any post-tensioning in this series of tests and thus to fundamentally obtain an understanding
of joints in the webs with the bottom and deck slabs being continuous at the same time. Furthermore, an in-situ concrete
layer was not applied to the bottom and deck slabs, instead they were designed to be continuous themselves with
continuous longitudinal reinforcement.
2 Results and discussion of the experimental investigations
Due to the selected reinforcement arrangement, it could be assumed that both girders would fail at the same torsional
moment. However, a different condition occurred. Figure 2 shows the torsional moment plotted against the angle of twist
measured at the load introduction area. For the sake of clarity, only two of the four tests are shown in the diagram. As can
be seen, the failure of the beam without joints in the webs occurred at a torsional moment of 551 kNm. The failure was
found to be due to the yielding of the longitudinal reinforcement at a distance of 1.20 away from the restraint, although
the bending moment occurring there as a result of self-weight is smaller than the one acting directly at the restraint. This
was attributed to vertical compressive stresses that counteract the torsional stresses locally in the area of the joint due to
the support of the compression struts resulting of torsion and shear force. The beam with unreinforced joints in the webs
had two additional bars in each corner, as described above. This meant that this beam had additional load-bearing reserves
in the area where the regular test specimen failed. In the area of the actually suspected weak point, namely the joint, the
vertical compressive stresses just mentioned helped to apply higher torsional loads. The failure finally occurred as a result
of shear failure in the joint. Based on this series of tests, the weakening due to the joint can be compensated by means of
additional corner reinforcement and the occurrence of vertical compressive stresses with regard to the LT bridge
construction method.
Figure 2: Results of the experiments [2]
Since it is also possible to produce the longitudinal girders of the LT-bridge construction segmentally, unreinforced joints
in the webs can also occur away from the connection to the pier segment. An experimental investigation of such joints is
being planned. Furthermore, the influence of post-tensioning is to be determined in this future series of tests.
3 References
[1] Untermarzoner, F., Rath, M., Kollegger, J. (2023). New Modular Construction Method for the Erection of Multispan Concrete Bridges. In: Ilki, A., Çavunt, D., Çavunt, Y.S. (eds) Building for the Future: Durable, Sustainable,
Resilient. fib Symposium 2023. Lecture Notes in Civil Engineering, vol 350. Springer, Cham.
https://doi.org/10.1007/978-3-031-32511-3_172
[2] Rath, M.; Untermarzoner, F.; Kollegger, J. On the Torsional Behavior of the Longitudinal Bridge Girders Used in
the LT-Bridge Construction Method. Appl. Sci. 2023, 13, 6657. https://doi.org/10.3390/app13116657
88
13th - Japanese-German Bridge Symposium, Osaka, Japan
Proposal on rigid connection between steel deck plate girder and RC abutment
in replacement project
Yasuo Tawaratani*1
Naomitsu Akashi*2
Mikinao Goto*3
Univ. Prof. Dr.-Eng. Osamu Ohyama*4
Univ. Assoc. Prof. Dr.-Eng. Yusuke Imagawa*5
*1
Sogo Engineering Inc., Tokyo, Japan, y-tawaratani@sogo-eng.co.jp
Sogo Engineering Inc., Nagoya, Japan, n-akashi@sogo-eng.co.jp
*3
Ota City, Tokyo, Japan., goto-m2310@city.ota.tokyo.jp
*4
Osaka Institute of Technology, Osaka, Japan, osamu.oyama@oit.ac.jp
*5
Osaka Institute of Technology, Osaka, Japan, yusuke.imagawa@oit.ac.jp
*2
Abstract:
In bridge replacement project over rivers, it is necessary to take into account the height from the river design water level
due to heavy rain. On the other hand, it is difficult to raise the bridge position in order to keep the daily life of the
surrounding bridges. Therefore, we are developing the jointless abutment structure using steel deck girders, which can
be constructed with a lower girder height than conventional bridges. The purpose of this study is focusing on the joints
between steel deck plate girder and RC abutments for a bridge replacement project with a length of about 10m in Ota
City, Tokyo. We report the results of verification by conducting loading experiments using a full-size model specimen.
Finally, we propose a rational connection design method for small and medium-sized bridges.
Keywords: Replacement Project, Composite Rigid Frame Bridge, Rigid Connection, Steel Deck Plate Girder
1 Introduction:
In recent years, many existing bridges in Japan are being replaced for the purpose of improving aging and seismic
performance. This paper describes a bridge that adopts a jointless abutment structure adopting a composite rigid frame
bridge, which is the object of bridge (NINOHASHI-Br.) replacement project in small and medium-sized rivers in dense
residential areas in Ota City, Tokyo [Fig.1~Fig.3]. In this project, the low girder height and a compact foundation were
required, so we proposed the composite rigid frame bridge with steel plate deck girder and RC abutments.
The conventional steel girder-abutment connection is steel girder with RC slab. The specification for highway bridges
stipulates the design method for the joints between steel girders and RC abutments [1], but it is limited to the
superstructure having the RC slab type. As shown in Fig.4, we proposed a resistance mechanism in which the rigid
joints are bent at right angle and headed studs (hereafter referred to as studs) are buried inside the abutment when a steel
deck plate girder structure with a low girder height is adopted for river bridge.
Studs cannot be
400 2000
10800
5000
500
welded
[Unit:mm]
500
2000 400
CL
[Proposal]
374
Tokyo Metropolis
Ota City
10@1000=10000 (Grder Pitch)
400
Steel deck plate Studs
girder type
Abutment
HANEDA
Airport
(Bridge construction site)
Fig.1 Location map
600
3000
Ro
ute
1
5
Tokyo
Bay
[Unit:mm]
1000
(RC Abutment)
High tide level A.P+2.500m
Embankment
Embankment
Steel Pipe Piles
Steel Pipe Piles
Fig.3 Bridge side view
3000
Ba
y
o
To
ky
10800
9800
1000
(RC Abutment) (Deck Plate Girder)
Uchikawa River
oka
ido
Lin
e
400
Girder
Fig.2 Cross section (Superstructure)
NINOHASHI-Br.
JR
T
Deck Plate
Studs can be
welded
[Conventional]
RC Slab
Composite
girder type
Girder
(Highway bridge
specifications method)
Back Side
(Stud Arrangement)
Studs
Abutment
Back Side
(Stud Arrangement)
Fig.4 Girder and abutment joint system
2 Verification of rigid connections using FE analysis:
The design of studs in highway bridge specifications divides the acting shear force evenly by the number of studs.
However, in this proposed type, the studs are arranged vertically on the abutment members. We considered that the
shear force at each stud position was different in the depth direction, so we verified it with FE analysis.
From analytical results, the shear force of all studs was lower than the design value. The shear force acting on the stud
was larger on the back side of the abutment than on the front one. In the vertical direction of the abutment, the shear
force increases toward the top for both the front side and back one of the abutment. In the horizontal direction, shear
forces are concentrated towards the inner studs close to the web position of the girder [2].
89
3 Verification of proposed model by loading experiment on actual scale:
To verifying the FE analysis results, we fabricated a test specimen of the actual
bridge model and carried out the loading test using the test equipment shown in
Fig.5. The result of applying load that exceeds the design load (66kN). we found
the following results.
(1) At the maximum load step of the rigid connection with stud arrangement
calculated the specifications for highway bridges, there was no gap occurred
between the top of RC abutment and the steel deck plate, and its bearing
capacity was about 3 times the design load.
(2) Measurements of the strain acting on the studs of each vertically arranged step
showed that the shear forces acting on them decreased with increasing
distance from the girder. This is the same tendency as the FE analysis.
Fig.5 Loading test (Full-size model)
Load
[kN]
12
320
300
A
a)Embedded Type
(with studs)
Rationalization
b)Embedded Type
(no stud)
B-B
B
912
924
12
a)Embedded Type
(with studs)
A-A
A
912
924
4 Proposal and verification of rational rigid connection:
As shown in the previous section, a loading test was performed on the
specimen Fig.6 a) for which the number of studs was calculated using the
formula in the highway bridge specifications. The results shown in Fig.7
indicate that the load carrying capacity exceeded the design load, but this was
not due to shear resistance between the steel girder and RC abutment, but to
buckling of the girder. Therefore, in order to reduced the number of studs and
verified a more rational joint method, we carried out two additional types.
They are embedded type, the stud omitted type [Fig. 6 b)], and the direct
connection type without the bent part of the girder [Fig. 6 c)].
250
Pmax=199
200
Pmax=208 kN
kN
320
300
B
c)Direct
Connection Type
(with studs)
150
b)Embedded Type
(no stud)
C-C
kN
C
362 12
Pmax=179
100
Design Load:66kN
660
300
50
C
0
-50
0
50
100
150
200
Displacement [mm]
c)Direct Connection Type
(with studs)
-50
Fig.7 Relationship between load and displacement
Fig.6 Proposed rational rigid connection
5 Conclusion:
(1) In a rigidly connected structure that integrates a steel deck girder and RC abutment, the load carrying capacity of the
proposed type in which the number of studs calculated in the specifications for highway bridges are arranged inside
the abutment has sufficient margin for the design load.
(2) Regarding steel deck girders, even if the arrangement of studs is simplified or omitted, small- and medium-sized
bridges have sufficient load carrying performance against the design load.
(3) When the girder height is lower, it was found that the bearing pressure between main girder and concrete is greater
than the composite effect of the steel girder and concrete by applying studs.
(4) In the future, it will be necessary to conduct repeated loading tests to verify the serviceability and durability for
fatigue resistance, respectively.
6 References:
[1] Japan Road Association: Specifications for highway bridges part4 substructures ver. 2017 (in Japanese).
[2] Tawaratani,Y. et al. :Mechanical properties of rigid connections between steel deck girders and concrete abutments
Bridge, Proceedings of The 77th Annual Conference of The Japan Society of Civil Engineers, pp. CS6-27- CS6-28,
September 2022 (in Japanese).
90
13th - Japanese-German Bridge Symposium, Osaka, Japan
Innovative developments of composite columns with high-strength steels
Michael Schäfers, M. Sc.*
Rudolf Röss, M.Sc.**
Prof. Dr.-Ing. Martin Mensinger***
*Technical University of Munich, Chair of Metal Structures, Germany, m.schaefers@tum.de
**Technical University of Munich, Chair of Metal Structures, Germany, r.roess@tum.de
***Technical University of Munich, Chair of Metal Structures, Germany, mensinger@tum.de
1 Introduction:
Regarding areas of high loads in building and bridge construction, composite construction with an efficient combination
of steel and concrete is a durable and extremely load-bearing solution. For composite columns, there is a significant
advantage in a fire-safe and slender construction method compared to conventional steel or concrete construction. The
slender construction method allows areas in multi-story construction to be rented out more efficiently or clearance
zones under bridges to be used. Furthermore, elegant bridge designs that highly blend with the surrounding can be
achieved.
In recent years, much research has been done in Germany on concrete-filled hollow section composite columns with
solid steel cores. Compared to cross-sections without steel cores, these show a high increase in ultimate load and a more
reliable behavior in case of fire. However, the disadvantages of solid sections are increased residual stresses and
reduced yield strengths with increasing diameters. Furthermore, solid-core profiles are only available with limited
diameters, which restricts the range of applications. In two research projects of the authors, novel column cross-sections
were developed to counteract these limitations. For the so-called bar bundle columns, high-strength reinforcing bars
with yield strengths of 670 MPa are inserted into a hollow section and grouted with mortar (Figure 1 a)). In the case of
the so-called laminated steel plate columns, individual plates with yield strengths of up to 960 MPa are flame-cut from
heavy plates, joined locally with the aid of bolts or welds to form a package of plates, and cast in a hollow section
(Figure 1 b)). By using these individual core cross-sections, residual stresses are reduced, and ultimate loads can be
increased. At the same time, high-strength steels are used for the core section, so the load-bearing capacities are further
increased.
a)
b)
Figure 1: Exemplary section design for a) bar bundle composite columns and b) laminated steel plate columns
The diameter of the columns can be adjusted individually to the required resistance, with the aid of the new construction
methods (see Figure 2). In Figure 2 b) buckling loads for laminated steel plate columns with buckling line b are shown
for an exemplary section of S355 and S960. In regions of low slenderness, the ultimate load can significantly be
increased by using high-strength steel. Since slenderness increases sharply with the columns’ lengths and elastic
buckling dominates the ultimate load, the use of high-strength steel is mainly suitable for nonslender components. Then,
the steel section significantly increases ultimate loads.
91
Figure 2: a) Plastic resistances of increased column diameters and b) Resistance to axial force for steel plate columns
(buckling line b)
2 Methodology:
The innovative approach of section design for composite columns with single components poses the hypothesis if a full
bond can be assumed between every component. Therefore, section designs have been developed by means of special
detailing to account for the individual challenges, each section entails. A set of 19 experimental buckling tests was
conducted. Based on the experimental tests, numerical models are developed, calibrated, and used to further examine
parametric influences on load-bearing behavior. As the section design is not covered by current standards, design
approaches are compared to the conducted behavior, and adjustments need to be elaborated, where necessary.
3 Results:
The presented design innovations for concrete-filled composite columns with steel cores have proven a significant
improvement in load-bearing compared to previous composite column designs. The use of high-strength steels of up to
960 MPa has enabled ductile failure and high load capacities. The effective bending stiffness could be evaluated by two
different approaches such as the moment-curvature-relation as well as the so-called Southwell’s method. While
evaluating the experimental results, eccentric core positions due to the manufacturing process have proven to have a
significant influence on the ultimate load.
Using the first approach, full sectional interaction at ambient conditions could reliably be derived for bar bundles as
core sections. A design proposal based on current European design standard EN 1994-1-1 could be developed for these
columns and confirmed with large parametric numerical studies. This includes a global imperfection of L/830 to
account for manufacturing imperfections of eccentric core positioning of up to 5 mm. Depending on the axial plastic
resistance of the chosen core configuration compared to the total axial resistance of the section the column can be
assigned to a buckling line of EN 1993-1-1. Here, either buckling line c or d can be regarded.
For the laminated steel plate columns with yield strengths of 960 MPa for the plates and yield strength of 890 MPa for
the tubular hollow section, the evaluation of the effective bending stiffness using the moment-curvature-relation showed
larger scattering. Southwell’s method, appeared as a more reliable and consistent approach to evaluate the behavior. The
limited connection between individual plates could be found of subsidiary influence, so the steel plate lamination can be
regarded as fully interacting. In contrast, current design approaches could not sufficiently describe the columns’
behavior under load. Former assumptions of the overall behavior such as a strain-limited normal-force bending-moment
interaction could be shown not to be applicable.
4 Conclusion:
Overall, a ductile failure of the developed columns could be observed and explained by the large bending capacity due
to the use of high-strength steel within the section. The high plastic strain capacity implies a deployment in combined
loading situations under compression and bending. Furthermore, considering the clamping effects of the connected
structures could minimize the buckling length and improve the higher load-bearing capacity for less slender structures.
Limitations of current design approaches were shown. Systematic investigations must be conducted using the developed
numerical models and further experimental tests. This will enable safe design proposals, which will be confirmed with
probabilistic evaluations.
92
13th - Japanese-German Bridge Symposium, Osaka, Japan
Precast Modular Bridge Structures
Current developments, pilot projects and experimental investigations
Univ. Prof. Dr.-Ing. Oliver Fischer 1), 2)
Dr.-Ing. Nicholas Schramm 2)
1)
Technical University of Munich, Chair of Concrete and Masonry Structures, Germany, oliver.fischer@tum.de
Büchting+Streit AG, Consulting Engineers VBI, Munich/Germany, nicholas.schramm@buechting-streit.de
2)
1 Introduction
The precast segmental bridge construction method has been well established worldwide for many years and in many
different ways as an efficient and fast bridge construction method. Additional ideas and options are currently emerging
towards resource-efficient modular construction methods. Besides that, industrial serial production facilitates both the rapid construction of new and the replacement of existing bridges with minimum disruption to traffic. Further, consistently
higher quality and accuracy can be ensured with concreting largely independent of weather conditions. In addition, new
technological possibilities and durable materials predestined for factory are available. In recent years, advanced modular
bridge construction methods with dry joints have been developed in Germany on the basis of various research and development activities, whereby the individual segments were not produced in a match-cast process but independently of each
other in a formwork. Hereby, the final segment geometry is achieved by subsequent high-precision grinding of the dry
joints. This concept was implemented with different detailed solutions in several pilot applications, which will be addressed in the paper, including the PTS bridges at Frankfurt/Main Airport as well as the airtight tubes made of high-performance concrete for the full-scale TUM Hyperloop prototype. Compared with conventional segmental construction, a
number of additional questions arise with the latter, e.g. with regard to the prestressing concept, load-bearing behaviour
and design assumptions, joint bearing capacity/tightness or durability aspects. The paper also reports on this and on corresponding theoretical, experimental and measurement investigations and discusses major results and findings.
2 Precast Segmental/Modular Bridge Construction
In precast segmental bridge construction, the structure is built by post-tensioning together precast elements with the bridge
being transversely divided into short segments. External bondless as well as internal longitudinal tendons are used. In
general, no continuous reinforcement is provided across the joints and in-situ concrete joints are only used in case of
longer continuous bridge girders to compensate for tolerances. Typically, the superstructure is composed of two to three
different types of precast segments, depending on the post-tensioning system used. Single as well as multiple-span bridges
have been built using PSB technology, mostly with box girder cross-sections. A significant acceleration of erection and
assembling of the segments can be achieved by using dry joints. Commonly, tendons crossing the joints are provided with
an additional circumferential protection, e.g. by plastic plug-in elements or rubber gaskets.
In Germany, the main bridge construction method with prefabrication has been the precast girder method with subsequent
in-situ concrete completion of the roadway slab. Until recently, only two road bridges were executed with the PSB
method. Presumably, international discussions in the early years of the PSB construction method, above all about the
quality of the glued joints and thus the reliability of the corrosion protection in the joint area, were the cause for not
pursuing the concept in Germany for quite some time. With the increased introduction of external prestressing around the
turn of the millennium, an attempt was made to reestablish PSB and to develop appropriate design and construction recommendations for segmental bridges with external tendons. However, own comparative analyses with conventional construction methods, among others in the course of preparing alternative proposals, showed that on the basis of these recommendations no technically feasible solutions were possible that could prevail in an economic competition. Therefore, the
very versatile segmental method has not yet been able to generally establish itself in Germany.
Currently, however, a clear development trend can be observed in research and engineering practice towards new types
of segmented and modular construction principles with a steadily increasing degree of automation of the entire production
chain – from structure to member/component level – to efficient element manufacturing and assembly principles as well
as process-integrated reinforcement concepts. Further, the current transformation towards modular concepts and industrial
production, which is also fostered by the German Ministry of Transportation, is being significantly promoted by digital
design and construction processes. In addition to new construction methods and enhanced detailing concepts, durable
high-performance materials such as corrosion-free reinforcement or UHPC are increasingly being used in modular construction approaches. One example is the first application of UHPC in a German railway bridge, where a deteriorated
existing deck was replaced – without any changes on both the substructure and track elevation – by a significantly more
slender prestressed UHPC precast superstructure.
In recent years, prefabricated components and modular system concepts have been increasingly used nationally/internationally and with a wide variety of approaches, especially to renew existing (damaged) structures in the shortest possible
core construction time. The utilization of prefabricated members shortens construction time on site and traffic disruption
93
is reduced to a minimum particularly in existing infrastructure measures. In Germany, too, various approaches to prefabrication and modularization have been proposed and implemented in pilot applications in recent years against the background of the extensive bridge renewal program required in a short period of time and with as little obstruction to traffic
as possible, especially for flyover structures and small/medium span bridges. Hereby, different approaches are taken, both
with regard to the degree of modularization and prefabrication, as well as with regard to the manufacturing and joining
technology of the elements.
2.1
Construction methods with full transverse segmentation of the superstructure
The following current application (PTS light rail connection bridges to Terminal 3 Frankfurt Airport, completion in 2022)
describes the renewed use of fully segmented prestressed bridges in Germany. For the first time, dry (non-profiled,
smooth) joints are used, which are prepared by precision grinding before assembly of the segments. In particular, because
of the high level of planarity and the accurate fit that can be achieved by grinding, as well as the residual compression of
the precompressed tensile zone in all SLS load combinations, it is possible to avoid gluing the joints here.
Depending on the span width (max. 40 m, total weight over 200 t), the statically determinate single-spans are manufactured in two or three elements, transported to the construction site and finally assembled there. The production of the
individual sections is not carried out with the match-cast method, but with formwork at the joints; the final geometry/accuracy is achieved by subsequent CNC precision grinding. Prestressing is provided by internal tendons with subsequent
bond (mortar grouting). Compared to conventional PC bridges, higher demands were placed on the compression under
SLS conditions. It had to be proven that all joints remain fully compressed under the rare load combination with sufficient
reserve (here: minimum compression 0.5 MPa, incl. prestress reduction by creep, shrinkage and relaxation effects) and
thus the superstructure acts like a monolithic girder. The ULS design approach was formulated on the basis of extensive
non-linear numerical analyses (as well as supplementary experimental results, e.g. on maximum friction that can be activated in joint contact surfaces), taking into account joint opening, the resulting profile deformation of the open “soft”
cross-sections and the combined effect of bending, shear and torsion. Due to the favorable joint arrangement in regions
with comparatively low shear forces, a maximum opening of up to 2/3 of the cross-sectional height h (i.e. remaining
compression zone xc > 1/3 h = xc,min) could be permitted under ultimate design loads. In addition to the SLS compression
reserve, the tendon duct diameters were enlarged in the joint areas in such a way that a full embedding in the cement grout
is ensured for the tendons. Further, all individual tendons are provided with a circumferential highly plastic sealing made
of butyl rubber applied to the joint contact surfaces. In order to check the tightness in advance, additional small-scale tests
were carried out on 20 x 20 cm grinded cubes clamped together and subjected to water/air pressure.
With the PTS project, an alternative segmental approach with subsequent CNC grinding of the joint surfaces (instead of
match-casting) was performed for the first time. Due to the achievable high precision and evenness of the contact joints,
this solution also offers a major potentials for the classic segmental construction method, makes it possible to dispense
with glued joints and also simplifies achievement of the exact geometry, the formwork technology and the manufacturing
process of the individual segments. In view of the additional effort involved, additional joint profiling that is common in
segmental bridge construction should be avoided as far as possible.
2.2
TUM Hyperloop: special type of a precast segmental bridge structure
Hyperloop refers to a ground-based, public transportation system operating at the speed of a commercial aircraft. For this,
so-called “Pods” are traveling in a near vacuum tube to reduce the air resistance. After founding a TUM Hyperloop
Program, a full-scale fully-functional 24 m technical prototype has been in development in combination with a detailed
concept analysis. An essential component of the concept is the precast segmental tube made of high-performance concrete
(C 100/115), which must ensure air-tightness in operation in addition to the requirements for load-bearing capacity and
serviceability of the bridge girders. For the prototype the Chair of Concrete and Masonry Structures of TUM was in
charge of the conceptional and detailed design of the tube and conducted several accompanying research activities, especially with regard to the load transfer mechanisms, the joint load-bearing capacity and the tightness.
Inside the tube (Ø 4.2 m), a high-precision guideway will be constructed, along which the Hyperloop vehicle will levitate.
The tube is manufactured in modular construction with circular tube segments (length 3.8 m each) with the 20 m prototype
span (overall tube length 24 m) being designed as a single-span girder prestressed longitudinally by bondless monostrands.
The contact between the individual segments is achieved by dry joints, whereby the segments are first produced in formwork and the final joint geometry and fitting accuracy is subsequently obtained by high-precision CNC grinding (similar
to the principle of the PTS). For the necessary airtight joint closure (underpressure compared to atmospheric air pressure
approximately 1 bar), special segment gaskets (similar to TBM tunneling) are set into the CNC-grinded prestressed joints.
In addition, a butyl tape is applied to the contact surfaces (between gasket and monostrands), which is squeezed into the
milled grooves of the ring segments.
To determinate the coefficient of friction, experimental investigations were performed. Furthermore permeation measurements and investigations on the airtightness of the gaskets were undertaken.
For the results of these tests and measurements please refer to the full paper. The same applies to the references.
94
13th - Japanese-German Bridge Symposium, Osaka, Japan
Analytical study on the mechanical behavior of the intermediate support
in the composite structure using bearing plates
Master’s Student Kenta Nakaoka *
Professor Takashi Yamaguchi**
Satoshi Kimura***
Taro Tonegawa****
* Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, si22234w@st.omu.ac.jp
** Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, yamaguti-t@omu.ac.jp
***Yokogawa NS Engineering Corp, sts.kimura@ynse.co.jp
****Yokogawa NS Engineering Corp, tru.tonegawa@ynse.co.jp
Abstract
In recent years, there has been a demand for rapid construction for the erection and replacement of bridges. The targeted
structure involves the installation of prefabricated members on the intermediate supports, transitioning from a simple
beam to a continuous beam. This structure significantly shortens the on-site construction period and enables erection even
in the restricted site space. However, the collapse process of the targeted structure remain unclear, and it remain unclear
whether the compressive force is adequately transmitted to poured concrete. In this study, finite element analysis (FEA)
was conducted to elucidate the collapse process and compressive stress characteristics. Two analysis cases were
considered: Type A with a girder height of 600mm and Type B with a girder height of 1000mm. The investigation covered
stress characteristics of the girder and concrete, crack formation, load distribution ratio of the reinforcing ribs, and
variation of the neutral axis in two girder types.
Keywords: Reinforcing ribs, Bearing plate, Composite structure, Nonlinear analysis
1
Introduction
Recently, large-scale deck replacement projects have been planned in Japan. However, during the deck replacement
process, it is necessary to remove the concrete from the main girders. In cases where the construction period is constrained,
the replacement of RC deck slabs becomes difficult. To address these challenges, a structure as shown in Fig. 1 is being
proposed. The Structure aims to shorten the on-site construction period by precasting main girder at a factory. The
compressive force from the lower flange is transmitted to the poured concrete through the bearing plates. However, the
collapse process of the targeted structure remain unclear, and the lower flange is discontinuous in the compression region.
Therefore, it remain unclear whether the compressive force is adequately transmitted to poured concrete through the
bearing plates and reinforcing ribs or not. In this study, finite element analysis (FEA) was conducted to elucidate the
collapse prosess and compressive stress characteristics. The differences in stress characteristics of ribs and concrete based
on different girder Types were also investigated.
2
FE analysis
The parameters are girder types: Type A (height is 1000mm) and Type B(height is 600mm). Fig. 2 (a)(b) show the side
view of Type A and Type B respectively. Type A has three ribs (upper, middle, and lower ribs), while Type B has one
rib. The analytical models are created by the DIANA 10.5 software. The boundary and loading conditions are shown in
Fig. 3. The support condition assumes that all cross-sections directly above the intermediate support are completely fixed.
As for the loading condition, a vertical downward enforced displacement was applied to the end of the cantilever beam
model. The structural elements are 4-node shell elements for the web plate, bottom steel plate, and cross girders, 8-node
solid elements for other steel members and concrete, and embedded reinforcement elements for the reinforcement bars.
An interface element has been applied between the upper flange and the splice plates to allow for relative displacement.
The steel material follows a ideal elastic-plastic model based on von Mises' yield condition. The crack model for concrete
is combination a plastic model and a multi-directional fixed crack model. The compressive behavior of concrete is
represented by the Drucker-Prager model, while the tensile behavior is represented by a softening model.
3
Analytical results and discussion
Fig. 4 shows relationship between the negative bending moment and vertical displacement for Type A and Type B. In
the same graph, the resistance bending moments for each failure mode are represented by horizontal lines. In Fig. 4(a)(b),
For both Type A and Type B, the analysis values significantly exceed the current design values D1 for failure modes. In
Fig. 4(c)(d), as for failure modes, the load stages and the sequence of occurrence are mostly consistent between analysis
values and design values D2 both Type A and Type B. This can be attributed to the effect of the reinforcing ribs and
bottom plate. The ratio of the sectional forces acting on lower flange and reinforcing ribs for Type A and Type B is shown
in the Fig. 5. In Type A, the ratio of compression forces acting on lower flange and each rib is as follows ; lower flange :
lower rib : middle rib : upper rib = 61% : 29% : 7% : 3%. In Type B, the ratio of compression forces is as follows; lower
flange : lower rib = 77% : 23%. The compression force acting on the lower rib is 29% in Type A and 23% in Type B. The
compression force acting on the lower rib with respect to the lower flange is larger in Type A than in Type B. It is
hypothesized that the position of lower rib with respect to the design neutral axis has an influence on the load distribution
ratio between the lower flange and the reinforcing rib. In Type A, middle rib and upper rib are also sharing the compression
force. The stress contour of longitudinal direction of bridge acting on the poured concrete at 3Md is shown in the Fig. 6.
95
For both Type A and B, the compressive stresses are distributed elliptically around the lower flange and more vertically
than assumed in the design.
4
1)
Conclusions
Both Type A and Type B, as for failure modes, the load stages and the sequence of occurrence are mostly consistent
between analysis values and design values D2 which consider some of the bottom steel plate and the reinforcing rib
to be effective.
Comparing the compression force acting on the lower rib in Type A and Type B, Type A shows a value of 29%, while
Type B shows a value of 23%. It is hypothesized that the position of lower rib with respect to the design neutral axis
has an influence on the load distribution ratio between the lower flange and the lower rib.
For both Type A and Type B, the compressive stresses are distributed elliptically around the lower flange and more
vertically than assumed in the design. The compression force was distributed by bearing plate, and smoothly
transmitted to the concrete.
2)
3)
Reinforcement
Splice plate
Concrete deck slab
Enforced displacement
Bottom steel plate
Design cross-section
Main girder
Reinforcing rib
Rigid plate
Poured-in-place
concrete
Cross girder
Bearing plates
Fix
Figure 3: The boundary and loading conditions
Pier
Rubber bearing
Fig. 1: Target structure on the intermediate support
解析値
設計値
破壊イベント名称
解析値
設計値
破壊イベント名称
解析値 設計値
設計値
破壊イベント名称
解析値
設計値
破壊イベント名称
解析値
設計値
破壊イベント名称
解析値
破壊イベント名称
破壊イベント名称
解析値
設計値
破壊イベント名称
解析値
設計値
破壊イベント名称
+
床版コンクリートのひび割れ
床版コンクリートのひび割れ
+ +++
床版コンクリートのひび割れ
床版コンクリートのひび割れ
床版コンクリートのひび割れ
床版コンクリートのひび割れ
+
床版コンクリートのひび割れ
+ +
床版コンクリートのひび割れ
床版コンクリートのひび割れ
Analysys
Design
Failure ivent
□
巻き立てコンクリートの支圧破壊
巻き立てコンクリートの支圧破壊
□□
巻き立てコンクリートの支圧破壊
巻き立てコンクリートの支圧破壊
□
巻き立てコンクリートの支圧破壊
巻き立てコンクリートの支圧破壊
□
巻き立てコンクリートの支圧破壊
巻き立てコンクリートの支圧破壊
□
巻き立てコンクリートの支圧破壊
Bearing failure of the poured cocrete
● □
上フランジの降伏
●
●
上フランジの降伏
●
上フランジの降伏
上フランジの降伏
● ●
上フランジの降伏
上フランジの降伏
●
Yieding of the upper flange
上フランジの降伏
上フランジの降伏
●
上フランジの降伏
鉄筋の降伏
××
鉄筋の降伏
鉄筋の降伏
×
鉄筋の降伏
Yieding of the reinforcement bars
鉄筋の降伏
×
鉄筋の降伏
×
鉄筋の降伏
鉄筋の降伏
×
△
下フランジの降伏
△△
下フランジの降伏
Yieding of the lower flange
下フランジの降伏
下フランジの降伏
下フランジの降伏
△
下フランジの降伏
△△
下フランジの降伏
下フランジの降伏
◇◇
摩擦接合継手のすべり
◇
摩擦接合継手のすべり
Slip in the frictional joint
◇
摩擦接合継手のすべり
摩擦接合継手のすべり
摩擦接合継手のすべり
摩擦接合継手のすべり
◇
摩擦接合継手のすべり
◇◇
摩擦接合継手のすべり
摩擦接合継手のすべり
Negative bending moment
（kN・m）
Negative bending moment
（kN・m）
5000
4000
3000
2000
1000
0
Ratio of the sectionnal force
（%）
(a) Type A girder（height of 1000mm） (b) Type B girder（height of 600mm）
Fig. 2: The side view of girder
3000
2500
2000
1500
1000
500
0
0
10
20
30
40
0
Vertical displacement（mm）
Negative bending moment
（kN・m）
Negative bending moment
（kN・m）
30
40
Type B（design value D1）
80%
Design value
60%
64.3%
57.0%
40%
20%
0%
TypeA
TypeB
Lower flange
lower rib
middle rib
upper rib
Fig. 5: The ratio of the sectional forces
3000
5000
4000
3000
2000
1000
2500
2000
1500
1000
500
0
0
0
10
20
30
40
Vertical displacement（mm）
(c)
20
Vertical displacement（mm）
Type A（design value D1） (b)
(a)
10
100%
0
10
20
30
40
Vertical displacement（mm）
CaseB
CaseA 1.5bf
CaseB 1.5bf
1.5bf
(b) Type
B
(a)
Type1.5bf
A CaseA
Type B（design value D2）
Fig. 6: The stress contour of the longitudinal
Fig. 4: Relationship between moment and displacement
direction of bridge acting on the poured concrete
Type A（design value D2） (d)
96
SESSION 3-B
Steel Structures 3
13th - Japanese-German Bridge Symposium, Osaka, Japan
Slip behavior between Cast iron deck module and Steel main girder
using High strength bolted frictional joints with Slotted hole
Yugo Shirai *
Prof. Takashi Yamaguchi **
Ryo Yamashita ***
Hironobu Tobinaga ****
* HINODE, Ltd., Saga, Japan, y-shirai@hinodesuido.co.jp
** Osaka Metropolitan University, Osaka, Japan, yamaguti-t@omu.ac.jp
*** HINODE, Ltd., Saga, Japan, r-yamashita@hinodesuido.co.jp
**** HINODE, Ltd., Saga, Japan, h-tobinaga@hinodesuido.co.jp
Keywords: cast iron deck, friction joint, slotted holes
1 Introduction
The application of cast iron deck to a bridge deck is explored. The cast iron deck was Produced by casting. The deck can
be of any shape without welding and thus reducing the likelihood of fatigue cracks[1]. The deck would be light,
approximately half the weight of an RC deck, to enhance the seismic resistance of a bridge. It is designed following the
Japanese design specifications for steel highway bridges. Incidentally, due to the difficulty of welding cast iron, the
connection between the cast iron deck and the main girder is achieved through high-strength bolt friction joints, using
filler plates for height adjustment and support members, as shown in Figure 1. In this study, the effect of the structural
type of the deck and support member between the deck and the girder on the slip behavior will be investigated by pushout tests conducted on test specimens designed for push-out tests of headed studs. Furthermore, the effects of casting draft
angles on slip behavior and the influence of enlarging bolt holes to accommodate construction considerations, such as
oversized holes or slotted holes, will be examined to understand their influence on slip behavior.
2 Experimental methodology
The tests were conducted using eight different specimens, with the shape of the support members and the bolt holes in
the cast iron deck serving as parameters. The support members were designed in three patterns labelled L-type, T-type ,
and T-type (improved connecting plate), as depicted in Figure 2. The bolt holes varied in design: tapered slotted holes
(26.5mm * 45 mm), untapered slotted holes (26.5mm * 45 mm), and oversized holes (26.5mm). The measured parameters
included the load and displacement of the testing machine, the relative displacement between the cast iron deck and the
main girder, the relative displacement between each joint, and bolt axial force. The loading device employed was a
2000kN testing machine manufactured by Shimadzu Corporation.
Cast Iron
Deck
Cast Iron
Deck Module
Main
Girder
Main
Girder
Figure 1: Joint structure between cast iron deck and main girder
(a) L-type
(b) T-type (improved connecting plate)
Figure 2: Experimental specimen
99
(Unit: mm)
(c) T-type
3 Results and discussions
3.1 Relaxation tests between the cast iron deck and the main girder
The residual axial force of bolts was measured immediately after tightening to 96 hours after tightening a high-strength
bolt friction bolted frictional joint between the cast iron deck and the main girder. The measurement results confirmed
that the axial force dropped significantly at 12 hours immediately after tightening and that the drop in axial force generally
subsided after 72 hours. No significant difference in the rate of axial force reduction was observed depending on whether
the cast iron deck was tapered or on the bolt hole (oversized hole or slotted hole).
3.2 Load-displacement relationship in the push-out tests
Focusing on the load-displacement relationship, the initial stiffness was higher in the order of L-type, improved inverted
T-type, and standard T-type. This suggests that the initial stiffness is influenced by the shape of the support members,
with structures joined via a connecting plate such as the inverted T-type exhibiting lower initial stiffness than the L-type.
Also, when the cast iron deck cross member has a tapered shape, it was confirmed that the initial stiffness decreases due
to the taper.
3.3 Slip coefficient
The specimens with L-shaped support members had a higher slip coefficient than those with inverted T-shaped support
members. For specimens featuring the inverted T-type support members, the improvement in the connecting plate
mitigated the relative displacement between the deck and the main girder, leading to an increased slip coefficient
compared to before the improvement. These findings suggest that enhancing shear stiffness effectively boosts the slip
coefficient of the cast-iron deck and main girder joint. No significant difference in the mean values of the slip coefficients
was observed between the oversized holes and the slotted holes, untapered in the transverse girders of cast-iron deck. This
implies that using slotted holes, untapered holes could achieve slip coefficients on par with those obtained using enlarged
holes.
3.4 Surface conditions of the joint interface after the experiment
The joint surface conditions were observed between the cast iron deck and the support members after the test. In all
specimens, the paint was destroyed within the range affected by the contact pressure from the bolt tightening. Slip marks
were also left around the lower bolt holes in both the upper and lower parts of the specimens, with the effect of the
presence or absence of taper being particularly notable. Paint film rupture was observed only around the bolt holes for
the specimens with a taper. In contrast, for those without a taper, slip marks were left in a downward direction (loading
direction) across the entire contact surface, confirming that the joint surfaces were in complete contact. Therefore, it is
believed that the specimens without a taper showed improved stiffness and slip resistance as a joint structure.
4 Conclusion
The relaxation characteristics of the elongated holes between the cast iron deck and the main girder showed a significant
decrease in axial force 12 hours after tightening, and the decrease in axial force generally converged after 72 hours.
Additionally, no significant differences in the axial force reduction rate were observed due to the presence or absence of
material taper and bolt-hole design (oversized or slotted holes). The initial stiffness calculated from the load-displacement
relationship was more considerable in L-type, improved inverted T-type, and standard inverted T-type. This indicates that
the type of support structure has a significant influence and methods that involve joining via a connecting plate result in
lower initial stiffness. Moreover, it was confirmed that the presence of a taper in the cross member of the cast iron deck
reduces initial stiffness. The slip coefficient was higher for specimens using L-type support than those with inverted Ttype support members. In the case of using inverted T-type support members, the improvement of the connecting plate
suppressed the relative displacement between the deck and the main girder, and the slip coefficient increased compared
to before the improvement. This suggests that enhancing shear stiffness effectively increases the slip coefficient in the
joint structure between the cast iron deck and the main girder.
5 References
[1] Hironobu TOBINAGA, Minoru MURAYAMA, Eiichiro SAEKI, Takashi TAMAKOSHI, Eiki YAMAGUCHI,
Chitoshi MIKI, FUNDAMENTAL STUDY OF APPLICATION OF SPHEROIDAL GRAPHITE CAST IRON
TO DECK SLAB FOR HIGHWAY BRIDGE, STEEL CONSTRUCTION ENGINEERING, 2017.
100
13th - Japanese-German Bridge Symposium, Osaka, Japan
Sufficient Choice of Steel Material for Bridge Bearings to Avoid Brittle Fracture
M.Eng. Natalie Hoyer *
Prof. Dr.-Ing. Bertram Kühn **
* University of Applied Sciences Mittelhessen, Chair of Steel, Composite and Bridge Construction, Gießen, Germany,
natalie.hoyer@bau.thm.de
** University of Applied Sciences Mittelhessen, Chair of Steel, Composite and Bridge Construction, Gießen, Germany,
bertram.kuehn@bau.thm.de
Abstract
Steel structures are usually designed assuming an upper shelf behavior of the steel toughness-temperature curve. In
order to account for the reduction of toughness properties in the ductile to brittle transition range, further safety checks
are required. Those are based on fracture mechanics considerations and allow a sufficient choice of steel material to
avoid brittle fracture.
In light of this context, proposals were already formulated in 2011 to regulate the suitable selection of steel grades for
bearing components of bridges. However, recent investigations showed that those proposals are no longer entirely up to
date. Consequently, a research project has been announced by the German Centre for Rail Traffic Research, with the
objective of defining a proposal for expanding the regulatory framework concerning appropriate material selection to
avoid brittle fracture in bridge bearings.
Keywords: bridge bearings; brittle fracture; choice of material; fatigue loads
1 Introduction
The Federal Republic of Germany is currently responsible for the duty to construct and maintain about 39,500 bridges
in the federal trunk road network. As a subsidiary of Deutsche Bahn AG, DB Netz AG also maintains over 25,000
railway bridges of various constructions throughout Germany [1].
To maintain life span and functionality of these structures, bearings are important structural elements regarding
stability. Those bearings ensure an appropriate load transfer from the superstructure to substructure, while also allowing
movements of the superstructure with low constraint.
2 Objective
The design of bridge bearings and bearing components according to DIN EN 1337 [2] in conjunction with the
corresponding parts of DIN EN 1993 often leads to large plate thicknesses for bridges with long spans. However, these
product thicknesses exceed the application limits of the national appendix of DIN EN 1993-2 [3] and the regulations
according to DIN EN 1993-1-10 [4] with regard to material toughness and through-thickness properties cannot be
applied without further modification. Therefore, these standards are not applicable in terms of the choice of material for
bearings without further information.
Against this background, a recommendation for the regulation of a suitable steel grade selection for bearing components
was developed in 2011 (see [5]) which has so far only been made available as a technical bulletin and has not yet been
fully incorporated into DB AG’s normative rules and regulations. In addition, recent findings have shown that some
bridge bearing components are exposed to high fatigue loads ([6], [7], [8]) which have to be taken into consideration in
the structural design, material selection and calculation. For this reason, the German Centre for Rail Traffic Research
called a research project with the aim to define a proposal to expand the standardization in order to implement a
sufficient choice of steel material for bridge bearings to avoid brittle fracture.
3 Working assumption
Relevant locations for the assessments to avoid brittle fracture of steel components of bearings were identified by
external member considerations and numerical simulations (Figure 1). On the one hand, standard details for bearing
components are to be investigated and on the other hand the so called hot spot stress locations where notch effects due
to geometrical detailing and due to welding are present and where tensile stresses occur can be identified.
The simulation models are then extended to include fracture mechanics considerations in order to derive appropriate
toughness requirements. In addition, these calculations are validated and calibrated via results of experimental
investigations including material analyses as well as tests on real bearing components, with the aim of inducing brittle
component failure.
101
For this purpose, a total of 15 tests are to be carried out. Those include five different steel components and two
different steel grades. An example of one of the specimens is shown in Figure 2. Then, a dynamic load is applied until
the crack has grown to a predetermined size. The specimen is cooled down with liquid nitrogen to an also
predetermined temperature at which brittle fracture failure is expected. Followed by the application of a static load
fracture of the components should occur. Subsequent parameter studies will serve as a basis to implement a new design
proposal in normative regulations taking fatigue loads into account.
Figure 1: FE-Simulation of the bottom component
(by ANSYS Workbench software)
Figure 2: Example of test specimen (milled top
component)
4 Acknowledgments
Sincerest gratitude to the German Centre for Rail Traffic Research for its support and financial funding in this research
project. Thanks are also due to all members of the working group accompanying the project, notably Maurer SE for
providing the test specimen, IWT Solutions AG for material analyses and the engineering firm Dr.-Ing. Markus Porsch
for analysis regarding fatigue effects.
5 References
[1] Kühn, B., Hoyer, N.: Geeignete Werkstoffwahl zur Vermeidung von Sprödbrüchen bei Brückenauflagern und
Festhaltekonstruktionen – Belastungen und Spezifika im Brückenbau. Interim Report AP2, DZSF research
project. Gießen, Germany, Dec. 2021.
[2] DIN EN 1337 Part 1 to 11: Structural bearings. Apr. 1998 to Jan. 2008.
[3] DIN EN 1993-2: Design of steel structures – Part 2: Steel Bridges. Dec. 2010
[4] DIN EN 1993-1-10: Design of steel structures – Part 1-10: Material toughness and through-thickness properties.
Dec. 2010.
[5] Feldmann, M., Eichler, B., Sedlacek, G. et al: JRC Scientific and Policy Reports: Choice of Steel Material for
Bridge Bearings to Avoid Brittle Fracture. Background documents in support to the implementation,
harmonization and further development of the Eurocodes. European Commission. Joint Research Centre, Ispra,
Jun. 2012.
[6] Hanswille, G., Heine, B., Porsch, M., Schmitz, C.: Lageraustausch an den Stabbogenbrücken im Zuge der
BAB A1 über den Dortmund-Ems-Kanal bei Ladbergen. Stahlbau Vol. 84, Issue 10, pp 721-743. Oct. 2015.
[7] Bewersdorff, S., Kina, J., Liebelt, M., Porsch, M., Schackenberg, R.: Entwicklung eines neuen Lagertyps für den
Eisenbahnbrückenbau. Stahlbau Vol. 88, Issue 2, pp 105-127. Feb. 2019.
[8] Porsch, M.: Gutachterliche Stellungnahme zur Frage der Normalkraftbeanspruchungen der Längsfesthaltungen
W/o – LR1/LR2 (EÜ Stockstadt, Überbau Nord-Ost, Widerlager Mainaschaff). Paderborn, 2020. unpublished
102
13th - Japanese-German Bridge Symposium, Osaka, Japan
MAURER Uplift Spherical Bearing
Dr.-Ing. Toshihisa Mano *
Dr.-Ing. Christian Braun **
Dr.-Ing. Torsten Ebert ***
* MAURER SE, Frankfurter Ring 193, 80807 Munich, Germany, t.mano@maurer.eu
** MAURER SE, Frankfurter Ring 193, 80807 Munich, Germany, c.braun@maurer.eu
*** MAURER SE, Frankfurter Ring 193, 80807 Munich, Germany, t.ebert@maurer.eu
Abstract:
MAURER MSM® Spherical bearing is one of the most commonly applied bridge bearings in Europe and in other
world. Thanks to MAURER’s special sliding material MSM® with high pressure resistance and very high durability,
MAURER Spherical bearings can be more compact in size in comparison with other types of bridge bearing and can
offer very reliable constant performance with long service life, independent of the climate condition. The further
development of the spherical bearing is its adaptation for the uplift force, which can be temporary loading case, such as
earthquake as well as permanent loading cases, while its high rotational capacity and mobility are kept.
Often the uplift force is accommodated by means of simple clamp-like construction on the side of the bearing. This
solution, for simple cases, fulfils its purpose satisfactorily, yet more concern should be raised for the cases, where the
rotation of the bridge is rather large, or impact uplift force is expected. MAURER Uplift Spherical Bearing can serve
more safety to the structure even for such cases.
Keywords: Sliding material MSM®, Uplift, Spherical bearing
1 Introduction:
MAURER Spherical bearing is one of the most applied bridge bearings in Europe and in other worlds. Thanks to
specially developed MAURER Sliding Material (MSM®) with high pressure resistance and very high durability,
MAURER Spherical bearings can be more compact in size in comparison with other types of bridge bearing and can
offer very reliable constant performance with long service life, independent of the climate condition. By applying the
lubrication for both sliding and rotation surfaces, very low friction is achieved, so that the reaction horizontal force and
the rotational moment to the structure is highly reduced. After many years of service, MAURER MSM® Spherical
bearing took a further development step, with which the uplift force can be accommodate, even with the permanent
uplift loading case, while its high rotational capacity and mobility are kept.
Often the uplift force is carried by means of an additional simple clamp-like construction on the side of the bearing.
This solution, for simple cases, fulfils its purpose satisfactorily, yet more concern should be raised for the cases, where
the rotation of the bridge is rather large, or impact uplift force is expected. MAURER Uplift Spherical Bearing can
serve more safety to the structure even for such cases.
In this paper, MAURER MSM® Uplift Spherical bearing is explained briefly.
2 MAURER Sliding Material MSM®:
For the sliding type of bearings, PTFE is most commonly used all over the world. This well-known sliding material has
served its purpose for a very long time, but it does not mean that there is no possibility of improvement. The wear
resistance, for instance, is increasingly more important, as the life of the modern structures is required to be as long as
50 years or even to such an extent of 100 years. The large bridges, such as, hanging bridges or cable stayed bridges, are
subjected to many cycles of large displacement due to the thermal expansion, traffic, and wind loading. That kind of
cyclic loading has to be enabled by the expansion joints and the bearings. MAURER Sliding Material MSM® was
proven to be wear-free after 50 km of accumulated sliding path under 60 MPa. Furthermore, MSM® has a higherpressure resistance and temperature resistance than PTFE. The higher-pressure resistance makes the sliding bearing
smaller and therefore more economical. Also, due to the high wear resistance, the life cycle cost of the sliding bearing
with MSM® is also reduced.
3 MAURER Spherical Bearing with MSM®:
The MAURER Spherical bearing can be categorized into three types, fixed type, guided sliding type and free sliding
type. The sliding and rotation mechanism are same for all types, and they are achieved by two lubricated sheets of
MSM®. Thanks to the advantageous properties of MSM® described in the previous chapter, the MAURER Spherical
bearing has a long service life and can be employed in a wide temperature range from -50 °C to 80 °C and its dimension
is compact. The quality of the MAURER spherical bearing with MSM® is assured with CE-marked according to
EN1337 and ETA-06/0131.
103
4 Conventional solution against uplift:
4.1 Common practice of restrainer construction:
Most of the existing bearing construction is not made against uplift force/movement, and therefore other types of
bearing have to be used or an external restrainer against uplift force has to be added, if the uplift force is expected to
happen in normal service condition. This is a common practice, and many similar examples are found. If the guide
bearing is designed with this clamp-like restrainers, the designer must assure that the uplift force is sufficiently carried
by these restrainers and at the same time the movement in the guide direction is not restrained. Also, the bridge bearing
often is subjected to the rotation. Considering these points, the clamp-like guide (restrainer) and the bearing should
possess a little bit of gap. If the width of the restrainer is not large or the bridge rotation is insignificant, the among of
gap is limited and then the restrainer comes to contact with a small uplift displacement. In such cases, this type of
restrainer can serve its purpose well. Otherwise, a larger uplift displacement is necessary for the bearing to come to
contact with the restrainer and it may cause the impact. As one possible solution to avoid such impact, a strip of sliding
material can fill the gap between the bearing and the restrainer. However, the rotation with respect to both longitudinal
and transversal axes causes the stress concentration towards edges. Combining this high edge pressure with sliding
movements during service life, an early wear damage can occur in the sliding material.
4.2 Actual case study of the bearing with clamp-like restrainer:
The Nipigon cable stayed bridge in Ontario Canada has opened for traffic in November 2015 but had to be closed
shortly after the opening due to the observed uplift. It was reported that the expansion joint of the west side of this bride
was lifted about 600 mm. In the early stage of the investigation, it was found that all 40 bolts, which connect the bearing
underneath this expansion joint and the superstructre, were torn. The bearing used here was so-called disc bearign with
the clamp-like guides and strips of teflon in the gap between the upper and lower restrainer in order to smoothen the
slidign movement. Further invetigation reveals an asymmetrical wear pattern of those sliding strips and also a sign of
low cycle fatigue tear of some of the outer bolts. Theese damage pattern and additonal FE analyses led to the possible
mechanis of the bearing failure with two main causes. First, the uplift force from the superstructre is transferred to the
bottom part of the bearing only throught the limited contact area or line, since the restrainer guide does not pursue the
rotation. This causes the cantilever effect and each time a line of bolts on one side carry much higher testion force.
Worse yet, the sole plate beinding causes another cantilever effect in the bridge transversal direction. The corner bolts
were highy loaded and they were torn or damaged first by low cycle fatigue most probably. Then line by line all bolts
were torn out. Second, the sliding strip of Teflon between the upper and lower restraint guide bars became increasingly
deformed and damaged at the both end due to the same rotation-induced cantilever effect. The gap between those
restraint guide bars increased with time and whenever the uplift force appeared, the impact effect became larger and
larger. This impact loading probably made the situation worse.
5 MAURER Uplift Spherical Bearing with MSM®:
The MAURER Uplift spherical bearing overcomes those possible drawbacks described above. In this bearing, the gap
between the bearing and restrainer is eliminated by MSM® rings but not only the gap is filled, the restrainer which is
integrated to the upper part of the bearing can follow the rotation of the bridge without constraint thanks to the internal
rotational calot. Therefore, no extreme edging effect on the sliding MSM® strip due to the rotation can occur. Also,
with this capability of the restrainer pursuing the rotation, and high wear resistance of MSM®, the widening of the gap
during the service can be prevented. Then no additional impact effect has to be concerned.
Because of these advantageous features, a number of MAURER uplift spherical bearings with MSM® have been
applied to the bridges in Germany and other countries. Also this is only one uplift bearing type which is technically
approved by DB (Deutsch Bahn = German Railway).
6 Conclusion
MAURER uplift spherical bearing is the solution even for the loading combination of uplift and rotation/displacement.
The integrated restrainer can pursue the rotation without restriction, and hence no extreme force distribution for the
connecting bolts and guide sliding material. Also, MSM® has a very high wear-resistance, thanks to which the
longevity of the bearing is assured, even if the large number of sliding and rotation cycles are expected.
7 References
[1] ETA06/0131: MAURER MSM® Spherical and Cylindrical Bearing, April 2019.
[2] DIN EN1337-2: Structural bearings – Part 2: Sliding elements. July, 2004.
104
13th - Japanese-German Bridge Symposium, Osaka, Japan
Significance of Treating Initial Imperfection in FE Simulation for
Compressive Behavior of Welded Steel Structural Members
Ph. D. Student, Master of Engineering. Yuxuan CHENG*
Master of Engineering. Shuhei NOZAWA**
Univ. Assoc. Prof. Dr. Eng. Mikihito HIROHATA***
* Osaka University, Department of Civil Engineering, Suita, Japan, y-cheng@civil.eng.osaka-u.ac.jp
** Osaka University, Department of Civil Engineering, Suita, Japan, s-nozawa@civil.eng.osaka-u.ac.jp
*** Osaka University, Department of Civil Engineering, Suita, Japan, hirohata@civil.eng.osaka-u.ac.jp
Keywords: Finite element analysis, welding residual stress, initial imperfection, buckling, load-carrying
performance
1 Introduction
For evaluating the influence of initial imperfections on the load-bearing capacities of steel structural members, a lot of
numerical analysis works have been conducted [1]. In numerical analysis using FE simulation, geometric and material
initial imperfections are normally treated independently. However, during the actual welding process in steel structural
member assembly, the welding deformation and the residual stress are always correlated with each other. Treating them
separately in numerical analysis may result in unbalanced forces, ultimately reducing simulation accuracy. To this
problem, the authors have investigated the possibility of continuous analysis through the welding process and the
loading process for high-accurate simulation without the unbalanced state by considering the perfectly related initial
imperfections.
In previous study [2], it was shown that the compression behavior of a cruciform column can be accurately reproduced
by using shell elements. However, the effect of improving the analysis accuracy on welding process has not been
verified. Furthermore, whether this method demonstrates superiority over traditional analysis method in achieving
higher precision predictions when independently introducing initial imperfections remains to be established.
Therefore, this paper aimed to examine the significance of treating initial imperfections in FE simulation for the
compressive behavior of welded steel structural members. A series of numerical analyses were conducted on welded
steel cruciform columns based on the proposed continuous simulation method for welding and loading processes. The
welding process of the cruciform column was simulated for obtaining the welding deformation and residual stress. After
that, the compressive loading process was continuously simulated.
2 Experimental procedure
2.1 Specimen and measurement points
The steel cruciform columns shown in Figure 1(a) were prepared as the specimens for this study. The plate material was
SM400B. The thicknesses of the flanges, web, and stiffeners were 19, 6, and 9 mm, respectively. The plates were joined
via gas metal arc welding using a welding wire specified by JIS YGW12.
The different measurement points are shown in Figure 1(b). Temperature history, out-of-plane deformation and residual
stress measurements are obtained from these locations.
2.2 Compressive loading experiment
Compressive loading experiments were conducted on two specimens. A static loading machine was used to apply a
monotonic compressive load along the axial direction of the specimen. Displacement in the loading direction was
measured at two diagonal positions between the edges of the upper and lower flanges.
3 Analysis models and analysis conditions
3.1 Analysis models and condition
The commercial software ABAQUS 6.14 was used for the thermal elastic–plastic analysis. The coupled temperature
and displacement functions were selected to simulate the welding process. By evaluating the temperature history from
thermocouples, the consistency between welding experiment and thermal elastic–plastic analysis was confirmed.
Furthermore, the other results of two analytical models with the different initial imperfection introducd methods were
compared.
The first analytical model is continuous model that to replicate the whole welding process by thermal elastic–plastic
analysis. Another is discontinuous model, which without simulating the welding process, but has the residual stress
distribution introduced into the model dependently. Figure 2 shows the analysis model assembled using the four nodes
of shell elements, also shows the simulation method was proposed using shell elements to weld T-section joints [3].
105
(a) Experimental specimen
(b) Measurement location on specimen
Figure 1: Experimental specimen and measurement location
Figure 2 : Compressive analysis setup
3.2 Residual stress introduced method on discontinuous model
By comparing the stress distribution between the continuous model and the discontinuous model, the initial stresses of
the two models were approximately consistent prior to the application of compressive loading can be observed.
4 Results and discussion
4.1 Out-of-plane deformation and strain
The deformation modes of each model obtained by the analysis when
both models almost reached at maximum load were obtained. From this
point of view, it can be said that the results of the continuous model are
most consistent with the experimental results on load-deformation curve
and also, the strain.
4.2 Load-vertical displacement relationship
Figure 3 shows the load-vertical displacement relationship of specimen
and each model obtained by experiments and analyses. For each model,
the elastic stiffness is generally consistent with the experimental value.
The maximun loading capacity of continuous model and discontinuous
model were approximately 813 kN and 915 kN, respectively.
Figure 3: Load-vertical displacement
relationship
4.3 Yield state
In order to confirm the difference in the load-bearing behavior of each model, the yield state of the vertical plate of each
model when the stiffness of the models sharply decreases (about 700 kN) was obtained.
5 Conclusions
A series of simulations on compressive behavior of cruciform using shell elements were carried out in cases where only
welding deformation and residual stress obtained by thermal elastic-plastic analysis was set as the initial condition.
By continuously simulating the welding process and compressive behavior, the out-of-plane deformation, strain, and the
load-displacement relationship during the compressive experiment could be accurately replicated compared to the case
where welding deformation and residual stress were introduced independently as initial imperfections. This might be
because not only all components of the deformation and residual stress generated in the welding process should satisfy
the self-balance state, but also the accumulated strain state during the welding process should be considered.
6 References
[1] Ueda, Y., Yasukawa, W., Yao, T., Ikegami, H., Ohminami, R. Effects of welding residual stresses and initial
deflection on rigidity and strength of square plates subjected to compression (Report II), Trans. JWRI 6 (1), pp.
33–38, 1977.
[2] Cheng, Y., Nozawa, S., Hirohata, M. High-accurate FE simulation on compressive behavior of steel cruciform
column with welding imperfection. Finite Elements in Analysis and Design, 221, 103960. 2023.
[3] Hirohata, M., Nozawa, S., Tokumaru, Y., Verification of FEM simulation by using shell elements for fillet
welding process. International Journal on Interactive Design and Manufacturing (IJIDeM), 16, pp. 1-13, 2005.
106
13th - Japanese-German Bridge Symposium, Osaka, Japan
Evaluation of Load Capacity of Temporary Bridges Using End-plate Connections
under Pure Bending Moments: A Proposal for a Simplified Calculation
Ph.D. Student Ruoxi LI *
Ph.D. Student Yu CHEN **
Isao MATSUDA ***
Hirotoshi AZUMA ****
Professor Takashi YAMAGUCHI *****
* Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, ruoxi_li@outlook.com
** Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, chenyu@omu.ac.jp
***Hirose &CO., LTD.,, Tokyo, Japan, i-matsuda@hirose-net.co.jp
**** Komaihaltec Inc., Tokyo, Japan, h.azuma@komaihaltec.co.jp
***** Department of Urban Design and Engineering, Osaka Metropolitan University, Japan, yamaguti-t@omu.ac.jp
Abstract:
Temporary bridges are crucial for disaster rescue and damage mitigation. In these bridges, high-strength bolts are
commonly used in end-plate connections to resist loads through axial forces efficiently, reducing construction time and
bolt usage [1]. To address the lack of a simplified load capacity estimation method for connections in temporary bridge
design and streamline the selection process for main girder cross-sections, we conducted a FEM analysis. This analysis
involved subjecting main girders to bending moments and monitoring the axial forces in the bolts, allowing us to calculate
load distribution ratios for each bolt and understand load-sharing characteristics. Based on these findings, we developed
a simplified method for accurately estimating the load capacity of end-plate connections under bending moments,
considering the cross-section of main girders. As a result, the proposed simplified calculation method showed a maximum
error of 8% compared to the FEM analysis. Additionally, utilizing a part model reduced the error to 3%, enabling its
application in parametric analysis.
Keywords: Temporary Bridge, End-Plate Connection, High Strength Bolted Tension Connection, Load Capacity
1 Introduction
The high-strength bolted tension connection is an effective method for
bridge connections, reducing the required bolts and enabling costeffective construction. As a result, it is commonly used for end-plate
connections in temporary bridges, which allows for shorter installation
periods, improving work efficiency and expediting temporary
operations (Figure 1). However, using FEM analysis for modelling
and computations has prolonged the design process for temporary
bridges because the current design lacks a simplified load capacity
evaluation method. To address this issue, we proposed a simplified
calculation method based on cross-sections and bolt arrangements to
quickly assess the load capacity of the main girders in temporary
bridges. This approach allows for a confident evaluation and
optimization of the end-plate connection's ultimate load under
different conditions.
Figure 1: Temporary bridge with an endplate connection
2 Analysis model and analysis cases
2.1 Target main girder
This study introduces an improved connection structure with an endplate overhanging, enhancing the previous research (Figure 2). The
connection was extended with a 500mm horizontal notch, a 100mm
overhanging, and an additional row of bolts. The I-girder had a 32mm
thick end plate and a span length of 12,000mm. The used bolts had an
axial force of 238kN for M24(S10T) designed, while the allowable
bolt axial force was calculated as 308kN.
Figure 2: The parts that make up the end2.2 Analysis model and analysis cases
plate connection.
We used Abaqus/Standard 2020 for 3D elastic-plastic FEM analysis.
To consider symmetry, a 1/2 model was used. For modelling, we used 3- or 4-node shell elements for the web and top
flange and 8-node solid elements for other parts. The bolt sets were cylinders, as shown in Figures 3 and 4, with the thread
part's cross-section area being 369 mm2 to simulate the bolt axial force increase accurately.
107
Figure 3: Analytical model and boundary conditions (Unit: mm)
Figure 4: Part model and bolt set of FEM analysis.
Figure 6: Comparison between the estimated
values and the simulated results of full models.
We selected four main girder cross-sections of temporary bridges to validate the evaluative calculation for the
serviceability limit state's load capacity. In addition to full models, we created part models that replicated the structural
details near the bottom flange of the girder connection, specifically including the lower three rows of bolts, as shown in
Figure 4. These part models reduced the number of elements and shortened the analysis time. Furthermore, we applied
three boundary conditions to compare their effects on stress transfer.
Figure 5: Bolts load-sharing ratio and effective cross-section area.
3 Simulation results
3.1 Original model and proposal of simplified calculation
We obtained each bolt's axial force increase curves based on the simulation results, allowing us to determine the bolt
load-sharing ratio at the ultimate load capacity. The bolt load-sharing ratio reflects the resistance from the bottom flange
and the web's acting forces. In the original model, the load-sharing ratio chart indicates that the first and second-row bolts
resist more than 95% of the applied forces when reaching the allowable bolt limit (308 kN), as shown in Figure 5. We
then calculate the effective cross-section area, considering the full cross-section of the bottom flange and the web crosssection between the second and third-row bolts. Based on this information, we propose a simplified evaluative calculation
to assess the load capacity of girder connections as Eq.(1).
𝑀0 =
𝑛∗ × 𝑁𝑎
𝐼𝑧
×
𝐴𝑑𝑓 + 𝐴𝑤 𝑦
(1)
where,
𝑛∗ : Equivalent number of resisting bolts
𝑁𝑎 : Allowable bolt axial force
𝐴𝑑𝑓 : Area of the bottom flange
𝐴𝑤 : Effective cross-section area of the web
𝐼𝑧 : Moment of inertia of the girder cross-section
y : Distance from the bottom side of the flange to the neutral axis
3.2 Simulation results
Figure 6 shows the estimated values and the analysis results for four main girder cross-sections. The error between the
estimated values and the simulation results is within 8%. The simulation results of the part model reduced the calculation
error to 3.04%, indicating its applicability for sensitivity analysis and further refinement of the evaluative calculation.
However, the current part model could be more conducive to evaluating the influence of the web's width-to-thickness
ratio, necessitating a reevaluation of the model's constituent parts.
4 Conclusion and future work
Under pure bending conditions, the variation in the axial force of tension bolts effectively shows the end-plate
connection's resistance to external loads. In the original model, the tension in each bolt indicates how well the end-plate
connection can withstand the bending force, and the sum of these values approaches 5.19 times the allowable bolt axial
force. By studying the changes in bolt tension and their underlying causes, we can further refine the resistance values of
the bolts under different conditions and gain insights into the maximum load that the end-plate connection can support
before reaching its serviceability limit state. The proposed simplified calculation method for estimating the load capacity
of end-plate connections, based on different cross-sections of main girder and bolt arrangements, proved to be accurate,
with the error between the estimated values and the analysis results being within 8%. By utilizing part models and
conducting sensitivity analysis, we were able to refine the evaluation calculation and reduce the calculation error further.
5 References
[1] Y Sugimoto, Y Mineyama, T Yamaguchi, Study on bending strength and evaluation of bolt axial force for the double
end plate connection with horizontal ribs, steel construction engineering, 27, 106, p. 106_61-106_74, 2020
108
SESSION 4-A
Composite Structures 2
13th - Japanese-German Bridge Symposium, Osaka, Japan
Bond Behavior of CFRP plates with tapered ends for steel structure reinforcement
Dipl.-Ing. Sakurai Shunta *
Dipl.-Ing. Hidekuma Yuya **
Univ. Prof. Dr.-Ing. Ohgaki Kazuo ***
Univ. Prof. Dr.-Ing. Okui Yoshiaki ****
* NIPPON STEEL Chemical & Material Co.,Ltd, Japan, sakurai.sh.msu@nscm.nipponsteel.com
** NIPPON STEEL Chemical & Material Co.,Ltd, Japan, hidekuma.3ae.yuya@nscm.nipponsteel.com
*** INSTITUTE OF TECHNOLOGISTS, Japan, ohgaki@iot.ac.jp
**** Saitama University, Japan, okui@mail.saitama-u.ac.jp
Abstract
In some cases, the load-bearing capacity of existing steel structures is insufficient due to the aging of girders, an increase
in the design load, and so on. The carbon fiber-sheet bonding method is applied to reinforce steel girders. A design and
construction manual and guidelines have been published on this method[1],[2]. Since the carbon fiber-sheet is lightweight,
the increase in dead load after reinforcement is insignificant. In addition, since carbon-fiber sheets are bonded to steel
girders using resin, there is no need to drill bolt holes, and accordingly, no damage to the girders occurs. However,
reinforcement with carbon fiber sheets requires the impregnation of each layer on site. This demands a large amount of
labor when the number of layers is large. Therefore, the authors investigated a reinforcement method using CFRP
prefabricated plate, which is easy to install on site. In this method, the amount of reinforcement is set according to the
thickness of the CFRP plate, and only a single bonding operation is required on site, thus saving labor. However, CFRP
plate reinforcement is likely to strip due to the stress concentration caused by the sudden change in stiffness at the edge
of a CFRP plate.
Previous studies have reported that taper processing on the end of CFRP plates improves stripping strength. However,
there are many unknowns, such as the optimum taper ratio and the stress reduction rate of the adhesive resin due to the
taper.
In this study, FEM analysis was performed to determine the optimum taper ratio of the end of the CFRP plate and to
confirm the effect of taper processing on reducing stress in the adhesive resin. Figure 1 shows Double Patch Tensile tests
specimen considered in this study. Based on the analysis results, double patch tensile tests were conducted on steel plate
specimens with CFRP plates with tapered ends bonded to both sides to investigate the stripping strength of CFRP plates
with tapered ends.
As a result, the tapered shape of the CFRP plate ends was determined to be 1/3 by parametric FEM analysis. It was
confirmed from the tensile tests that the tapered ends improved the stripping limit of the CFRP plate.The carbon fiber
sheet bonding method is applied to reinforce steel girders. Since the carbon fiber sheet is lightweight, the increase in dead
load after reinforcement is insignificant. However, reinforcement with carbon fiber sheets requires the impregnation of
each layer on site. This demands a large amount of labor when the number of layers is large. Therefore, the authors
investigated a carbon fiber reinforced polymer (CFRP) prefabricated plate for reinforcement of steel girders, which is
easy to install on site. In this method, the amount of reinforcement is decided according to the thickness of the CFRP
plate, and only a single bonding operation is required on site, thus saving labor. However, CFRP plate reinforcement is
likely to strip. In this study, double patch tensile tests and FEM analysis were conducted on steel plate specimens
reinforced with CFRP plates with tapered ends, which intended to reduce the stress concentration at the ends of CFRP
plates. The tapered shape of the CFRP plate ends was determined to be 1/3 by parametric FEM analysis as a result. It was
confirmed from the tensile tests that the tapered ends improved the stripping limit of the CFRP plate.
Keywords: steel, CFRP plate, taper, bond, reinforcement
P
P
CFRP Plate
Steel Plate
Adhisive Resin
Tapered
Figure 1: Double patch tensile tests specimen
1 References
[1] Nippon Expressway Research Institute Company Limited: Design and Construction Manual for Repair and
Strengthening Steel Structures with Carbon Fiber Sheets, July. 2020 in Japanese.
[2] JSCE :Guidelines for Repair and Strengthing of Structures using Externally Bonded FRP, July, 2018 in Japanese.
111
112
13th - Japanese-German Bridge Symposium, Osaka, Japan
The effect of prestressing on the shear capacity of post-tensioned concrete beams
Sebastian Lamatsch, M.Sc. *
Univ. Prof. Dr.-Ing. Oliver Fischer *
* Technische Universität München, Chair of Concrete Structures and masonry, Germany, sebastian.lamatsch@tum.de
* Technische Universität München, Chair of Concrete Structures and masonry, Germany, oliver.fischer@tum.de
1 Introduction:
As traffic loads continue to increase, bridges today are being proactively designed to be robust and prepared for future
traffic scenarios. However, the majority of existing post-tensioned concrete bridges in Germany were built from 1965 1985 and could not be dimensioned for the subsequent extreme increase in heavy traffic. Recalculations of existing
bridges with new traffic load models often result in significant deficiencies in the shear capacity check.
Developing new analytical models to better describe shear failure, in recent years, few tests on continuous post-tensioned
beams were carried out. To cover realistic construction details and extend the test data set, the tests' results on eight posttensioned beams with a low amount of shear and longitudinal reinforcement, different prestressing levels, and crosssections subjected to a single point load are presented. The ultimate shear capacity is evaluated using analytical models.
2 Experimental program
A series of eight tests on substructures of continuous post-tensioned beams is presented in this section. The scope of the
series was to investigate the shear capacity of realistic (h = 1.2 m) continuous post-tensioned beams at the inner support
with respect to the effects of the prestressing level. The dimensions and reinforcement layout of the beams are shown in
Figure 1a). For continuous post-tensioned bridges, the critical area for shear failure is often at the inner support. The
internal forces, the extracted substructure element and the chosen M/V ratio, are shown in Figure 1b). An innovative test
rig at the Technical University of Munich is used to apply the internal forces of the reference beam.
(a)
(b)
Figure 1: Derivation of the substructure element: a) Fictitious continuous bridge girder with internal forces and
extracted substructure element; b) Dimensions, reinforcement and tendon profile of beam elements.
In addition to conventional measurements of forces at the loading plate and the tendon anchorage, vertical deformations
of the beam, and horizontal joint opening at the load plate, an extensive measurement setup was used. Distributed fiber
optical sensors (DFOS) were mounted on the reinforcement and in the duct to investigate strains in both the reinforcement
and the tendon. In direct contact with the grout adjacent to the strand, an assessment of the crack-induced strain increase
is possible.
2.1 Main results
In all specimens, shear failure occurred despite the low amount of longitudinal reinforcement with simultaneous rupture
of almost all the stirrups in the test area. As the load increases, shear cracks develop in all beams, starting from flexural
cracks. Single T-beams, however, had shear tensile cracks. The crack kinematics were mainly focused on one critical
flexural shear crack, and failure occurred by fracture of the compression zone near the inner support.
113
An example of the crack pattern after failure can be seen in Figure 2a). As also illustrated, the DFOS within the duct
shows good agreement with the crack pattern. Occurring strains are therefore plotted orthogonal to the duct axis. The
measured strains increase strongly with the development of new shear cracks that cross the tendon and fade as the higher
stresses are introduced into the concrete until a new crack area interferes. The measured crack widths of the critical shear
crack at the centroidal axis are shown in Figure 2b) and match the highest strain peaks of the DFOS measurements with
the highest crack width as well.
a)
R-L5-S1.7
b)
R-L5-S3.1
Figure 2: Main test results: a) Crack pattern after failure and strain profile in the duct; b) Critical crack width.
2.2 Detailed evaluation of the effect of prestressing and shear strength models
Due to the low amount of longitudinal reinforcement, the stiffness of the tension chord is strongly influenced by the
number of strands and the initial stress in the strands. A higher level of prestressing results in significantly higher failure
loads and affects cracking as well as the direction of principal stresses and the remaining load delta after the first shear
crack occurs until the maximum shear force is reached.
The ultimate strengths of all the tests presented are therefore shown in Figure 3a). For better comparison, the ultimate
shear force is adjusted for the effects of concrete strength, and shear area and trend lines are plotted for both cross-sections.
Both cross sections show the same increase in ultimate shear strength as a function of the initial prestressing level.
The measured shear strength of the presented test specimens is compared to the prediction of design codes in order to
investigate their accuracy and the effect of the prestressing (see Figure 3b). The selected design codes are Eurocode 2, fib
Model Code 2010 with its highest Level of Approximation and the draft of the new recalculation guideline in Germany
BEM-ING/T2. The shear design models considered give conservative results for all tests. Eurocode and Model Code give
similar results, despite their different theoretical backgrounds. The calculations based on the draft of BEM-ING/T2 predict
the maximum shear capacity best with an average of 1.35 and a very low coefficient of variation. The shear strength
model is based on a truss model with an additive concrete contribution and was only proposed for recalculations.
(a)
(b)
Figure 3: Detailed investigations on the shear strength: a) Adjusted shear capacity compared to the prestressing,
b) Statistical evaluation of the shear strength determined with the considered analytical models.
114
13th - Japanese-German Bridge Symposium, Osaka, Japan
Crack prevention methods of pre-flexed beam prefabricated by segmental method
Hiroaki FUJIBAYASHI*
Naoki NORO**
Shota TSUZI***
Prof. Dr. Eng. Osamu OHYAMA****
Prof. Dr. Eng. Shigeyuki MATSUI*****
*Kawada Industry, Inc. Osaka, Japan, h.fujibayashi@kawada.co.jp
**Kawada Industry, Inc. Osaka, Japan, naoki.noro@kawada.co.jp
***Kawada Industry, Inc. Osaka, Japan, shota.tsuzi@kawada.co.jp
****Osaka Institute of Technology, Osaka, Japan, osamu.oyama@oit.ac.jp
*****Osaka Institute of Technology, Osaka, Japan, shigeyuki.matsui@oit.ac.jp
Abstract:
A pre-flexed beam (hereafter referred to as PREBEAM) is a double composite girder bridge which consists of steel girder,
prestressed lower flange concrete and floor slab concrete.
In recent years, it has been observed that cracks occurred in the lower flange concrete on the segment division range of
the beam as the span length increases. In this paper, we investigated the mechanism of crack occurrence in the lower
flange concrete in the segmented construction method of PREBEAM and proposed crack prevention measures. In addition
to the method of attaching a cushioning material to the steel flange side, there are two methods that involve dividing the
shear connectors of rectangular bars at the block ends or combining them with headed studs. The effectiveness of these
crack prevention measures was verified through finite element analysis and static load tests on columnar specimen.
Keywords: pre-flexed beam, segmental prefabrication method, crack mechanism,
crack prevention method
1 Introduction
Figure 1 shows the structural overview of PREBEAM [1]. This type of
girder applies sustained four-point bending load to a steel I-beam,
concrete is casted around the lower flange, and after the concrete hardens,
the bending load is released for introducing prestress. The segmented
construction method [2] was developed to facilitate on-site construction,
reduce labor, and simplify the transportation and installation of long-span
girders. PREBEAM, which has introduced prestress to the lower flange
concrete at the factory, is divided to 2 or 3 blocks for transportation and
then reconnected on-site before the floor slab is constructed.
In recent years, there are several cases of cracking occurring at the block
ends of the lower flange concrete where is discontinuous at the connection
points of the steel girders during prestressing by adopting the segmented
construction method. The crack pattern on the underside of the lower
Figure 1: Structural overview of PREBEAM
flange concrete is shown in Figure 2.
Therefore, we carried out the finite element analysis at the block end of
the PREBEAM where cracks occurred. In the analysis, the released
bending moment was applied at the end of the model as a load. Based on
the FE analysis results, we made clear the causes of the two types of cracks
could be inferred.
The cause of Type 1 cracks is shown in Figure 3, it illustrates the
deformation behaviour of the lower flange due to the Poisson effect during
release. When the tensile stress in the steel girder is released, prestress is
introduced to the concrete through the rectangular bar connectors. As the
Figure 2: Condition crack occurrence
tensile stress in the steel girder decreases, the Poisson effect causes
expansion deformation in the steel cross-section. This leads to the
Poisson effect causes
expansion deformation
behaviour where the steel girder pushes and expands the concrete.
cracks
The cause of Type 2 cracks is shown in Figure 4, it illustrates the
mechanism of secondary stress generation due to the compressive
stress from the rectangular bar connectors. The bearing pressure
from the rectangular bar connectors at the block ends acts at the
position shifted from the centre of the concrete thickness. As a
result, eccentric bending in the axial direction of the bridge,
115
large
small
T ensile action of concrete
Block End section
General section
Figure 3: Tensile action with pushing out behavior
Lower flange concrete
Rectangular bar connector
Steel girder flange
causing the concrete section to be pushed downward vertically and
leading to deformation in the delamination direction. It is believed
that this delamination deformation leads to the generation of tensile
stress in the axial direction of the bridge.
Vertical direction
2 Countermeasures of cracks and verification experimentation
Type 1 cracking measures address the expansion strain in the
perpendicular direction of the bridge axis caused by the Poisson
effect resulting from the release of tensile stress in the steel girders.
As a mitigation measure, we proposed that the installation
cushioning material on the sides of the steel girder flange and the
rectangular bar connectors at the block ends (Solution A). By
adopting cushioning material, it is possible to absorb the differential
deformation between the steel girder and the concrete.
e
Bridge axis direction
3 Conclusions
This paper elucidates the mechanism of crack formation in the
segmented construction method of precast beams, validate the
effectiveness of preventive measures through FE analysis and
full-scale static loading tests, respectively. The conclusion
obtained from this study are summarized as follows:
1) For the mechanism of Type 1 crack formation in the
concrete along the steel girder lower flange, Solution A,
which involves attaching a cushioning material to the side
of the steel girder flange and rectangular bar connectors,
was validated to prevent the crack formation.
Figure 5: Condition of attached
cushioning material
100 1 0 0
255
Solution B
160
60
160
10
Solution C
40
320
150
Strain gauge
1100
150
150
150
150
10
175
15 25
60
420
1000
On the other hand, to verify the effectiveness of mitigation Solution
B and C, axial compression tests were conducted using columnar
specimen in Figure 6 that only modelled the end portion of the block.
The comparison of the effectiveness of the countermeasures is
presented in Table 1.
Deformation at underside of concrete
Figure 4: Bending action by rectangular
bar connector
Type 2 cracking measures address the fact that bearing pressure
applied eccentrically from the rectangular bar connectors deform the
concrete cross section in the delamination direction and generates
large tensile stresses in the center of the flange. Mitigation measures
are to disperse the bearing pressure from the rectangular bar
connectors and to reduce the eccentric height. As a measure to
disperse the forces, we proposed to separate the rectangular bar
connectors at the ends (Solution B) and also to install additional
headed studs between the rectangular bar connectors to reduce the
deformation of the lower concrete in the delamination direction
(Solution C).
To verify the effectiveness of mitigation Solution A, the beam
specimens were fabricated, and strain measurements of concrete in
the transverse direction of the bridge axis were conducted during
stress application applying the release. The cushioning material of
the beam specimen is shown in Figure 5.
Bending action
Bearing pressure
40
400
45
600
270
Figure 6: Columnar specimen shape
Table 1: Comparison of mitigation effectiveness
Solution
A
Cushioning material
Cushioning material
A+C
+ Headed stud
Cushioning material
A+B+C + Headed stud + separated
rectangular bar connectors
Tensile stress on the concrete
2
surface (N/mm )
Ratio
7.8
1.00
2.9
0.37
1.8
0.23
2) For the mechanism of Type 2 crack formation at the central part of the flange, Solution B, which divides the
rectangular bar connectors at the block ends where the bearing pressure is concentrated, and Solution C, which
headed studs between the rectangular bar connectors welded between second to fourth positions, were proposed.
Both measures are effective, and their combined use enhances the effectiveness from Table 1.
4 References
[1] Design and construction guidelines for PREBEAM composite girder bridge 4th: Japan Institute of Country-ology
and Engineering, 2018.8(in Japanese).
[2] Matsui, S. Kurita, A. Watanabe, H. and Yamagishi, T.: Segmental prefabrication method for pre-flexed beam,
Proceedings of the Symposium on Research and Application of Composite Structures, pp.159-164, 1986.9 (in
Japanese).
116
13th - Japanese-German Bridge Symposium, Osaka, Japan
Imaging of Ultrasonic Echo Measurements for Reconstruction of Technical Data of Bridges –
Possibilities, Limitations and Outlook
Dr.-Ing. Stefan Maack *
Dr. rer. nat. Ernst Niederleithinger **
* Bundesanstalt für Materialforschung und -prüfung (BAM), Scientist, Germany, stefan.maack@bam.de
** Bundesanstalt für Materialforschung und -prüfung (BAM), Head of Division, Germany,
ernst.niederleithinger@bam.de
1 Extended abstract
Since the early beginnings (1920s and 1930s) of the use of non-destructive testing methods (NDT) to obtain information
about concrete structures, the acceptance of these methods in practice has increased continuously. While in the
beginning the focus was on the determination of material parameters and thus the quality of the materials used, today
complex testing tasks can be solved, such as the geometrically exact imaging of the internal structure of structures.
Today, the determination of the integrity of concrete structures as well as the determination of the internal structure is
carried out according to the state of the art with test methods based on the physical principles of electromagnetic and
acoustic wave propagation such as active and passive thermography, Ground Penetrating Radar (GPR) and ultrasonic
methods [2], [4], [5]. Statements about the integrity of a concrete structure, such as the presence of honeycombs or even
deep-lying delamination’s, are often carried out using the GPR or Ultrasonic methods, depending on the task [1], [2],
[3]. The advantage of these two inspection methods is that they can be used as pulse-echo methods. This means that
only one-sided access to the structural component is required. In addition, geometric dimensions, such as the thickness
of a component or even the precise position of tendons, can be determined with these test methods [2], [3], [4].
Figure 1: a) Case study -calculation of dead weight- bridge element of the “Köhlbrand” bridge in Hamburg; b)
Technical drawing of the profile of the bridge with the LIDAR point cloud of the hollow box girder c)
Standard cross section of the bridge [5].
A case study shows how the dead-weight of a bridge element can be calculated using non-destructive testing methods
[5]. The object investigated is the Köhlbrand Bridge (Hamburg, Germany). The bridge structure consists of two ramps,
which were built as prestressed concrete continuous beam with hollow box cross-section. The calculation guideline for
the recalculation provides for the possibility of reducing the partial safety factor of the dead-weight from G = 1,35 auf
G = 1,2 if this can be determined more precisely. Annular ultrasonic echo measurements were carried out inside the
box girder at equidistant positions. The volume was then calculated and the dead-weight of the box girder determined
with the help of the drill cores taken. A particular challenge in this application is the different angles between the
component surface and the component back wall.
With exact knowledge of the input parameters for a reconstruction calculation of ultrasonic data, the component
thicknesses of concrete structures with back walls plane to the component -measuring- surface can be determined with
117
high accuracy. Depending on the selected type of calibration of the ultrasonic velocity (c), the deviation for the case
investigated in [4] is less than 1% in relation to the actual component thickness. A desk study was conducted to record
typical angular relations on a total of 30 bridges of different concrete construction types. Based on these results, a series
of specimens with different angles between 0° and 25° was designed and manufactured. Figure 6 b) shows a B-scan of
the raw data along a measurement line (Figure 6 a); blue dashed line) in the area without reinforcement. The B-scan
clearly shows the backwall echo (red arrows). In addition, the actual location of the backwall echo at an angle of 25° is
plotted in Figure B (green dashed line). It can be clearly seen that the ultrasonic signal deviates from the actual position
of the backwall echo. In this case, the angle resulting from the position of the backwall echo is approx. 2.5° larger than
the actual angle of the backwall.
Figure 6: a) Geometric dimensions of specimen Tk5-25° in m; b) B-scan (raw data) along the measuring line in the
unreinforced area of Tk5-25° (green dashed line: component back wall; red arrows: back wall echo) [6].
In further series of investigations, it will therefore be evaluated how the reconstruction of sloping back walls can be
improved on the basis of ultrasonic data using the SAFT algorithm. Furthermore, at BAM it is investigated to what
extent the possibilities of Reverse Time Migration (RTM) can be transferred to use cases in civil engineering [7], [8].
The RTM method is an iterative correlation method in which simulated data of elastic wave propagation and measured
data of the component are processed. No prior knowledge of the component is required for modeling the synthetic
measured data in the first iteration steps. With each iteration step, only the synthetic model is adjusted. The result is a
geometric representation of the internal structure of the component under investigation. This is a significant advantage
over the SAFT algorithm.
2 References
[1] Reinhardt, H.W. and Grosse, C.U.: Setting and Hardening of Concret continuosly monitored by Elastic Waves.
http://www.ndt.net/article/grosse1/grosse1.htm, 1996.
[2] Bergmeister, K. and Rostan, S.: Monitoring and safety evaluation of existing concrete structures: bulletin 22.
state-of-the-art report. Bulletin 22, 2003.
[3] Beushausen, H. and Fernandez Luco, L. (eds): Performance-Based Specifications and Control of Concrete
Durability: State-of-the-Art Report RILEM TC 230-PSC, 1st edn, Springer, Dordrecht, 2015.
[4] Maack, S., Küttenbaum, S., Niederleithinger, E.: Practical procedure for the precise measurement of geometrical
tendon positions in concrete with ultrasonic echo. MATEC Web Conf., 364 (7-8), 3008, 2022.
[5] Maack, S., Knackmuß, J., Creutzburg, R. Comparative visualization of the geometry of a hollow box girder using
3D-LiDAR – Part 2: Reconstruction of 3D Geometric Model, pp. 54–64, 2017.
[6] Winkelmann, P.: Systematische Untersuchungen zur Dickenmessung mit Ultraschallecho an geneigten
Bauteilflächen. (engl.: Systematic studies on thickness measurement with ultrasonic echo on inclined component
surfaces.) Hochschule für Wirtschaft und Technik Berlin. Masterarbeit, 2019. (German)
[7] Grohmann, M., Müller, S., Niederleithinger, E., Sieber, S.: Reverse time migration: Introducing a new imaging
technique for ultrasonic measurements in civil engineering. Near Surface Geophysics, 15 (3), 242–258, 2017.
[8] Grohmann, M., Niederleithinger, E., Maack, S., Buske, S.: Application of Elastic P-SV Reverse Time Migration
to Synthetic Ultrasonic Echo Data. Journal of Nondestructive Evaluation, 2023. (ACCEPTED)
118
13th - Japanese-German Bridge Symposium, Osaka, Japan
Effect of Fire Damage on Residual Prestress and Load Carrying Capacity of Pretensioned
Prestressed Concrete
Dennise
Prof. Yasuhiro Mikata *
Prof. Susumu Inoue **
Osaka Institute of Technology, Indonesia, m1m22107@oit.ac.jp
* Osaka Institute of Technology, Japan, yasuhiro.mikata@oit.ac.jp
** Osaka Institute of Technology, Japan, susumu.inoue@oit.ac.jp
Abstract:
In recent years, fire damage to bridges and viaducts has been reported. Thus, increasing the necessity to clarify the
relationship between the heat received in the PC members and the residual strength. This study aims to investigate the
effect of heating time and temperature on the spalling conditions of the cover concrete. In addition, the maximum heatreceiving temperature of the internal PC steel and the effects of heating time and heating temperature on the residual loadflexural behavior by comparing the loading test results of specimens heated at 900°C for 60 minutes with the sound
specimen’s results. In the heating tests, the various strengths of concrete subjected to high temperatures were greatly
reduced. In the loading test, the maximum load was reduced by about 14% in specimen 60HC900-1 compared to sound
specimen N-2, due to the decrease of prestress suggesting that the high-temperature history at 900°C significantly affects
the residual load-carrying capacity of the concrete.
Keywords: Load Carrying Capacity, Residual Prestress, Fire Damage, Spalling and Prestressed Concrete
1 Introduction:
In Japan, the number of cases in which reinforced concrete or steel structures, except for tunnel structures, have collapsed
or suffered other serious damage due to fire is only a few, so the performance verification of these structures against fire
has not always been considered important. However, due to the increase in the number of cases in which the rise to the
maximum temperature during a fire exceeds the conventional assumptions, and the impact on society when infrastructures
with a high public nature are damaged by fire due to increased traffic has been increasing. Thus, the need to reconsider
the performance of concrete structures in the civil engineering field against fire and its verification methods has been
increasing.
2 Purpose
This study aims to investigate the effect of heating time and temperature on the spalling conditions of the cover concrete.
In addition, the maximum heat-receiving temperature of the internal PC tendons and quantitatively evaluate the effect of
these relationships on the residual load-carrying capacity.
3 Specimen Outline
The PC girder details were as follows: top width 640 mm, bottom width 700 mm, total length 5300 mm, and pre-tensioned
PC girder 1) (pre-tensioned PC girder (AS-05) as specified in JIS-A-5373). The design compressive strength of the
concrete is 50 N/mm2.
Figure-2 Cross Section of Concrete Specimen
Figure-1 Side Section of Concrete Specimen
4 Heat test
A horizontal heating test furnace (4m x 3m) was used. For the fire curve, a heating curve (HC900) was selected, in which
the maximum temperature of the hydrocarbon curve HC curve (1100°C) specified by Eurocode3 2) was modified to
900°C. The heating time was 60 minutes. Heating tests and load-bearing tests were conducted on specimen 60HC900-1.
For comparison, a specimen N-2 without a heating test was used. The bottom surface of the heated PC girder is shown in
Figure-3 after natural cooling and removal. The concrete has spalled to such an extent that the entire transverse stirrups
and the PC steel are exposed. The maximum depth of the spalling was 61mm. The depth of the spalling was smaller in
the center than at both ends.
119
5 Loading test
The static loading test was conducted using the two-point concentrated load method with the bending span of 1000 mm
and the shear span of 1850 mm, and monotonically increasing loads were applied until failure.
Figure-5 and 6 show the crack of each specimen after the loading test. It also shows flexural cracks occurred from the
bottom edge near the center of the span in both specimens, and the concrete in the compression zone was crushed. In case
N-2, flexural cracking occurred at a load of 220 kN. Finally, flexural tension failure occurs at a load of 380 kN, and the
flexural compression zone is crushed. In the case of 60HC900-1, flexural cracking occurred at a load of 180 kN and
flexural tension failure occurs at a load of 330 kN.
Figure-7 shows the load-displacement relationships. In the case of comparing the measured maximum load values of each
specimen where 60HC900-1 is 332.5 kN and N-2 is 387.9 kN, a decrease of 14% can be observed and it also shows that
the initial stiffness and displacement of maximum load decreased due to heating.
The reason for the decrease in the maximum load and initial stiffness of 60HC900-1 are the decreasing cross-sectional
area and decreasing prestress due to spalling of the cover concrete.
Figure-3 Explosion Spalling condition after heat
Figure-4 Explosion Spalling condition after heat
Figure-3 Crack Condition after Loading Test (60HC900-1)
Figure-6 Crack Condition after Loading Test (N-2)
6 Conclusion
The conclusions obtained in this study are as follows. The
concrete has spalled to such an extent that the entire transverse
stirrups and the PC tendons are exposed. The spalling depth of
beam end was 61mm and the spalling at the beam center was
26mm. This is due to the spalling of the cover concrete causing
the heat sensor to be directly caught under fire and increased
exponentially. In static loading tests, the maximum load was
reduced by about 14% in 60HC900-1 compared to N-2, and the
specimens used in this study showed a large change in bearing
capacity at the maximum temperature of 900°C. The effect of
time was also significant. On the other hand, the calculated
bending failure load was calculated on the safe side of the
measured value, and the residual bearing capacity of the PC
girders after heating could be evaluated by considering the rate
of decrease of prestress and bond strength.
Figure-7 Load - Displacement Relationships
References
1) European Committee for Standardization (CEN)(2002): Eurocode 1: Actions on structures – Part1-2: General actions
– Actions on structures exposed to fire (EN 1991-1-2)
2) Japan Society of Civil Engineers (JSCE) (2012): Standard specifications for concrete structures – 2012 Design
3) Osamu OHYAMA, Akimitsu KURITA, Proposal of prediction equation of concrete depth due to explosive spalling,
vol.34, no.1, pp.1132-1137, 2012
4) Susumu Inoue, Yosuke Tabuchi: Effect of Fire Damage on the Residual Prestress and Load Carrying Capacity of
Pre-tensioned Prestressed Concrete Bridge Girders, Proc. Of the 5th International fib Congress, ID81,2018.10
Acknowledgement
This work was supported by Oriental Shiraishi Corporation
120
SESSION 4-B
Vibration and Monitoring
13th - Japanese-German Bridge Symposium, Osaka, Japan
Study of a monitoring plan and behavior analysis to verify the performance
of an integrated column by multiple steel pipes
Shinsuke AKAMATSU *, Masahiro HATTORI1 **, Yasumoto AOKI ***,
Yoshiki TANIGUCHI ****, Kunitomo SUGIURA *****
* Hanshin Expressway research institute for Advanced Technology, Japan, shinsuke-akamatsu@hit.or.jp
** Hanshin Expressway research institute for Advanced Technology, Japan, masahiro-hattori@hit.or.jp
*** Hanshin Expressway Company Limited, Japan, yasumoto-aoki@hanshin-exp.co.jp
**** Hanshin Expressway Company Limited, Japan, yoshiki-taniguchi@hanshin-exp.co.jp
***** Graduate School of Engineering, Kyoto University, Japan, sugiura.kunitomo.4n@kyoto-u.ac.jp
1 Introduction
In the Hanshin-Awaji Earthquake of 1995, piers collapsed or were severely
damaged, and it took a long time to restore them. Therefore, we developed an
integrated column by multiple steel pipes that have higher earthquake resistance
than reinforced concrete or steel piers, and can be restored easily and quickly. As
shown in Figure 1, this structure is composed of a single column by connecting
four steel pipes with shear links that incorporate shear panels with historical
damping functions. In the event of large seismic motion, the damage is
concentrated in the shear panels so that the steel pipe columns remain sound.
However, it has not been verified whether the columns behave according to the
concepts in the actual structure.
In this study, based on the structural characteristics of this column, the performance
required of this column during earthquakes and other events was summarized. A
monitoring plan was developed to verify the performance of the column, and
measurements were conducted based on the plan. The obtained measurement
results were used to analyze the behavior of the column.
2 Develop a monitoring plan
To verify the performance through measurements, a
monitoring plan was developed according to the
procedure shown in Figure 2. Figure 3 shows an image of
the evaluation metrics and measurement points for each
member. The behaviors assumed in the design were
listed, and for each assumed behavior, the behaviors that
could occur in the actual structure, the so-called doubts,
were extracted as evaluation metrics. The items to be
captured to verify the extracted evaluation metrics were
examined, and the measurement points and measurement
items were organized. The conditions assumed in the
design are for ordinary, Level 1 earthquake, Level 2
earthquake, and scenario earthquake larger than Level 2
earthquake, and the evaluation metrics and measurement
points of the PD4 for each condition were extracted.
A feasible measurement method is considered for the
evaluation indexes and measurement points extracted.
The measurement period is assumed to be 5 years, which
is a long-term measurement period, and the selection of
equipment was made in consideration of the deterioration
of the measurement equipment due to rain and wind in
outdoor measurement.
If the selected instruments were installed on all members
of the structure, the behavior could be captured in detail,
but the measurement cost would be very expensive.
Therefore, several measurement plans were created, and
an appropriate measurement plan was selected after
sorting out the merits and demerits of each plan.
123
Figure 1: Structural outline
Start
Organizing the structural
characteristics of the target structure
Consideration of evaluation metrics and
measurement points for each member
Consideration of measurement method
Consideration of evaluation method of
measurement results
Consideration of data receipt and
recording methods
End
Figure 2: Flow chart for developing a monitoring plan
Extraction of
measurement points
Behaviors assumed in design
Consideration of
measuring points
Behaviors to capture to confirm
design assumptions
Behaviors that could occur in
the actual structure
Measurement
points
Measurement
items
Figure 3: Consideration of evaluation metrics and
measurement points for each member
3 Monitoring Measurement
The measurements based on the
monitoring plan in Section 2 have been
taken since April 2021, and Figure 4
shows the locations of the measurement
equipment and the names of the measuring
points.
The measurement cross section of the
strain gauges was set at the base of the
piers at a distance of 100 mm from the
protective concrete, and at the top of the
piers at a minimum distance of 100 mm
from the fillet between the superstructure
and the piers to avoid the effect of the
stress concentration. Strain gauges were
attached to each steel pipe at four points in
the longitudinal and transverse direction
and at 16 points per cross-section, and the
Figure 4: Locations of the measurement equipment and names of the
direction of measurement was vertical
measuring points
direction. Accelerometers were set 100
mm below the ground surface and above each shear link of Pipe B, for a total of five locations. The measurement
directions were the longitudinal direction, the transverse direction, and the vertical direction with regard to the
integrated column by multiple steel pipes.
4 Analysis of measurement results
The following is an example of performance evaluation based
on measured data for behavior under temperature change. The
Integrated column by multiple steel pipes is assumed to "Four
steel pipes resist horizontal loads as a single column". In
contrast, since it is suspected that each steel pipe resists
discretely, the behavior was analyzed using measured data
during temperature change.
Among the measured data, the lowest outside temperature was
3.6°C at 5:38 on January 23, 2022, and the highest outside
temperature was 39.1°C at 16:04 on July 23, 2022. Figure 5
shows the amount of change in pipe strain resulting from this
temperature change (+35.5°C). Focusing on the strain at the
base of steel pipes, the strain at the outer transverse direction
of pipes B and D (B-b-2 and D-b-2) changed by +617 με and
+649 με, respectively, and the strain at the outer transverse
direction of pipes A and C (A-b-4 and C-b-4) changed by -350 Figure 5: Change in steel pipe strain with temperature
με and -524 με, respectively. This is considered to capture the
change (+35.5°C)
behavior of PD4 being pushed outward out of the curve due to
the extension of the superstructure caused by the +35.5°C temperature change. And the strain at the outermost edge in
the direction of extrusion is the maximum and minimum, respectively. In this manner the strain distribution behaves
like a single column, so it was confirmed that the four steel pipes resist as a single column.
5 Conclusion
In this study, a monitoring plan for verifying the performance of an integrated column by multiple steel pipes was
developed based on the structural characteristics. And the measurement data obtained from the measurements based on
the plan were analyzed. The results of this study are as follows.
1) Based on the structural characteristics of the integrated column by multiple steel pipes, measurement points and
items were organized to confirm the assumptions in the design.
2) Measurement methods for the organized measurement points and items were studied, and the measurement plan was
selected after listing the measurement plans.
3) The strain in the steel pipes under temperature change obtained from ordinary condition measurements was analyzed,
and it was confirmed that the four steel pipes behaved as a single column.
In the future, it will be necessary to verify the accuracy of the seismic response analysis of the integrated column by
multiple steel pipes based on the obtained seismic observation data, and to extract the response of each measured point
in the damage event of this structure by pushover analysis, and to grasp the response as a guide to decide in case of
emergency.
124
13th - Japanese-German Bridge Symposium, Osaka, Japan
Application of Bridge Weigh-in-Motion on a Bridge with Prestressed Concrete Girders
Marcel Nowak ∗
Oliver Fischer ∗∗
∗
Technical University of Munich, Chair of Concrete and Masonry Structures, Germany, marcel.nowak@tum.de
Technical University of Munich, Chair of Concrete and Masonry Structures, Germany, oliver.fischer@tum.de
∗∗
1 Introduction
A reliable, robust, and comprehensive data basis for all relevant traffic and vehicle parameters is essential for traffic load
modeling with consideration of local traffic characteristics. However, reliable measurement data are often unavailable for
dominant parameters such as vehicle weight, vehicle headway, or congested traffic. Missing parameter information has to
be replaced by corresponding assumptions, often based on measurements at other locations of the road network or from
literature, usually leading to additional inaccuracies in the traffic load modeling. Within this context, in 2019 and 2020,
a long-term monitoring campaign over one year on a selected bridge structure on Federal Highway A92 was carried out
to create a comprehensive data basis for all parameters relevant to traffic load modeling. The core component of this
monitoring concept is structural monitoring on the bridge with an application of brigde weigh-in-motion (BWIM). Based
on different data analysis strategies, an automated algorithm is developed, allowing for data from multiple sensors to be
evaluated towards relevant parameters from vehicles of the passing road traffic.
2 Monitoring Concept
To implement the monitoring campaign, bridge structure 29/1 (here: northern superstructure for the direction of travel
to Munich) was selected, located northeast of Munich between the exits of Freising-Ost and Erding on Federal Highway
A92. The bridge crosses a receiving ditch in five spans with nearly uniform widths of about 16.0 m. The bridge consists
of two identical superstructures, each supporting the roadway for one driving direction with a standard configuration of
two lanes plus emergency lane. For the structural monitoring, 29 strain gauges and two temperature sensors are installed
on the bottom side of the northern superstructure, which carries traffic in the driving direction of Munich. The sensor
layout designed for an application of BWIM mainly consists of two components: sensors for global and local structural
response (Figure 1).
strain gauge (global)
8.00
1.80
2.40
strain gauge (local)
thermo sensor
3.80
80 1.26 80 1.00 86 1.00 1.00 86 1.00 80 1.26 80
4-5
4-28
4-18
4-27
4-17
4-26
4-25
4-3
4-3T
4-24
4-23
4-2
4-22
4-16
4-15
Ϯϱ
4-14
4-13
ϭϱ
4-1
4-21
ŐůŽďĂůsĞŶƐŽƌϰͲϯ
ůŽcĂůĞ sĞŶƐŽƌϰͲϮϯ
ϮϬ
ε · 10 6
4-4
ϭϬ
4-12
ϱ
4-11
Ϭ
ϭϮ͗Ϭϭ͗Ϯϭ͘Ϭ
4-1T
ϭϮ͗Ϭϭ͗Ϯϭ͘ϱ
time
ϭϮ͗Ϭϭ͗ϮϮ͘Ϭ
ϭϮ͗Ϭϭ͗ϮϮ͘ϱ
Figure 1: Scheme of sensor layout in top view (left) and cross section (right, top) of span 4, sample signal of vehicle
crossing for global and local sensor (right, bottom).
3 Application of Bridge Weigh-in-Motion
BWIM is one of the methods of indirect impact monitoring, which describe the load process of the traffic passing over a
bridge by solving an inverse problem. The basic principle is to infer the causative action from traffic-induced measurement
signals of the structural response with the help of findings from defined proof load tests. The basis for the application of
125
the BWIM algorithm is the sequence of bridge loading events (BLEs) obtained by evaluating the measured data of the
structural monitoring, as well as the reference influence lines (RILs) determined based on the signals of the proof load
test. If a single vehicle axle passes the bridge structure in one of the traffic lanes, this usually leads to a distinctive peak
in the signals of at least one of the local sensor pairs of this lane. The time interval of this peak between the two signals
of the sensor pair depends on the spacing of the sensors of a pair (fixed value 2.4 m by the installation of the sensors)
and the velocity of the passing vehicle axle (variable per vehicle crossing). Accordingly, the crossing of a vehicle leads
to a peak sequence at both sensors, whose time intervals between the sensors are (approximately) equal. By identifying
such decisive peak sequences in the signals of the local sensor pairs, individual vehicles are detected within the BLE. The
number of peaks within a sequence corresponds to the number of axles of the vehicle. The velocity can be determined
from the time intervals of the peak sequence between the sensors of a pair and their spacing. Finally, with the help of
the now-known velocity and the time intervals of successive peaks of a sequence on the signal of the same sensor, the
axle spacings can be inferred (see Figure 2). The local sensor pairs are assigned to the possible transverse positions for
each traffic lane configuration, allowing a conclusion for the transverse position of the identified vehicles. The parameters
determined for the detected vehicles based on the local sensors form the basis for determining the axle weights. For this
purpose, an optimization problem is defined whose optimization variables are the individual axle weights. The objective
function corresponds to the sum of squared errors between the decisive global measurement signal due to the crossing of
the (unknown) vehicle and the approximation by scaling and superposition of RILs of the global sensor according to the
previously determined axle layout and the velocity (see Figure 2). The optimum for the axle weights is determined by
minimizing the objective function.
ƐŝŐŶĂůϰͲϮϯ
ƐŝŐŶĂůϰͲϭϯ
ŝĚĞŶƚŝĨŝĞĚǀĞŚŝĐůĞĂdžůĞƐ
ε4 23/4 13 · 10 6
ϮϬ
ϭϱ
ϭϬ
ϱ
Ϭ
xloc1/2 = 2.4ŵ
nax = 5
v = 2.4/0.098 = 24.5ŵͬƐ
xax1/2 = 24.5 · 0.160 = 3.94ŵ
xax2/3 = 24.5 · 0.203 = 4.92ŵ
xax3/4 = 24.5 · 0.055 = 1.34ŵ
xax4/5 = 24.5 · 0.052 = 1.29ŵ
Ϭ͘ϬϬ
Ϭ͘Ϯϱ
0.052Ɛ
0.160Ɛ 0.203Ɛ0.055Ɛ
0.098Ɛ
0.098Ɛ
0.098Ɛ 0.100Ɛ
0.098Ɛ Ϯ
ϯ ϰ ϱ
Ϭ͘ϱϬ
Z/>ͲƐƵƉĞƌƉŽƐŝƚŝŽŶ
ĂƌďŝƚƌĂƌLJŝƚĞƌĂƚŝŽŶƐƚĞƉ
ϭ͘Ϭ
ϱϬ
Ϭ͘ϴ
ϰϬ
Ϭ͘ϲ
ϯϬ
Ϭ͘ϰ
ϮϬ
ε4 3 · 10 6
Ϯϱ
ƐŝŐŶĂů
Z/>͕ƐĐĂůĞĚďLJĂdžůĞǁĞŝŐŚƚ
ϭ
Pax1 = 5.51ƚ
Pax2 = 3.04ƚ
Pax3 = 4.17ƚ
Pax4 = 5.62ƚ
Pax5 = 3.34ƚ
ŽƉƚŝŵƵŵƐŽůƵƚŝŽŶ
Pax1 = 6.11ƚ
Pax2 = 7.19ƚ
Pax3 = 6.87ƚ
Pax4 = 7.31ƚ
Pax5 = 7.64ƚ
Ϭ͘Ϯ
ϭϬ
Ϭ͘ϳϱ
ϭ͘ϬϬ
Ϭ͘ϬϬ
Ϭ͘ϬϬ͘ϬϬ
ϭ͘Ϯϱ
ĚƵƌĂƚŝŽŶŽĨ>;ƐͿ
Ϭ͘Ϯ
Ϭ͘ϱϬ
ϭ͘ϬϬ
Ϭ͘ϰ
Ϭ͘ϬϬ Ϭ͘ϲ
Ϭ͘ϱϬ Ϭ͘ϴ
ϭ͘ϬϬ
ϭ͘Ϭ
ĚƵƌĂƚŝŽŶŽĨ>;ƐͿ
Figure 2: Vehicle parameter determination based on local sensor signals (left) and optimization of axle weights based on
global sensor signal (right).
4 Conclusion and Discussion
This paper presents the application of Bridge Weigh-in-Motion (BWIM) during a long-term monitoring campaign at a
selected bridge structure on Federal Highway A92 over one year to comprehensively describe local traffic characteristics.
An automated algorithm is developed based on suitable strategies and concepts for collecting, processing, and evaluating
all available measurement data, allowing for a comprehensive analysis of data from multiple sensors towards relevant
parameters from vehicles of the passing road traffic. By an intelligent arrangement of the measurement sensors and the
time-synchronous evaluation of measurement signals for local and global structural responses, some of these parameters
can be determined directly from the measurement data. As a result, the number of optimization variables and, thus, the
complexity of the optimization problem in the BWIM algorithm is reduced. BWIM proves to be an economical, robust,
and powerful monitoring option. The data acquisition - measured by accuracy standards of common engineering practice is highly reliable, even for a wide range of boundary conditions (different lane configurations and loading constellations),
and the obtained data set covers a large part of the relevant parameter spectrum of traffic load modeling. Overall, the
monitoring campaign provides an comprehensive and valuable data basis for the detailed and realistic modeling of local
traffic characteristics. However, further investigations are required to quantify the actual benefit.
Acknowledgments
The research this work is based on is funded by Die Autobahn GmbH des Bundes (branch Southern Bavaria).
126
13th - Japanese-German Bridge Symposium, Osaka, Japan
Natural Frequency of Lightweight Foamed Concrete Composite Slabs (LFCCS)
Zainorizuan Mohd Jaini *
Kunitomo Sugiura **
Sakhiah Abdul Kudus ***
* Faculty of Civil Engineering and Built Environment, Universiti Tun Hussein Onn Malaysia, Malaysia,
rizuan@uthm.edu.my
** Department of Urban Management, Graduate School of Engineering, Kyoto University, Japan,
sugiura.kunitomo.4n@kyoto-u.ac.jp
*** School of Civil Engineering, College of Engineering, Universiti Teknologi MARA, Malaysia,
sakhiah@uitm.edu.my
Abstract:
Recently, lightweight foamed concrete composite slabs (LFCCS) have become increasingly popular. When considering
LFCCS for urban buildings and pedestrian bridges, the serviceability with regard to vibration behaviour must be taken
into account. Due to the density of foamed concrete, LFCCS are 40% lighter, making them more susceptible to damage
and uncertainties in structural resonance. In addition, LFCCS are vulnerable to dynamic loading where the natural
frequency needs to be monitored to avoid discomfort issue. This study investigates the natural frequency of LFCCS by
means of experimental study and numerical modelling. In experimental study, LFCCS were prepared for the hammerimpact test. The slab thickness ranges from 100 mm to 200 mm. In numerical modelling, LFCCS was modelled in
SAP2000 using a special technique called the simplified equivalent plate model. The effective material properties were
derived from the rule of mixtures and depend entirely on elastic properties with strength characteristics. It has been found
that the natural frequency decreases with slab thickness, signifying that the natural frequency is dominated by mass rather
than stiffness. Overall, the natural frequency of LFCCS is around 27.23Hz to 31.45Hz.
Keywords: Natural frequency, lightweight foamed concrete composite slabs, hammer-impact test, SAP2000
1 Introduction
The lightweight foamed concrete composite slabs (LFCCS), specially developed for flooring systems, offer a better
alternative to the conventional composite slabs. LFCCS consist of foamed concrete as the topping material and a steel
deck serves as the tensile reinforcement. The composition of foamed concrete and steel deck are environmentally friendly
and sustainable materials, suitable for urban buildings and pedestrian bridges. It is the novel solution to the dead weight
disadvantage, which is always a major concern with conventional composite slabs. Foamed concrete is known to be lower
in density and typically 40% lighter than normal concrete. Similar to conventional composite slabs, the use of LFCCS is
practical and economical. Foamed concrete has gained wide acceptance as a construction material, although its use in
flooring systems is still rare.
Investigations on the structural behaviour of LFCCS under static loading were conducted by Flores-Johnson & Li
[1] and Jaini et al. [2]. LFCCS has been found to perform excellently. With regard to dynamic loading, Rum et al. [3]
examined the vibration behaviour of LFCCS subjected to the hammer-impact test. The natural frequency of LFCCS has
been found to be around 29.70Hz to 33.78Hz. It is well above the vibration limit. For the damping ratio, the value is
around 3% to 5%. A finite element analysis of the serviceability of LFCCS under human excitation was performed by
Nurhalim et al. [4]. In the transient state, the natural frequency is linearly related to thickness. Although the natural
frequency is above the vibration limit, LFCCS with a slab thickness of more than 175 mm exceeded the permissible
deflection.
Under dynamic loading, very low inertia forces can occur due to the poor stiffness or low mass. Therefore,
understanding the vibration behaviour of LFCCS is crucial to ensure that its application for urban buildings and pedestrian
bridges meets comfort criteria. As LFCCS becomes more popular, the study on vibration behaviour is of paramount
importance to rule out any possibility of damage and uncertainties in structural resonance. In view of this problem, this
study presents the investigation of the natural frequency of LFCCS within the framework of experimental study and
numerical modelling. The focus is on the effects of slab thickness on the natural frequency.
2 Methodology
A total of 15 slab specimens were prepared for the experimental study. The dimension of slab specimens is 840 mm 
1800 mm and the slab thickness ranges from 100 mm to 200 mm. The density of foamed concrete is 1800 kg/m3 and the
steel deck is based on PEVA45. The instruments such as hammer, data logger and accelerometers were used in the
hammer-impact test. On the other hand, the numerical modelling of LFCCS was performed with SAP2000. The physical
properties of LFCCS were constructed three-dimensionally using shell elements. Since LFCCS have a corrugated shape,
a special technique called simplified equivalent plate model proposed by El-Dardiry & Ji [5] was employed in the
numerical modelling. Norhalim et al. [4] proved that the simplified equivalent plate model is suitable for determining the
127
vibration behaviour of composite slabs. The effective material properties for LFCCS were derived from the rule of
mixtures and depend entirely on elastic properties and strength characteristics.
3 Results and Discussion
Figure 1 shows the natural frequency of LFCCS from experimental study and numerical modelling. The discrepancy of
results is less than 3.5%. This is apparent evidence that numerical modelling with SAP2000 gives reasonably good
agreement despite the lack of conventional modelling technique for composite slabs. It can be observed that the natural
frequency decreases with slab thickness. This means that the vibration behaviour of LFCSS is dominated by mass and
not stiffness. This finding is similar to that observed by Rahimi et al. [6] for the precast hollowcore slabs. The plot of
natural frequency versus length-to-thickness ratio provides a clear picture on the vibration behaviour of LFCCS in terms
of geometry properties and design aspects. The relationship can be represented as:
Natural Frequency, f (Hz)
2
f = − p 0.0617 ( L h ) − 2.0024 ( L h ) − 14.892
(1)


where L is the length, h is the slab thickness and αρ is the multiplying factor related to the density of foamed concrete. For
foamed concrete with a density of 1800 kg/m3, the multiplying factor can be taken as 1.0.
35
34
33
32
31
30
29
28
27
26
25
Experimental Study
Numerical Modelling
100
120
140
160
180
200
Slab Thickness, h (mm)
Figure 1: The natural frequency corresponds to the slab thickness.
4 Conclusion
•
•
•
•
The natural frequency decreases with slab thickness, signifying that the natural frequency is dominated by mass rather
than stiffness.
The numerical modelling of LFCCS was established on the basis of the simplified equivalent plate model in
SAP2000. The material properties were defined as effective material properties derived from the rule of mixtures.
The results of numerical modelling showed similar tendencies as experimental study.
The simplified equivalent plate model was able to produce convincing results despite the lack of conventional
modelling technique for composite slabs.
An empirical formula that relate the natural frequency with the length-to-thickness ratio was established using the
mean data of results. However, this empirical formula requires further validation.
5 References
[1] Flores-Johnson, E. A. & Li, Q. M. Structural behaviour of composite sandwich panels with plain and fibrereinforced foamed concrete cores and corrugated steel faces. Composite Structures, 94, 1555-1563, 2012.
[2] Jaini, Z. M., Rum, R. H. M., Hakim S. J. S. & Mokhatar, S. N. Application of foamed concrete and cold-formed
steel decking as lightweight compoiste slabs: Experimental study on structural behaviour. International Journal of
Integrated Engineering, 15, 181-193, 2023.
[3] Rum, R. H. M., Norhalim, A. F., Jaini, Z. M. , Abd Ghafar N. H. & Kozlowski, M. Density and strength of foamed
concrete: The influence on dynamic characteristics of lightweight profiled composite slabs. International Journal
of Integrated Engineering, 11, 285-295, 2019.
[4] Norhalim, A. F., Jaini, Z. M., Ghafar, N. A., Majid , M. A. & Ahwang, A. Dynamic serviceability of lightweight
composite deck as floor system under human excitation. IOP Conf. Series: Materials Science and Engineering, 713,
012026, 2020.
[5] El-Dardiry, E. & Ji, T. Modelling of the dynamic behaviour of profiled composite floors. Engineering Structures,
28, 567-579, 2006.
[6] Rahimi, M. I., Ghafar, N. H. A., Ibrahim, Z., Jaini, Z. M., Aziz, N. Z. A. & Gamadi, N. E. M. The effect of concrete
topping thickness on the vibration response of prestressed and precast hollow core floor systems. International
Journal of Integrated Engineering, 12, 356-364, 2020.
128
13th - Japanese-German Bridge Symposium, Osaka, Japan
Geo-referenced localisation of SHM sensors on new bridge construction based on the example
of the digital bridge Schwindegg (Germany)
Johannes Wimmer M.Eng. ∗
Univ. Prof. Dr.-Ing. Thomas Braml ∗∗
∗
University of the Bundeswehr Munich, Institute for Structural Engineering, Germany, johannes.wimmer@unibw.de
∗∗
University of the Bundeswehr Munich, Institute for Structural Engineering, Germany, thomas.braml@unibw.de
Keywords: Digital Twin, Structural Health Monitoring, Sensor, Asset Administration Shell, Digitalisation
1 Introduction
Germany is advancing in digitalization, as indicated by the annual Digitalization Index published by the Federal Ministry
of Economics and Climate Protection. However, the construction industry ranks lower compared to sectors like automotive and automation. To address this, Industry 4.0 and the use of data for monitoring product lifecycles have been
introduced.
The University of the Bundeswehr Munich developed BBox, a configurator for Asset Administration Shells (AAS) used
for bridges. AAS creates digital twins that interact with the product. A global coordinate system is necessary to connect
multiple AAS for various structures. Currently, Building Information Modelling models mostly utilize local coordinate
systems.
Structural Health Monitoring (SHM) is commonly used to monitor structures, but it often has limited data collection
or focuses on later stages of a bridge’s life. To overcome this, approximately 140 sensors were installed during the
construction of a road bridge in Schwindegg, Germany. These sensors continuously measure data, allowing for the
generation of a digital twin and practical experience with sensor technology for bridge monitoring according to Industry
4.0 standards
2 Pilot project digital bridge
The bridge was designed as a deep-grounded single-span prestressed concrete frame bridge. The cross-section consists of
four prestressed, precast beams with in-situ concrete supplement. The abutments are founded on drilled piles, they were
backfilled with lightweight material. The goal of the digital twin is the aquisition of Big Data with different sensors over a
long period. For this reason, the environmental conditions of the infrastructure, the measurement tasks and the geometric
and static conditions must be taken into account.
The former includes power and internet connectivity. As the sections are laid over the new bridge structure, it was
possible to provide power to the structure. A fibre optic connection for the internet was also possible. In order to define the
measurement objective, the static system of a frame bridge was investigated and typical details such as the bored pile head,
the back-filled abutment, the frame corner and the span were examined in a finite element model. Static influences on this
design include traffic and daily and seasonal temperature variations. Based on these assumptions, monitoring areas were
defined from which the measured variables of strain, temperature, inclination, acceleration, deformation, earth pressure
and weather influences such as air temperature, humidity, wind speed, wind direction, solar radiation, precipitation and
air pressure were selected. Slowly changing values such as building temperature and weather were sampled every five
minutes. Strains etc. are measured at 10-20 Hz and accelerations at around 200 Hz. During the design and installation of
the monitoring system, it became clear that the conditions varied depending on the location (precast plant or construction
site). In addition, the sensors were partly installed in the components. This affected the design and installation of the
cables as well as the construction processes. With structured planning and execution of the installation and surveying,
delays due to sensor installation could be avoided.
The acquisition of data is carried out by data loggers in the technical block with on-site buffer memory. The data can
also be pre-processed there. The data is transferred to the University of the Bundeswehr Munich network using MQTT
via the mobile network and the fibre optic network. The data is stored there using the Asset Administration Shell (AAS)
BBox. From there, the data can be displayed on the dashboard, processed on the cluster or used to train machine learning
algorithms. The sensors had to be located accurately in order to locate the measurement data and, for example, to ensure
a later application with finite element models. This was achieved by geo-referencing the sensors.
129
3 Geo-referencing
In order to know the position of the sensors even after decades of bridge operation, they must be globally calibrated, as
local coordinate systems can be lost, e.g. due to changes in the structure or loss of the coordinate origin. The latter can be
counteracted, for example, by using sustainable data storage such as the AAS BBox. To implement a global coordinate
system, an ellipsoidal system must be selected. While the UTM system is based on WGS84 and uses 6° meridian strips,
the Gauss-Krueger (GK) system uses 3° wide strips. The latter was chosen because of the usual location of construction
sites and finished structures in Germany.
The most common method of global positioning today is to use a total station. This can be used to measure individual
points on the structure once, or to monitor the displacement of multiple points over time. In addition, point clouds produced by a 3D laser scan can be accurately placed in a global coordinate system. This means that a large number of points
on a structure can be located and a 3D model derived in a short period of time.
As the sensors in this project were partly embedded, partly buried and partly retrofitted, different methods were used.
Sensors installed on the back of the abutment could be calibrated using the total station. Sensors embedded in the precast
beam were measured on the reinforcement and formwork in a local coordinate system. Once the installed beam had been
geo-referenced, the local coordinates could be transferred to the global coordinate system. Due to long installation times
and unfavourable shading caused by high formwork and dense reinforcement, direct calibration is often not possible.
One solution is to measure a sufficient number of reference points against which the distance to the sensor is measured.
Trilateration, the geometric determination of the intersection of three spheres in space, can be used to determine the
position of the sensor. It was used to calibrate the sensors in the bored pile head, the frame corner and the carriageway
slab, where calibration with a total station was not possible for the reasons given above. However, this method is only a
stopgap solution due to the accumulation of different measurement errors. A better solution needs to be found for future
projects. Another option is to create a point cloud with a 3D laser scan. This was carried out on the structure after the
abutments had been removed and the bridge completed. As with total station surveying, this is an optical method, which
means that shaded areas cannot be captured. It was therefore not possible to use the sensors that were to be embedded in
the concrete, but the sensors that were subsequently installed on the underside of the bridge could be used. A 3D model
of the bridge was then derived from the point cloud and used to check the plausibility of the calculation of further sensor
coordinates.
4 Experiences
New insights have been gained in the application of the methods presented. On the one hand, the main time consuming
task is the calibration of the total station at each station. Therefore, it makes sense to be able to measure as many points
as possible from one station. One point per station is not practical. If the system is used efficiently, it will quickly provide
good and sufficiently accurate values. Trilateration allows many markers to be measured quickly from a few points with
a total station. Calibrating the sensors on (at least) three markers was also quick. However, measurement errors add up
when several measurements are taken in succession, making the method accurate in theory but much less so in practice.
The 3D laser scan is well suited to capturing the actual shape of the structure. Software could be used to quickly derive 3D
models. Unfortunately, this does not work for sensors, as the resolution is too low for the large format for data reduction,
and no algorithms for sensor recognition could be found. Furthermore, shaded sensors are not even captured with a point.
However, the sensors that were not in the shade could be found quickly and easily.
130
13th - Japanese-German Bridge Symposium, Osaka, Japan
Standardisation in Structural Health Monitoring (SHM) - a concept proposal
Univ. Prof. Dr.-Ing. Thomas Braml ∗
Johannes Wimmer M.Eng. ∗∗
Fabian Seitz M. Sc. ∗∗∗
Univ. Prof. Dr.-Ing. Max Spannaus ∗∗∗∗
∗
University of the Bundeswehr Munich, Institute for Structural Engineering, Germany, thomas.braml@unibw.de
University of the Bundeswehr Munich, Institute for Structural Engineering, Germany, johannes.wimmer@unibw.de
∗∗∗
University of the Bundeswehr Munich, Institute for Structural Engineering, Germany, fabian.seitz@unibw.de
∗∗∗∗
University of the Bundeswehr Munich, Institute for Structural Engineering, Germany, max.spannaus@unibw.de
∗∗
Keywords: Digital Twin, Sensor, Bridge maintenance, Digitalisation, Lifecycle management
1 Introduction
The collection of sensory measurement data on a structure is an increasingly popular way to support verification. At
present, the preferred methods are proof loading, non-destructive testing (NDT) or structural health monitoring (SHM).
The data is either stored on-site on the measurement computer and hard drives in the manufacturer’s proprietary format,
or made available online via the manufacturer’s cloud. In most cases, software tailored to the instrument is used for
evaluation and display. If multiple systems are used, then multiple of these data processing chains must be used. Based
on our experience of past and future monitoring projects, we have developed a proposal for standardising data collection,
transmission, storage, analysis and visualisation, cf. Figure 1.
Figure 1: An overview of the contents of all the chapters in this document.
2 Data Acquisition
At the top of an SHM system are the sensors. These are attached to the object to be measured. There are many methods
of attaching sensors to the structure, such as bonding, screwing, magnetism and embedding in concrete. Depending on
the application, there are different robust sensors. They must be able to withstand the harsh environmental conditions
of the installation site. The sensors are the beginning of the measurement chain according to DIN 1319-1, the end of
which is the data logger. Within the measurement chain, the measurement data is recorded, the mostly analogue signal is
amplified, digitally converted and stored. The data can be processed on site (’edge’) and prepared for further transport,
which is particularly necessary for methods such as acoustic emission due to the enormously high sampling rates. The
preparation for further transport is done by converting the data into a suitable transmission protocol. In the case of the
131
proposal presented here, the Message Queuing Telemetry Transport (MQTT) protocol was agreed upon. The data is
forwarded from the Edge Processing Unit to the Data Transfer.
3 Data Transfer
In Data Transfer, a distinction is made between the transfer protocol and the transfer medium. For the former, the industrial
protocols OPC UA and MQTT were presented. The latter seems to be a good choice for data transfer in SHM due to its
high flexibility, real-time capability, edge compute support and integration of an operational dashboard. The choice of
transmission medium is more difficult. The selection criteria are the amount of data to be transmitted and the desired
transmission speed. Transmission technologies such as LoRaWAN, FluidMesh, WLAN, SatCom, LTE, 5G, Ethernet and
fibre have been compared with possible applications and their advantages and disadvantages highlighted. The choice is
made taking into account the conditions at the object to be monitored.
4 Data Security
The security and trustworthiness of data is of great importance throughout the entire value chain. In a brief excursus,
we touched on encryption, authentication and authorisation, network segmentation, security monitoring and incident
response, data protection and preservation of evidence, and data traceability.
5 Data Storage
The goal of secure data transfer lies in data storage. Several methods have already been investigated in different research
projects. In the BrAssMan, DiMaRB and OSIMAB projects, collaboration platforms have been investigated in varying
degrees of detail. The proposed method is based on the Asset Administration Shell (AAS), which is already used in
Industry 4.0. With the Bridge Box (BBox), an AAS configurator for bridge structures has been created that can process
and store both static (documents, structure properties, BIM models) and dynamic sensor data. The data lake behind the
software can be a local storage (here: S3) or a cloud solution. It offers interfaces to provide data for analysis (statistical
or with artificial intelligence) with all common frameworks.
6 Data Analysis
Depending on the type of SHM, different evaluation methods are used. Therefore, standardisation of the method is not
trivial. Evaluation software already supports the most common methods, such as Fourier transforms or time series peak
finding. Other novel methods, such as Structural Health Information Patterns (SHIPs), still need to be developed. With
regard to the digital twin, the aim is to create construction kits adapted to bridge structures, which can be used depending
on the type of data collection.
7 Data Visualisation
The most common type of visualisation is the dashboard view of the measurement computer’s control software. This is
useful for short tests and for calibrating sensors. However, if all the data from the bridge is to be displayed holistically
over the entire life cycle of the bridge, a unified system is advantageous. Files, models and measurement data can all be
displayed in this system. Due to the structure according to the German ASB-ING, all information of the bridge can be
found in a structure known to the client. The display format must be adaptable depending on the current life phase of the
structure. BBox works here in the form of widgets.
8 Practical Example and Outlook
At the German Isen Bridge in Schwindegg, the sequence shown in Figure 1 was carried out. Approximately 140 sensors
were installed and sampled in four measurement chains. These are transmitted via the 4G mobile network to the AAS
BBox using MQTT. The measurements have been running since the bridge was opened to traffic and are primarily used to
test new findings on the use of SHIPs on a bridge. Statistical evaluations are performed in parallel. The BBox dashboard
is being tested on the basis of the live data.
More bridges are to follow in the future, including a collaboration with Deutsche Autobahn GmbH on a 930m bridge over
the Danube.
132
SESSION 5-A
Advances in Bridge Engineering and Technologies 1
13th - Japanese-German Bridge Symposium, Osaka, Japan
IMPROVING DAMAGE PREDICTION BY ASSESSING STRUCTURAL DAMAGE
THROUGH SENSOR MEASUREMENTS IN COMBINATION WITH
VIRTUAL BUILDING MODELS
Nathalie Nießer *
Prof. Dr.-Ing. Geralt Siebert **
* University of the Bundeswehr Munich, Chair for structural design and building physics, Institute and laboratory for
structural engineering, Germany, nathalie.niesser@unibw.de
** University of the Bundeswehr Munich, Chair for structural design and building physics, Institute and laboratory for
structural engineering Germany, geralt.siebert@unibw.de
Extended abstract:
In the construction sector, the importance of structural monitoring to assess potential damage has increased. In general,
damage is defined here as a present or future adverse change to the system. In the case of a building, these are mainly
changes in the material or geometric properties. The most common measurement method for geometric data collection
and documentation of a building is the terrestrial 3D laser scan. It captures precise data through volume-based object
modelling, which can detect changes in even the smallest structures and thus enables complex recording of the geometries
of individual building elements. [1, 2]
This type of damage analysis is tested on a 540 m long hall made of wood with a glazed roof. The repeating structures of
the roof are suitable for comparative investigations of geometries. Repeating geometries can also occur on façade elements
or bridge sections, which is why this object was chosen as a general example.
Since the main interest was ultimately on the timber structure of the building, it is first necessary to concentrate the data
on the essentials. It can also be helpful to divide large structures into smaller sections or clusters. This approach helps to
significantly reduce the amount of data and allows to focus on specific areas or components for a more targeted analysis.
Parallel to the analysis of the point cloud, a CAD model of the structure is also created, in which all additional data of the
structure can be captured. The aim is to use the 3D point clouds and other collected data, e.g. information about leaks in
the roof structure, to expand the existing 3D model in order to decide which areas are best suited for sensor deployment
in the example building. For this purpose, the building was divided into 35 geometrically corresponding sections. In order
to identify possible deviations in the structures, the sections were compared with each other. A low deviation indicates a
smaller geometric change in the structure, while a high value can indicate possible damage or irregularities.
In the course of the damage analysis, it was established that geometric deviations occur especially in the entrance area of
the hall. These deviations also correlate with the visible damage detected at these measured points. Through this analysis
of the geometric changes, sensors can subsequently be used in a targeted manner and integrated into the monitoring
process. The potential of such sensor information systems lies in the continuous availability of measurement and object
data. Later, measured and calculated values can be continuously compared and adjusted within a simulation model. The
aim is to minimise discrepancies between the model and the monitoring results. This approach involves installing various
sensors, such as pressure sensors, humidity sensors, weather stations and more, at strategic locations in the building so
that these sensors continuously collect data. [3]
The proposed concept therefore combines sensor measurements, virtual building models and finite element methods to
enable a more accurate assessment of (structural) damage. Artificial intelligence plays a supporting role in analyzing
patterns in big data, increasing the overall efficiency of the monitoring and analysis process.
Keywords: Structural monitoring, Virtual building model, 3D laser scanning, Point cloud, Digital twin
References
[1] Ehm, M., Hesse, C.: 3D-Laserscanning zur Erfassung von Gebäuden - Building Information Modeling (BIM),
Bautechnik, 91(4), 2014.
[2] Freeden, W, Rummel, R.: Handbuch der Geodäsie. 6 Bände. Living Reference Work, continuously updated
edition. Berlin, Heidelberg: Springer Spektrum, 2016.
[3] Krawtschuk, A., Strauß A., Haider K., Zimmermann T., Bergmeister, K.: Ermittlung von Modellunsicherheiten
bei Stahlbetonstrukturen. Beton- und Stahlbetonbau, 107(12), 2012.
This extended abstract is funded by dtec.bw – Digitalization and Technology Research Centre of the Bundeswehr which
we gratefully acknowledge as part of the project RISK.Twin.
135
136
13th - Japanese-German Bridge Symposium, Osaka, Japan
“SmART Strand” Prestressing Steel Strand
with Optical Fiber Sensor for Tension Monitoring
Masashi Oikawa*, Shinji Nakaue**, Naoki Sogabe***, Michio Imai****
* Sumitomo Electric Industries, Ltd, Tokyo, Japan, oikawa-masashi@sei.co.jp
** Sumitomo Electric Industries, Ltd, Itami, Japan, shinji-nakaue@sei.co.jp
*** Kajima Corporation, Tokyo, Japan, n-sogabe@kajima.com
**** Kajima Corporation, Tokyo, Japan, michio@kajima.com
Abstract
SmART Strand (Figure 1 and 2) is a prestressing steel strand equipped with an optical fiber to accurately measure its
tension force along the entire length of the prestressing steel cable. For prestressed concrete structures and ground anchors,
it is important to be able to verify that the required tension of prestressing steel cable is applied and maintained. In order
to directly evaluate the tension force of prestressing cable during right after stressing and in service, a new measuring
method using SmART Strand has been developed. SmART Strand can be used for the maintenance of prestressed concrete
structures and ground anchors through their prestressing force management.
1 Introduction
To allow a prestressed concrete structure to demonstrate its performance, it is essential to apply prescribed tension to
prestressing steel cables, which serve as tendons, and to maintain soundness during the service life of a structure.
However, the tension of a prestressing steel cable can only be managed based on the hydraulic pressure applied by tension
devices (hydraulic jacks), which are installed at the end of prestressing steel cables when applying tension, and the tension
elongation amount of prestressing steel cables. It is difficult to accurately identify the tension at each position of a
prestressing steel cable that is arranged in a curved profile in a structure. There is no established technique to monitor
tension after completion of construction over the long term.
For ground anchors, it is known that the remaining tension of prestressing steel cables changes due to modification of
geographical features in the vicinity and weathering and deterioration of anchor bodies. Excessive changes cause slippage
of anchor bodies, rupture of prestressing steel cables, and a decrease in resistance against landslides. Such changes are
likely to result in serious events, such as collapse of slopes or structures, or protrusion and falling of anchor heads. It is
required to quickly detect abnormalities and deformation of prestressing steel cables and implement effective measures.
However, the tension distribution in the ground cannot be measured, and it is difficult to accurately estimate the change
factors.
The authors have developed a technique that can solve these issues: to measure and maintain the tension distribution of
prestressing steel strands using SmART Strand, a prestressing steel strand with embedded optical fibers. This
measurement technique makes it possible to detect and evaluate changes in the tension at any position of prestressing
steel strands, which are buried in concrete structures or in the ground, as well as abnormalities of prestressing steel strands
and the scope of influence of such abnormalities. Optical fibers, which are mainly made from glass, are highly resistant
to deterioration due to age, such as corrosion. Optical fibers used for measurement can be extended to a location that is
easily accessible to measurers so that measurement can be conducted safely as needed. This measurement technique is
considered to be suited for long-term monitoring to maintain structures.
Figure 1: Image of embedded optical fibers
Figure 2: ECF type SmART Strand
2 Measurement Result of application to actual structrue
Application to construction of a Prestressed concrete bridge superstructure
To verify the applicability of this measurement technique to actual prestressed concrete structures, SmART Strand was
applied to various prestressing steel cables used for construction of the superstructure of prestressed concrete bridges,
and measurement was conducted. As one example, the measurement results of the internal cable for the Pier 1 (P1)
137
2
capital (Figure 3) and the external cable between Pier 2 (P2) and Pier 4 (P4) (Figure 4) are presented. Anchorages and
hydraulic jacks were installed in the same procedure for construction of ordinary cables, and tension was applied. The
15.2 mm bare wire type SmART strand was used as the internal cable for the P1 capital, and the 15.2 mm ECF type
SmART strand (smooth surface type) was used as the external cable between P2 and P4. For both types of cables, the
tension distribution over the entire length was measured. The influence of tension loss due to friction and due to
reduction of elongation of a prestressing steel cable caused by the biting of a wedge for anchorage was also measured. It
was confirmed that the results exceeded the tension required in the design over the entire length. Measurement was
conducted again 26 months and 19 months later, respectively. It was confirmed that it was possible to conduct
measurement properly.
Figure 3: Measurement results
Figure 4: Measurement results
(internal cable of the P1 capital)
(external cable from Pier 2 to Pier 4)
Application to ground anchors
To verify the applicability of this measurement technique to construction of ground anchors, the technique was applied
to construction of ground anchors for slope reinforcement, and measurement was conducted. The 15.2-mm ECF type
SmART strand (PE-sheathed grit type) was used for the construction. Part of the results is shown in Figure 5. It was
confirmed that the 2 m part (12 to 14 m) on the free length side of an anchor body supported the tension and that tension
was not transmitted to the end. It was also confirmed that tension was almost constant for the free length part, which
transmits tension to an anchor body, and that the tension distribution was as expected in the design.
Measurement was conducted 6 months and 18 months after applied tension, and it was confirmed that soundness was
maintained. For application to ground anchors, deformation model experiments, such as slippage of anchor bodies and
landslides, were conducted separately. A study is being conducted on the possibility of conducting causes analysis
based on the tension distribution profile.
Figure 5: Example of measurement results (ground anchor)
3 Conclusion
We developed a tension distribution measurement technique for prestressing steel strands using SmART Strand. The
technique was applied to prestressed concrete bridges and ground anchors, and its effectiveness was confirmed.
This measurement technique makes it possible to confirm the prestressing steel cable tension at any position, in
concrete structures or in the ground, which was previously difficult. It also makes it possible to select and design the
countermeasure construction properly depending on the amount of deformation, which is estimated based on the
measurement results. Recently, its applications are expanding gradually, for instance, it has been applied to tension
measurement of stay cable on a construction for extra-doused bridge and tension monitoring after construction of
external cable for long span girder bridge. The technique is expected to contribute to advancement of maintenance
operations.
138
13th - Japanese-German Bridge Symposium, Osaka, Japan
Use of data from BIM Method for new and existing concrete bridges Practical report and possible improvements
Dipl.-Ing. Christian Kainz *
M.Eng. Gertraud Wolf **
* University of the Bundeswehr Munich, Chair of Structural Concrete, Germany, christian.kainz@unibw.de
** University of the Bundeswehr Munich, Chair of Project Management, Germany, gertraud.wolf@unibw.de
Extended Abstract
Building Information Modeling (BIM) is considered a methodology that enables the continuous use of digital information
throughout the entire lifecycle of a construction project. The focus lies in the integration of information and building
elements. Even though the use of the BIM-method is mandatory in Germany for all new bridges in the highway system
since 2020 and for federal buildings since 2023 many other projects are currently still realized without the use of BIM.
While there is a strong administration on a federal level, especially local communities with only few buildings under their
management and rare points of contact to construction struggle to implement modern planning methods. While there are
about 40.000 bridges in the federal highway system and about 25.000 railway bridges, the majority of bridges with a
number of about 90.000 are managed by municipalities, states or rural districts in Germany [1-3]. BIM changes the way
bridge design is approached, offering advantages throughout the entire construction process by enabling seamless
collaboration between various stakeholders and disciplines involved in the project. In a comprehensive platform
engineers, contractors, and other professionals can work in a coordinated manner. Through a 3D model, BIM allows for
better visualization and understanding of the bridge's design, leading to improved information workflow and reducing the
likelihood of errors during construction. Two-dimensional views (2D) are derived from three-dimensional (3D)
geometries for the representation and transmission of information. It provides many potentials like advanced co-working,
data exchange, digital archiving and error detection between all project partners and over the whole lifecycle of the
structure. Figure 1 illustrates the data exchange of a bridge structure when multiple planning stakeholders work in an
open BIM workflow and their domain-specific models are integrated into a coordination model. The coordination model
provides the opportunity for a unified and transparent communication basis.
Figure 1: Coordination Model for a rural bridge with the BIM method
In the next years, Germany will continue to invest in infrastructure to cope with the growing traffic demands. The German
procurement procedures for construction services are legally regulated and aim to make the allocation of public contracts
transparent, fair, and competitive. They are primarily based on the Act against Restraints of Competition (GWB) [4] and
the Procurement Regulation (VgV) [5]. Open questions about the scope of service of BIM concern structural engineers,
construction companies and administrations. When BIM is not mandatory, the methodology is only sporadically applied
in practice, limited to specific disciplines, or selected processes. The added value is often restricted to optimizing only
the considered processes, without considering how, for example, collaboration with other specialized planer functions. In
most projects, the scope of services still primarily involves 2D planning, and accordingly, the compensation is also based
139
on this level of service. BIM services are still considered as so called special services according to the Official Scale of
Fees for Services by Architects and Engineers (HOAI) [6]. They must be separately tendered and compensated [7].
Additionally, there is often a change in the planning team between project phases 3 and 5 (according to the HOAI) for
object planning and structural calculations due to verification concerns. When 2D data is the data exchange format,
preparing and processing the data is time-consuming and error-prone due to manual input processes. Although project
knowledge increases with the progress of the project, there are still breaks in the flow of information.
On the example of the design of a small, yet geometrically challenging concrete bridge a practical report is given. The
basis is a 3D-model, see figure 2. Information about the structure is very important, e.g. for future reassessment processes,
bridge inspections or for the system reliability. The existing road bridge over a small river in the midst of a local
municipality is rebuilt as part of the general village renewal. The bridge serves as a connection between different parts of
the village and provides access to the regional road network (connecting to an interstate road). The bridge is used by cars,
heavy vehicles, bicycles, and pedestrians. The water crosses the bridge in a west-east direction at an angle of
approximately 50 degrees. The return walls are aligned with the course of the crossing creek or with the course of the
road. In the longitudinal section the bridge is located in a rounding area of the gradient, causing the longitudinal slope to
vary. The gradient ends before and after the bridge and connects to the planned access road. In the top view, the bridge is
partly located in a radius of R = 25 m. Additionally, the bridge partially covers the intersection of the adjoining road,
leading to variations in the roadway width and the bridge's width. The total width of the structure between the railings is
a minimum of 13.22 m. This single span reinforced concrete frame bridge has a span of 4.30 m (measured perpendicular
to the abutment). It is designed in cast-in-place concrete and is executed with a clear height of at least 1.42 m. The
abutment walls are founded on 6 m long Ø75 cm piles without a pile cap.
Figure 2: 3D reinforcement of the abutments of the bridge
Currently, more than two-thirds of the investments in the Federal Transport Infrastructure Plan (BVWP) focus on
maintaining the existing network, with an emphasis on modernizing and expanding the infrastructure [8]. Most bridge
structures are being renovated to cope with increasing traffic loads and technical requirements [1]. Digital building
information models could bring significant added value in this context. For the assessment of structures, such as their
stability, a digital building model with information about materials, reinforcement layout, etc., can be consulted for
verification. Damages to the structure can be documented and considered on a component-oriented basis.
Keywords: BIM, building information modelling, 3D-reinforcement, rural bridges, concrete bridge
References
[1] W.-H. Arndt, Ersatzneubau kommunale Straßenbrücken. Endbericht. Deutschland, Berlin, 2013. [Online].
Available: https://repository.difu.de/jspui/handle/difu/255098
[2] Deutsche Bahn AG, Brücken bei der Deutschen Bahn. [Online]. Available:
https://www.deutschebahn.com/de/presse/suche_Medienpakete/medienpaket_bruecken-1191268
[3] B. Grabow, Ed., Ersatzneubau kommunale Straßenbrücken - vorläufige Ergebnisse: Projekt zur Abschätzung des
Ersatzbedarfes im Bereich der kommunalen Infrastruktur am Beispiel der Straßenbrücken, Jul. 2013.
[4] Gesetz gegen Wettbewerbsbeschränkungen (GWB) in der Fassung vom 29. Juli 2022
[5] Verordnung über die Vergabe öffentlicher Aufträge (VgV) in der Fassung vom 02. August 2021
[6] Honorarordnung für Architekten und Ingenieure – HOAI, 2021
[7]
AHO-Arbeitskreis: Leistungen Building Information Modeling – Die BIM-Methode im Planungsprozess der
HOAI, 2019
[8]
Verband Deutscher Verkehrsunternehmen e. V. (VDV): Bundesverkehrswegeplan 2030, Köln, 2016
140
13th - Japanese-German Bridge Symposium, Osaka, Japan
Study on Damage Detection of Simply-supported Bridges Using Structural Responses of
Girder Ends
Phyoe W. Hein1, Yoshinao Goi2, Yasuo Kitane3, Kunitomo Sugiura4
Student, Dept. of Urban Management, Kyoto University, Japan, phyoe.hein.47c@st.kyoto-u.ac.jp
2
Asst. Prof., Dept. of Civil & Earth Resources Engineering, Kyoto University, Japan, goi.yoshinao.2r@kyoto-u.ac.jp
3
Professor, Dept. of Civil & Earth Resources Engineering, Kyoto University, Japan, kitane.yasuo.2x@kyoto-u.ac.jp
4
Professor, Dept. of Urban Management, Kyoto University, Japan, sugiura.kunitomo.4n@kyoto-u.ac.jp
1
1. Introduction
In recent years, an increasing trend of damage identification based on static responses has been observed using deflection,
curvature, strain, rotation, influence line, and neutral axis [1]. The established damage identification methods based on
static responses still need to be enforced to overcome the constraints of response features, excitation hurdles by test
vehicles, and limitations for sensor installation. Hence, the research hypothesis is formulated that the measurements of
rotational and longitudinal displacements at girder ends can provide a reliable monitoring system by the reduction of
sensors and no compulsory application of test vehicles. In this study, a novel bridge damage detection approach for a
simply-supported bridge is proposed by using rotation at the hinged support, shortening at the roller support, and
deflection at the point of load application, and the conjugate beam method is applied for theoretical formulations of those
responses under both intact and damaged conditions by defining the reduction in flexural rigidity as a damage indicator.
Moreover, three damage parameters, such as damage location, damage severity, and damage extent, are deliberated. The
succeeding content of the paper is structured into the proposed method, feasibility of damage detection, and conclusions.
2. Proposed Method
The Euler-Bernoulli model of a simply-supported beam with a span of L is considered as shown in Fig. 1(a). Point A is
defined as a hinged support, and Point B as a roller support. Based on quasi-static responses, the traffic load is applied as
a concentrated moving force at a distance of αL from the hinged support, and the point of load application is assigned as
Point P. Rotational angle at the hinged support, deflection at the point of load application, and shortening at the roller
support are targeted and denoted as θA, vP, and uB respectively. When damage occurs at Point C which is at the distance
of βL from the hinged support, the stiffness of the structural system under the damaged area (ηL) is reduced from healthy
flexural rigidity to damaged flexural rigidity (ψEI). In this way, β for damage location from the hinged support, η for
damage extent (area), and ψ for damage severity are introduced for damage detection. The location of the damage can be
detected by checking the shapes of responses since those by the damage are larger than the ones under the intact condition.
(a)
(b)
Figure 1: Proposed Method when α < β (a) real beam; (b) additional elastic load by damage in conjugate beam.
The Conjugate Beam method is primarily applied in formulating general expressions of rotation, deflection, and
shortening of intact and damaged beams, as illustrated in Fig. 1(b). In order to overcome the challenges of figuring out
how heavy vehicles are running over bridges, normalization is proposed so that each displacement response is divided by
its maximum value and expressed as its normalized displacement response. Those normalized values are as below.
𝜃𝐴
𝜃𝐴,𝑚𝑎𝑥
𝑣𝑃
𝑣𝑃,𝑚𝑎𝑥
𝑢𝐵
𝑢𝐵,𝑚𝑎𝑥
=
3√3
2
𝛼(1 − 𝛼)(2 − 𝛼)
(1)
= 16 𝛼 2 (1 − 𝛼)2
=
32
3
(2)
𝛼 2(1 − 𝛼)2 (−2𝛼 2 + 2𝛼 + 1)
(3)
Incremental values of normalized rotation at the hinged support (∆𝜃𝐴 ), normalized deflection at the loading point (∆𝑣𝑃 ),
and normalized shortening at the roller support (∆𝑢𝐵 ) by the damage are derived as follows.
∆𝜃𝐴
𝜃𝐴,𝑚𝑎𝑥
1
= 9√3 ( − 1) 𝜂𝛽(1 − 𝛼)(1 − 𝛽)
𝜓
1
= 9√3 ( − 1) 𝜂𝛼(1 − 𝛽)2
𝜓
141
if 𝛼 > 𝛽
(4)
if 𝛼 < 𝛽
(5)
∆𝑣𝑃
𝑣𝑃,𝑚𝑎𝑥
1
= 48 ( − 1) 𝜂𝛽(1 − 𝛼){𝛼(1 − 𝛽) − (𝛼 − 𝛽)}
𝜓
1
= 48 ( − 1) 𝜂𝛼 2 (1 − 𝛽)2
𝜓
∆𝑢𝐵
𝑢𝐵,𝑚𝑎𝑥
1
= 160 [( − 1) 𝜂𝛽 2 (1 − 𝛼)2 {−𝛼 2 + 2𝛼 − 𝛽 2 }]
if 𝛼 > 𝛽
(6)
if 𝛼 < 𝛽
(7)
if 𝛼 > 𝛽
(8)
if 𝛼 < 𝛽
(9)
𝜓
1
+ 480 [( − 1)2 𝜂 2 𝛽 3 (1 − 𝛼)2 (1 − 𝛽)]
𝜓
1
= 160 [( − 1) 𝜂𝛼 2 (1 − 𝛽)2 {−𝛼 2 + 2𝛽 − 𝛽 2 }]
𝜓
1
+ 480 [( − 1)2 𝜂 2 𝛼 2𝛽(1 − 𝛽)3 ]
𝜓
3. Feasibility for Damage Detection
The applicability of the proposed method for damage detection is determined by scenarios when a single damage is located
near the roller support, at one-fourth of the span, and at the mid-span. Four damage severities (i.e., ψ = 90%, 75%, 60%,
and 50% of EI) at two damage areas (i.e., η = 5% and 10% of L) are considered in each scenario. The results of focused
responses by a damage at the mid-span are mentioned as an example in Fig. 2.
Figure 2: Incremental normalized values of responses at points of interest due to the damage (β = 0.5L, η = 0.1L)
The locations of the maximum in total responses of damaged beams are consistent with those of intact conditions in which
the largest total normalized deflection and shortening occur at α = 0.5L. The variation in the location of applied loading
for the maximum rotation (i.e., 0.42L ≤ α𝜃𝐴,𝑑,𝑚𝑎𝑥 ≤ 0.45L) indicates that the results by rotation measurement at only one
support are relatively inaccurate. The increase in the normalized rotation and deflection due to damage are at their peaks
when both damage and load are at the same position. However, for increased values of normalized shortening, it is slightly
shifted from the point of β = α, except for the mid-span damage. Such a phenomenon may be because the incremental
value of shortening is composed of two terms due to the non-linear relation between shortening and rotation, as mentioned
in Eq. (8) and Eq. (9). Moreover, shortening at the roller support is found to be the most sensitive to the damage. Increased
percentage of rotation at the hinged support due to damage tends to be higher when damage occurrence is approaching
its support, whereas that of deflection at the point of load application remains constant in all damage cases. An increased
percentage of shortening at the roller support by the damage is the maximum when damage is located at the mid-span and
decreases at the same rate when getting closer to respective supports. When the damage area becomes two times larger,
normalized responses tend to be increased about twice. When flexural rigidity is reached at 0.5EI from 0.75EI, the factor
of increase in that rotation and deflection becomes three times, and that of shortening is about 3.18 times that at the 0.75EI
stage, regardless of damage extent and load position. In this way, damage detection is verified by the shapes of responses.
4. Conclusions
The rotational and longitudinal displacements at girder ends are proven to be damage-sensitive responses for detecting
damages in simply-supported bridges. Without regard to damage severity and damage area, the coincidence of damage
location and loading position results in the largest responses. When the damage becomes closer to the hinged support, its
rotation becomes the most significant, and shortening at the roller support is at its peak when the damage is located at the
midspan. When flexural rigidity is reduced from 0.75EI to 0.5EI, the rotation and deflection increase in the factor of 3,
while the multiple of increase for the shortening is about 3.18, irrespective of damage extent and load position. This study
highlights that shortening at the roller support is the most susceptible to the damage among the three focused responses,
and recommends considering it in the implementation of bridge damage detection. It also agrees with previous studies
that sensor installations at girder ends are the most reliable for simply-supported bridges due to considerable variation of
the increased percentage of rotation and shortening, and verifies the efficiency of sensor installation at girder ends.
5. References
[1] Zhang, L., OBrien, E. J., Hajializadeh, D., Deng, L., & Yin, S. (2023). Bridge Damage Identification Using Rotation
Measurement. Journal of Bridge Engineering, 28(5), 04023015.
142
13th - Japanese-German Bridge Symposium, Osaka, Japan
Reuse of Structural Steel Products
Christoph Ehrenlechner, M.Sc. *
Dr.-Ing. Christina Radlbeck; Univ.-Prof. Dr.-Ing. Dipl. Wirt.-Ing. (NDS) Martin Mensinger *
Matthias Müller, M.Sc.; Univ.-Prof. Dr.-Ing. Thomas Ummenhofer **
* Technical University of Munich (TUM), Chair of Metal Structures, Germany, christoph.ehrenlechner@tum.de
** Karlsruher Institute of Technology (KIT), Steel and Lightweight Structures, Germany, matthias.mueller@kit.edu
Abstract
Construction entails massive consumption of resources and energy and thus significantly impacts the environment.
According to the International Energy Agency the building and construction sector is responsible for approx. 40 % of
carbon dioxide (CO2) emissions. Obviously, circular and environmental-friendly construction is indispensable in order to
achieve climate policy goals. Consequently, the cautious use of materials, the exploitation of recycling potential, and the
reuse of suitable building products, such as steel parts, becomes increasingly important.
Reuse offers potential to reduce primary material consumption, reintroduce secondary material into the material cycle
and avoid material bottlenecks. However, the challenge is to evaluate used steel parts in terms of design requirements and
derive their respective reuse potential. Therefore, the foundations for the reuse of steel building products in Germany are
laid within a governmental funded research project.
Keywords: sustainability, reuse, structural steel, evaluation, deconstruction
1 Introduction
Many people in today’s generation are widely aware that the construction sector significantly contributes to climate
change and environmental degradation through emissions and resource consumption. For example, in 2018 the buildings
and construction sector accounted for 39 % of energy and process-related carbon dioxide (CO2) emissions in 11 % of
which resulted from manufacturing building materials and products such as steel, cement and glass [1, 2]. As part of this,
a governmentally funded research project is currently being approved at the TUM and KIT that will focus on the reuse of
steel and wooden construction components. In the following, the part of the project dealing with steel is presented.
2 State of Research and Standardization
Since the political and ecological pressure is high, there are already some studies [3, 4] dealing with the reuse of steel
components. The focus is on feasibility, practicability and boundary conditions. [4] also contains some concrete case
studies. With [5] a document for the regulated execution of steel structures with used steel components is in prospect.
This is intended to close gaps in the determination of relevant properties for the reuse of structural steel and in the design
of 'secondary' steel structures according to the rules of EN 1993. However, since the national introduction of these
regulations is not yet in sight, a technical guide to reuse is to be developed as part of the ‘ReUse’ research project. Its
application is intended to show a practicable way to reuse certain steel construction products (Figure 1).
Figure 1: Typical steel profiles (left) and common load-bearing steel structure (right) for potential reuse.
3 Research Project ‘ReUse’
The prime objective is to significantly reduce the consumption of building materials and the amount of waste generated.
The aim of the ReUse project therefore is to develop such an innovative approach based on the reuse of steel products in
buildings. Figure 2 shows a sketch of the decisive process steps related to the life cycle of steel building products with
subsequent reuse. This so-called ‘cycle of reuse’ includes reusability assessment, deconstruction, testing as well as - if
necessary – conditioning and reassembly.
143
Figure 2: Cycle of ReUse for steel products from existing buildings
4 Technical Guideline
An evaluation procedure in form of a technical guideline for the reuse of steel components is outlined, which should meet
both safety and reliability requirements. The guideline therefore will contain the following procedure for determining
material properties as the basis for a design according to EN 1993 and in accordance with EN 1090 for obtaining a
corresponding certificate of usability:
1.
2.
3.
4.
5.
6.
Data collection;
Reusability assessment;
Labeling;
Deconstruction;
Testing, and;
Conditioning;
5 Conclusion
The principle of reuse has not yet been comprehensively implemented in construction practice. Basically, there is a lack
of specifications, e. g. in the form of standards with clear procedures. The ReUse research project is intended to form the
basis for a practice-oriented technical guideline. In it, the essential process steps for successful reuse of structural steel
are outlined and corresponding evaluation options are shown. The decisive process steps of the ‘ReUse-cycle’ include the
assessment of reusability, deconstruction, testing and conditioning for new applications. In general, steel construction
offers the best conditions for circular building due to its materiality and design principles. In particular, steel structures
with bolted component connections and standardized section cross-sections are ideal for reuse.
6 Outlook
The existing building stock requires complicated deconstruction and as a result still makes reuse difficult. Thus planning
for deconstruction (e. g. demountable connections for beams, columns and facades; shear stability through horizontal
steel beams instead of floors; modular construction, etc.) and the integration of BIM should therefore be standard in the
future.
7 Acknowledgements
This paper is based on the research project ‘Preparation of the reuse of certain building products of timber and steel
construction’, which is funded by the Ministry of State Development and Housing Baden-Württemberg. Special thanks
is expressed to the client, represented by E. Kühnemann, and the research partners at TUM and KIT.
8 References
[1] Global Alliance for Buildings and Construction. 2019 global status report for buildings and construction: Towards
a zero-emission, efficient and resilient buildings and construction sector. International Energy Agency 2019.
[2] Edwards B. Rough Guide to Sustainability: A Design Primer. London: RIBA Publishing 2014.
[3] Coelho A, Pimentel R, Ungureanu V, Hradil P, Kesti J. European Recommendations for Reuse of Steel Products in
Single-Storey Buildings. Coimbra; 2020.
[4] Brown DG, Pimentel RJ, Sansom MR. Structural steel reuse: Assessment, testing and design principles. Ascot,
Berkshire: SCI 2019.
[5] CEN/TC 135. PrCEN/TS-1090-xxx:2022: Execution of steel structures and aluminium structures - Steel structures
- Part xxx: Reuse of structural steel.
144
SESSION 5-B
Fatigue 1
13th - Japanese-German Bridge Symposium, Osaka, Japan
Analytical Study on the Reinforcement of
Intersections Structure in Orthotropic Steel Deck by U-rib Cutting Method
Qihang Shen *
Prof. Dr.-eng. Takashi Yamaguchi **
* Kawada Industries, Inc., Osaka, Japan, keiko.shin@kawada.co.jp
** Engineering Department, Osaka Metropolitan University, Osaka, Japan, yamaguti-t@omu.ac.jp
Abstract:
The orthotropic steel deck, which is widely used in Japan, has fatigue cracks. To repair and strengthen the steel deck, the
U-rib cutting method is proposed, which involves cutting the weld between the U-rib and the deck plate and installing
reinforcement plates to prevent fatigue cracks. The aim of this study is to reduce the stress concentration in the welds
between the deck plate, U-rib and transverse rib and to propose a rational design. The FE analysis was carried out to
clarify the influence of the patch plate at the transverse rib and the stop hole shape on the welded parts. As a result, the
stress concentration of the welded parts between the deck plate and the transverse rib can be reduced by changing the stop
hole shape instead of using a reinforcing patch plate.
Keywords: Orthotropic steel deck, Fatigue, U-rib cutting method, Patch plate reinforcement
1 Introduction:
Orthotropic steel deck is lighter than concrete slabs and are therefore widely used for long-span bridges, so they are often
used for viaduct bridges in urban areas where the dead load is required to be reduced. However, many fatigue cracks have
been reported in the welds between deck slabs and U-ribs (deck-U-rib welds), including fatigue cracks extending from
the root of the corner weld to the top surface of the deck slab. Due to the urgency of these cracks, many researchers are
still investigating the repair and strengthening of these fatigue cracks.
As a preventive maintenance measure against fatigue cracks in the deck and U-rib welds, a method of cutting the U-ribs
of orthotropic steel decks has been proposed which can only be carried out from the u nderside of the steel deck. As shown
in Figure 1, the welded joint between the deck and the U-ribs is cut and instead the deck and the U-ribs are reinforced
with patch plates, which are tightened with high strength studs and single sided bolts. The patch p lates absorb the local
bending deformation of the deck and U-ribs to prevent fatigue cracking in the welds.
Figure 1: U-rib Cutting method
However, due to the complex structure of the intersection, even the reinforced structure by the U-rib cutting method, the
fatigue crack will occur in the welded parts between the transverse rib and the deck plate.
This study investigates the mechanism of stress concentration at the deck and transverse rib welds by developing an FEM
model of a normal orthotropic steel deck reinforced by the U-rib cutting method. Then, the factors such as the shape of
the patch plate, the presence of the patch plate at the intersection, the cutting length of the stop hole, the presence of bead
cutting, the support length of the patch plate are varied to investigate how these factors affect the stresses of the welds
between the transverse rib and the deck plate.
2 FE analysis model
The FE analysis model is shown in Figure 2. The analysis model was constructed using Abaqus 2020 finite element
analysis software for three-dimensional elastic analysis. In constructing the model, the area around the intersection of the
transverse ribs, which is the reinforcement area, was modelled using an 8 -node reduced-integral solid element, while the
other area was modelled using a 4-node reduced-integral shell element. The common element edge length of the focal
weld was set to 1 mm. The element length of the general part was set to 10 mm. For the modelling of the bolts, the nodal
degrees of freedom of the contact surfaces on the steel deck side and the patch plate side, where the contact pressure due
to the tightening of the bolts acts. The contact surfaces between the patch plate and the steel deck were given contact
147
conditions to allow slippage and separation, and the coefficient of friction was set to 0.4 with reference to the Specification
for Highway Bridges.
To relieve the stresses in the deck and transverse rib welds after U-rib cutting, the rigid support model FCD, which
limits the deck deflection by installing the patch plate on the transverse rib (Fig.2 (b)), and the flexible support model
SH, which does not support the deck and reduces the stiffness of the transverse rib (Fig.2(c)), are considered.
Solid element
Shell element
Ls= 78
Steel deck
6
2
Steel deck
Metal Touch
24
Stop-hole
Roller
U rib
Transverse
rib cutting
Transverse rib
Transverse
rib cutting
Reinforced area
by patch plate
Ø24.5
U-rib
Pin
(a) Overall (FCD and SH model)
(b) the detail of Intersections
structure of FCD model
(c) the detail of Intersections
structure of SH model
Figure 2: Finite Element analysis model
ORG
FCD3
CUT
SH-1
FCD1
SH-2
FCD2
SH-3
0
-100
Stress (MPa)
3 Result and Conclusion
The distribution of stress in each case for elements along the deck
side of the deck and the transverse rib weld is shown in Figure 3.
For the rigidly supported FCD-1, the minimum stress is 150 MPa
lower than for the ORG. This is because compressive stresses
predominate at the sides of the deck and rib welds and, in addition,
stress concentration occurs at the welds resulting in high
compressive stresses at the weld tips. In the flexibly supported SH
model, the minimum stresses at the weld tip were 35 MPa and 37
MPa lower than in SH-1 for SH-2 where the weld bead was
machined. This indicates that the stress concentration at the deck
and rib welds can be reduced by flexibly supporting the deck and
slowly changing the deflection of the deck.
-200
-300
tip
-400
Side
-500
side
-600
0
5
Side
curve
10
15
curve
tip
20
25
30
35
side
40
45
50
Distance from the weld root (mm)
Figure 3: Stress of weld detail (deck and
From other results of the FE analysis, the U-rib cut results in a
transverse-rib weld )
larger deck support spacing and greater deck deflection than the
original model, which significantly increases the deck side
stresses at the deck and rib welds to approximately 330% of the original model.
The rigidly supported model, where the patch plate support area was increased, had no effect on the deck and transverse
rib weld reinforcement. However, cutting the bead at the deck and rib welds or increasing the cut length of the stop hole
and decreasing the patch plate support distance reduced the deck-side stresses at the deck and rib welds and showed a
strengthening effect.
In summary, for the rigidly supported model FCD, FCD-2 with increased contact area compared to FCD-1 and FCD-3
with reduced contact area and scallop support showed no strengthening effect on the deck and transverse rib welds. For
the flexible model SH, a reduction in deck-side stresses at the deck and rib welds was observed by cutting the bead at the
deck and rib welds or by increasing the cut length of the stop hole and increasing the support spacing of the support plates
at the supports, showing a strengthening effect.
4 References
[1] Kota Morishita, Takashi Yamaguchi, Akiko Tabata, Manabu Okumura, Tetsuro Hidaka: A study on the retrofitting
method with spheroidal graphite cast iron patch plate for steel deck with trough rib (In Japanese), Journal of Structural
Engineering, Vol.63A, Pages 1331-1342, 2017.3
[2] Kota Morishita, Takashi Yamaguchi, et: Analytical study on the range of reinforcing method for steel deck with
trough rib from underneath (In Japanese), Journal of Structural Engineering, Vol. 64A, 583-593, 2018.3
148
13th - Japanese-German Bridge Symposium, Osaka, Japan
Steel Castings in Infrastructure Projects
Dr.-Ing. Sven Nagel*
Univ. Prof. Dr.-Ing. Max Spannaus**
* IGESS Ingenieurgesellschaft für Stahlbau und Schweißtechnik mbH, Karlsruhe, Germany, nagel@igess.de
** University of the Bundeswehr Munich, Institute for Structural Engineering, Germany, max.spannaus@unibw.de
Keywords: steel casting; casting defects; fracture mechanics; fatigue
1 Introduction
In theory, cast steel components are ideal for use in fatigue-stressed constructions due to their high mechanical strength
and the almost free shaping possibilities to reduce stress peaks. However, the negative effects of casting defects on the
fatigue resistance are, though well known, not sufficiently quantified and not yet transferred into generally applicable
design approaches. Currently, engineers have no other choice than to draw on experience, set unquantified but high overall
demands on production quality and carry out component tests. On the basis of literature data, experimental and numerical
investigations, a simplified design method that couples the manufacturing quality to technically required FAT classes,
using newly defined resistance categories, has been developed in [1]. This article highlights the essential elements of this
simplified concept that was designed for the special requirements of the construction industry. The resulting maximum
permissible defect sizes for the targeted fatigue resistances are in the order of magnitude of the current quality classes.
2 Problem definition
A fatigue-optimized component design tries to keep local stress peaks due to geometric influences as low as possible. In
no other manufacturing process can this be realized as effectively as in the case of cast components. In addition to the
optimized component geometry in terms of stress, the material properties can be adapted to the specific requirements,
significant residual stresses can be avoided through necessary heat treatments and additional stresses resulting from
eccentricities can be reduced. The benefits referred to regarding fatigue-optimized component design are offset by the
casting defects that are unavoidable in cost-effective production. In general, these defects occur both inside the component
and on the surface as geometric deviations, volume defects (e.g., blowholes, inclusions) or two-dimensional, crack-like
defects.
The aim of this paper is to present an easy-to-use design method that links the local fatigue loading capacity of cast steel
components to a maximum allowable defect size. The design concept and the scientific principles have been developed
within the framework of an extensive research project [2] and a dissertation [1]. Within the project, extensive
investigations were carried out on large components, components with real casting defects and on the base material.
Figure 1 shows the investigation matrix.
Figure 1: Investigation matrix, represents all relevant scales of observation for an engineering model [2]
149
3 Backgrounds
The influence of real casting defects on fatigue strength was investigated in [2][1] on the fatigue tensile specimens. The
strongly scattering test results (nominal stress ranges) are shown in Fig. 2a, separated by material and specimen type.
Fractographic examinations revealed different failure starting points. The fracture surfaces for defects that extended to
the surface of the test geometry differed significantly from those of the internal defects. If the stress ranges are related to
the load-bearing cross-section in the failure stress zone, as in Fig. 2b, two populations can be identified. Here, the
regression line of the surface defects not only shows a significantly lower value for the reference value at 2·106 load
cycles, but the gradients for both populations are significantly different.
Figure 2: Test results of the nominal stress vibration amplitudes separated by material and specimen type, b) joint
evaluation of the test results for both materials separated by defect location, stress ranges are related to net
cross-section (from [1])
Based on such extensive knowledge, different influence factors could be determined experimentally. Based on this,
numerical and analytical models could be developed, with the help of those, a further field of parameters could be
investigated. These findings are the basis for the derivation of a general design concept that is easy to use for the structural
engineer.
4 Design Concept
The design concept provides a consistent link between fatigue strength and quality requirements of cast steel components.
The newly defined maximum allowable defect sizes are directly linked to the resistance and take the different mechanical
effects of surface and internal defects into account. This improves the communication between designers and foundries
and motivates to define local quality requirements depending on local stresses and thus to facilitate castability. Decisive
influencing factors such as stress gradients, interaction of neighboring defects, mean stress dependence, geometric
tolerances or brittle fracture are considered by reduction factors. The design is carried out by limiting the local stress
ranges determined by a linear elastic analysis. The resulting defect sizes are in the order of magnitude of existing quality
levels, but so far, they represent mechanical requirements independent of the test method used and are not to be understood
as NDT display characteristics. The method is optimized for the use in civil engineering but can be transferred to other
fields of application and the prevailing special requirements to permit less conservative fatigue resistances (material, surface condition, stresses, and safety concept).
5 References
[1]
Nagel, S. (2021) Design of Cast Steel Components under Cyclic Loading [Dissertation]. Karlsruher Institut für
Technologie, Stahl- und Leichtbau, Karlsruhe.
[2]
Nagel, S.; Ummenhofer, T.; Jung, M. et al. (2021) Abschlussbericht zum Forschungsvorhaben IGF Nr. 19691N
Bemessung ermüdungsbeanspruchter Stahlgussbauteile unter Berücksichtigung herstellungsbedingter Ungänzen.
Forschungsvereinigung Gießereitechnik e.V., KIT Versuchsanstalt für Stahl, Holz und Steine; Fraunhofer IWM;
Fraunhofer IZFP, Düsseldorf, Karlsruhe, Freiburg, Saarbrücken.
150
13th - Japanese-German Bridge Symposium, Osaka, Japan
An Investigation on Prevention of Weld Root Fatigue Crack
by Assistance with Adhesive Bonding
Yifei XU
Mikihito HIROHATA *
Jiahao MAO**
Division of Global Architecture, Graduate School of Engineering, Osaka University, Osaka, Japan,
y-xu@civil.eng.osaka-u.ac.jp
* Division of Global Architecture, Graduate School of Engineering, Osaka University, Osaka, Japan,
hirohata@civil.eng.osaka-u.ac.jp
** Division of Global Architecture, Graduate School of Engineering, Osaka University, Osaka, Japan,
j-mao@civil.eng.osaka-u.ac.jp
Keywords: Fatigue, Steel Bridge, Weld Root, Fillet Welding, Adhesive Bond
1 Introduction
Enhancing the service life and minimizing damage caused by fracture or deterioration through appropriate maintenance,
repair, and reinforcement are crucial concerns in infrastructure, particularly for structures like bridges. Fatigue crack
initiation and propagation pose a significant risk to steel structures, with weld toes being a common site of fatigue crack
formation in steel bridges. Strategies such as grinding, TIG dressing and peening treatment can capably prevent fatigue
cracks originating from weld toes. However, fatigue cracks arising from weld roots present a greater challenge as
mechanical treatments are not feasible at the inside part of weld bead 1). Detecting and repairing these hidden cracks is a
complex task. Therefore, it is necessary to develop approaches that mitigate stress concentration near weld roots and
prevent crack formation during the fabrication phase.
Previous studies 2) have verified the effectiveness of utilizing adhesive bonding in combination with fillet welding in
cover plate joints for mitigating stress near weld roots and potentially enhancing fatigue life. In this study, in order to
investigate the practical applicability of combining fillet welding and adhesive bonding in actual steel structural members,
the joints between the lower flange and the sole plate at a girder end in a steel bridge were selected as the target structures,
FE simulation was conducted to verify the correctness of the previous fatigue test. By comparing stress mitigation and
stress distribution near the weld root in the assisted bonding group, we tried to perform a numerical analysis about
mitigation effect and assess the effectiveness of fatigue prevention provided by the bond insertion.
2 Experimental specimens
2.1 Specimens and materials
The dimensions and geometry of the experimental specimens are shown in Fig. 1. The steel materials used were SM400A
of 22 mm (sole plate) and 9 mm (flange). A 490 N/mm2 class wire (1.2 mm diameter) specified by JIS Z3312 YGW12
was used as a welding material to join the sole plate and flange. The mechanical properties of the steel plates and the
welding wire are shown in Table 1.
Fig. 1 Geometry and dimensions of the specimens
Table 1 Properties of weld material
2.2 Fabrication of specimens
In this study, specimen groups were mainly divided into only-weld specimens (W) and weld and bond specimens (WB).
W specimens used fillet welding to join the sole plate and flange. The original width of the sole plate and the flange was
200 mm. The joined plate was cut into 40 mm wide for each specimen. As for WB specimens, before welding, the sole
plate and the flange were firstly joined by the epoxy resin. After curing for 24 hours, the sole plate and flange were joined
by fillet welding at the edge of the sole plate. Heat-resistant rubber was placed at both ends of the groove to prevent
adhesive bond leakage during curing. The thicknesses of the rubber were 0 mm (without rubber: WB0), 0.5 mm (WB0.5),
and 1 mm (WB1.0). Each group (W, WB0, WB0.5, and WB1.0) has 9 individual specimens, and the total number of
specimens is 36.
151
3 Fatigue experiment
3.1 Experimental conditions
Considering the defects, specimens of W, WB0, and WB0.5 which exhibited no visible weld defects were subjected to
the fatigue experiment. Fig. 2 shows the experimental setup. Based on the applied stress state at the welded parts of the
sole plates in the steel bridges, a four-point bending load pattern was selected. A cyclic load ranging from 2–6 kN was
applied for generating tensile stress around the weld roots. The ratio of the minimum to the maximum load was 0.1 which
corresponded to a stress ratio.
3.2 Results and discussion
Fig. 3 shows the relationship between the applied load range and the fatigue life. The fatigue life was defined as the
number of repetition when the fatigue crack was detected around the weld root visually. The arrows on some cases
represent that there was no crack or fatigue failure even after 2 million cycles of loading. When the load range was higher
than 4 kN, the fatigue life of WB specimens became 5–7 times longer than that of W specimens. The fatigue life of WB
specimens was 3–4 times longer than W specimens when the load range was between 3–4 kN. When the load range was
2 kN, W specimens showed a fatigue crack before 2 million cycles of loading while fatigue cracks were not detected in
WB0 and WB0.5 specimens.
Fig. 2 Experimental setup
Fig. 3 Results of fatigue experiment
4 Elastic FE Analysis
4.1 Analysis model information
According to specimen, three steel material properties and one bond material property same as fatigue test. The elastic
modulus and Possion’s ratio of steel material were 2.06×105 N/mm2 and 0.3. The elastic modulus and Possion’s ratio of
bond material were 3.8×103 N/mm2 and 0.35. In order to simplify the calculation procedure, the mode was divided into
1/4 of the original specimen according to symmetry. As for the gap height between sole plate and flange, when there was
no rubber appication, it was 0.2 mm, which changed to 0.5 mm with a 0.5 mm thick rubber sheet insertion in WB0.5
cases. Load was set at 10 mm away from the edge of flange. Constraint was set at 30 mm away from the edge of sole
plate. Section type was solid and homogeneous. The element type was 8-nodes and 3D stress.
4.2 Stress mitigation provided by bond insertion
The 4 kN loading pattern was chosen to be analyzed, the analysis revealed that the maximum stress occurred near the
center of the weld root in the weld bead, which could potentially lead to crack formation during the fatigue test. The
direction of the maximum principal stress at the root was predominantly aligned with the vertical direction.
Compared with the W specimen cases (498.452 N/mm2), the maximum principal stress value of WB0 was about 2.5 times
smaller (191.108 N/mm2), and the value of WB0.5 case was roughly equal to WB0 (205.909 N/mm2). These findings
suggest that the additional rubber materials used in the bonding process only contributed to defect-free specimens during
the manufacturing stage. The results indicate that even under high loads, the presence of the bond mitigated the opening
caused by tensile stress, thereby improving the performance of the experimental specimens in terms of fatigue resistance.
5 Conclusion
A series of experiments and analysis simulation were conducted and the main results obtained are as follows:
(1) A four-point bending fatigue experiment was conducted, which showed that under several loading patterns, 3–7 times
fatigue life elongation using adhesive bonding was confirmed.
(2) The FE analysis revealed that in the vicinity of the weld root area, the maximum stress value in the cases with
adhesive bond was 2.5 times lower compared to only-welded cases. Besides, due to the ability of the bond to limit
opening displacement, the specimens with adhesive bond were expected to exhibit prolonged fatigue life.
6 References
[1] J. W. Fisher, S. Roy: Fatigue damage in steel bridges and extending their life, Advanced Steel Construction, 11-3
(2015), 250-268.
[2] Y. XU, M. Hirohata, T.Suzuki, H. Konishi, S.Tominaga : Assistive Bonding Assisted Prevention of Weld Root
Fatigue Cracks, Welding Letters, 40-4(2022), 5-8.
152
13th - Japanese-German Bridge Symposium, Osaka, Japan
Crack propagation calculations with scattering material parameters for the assessment of
welded bridges
Dorina Siebert *
Dr.-Ing. Christina Radlbeck *
Univ. Prof. Dr.-Ing. Martin Mensinger *
* Technical University of Munich, Chair of Metal Structures, Munich, Germany, dorina.siebert@tum.de,
c.radlbeck@tum.de, mensinger@tum.de
Abstract:
The residual service life of existing steel bridges can be assessed in detail by fracture mechanical methods. An initial
crack is assumed, and its growth is calculated by applying the Paris equation with the respective material parameters.
Fracture mechanical material properties can be determined experimentally, or literature values may be assumed
alternatively. Ideally, the scattering of material parameters should be considered, e.g. by assuming distribution functions
for the input parameters. Monte Carlo simulations can evaluate the resulting distribution functions efficiently. This paper
focuses on assessing welded structural details of existing railway bridges. Within the framework of fracture mechanical
methods, we consider scattering material parameters by implementing a multiprocessing Monte Carlo approach for
calculating crack propagation. Background and procedure are presented, and exemplary results are shown and discussed.
The presented method allows the efficient evaluation of crack propagation calculations in the frame of linear elastic
fracture mechanics as a basis for assessing welded railway bridges.
Keywords: assessment of welded railway bridges; linear elastic fracture mechanics (LEFM); crack propagation
calculations; scattering material parameters; Monte Carlo simulations
1 Introduction
The residual service life of existing steel bridges is usually determined by the nominal stress concept. Even though this
approach is comparatively simple, residual service lifespans are often underestimated. Fracture mechanics (FM) offer a
more detailed assessment by crack propagation calculations. In the frame of linear elastic fracture mechanics (LEFM),
the crack growth of an assumed initial crack can be described by the Paris equation with the material parameters C and
m. Furthermore, a threshold value of the stress intensity factor range and the fracture toughness are required. These
fracture mechanical material properties describing the Paris curve can be determined by standardized testing.
Alternatively, guidelines or general technical literature summarize material parameters, which can be assumed. Ideally,
the present scattering of material parameters should be considered. For this purpose, distribution functions for the input
parameters can be assumed and evaluated efficiently by Monte Carlo simulations (MCS).
This paper focuses on assessing welded structural components of railway bridges. Within the framework of LEFM, we
consider scattering material parameters by implementing MCS for calculating crack propagation. Multiprocessing is
proposed for a performant simulation and efficient evaluation of the resulting distribution functions. The method is shown
exemplarily for a semi-elliptical surface crack, as it is part of reference models for welded structural details. The focus
lies on implementing MCS in LEFM. Firstly, we summarize some basics of MCS and FM and then introduce the input
parameters needed. For these sections, the reader is referred to the full paper. Subsequently, we present the process of
calculating crack propagation with scattering material parameters, finally discussing exemplary results.
2 Method for crack propagation calculations with scattering material parameters
Crack propagation calculations are implemented in the programming language Python. The growth of an initially assumed
crack is determined by integrating the Paris equation stepwise. In the frame of MCS, the simulations need to be performed
several times with random samples of the input parameters, which are described by distribution functions. Thereby
scattering of the material parameters is considered. The relationship between the number of calculations and the accuracy
of the results determines the needed number of samples. Generally, a quick and automatic generation of the samples and
respective results is aimed. Through a multiprocessing approach in Python, performant simulation and evaluation is
achieved. Multiple simulations are run in parallel with the input parameters represented by vectors, with each entry
reserved for one simulation. Randomly sampled parameters are created by the methods in numpy.random, e.g.,
numpy.random.lognormal for the Paris parameter C. The results are also evaluated in Python. In the further course of the
project, a subsequent evaluation of multiple input parameter combinations by regression analyses is planned.
3 Results and Discussion
This section in the full paper presents the results of multiprocessing MCS for selected parameter combinations. The
number of days until predicted failure is the most demonstrative output parameter and therefore used for visualization. In
[1], three different distribution functions for the one-stage Paris curve are given (for metals in air). These are compared
for one specific crack size and fictional section modulus. Figure 1 compares the results for Paris parameters for weld
material “DNV” according to [2] and for weld material “SNI-W” and parent material “SNI-PM” by [3]. It gets clear that
153
the distribution is similar and follows approximately a normal distribution or, more precisely, a Gumbel distribution.
About 250 samples represent a reasonable compromise between the accuracy of the results and calculation effort. Drawing
a comparison of the results shows that the deviations are the highest for "DNV". This is consistent with the highest
standard deviation among the three parameter sets. Accordingly, the smallest deviations occur for "SNI-PM". For the
assumed input parameters, "DNV" tends to have the longest lifetime, followed by "SNI-PM". Here, the tendency seems
to be determined by C in the case of similar m (3.1 and 3.07), so larger values of C lead to shorter lifetimes. For larger
differences in the slope m (2.8 compared to 3.07), the lower m is the determining factor leading to longer lifetimes, even
if the C is four times higher. However, this does not apply to even higher differences in C, as it is the case for “DNV”
compared to “SNI-PM”, where “DNV” results in longer lifetimes with a higher m.
Figure 1: Standard deviation for the number of days until predicted failure for different Paris parameter sets “DNV”,
“SNI-W” and “SNI-PM” (left) and histograms (right) for initial crack depth a 0 = 5 mm, initial crack width
2c = 25 mm, W = 20 dm³, dimensions according to Table 2 in the full paper and 250 simulations.
4 Summary, Conclusion and Outlook
Crack propagation calculations allow the residual lifetime assessment of existing bridges. The input parameters needed
are mostly not deterministic but scattering, i.e. statistically random variables following distribution functions. Assuming
all parameters on the safe side would result in very conservative results leading to unnecessarily short inspection intervals or residual lifetimes, respectively. This paper presents a multiprocessing MC approach to implement random input
parameters in crack propagation calculations in the frame of LEFM. Exemplarily, the Paris parameters C and m are
considered, wherefrom C is assigned a lognormal distribution according to [1]. MCS approximate the probability
distribution of the complex analytical equations for crack propagation by repeating the calculation for several input
parameter samples. The needed number of samples to obtain accurate results is determined to about 200-250 for the input
parameters investigated. A multiprocessing approach in Python facilitates a performant simulation and evaluation.
Different resulting distribution functions are presented and compared for different input parameter sets. In all cases, the
resulting distribution function follows a Gumbel distribution. The standard deviation of the resulting distribution functions
depends on the standard deviation of the input parameter. Furthermore, the lifetime is determined by C for similar slopes
m. Higher difference in the slope m leads to a lifetime determined mainly by m as far as the difference in C is not too
high. Hence, C and m influence the crack propagation in combination and need to be considered as pair of material
parameters. Assuming the mean values of the selected distribution functions as deterministic results in shorter lifetimes
compared to the mean value from MCS. A good agreement is obtained for lower bound of the 90 % credible interval of
the distributed results for Paris parameters for welds with the upper bound recommendation in [1] for the simplified Paris
law. The presented results show the potential of assessing bridges based on statistical crack propagation calculations.
Future investigations deal with the adequate evaluation of the shown distribution functions also with respect to the safety
level to be reached. Furthermore, distribution functions for other (material) parameters are planned to be implemented
and evaluated to develop a probabilistic concept for assessing welded bridges. As a matter of case, this is also
accompanied by an extension of the assumed input parameters (structural details, dimensions of structural detail and
bridge, traffic) to be able to evaluate a wide range of different bridge details.
5 Funding Acknowledgements
Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – 506471463
6 References
[1]
The British Standards Institution, Guide to methods for assessing the acceptability of flaws in metallic structures:
British standard BS 7910:2019, BSI British Standards Institution, London, 2019.
[2]
DNV, Fatigue strength analysis for mobile offshore units, Det Norske Veritas, Høvik, 1984.
[3]
H.H. Snijder, F.B.J. Gijbers, O.D. Dijkstra, F.J. ter Avest, in: C. Noordhoek, J. de Back (Eds.), Steel in marine
structures: Proceedings of the 3rd International ECSC Offshore Conference on Steel in Marine Structures (SIMS
'87), Delft, The Netherlands, June 15 - 18, 1987, Elsevier, Amsterdam, 1987.
154
13th - Japanese-German Bridge Symposium, Osaka, Japan
Experimental investigation of the ultra-low-cycle-fatigue (ULCF) behaviour
of full-scale steel components
Sergey Chernyshov M. Sc. *
Prof. Dr. Andreas Taras **
* Institute of Structural Engineering, University of the Bundeswehr, Germany, sergey.chernyshov@unibw.de
** Dept. of Civil, Environmental and Geomatic Engineering, ETH Zürich, Switzerland, taras@ibk.baug.ethz.ch
Keywords: Experiments, ULCF, Anti-Seismic Devices, Strengthening and refurbishment
1
Introduction
Generally, the variety of steel hysteresis dampers can be differentiated based on their yielding mechanisms. In practice,
hysteretic steel dampers with axial tension/compression, shear, or moments action have proven their effectiveness.
Depending on the loading type, a corresponding state of stress and strain is established in the dissipation element, which
greatly influences the load-bearing behavior and service life of the steel elements during earthquake-induced loading. The
influence is even more significant when the deformation loads are stronger. This phenomenon is referred to as ultra-lowcycle-fatigue (ULCF) with a service life of up to approximately 1000 cycles. In this paper, we present and discuss the
results of laboratory tests conducted on steel samples in component size under predominantly global cyclic axial
tension/compression and shear loading.
2
Materials and geometry of test specimens
Within the scope of this study, three different steel grades were tested: one grade produced in a conventional manner, a
thermo-mechanically rolled plate, and a grade with a particularly high ratio between ultimate and yield stress. All steel
grades nominally fulfilled the requirements of S355 steel according to EN 10025. From the selected steel plates with a
nominal thickness of 26 mm, tension/compression and shear full-scale specimens were prepared. Figure 1 shows the
stress-strain diagrams and the geometry of the specimens.
b)
engineering stress [N/mm²]
600
S355-n.pl.
c)
S355-TM
S355-Std.
400
200
d)
0
0
10
20
30
engineering strain [%]
a)
Figure 1: Stress-strain diagrams of the materials (a), geometry of the specimens for the standard monotonic tensile
tests (b), full-scale specimens for the tension/compression (c) and shear tests (d)
3
Experiments
In the first step, a tension/compression specimen was pulled apart monotonically for each steel grade to determine the
elastic limit value of the actuator displacement dy (yield displacement). Next, the tension/compression specimens were
subjected to hysteretic displacement controlled with progressively increasing displacement amplitudes in accordance with
the loading history as recommended by the European Convention for Constructional Steelwork [1]. The slender
tension/compression specimens buckled after a few cycles. To prevent the global buckling effect, a support structure
similar to a buckling restrained brace (BRB) was designed and implemented. The stabilized tension-compression
specimens were then subjected to testing following the ECCS test protocol until failure. Subsequently, ULCF tests were
conducted using three different amplitudes of the actuator displacement dmax. In accordance with the testing concept, the
155
ULCF behavior of the different steel grades under consideration was investigated under shear loading. To achieve this, a
shear specimen with a corresponding specialized load application construction was developed. Figure 6 illustrates the
experimental setups and the specimens at the point of failure obtained from the cyclic shear and tension/compression
tests.
a)
b)
c)
d)
e)
f)
Figure 2: view of the experimental setups and the specimens at the point of failure obtained from the cyclic
shear (e, f), supported (c, d) and unsupported (a, b) tension/compression tests
4
Test Results and Discussion
The force-displacement diagrams obtained from the cyclic experiments were evaluated regarding the plastic dissipation
energy accumulated over the service life and the associated accumulated total deformations. This analysis revealed
significant differences in the load-bearing behavior, ultra-low-cycle-fatigue behavior, and residual load bearing capacity
of the investigated steel grades depending on the loading type.
In the unsupported axial tension/compression tests, all steel grades exhibited approximately the same total plastic
deformation energy until failure. By applying global buckling stabilization to the tension/compression specimens, the
specimens made of S355-Std. and S355-n.pl. were able to double their dissipation capacity. In contrast, the specimen
made of S355-TM exhibited a particularly disadvantageous behavior. The results and relationships obtained from the
cyclical pre-tests according to ECCS were further confirmed during the evaluation of ULCF tests. The evaluation of
ULCF tests revealed the following: as the fatigue magnitude increased, the effects of the differences became more
pronounced. With cyclic shear loading, both local and global stability effects, as well as the necking effect, could be
largely eliminated due to the orientation of the main strains and the dimensions of the cross-section used. Probably for
this reason, the specimen made of S355-TM did not exhibit a worse, but rather a significantly higher dissipation capacity
and better fatigue behavior compared to the other two steel grades.
5
Summary, Conclusions, and Outlook
This paper presented experimental investigations on full-scale components made from three different steel grades under
cyclic tension/compression and shear loading. The conclusions of are summarized as follows. The loading type and the
choice of materials have a significant influence on the load-bearing capacity and fatigue behavior of steel components
under plastic cyclic loading, particularly in the context of ULCF. Skillful material selection can greatly enhance the
performance of a steel hysteresis damper. With increasing fatigue amplitude, the differences in load-bearing capacity and
ULCF behavior between the investigated steels became even more pronounced. Therefore, careful consideration of the
load-bearing capacity of a steel hysteresis damper is crucial when selecting appropriate materials.
According to current standards, earthquake protection devices made of steel undergo testing on scale prototypes,
considering the maximum deformation amplitudes expected during seismic events [2]. The material capability must be
evaluated based on specified requirements. In the opinion of the authors, there is a need for further research to optimize
the current standards. It is proposed that the materials in question be tested, if possible, before conducting full-scale
component investigations, specifically focusing on dissipation and ULCF behavior. To achieve this, suitable resourcesaving test methods using small material specimens for different types of loading should be developed. Such
advancements will contribute to enhancing the understanding and performance evaluation of earthquake protection
devices made of steel.
6
References
[1] European Convention for Constructional Steelwork – Technical Committee 1: Recommended Testing Procedure
for Assessing the Behavior of Structural Steel Elements under Cyclic Loads. ECCS-Publication No. 45, 1986.
[2]
DIN EN 15129. German version EN 15129:2009. Anti-seismic devices. 2010.
156
SESSION 6-A
Advances in Bridge Engineering and Technologies 2
13th - Japanese-German Bridge Symposium, Osaka, Japan
Real-time damage assessment of bridge structures based on reduction of natural frequency
under ambient vibration measurement
Khuyen Trong Hoang1, Hiroyuki Uchibori 2, Naoki Nagamoto3
1
Sumitomo Mitsui Construction Co., Ltd., Tokyo, Japan, h-khuyen@smcon.co.jp
Sumitomo Mitsui Construction Co., Ltd., Tokyo, Japan, huchibori@smcon.co.jp
3
Sumitomo Mitsui Construction Co., Ltd., Tokyo, Japan, nagamoton@smcon.co.jp
2
Abstract
In this paper, a novel method using nonlinear incremental modal pushover analysis is proposed to construct a declined
diagram of natural frequency versus structural damage to enable a quantitative assessment of bridge structural damage
conditions. The accuracy of the natural frequency declined diagram was validated by eigenvalue analysis and nonlinear
dynamic response analysis using different inputs of earthquake acceleration. A quantitative method for post-earthquake
structural damage assessment using natural frequency decline was proposed. An integrated system integrating wireless
accelerometers, cloud system and edge computing and this quantitative assessment method were also developed to
automatically assess the post-earthquake bridge condition. By tracking the change in natural frequency after the
earthquake, this system can automatically assess and report the condition of bridges in sub-real time.
Keywords: Damage assessment, wireless accelerometer, natural frequency, ambient vibration, digital
transformation.
1 Introduction
When a structure is damaged, the stiffness of the structure is reduced, which causes the natural frequency of the structure
to decrease. Therefore, the decline in natural frequency can be a quantitative indicator to evaluate the damage condition
of structures. However, the criteria for judging the level of structural damage or the condition of bridge structures using
the frequency decrease are still questionable for engineers. In this paper, a method for constructing a natural frequency
decline diagram using nonlinear incremental modal pushover analysis was proposed. A quantitative method for damage
level assessment of bridge structures using the natural frequency decline diagram was also proposed. This paper also
proposed a monitoring system for bridges and buildings to detect earthquakes and evaluate the structural damage after
these events. The system includes measurement system using wireless accelerometers integrated edge computing and
cloud system for data management and processing. The system can detect hazardous events such as earthquakes and
evaluate in sub-real time and report the damage condition automatically.
S a (m/s 2 )
2 Proposed method for establishing declined diagram of natural frequency
In order to estimate the frequency of a structure due to damage, a method that can account for the nonlinear behavior of
structures is required. This study presents a procedure using nonlinear incremental modal pushover analysis to obtain the
reduced diagram of natural frequency versus damage level of structures. Focusing on rigid-frame prestressed concrete
bridges, the P-Δ curve as shown in Figure 1a, as an example, is obtained from the incremental modal pushover analysis
with the lateral load distribution following the shape of the natural mode with the largest effective mass ratio. The
equivalent frequency is then calculated using the secant stiffness of the P-Δ curve obtained from the pushover analysis.
The declined diagram of natural frequency is obtained by frequency ratio and ductility factor in Figure 1c. That frequency
ratio is obtained by normalizing the frequency against its initial value and the ductility factor is calculated from the ratio
of the representative displacement Δ to its value at rebar yield.
(a) Sa-Δ curve( P-Δ curve)
(a) declined diagram
Figure 1: Declined diagram of frequency obtained from the proposed method
159
3 Validation by eigenvalue analysis and nonlinear time history analysis
Eigenvalue analysis was performed to validate the natural frequency calculated from the initial secant stiffness of the PΔ curve. As shown in Table 1, the result showed a good agreement between the natural frequency calculated by the
proposed method and the eigenvalue analysis for the studied cases of three different bridges. In addition, the nonlinear
dynamic response using earthquake acceleration inputs was performed to verify the frequency change due to damage.
Two different earthquake inputs were used to capture the dominant frequency in given periods during an earthquake using
the hysteresis law. The earthquake inputs were scaled up to rebar yielding, so that the frequency reduction due to rebar
yielding is then compared to the natural frequency calculated from the proposed method. The reduction in natural
frequency due to rebar yielding is consistent between pushover analysis and nonlinear time history analysis for all 5 cases
studied as shown in Table 2. It was also found that in each bridge, the same rebar yielding, and similar damage pattern is
confirmed even though the earthquake acceleration inputs are different. The damage pattern in the nonlinear time history
analysis is found to be similar to the damage pattern obtained from the proposed modal pushover method under the load
distribution following the shape of the mode with the largest effective mass ratio.
Table 1. Comparison of natural frequency results between the proposed incremental pushover analysis method and
eigenvalue analysis
Studied
Order
Mass effective
Eigenvalue analysis
Proposed method
Error
cases
(%)
(Hz)
(Hz)
(%)
Bridge A
1st
40
0.735
0.746
1.5
3rd
23
0.959
0.977
1.8
Bridge B
1st
44
0.366
0.385
5.1
3rd
20
0.723
0.685
-5.3
Bridge C
1st
60
0.262
0.248
-5.3
Table 2. Comparison of natural frequency results between dynamic analysis and the proposed incremental modal puhsover
analysis
Description
Dynamic analysis
Proposed
Error
method
Bridge
Earthquake
Damage
Before
During
Freq ratio
Freq ratio
input
dominant
earthquake
earthquake
(Hz)
(Hz)
(%)
(%)
(%)
Bridge A
Input 1
Mode 1
0.738
0.466
63.1
63.3
0.2
Input 2
Mode 1
0.738
0.467
63.3
63.3
0.0
Bridge B
Input 1
Mode 1
0.367
0.250
68.1
66.9
-1.2
Input 2
Mode 1
0.367
0.229
62.4
66.9
4.5
Bridge C
Input 2
Mode 1
0.268
0.201
75.0
71.4
-3.6
observed nataural frequency
(in ambient vibration)
time
(a) Anomaly, damage detection
Frequency ratio
frequency
frequency
4 Quantitative damage level assessment using natural frequency declined diagram
This study proposes a quantitative method for damage assessment of bridge structures using decline of natural frequency.
Daily ambient data are collected to construct a “as normal” distribution of the structure’s natural frequency. After such a
hazardous event such as an earthquake, ambient vibration data are also measured to capture the natural frequency of the
structures after the earthquake. If the earthquake causes the statistical distribution model to change, possible damage or
anomalies are identified (Figure 2a). The level of damage in the structure is then automatically assessed by comparing
the decline in the natural frequency to thresholds which are defined using a decline diagram as shown in Figure 2b. This
diagram can be archived using the nonlinear pushover analysis method described in Section 2. An integrated monitoring
system incorporating this quantitative damage assessment has also been developed to automatically assess the postearthquake bridge condition in sub-real time.
distribution
Freq. decline digaram
Freq. in earthquake
Ambient vibration freq.
Threshold 1
Rebar yield limit
Threshold 2
Ultimate limit
Ductility factor
(b) Damage level judgement
Figure 2: Damage assessment using the decline diagram of natural frequency of structures
160
13th - Japanese-German Bridge Symposium, Osaka, Japan
Influence of the Longitudinal Reinforcement Ratio of Prestressed Beam Elements on the
Development of Strain and Compression Softening in the Cracked Web
Sebastian Thoma, M.Sc. ∗
Univ. Prof. Dr.-Ing. Oliver Fischer ∗∗
∗
Technische Universität München, Chair of Concrete and Masonry Structures, Germany, sebastian.thoma@tum.de
Technische Universität München, Chair of Concrete and Masonry Structures, Germany, oliver.fischer@tum.de
∗∗
1 Introduction
The shear capacity of prestressed beam elements is a relevant topic given the ageing bridge sector in many countries, which
is why various research projects have been initiated in recent years with the aim of developing refined approaches for the
realistic assessment of the load-bearing behaviour. This paper presents selected results of a series of tests investigating
the shear behaviour of prestressed beam elements under a considerably reduced degree of longitudinal reinforcement
(compared to the available experimental data) and thus links to economically designed bridge cross-sections and expected
longitudinal strain in the ultimate limit state. Based on the measurement data from digital image correlation, the strain
and cracking processes in the plane of the web can be continuously evaluated over the duration of the experiment.
For this purpose, fictitious square panels are spanned in field and support areas, and principal strains are evaluated in
order to estimate the concrete compressive strength under compression softening, i.e., the reduced strength with respect
to transversal tensile strain. Finally, the formulation of an effective concrete compressive strength in the cracked web
(compression softening) is discussed by means of the principal strains and, relying on existing approaches, the vacant
correlation of concrete degradation with final shear force failure in realistic bridge systems is considered.
2 Experimental Investigations
The institute conducts tests on prestressed beam elements, known as substructures, using a specialized setup. These
substructures are analyzed for shear behavior between the load introduction and central support of a reference continuous
beam.
M
V
Figure 1: Experimental Investigations of continuous beam systems using a substructure approach
The test series examines the impact of gradually reducing the degree of longitudinal reinforcement on rectangular crosssections and T-beams. All beam elements feature a minimal amount of shear reinforcement, approximately 90% of the
161
minimum shear reinforcement specified in Eurocode 2. Further structural detailing of the design remains unchanged.
All beams in the test series experienced shear failure. It was observed that a decrease in the amount of longitudinal
reinforcement did not have a detrimental effect on the shear capacity. This was made possible by a significant increase in
strain of the initially moderately prestressed tendons. The failure of the internal force equilibrium depended on the overall
deformation of the system, the stiffness ratios of the tension chords, the cracked compressive stress field in the web, and
the interaction with crossing reinforcement and tendons.
3 Thoughts on Compression Softening
3.1 Digital Image Processing Analysis
Optical measurements offer an extension to conventional measurement technology such as displacement transducers or
strain gauges. For this purpose, images of the measurement field are generated at regular intervals. These are then
correlated or referenced during post-processing of the measurement data, so that statements can be made about, for
example, the 2D displacements of individual measurement points, principal deformation changes or the strains on the
surface of the object under observation. In this way, incipient cracking in the web plane of the prestressed beam elements
can be recorded and evaluated over the entire test period. While the orthogonal parts of the strain are determined in the
course of the digital image correlation in resolution of the selected pixel density, so-called regions of interest (ROI) are
defined as square subsets for the considerations presented below, see Figure 2. The strains in these comparatively large
Figure 2: Schematic representation for the extraction of a fictitious square panel at any position in the web of investigated
beam elements and evaluation of the distortions on the basis of digital image correlation
sections of the measuring field are smeared so that the calculated principal strains per ROI now result in a scalar value.
This approach corresponds to the shear panel experiments, which also only output an average value per component across
their structural dimensions and the crack pattern.
3.2 Analysis of principal compressive stress in the cracked web
For the present consideration, two squares are defined for each field or support area, which correspond in their global
position over the beam height for all test specimens. Different stress and strain states thus result primarily from different
degrees of longitudinal reinforcement and the associated mean longitudinal strain. In the case of a low degree of longitudinal reinforcement and the associated early increase in strain in the chords and resulting bending and shear cracking,
the assumed effective concrete compressive strength drops quickly and significantly when compression softening is taken
into account. This circumstance is not adequately taken into account by constant factors in design standards; the factors
can thus partly be on the uncertain side, especially with respect to the reassessment of existing bridges.
162
13th - Japanese-German Bridge Symposium, Osaka, Japan
A Simulation Model for Heating Correction on I-Shaped Welded Steel Bridge Members
Xiaoyu GUAN *
Univ. Associate Prof. Dr. Mikihito HIROHATA **
Satoshi MUKAWA ***
Dr. Seiji OKADA ****
* Graduate School of Engineering, Osaka University, Japan, x-guan@civil.eng.osaka-u.ac.jp
** Graduate School of Engineering, Osaka University, Associate Professor, Japan, hirohata@civil.eng.osaka-u.ac.jp
*** IHI Infrastructure Systems Co., Ltd., Tokyo, Japan, mukawa7877@ihi-g.com
**** IHI Infrastructure Systems Co., Ltd., Tokyo, Japan, okada0269@ihi-g.com
1 Introduction
Heating correction is generally used to reduce welding deformations because of its
high workability. Finite element method is a powerful tool to simulate the heating
correction process. Currently, most of research focus on spot heating correction [1].
However, linear heating correction is widely used for actual structural members.
The purpose of this research is to conduct a series of basic studies to build a heat
input model applied to linear heating correction simulation.
2 Experiment
The shape of the experimental specimen assumed an I-section plate girder bridge.
The specimen was assembled by two flanges and a web. These members were
joined by fillet welding. Then a horizontal stiffener was welded to the web at 100
mm from the upper flange by fillet welding with two passes in total. The dimensions
of all parts were shown in Figure 1. The number of specimens is two. The material
of all test steels was SM400A. The average welding leg length of stiffener was
7.3mm. After welding, the welding out-of-plane deformations in the web of the
specimens were corrected by heating. The correction method was linear heating by
using a gas burner. As shown in Figure 2, the burner was moved along the length
of the stiffener while heating the back surface of the horizontal stiffener. The length
of the gas flame was 60 mm, the distance between the flame and the horizontal
stiffener was about 40 mm, and the target temperature of the steel plate was 600 °C.
The moving speed of the gas torch was 15.6 mm/s on average.
The temperature histories of the specimens were measured with three
thermocouples. The positions of the thermocouples are shown in Figure 2. Before
and after the heating correction, the out-of-plane deformations were measured by a
dial gauge. The out-of-plane deformation of the specimens was measured in the
direction of the reference line (x). To clarify the out-of-plane deformation caused
by welding of the horizontal stiffener, the out-of-plane deformation was measured
at the time when the welding of the upper and lower flanges was completed. The
out-of-plane deformation was also measured before heating, and the out-of-plane
deformation due to heating was obtained as the difference in deformation before
and after heating.
Figure 1: The shape and
dimensions of the specimen
Figure 2: Heating position
and direction
Figure 3: Analysis model
3 Thermal elastic-plastic analysis
The heating correction procedure was simulated by thermal elastic-plastic analysis. Figure 3 shows the analysis model
based on the experimental specimen. The coupled temperature-displacement analysis was used in the Abaqus software.
The analysis model was constructed using four-node shell elements. Normally, the analysis should start with the welding
process of the specimen to reproduce the residual stress and deformation caused by welding, and then the heating
correction process should be simulated. However, to simplify the simulation, only the heating correction process was
reproduced. The welding of the upper and lower flanges to the web and the stiffener to the web were not modelled, and
the steel plates were directly joined without considering the weld bead. The mesh size was set to 8.8 mm × 10 mm in the
region where heat input was applied by heat correction. The initial residual stresses were not included in the model
because the welding process was not considered. The initial shape of the web was simplified from the out-of-plane
deformation of the web measured by the experiments. Since the tendency of welding out-of-plane deformation was similar
for both the specimen 1 and the specimen 2, the initial shape was based on the welding out-of-plane deformation from
163
the specimen 2. The heat transfer from the model surfaces into the air was
considered as the thermal boundary condition. Fixing the rigid-body displacement
was considered as the mechanical boundary condition.
Previous studies have shown that the positions of the maximum temperature may
not be consistent with the heating center [2]. Under the experimental condition, the
maximum temperature was observed at approximately 17 mm from the center of
the steel plate as shown in Figure 8. The temperature of the shaded area in Figure
8 is lower because the shaded area is the stiffener area, and the infrared camera
could not accurately measure the temperature of this area. The thermal influence
range of the heating correction was a radius of 44 mm based on the experimental
results, as shown in Figure 9. The heat source was based on Gaussian heat input
method which was proposed by Friedman [3]. Since the maximum point of the heat
distribution is not at the center, a Gaussian combined heat source model is used.
Figure 4: The temperature
distribution of experiment
The heat input formula shown in Eq. (1) was applied to the I-section model.
𝑞(𝑟) = 𝑞1 ∙ 𝑒
−
𝑟2
𝑅1 2
− 𝑞2 ∙ 𝑒
−
𝑟2
𝑅2 2
(1)
Where 𝑞(𝑟) is the heat energy, 𝑞1 is the maximum heating energy in the heating
center without being corrected, 𝑞2 is the maximum corrected heating energy in the
heating center, 𝑅1 is the radius of the thermal influence range, 𝑅2 is the distance
from heating center to the maximum tempaerature point, 𝑟 is the distance from
heating center. These values were arranged for acculately simulating the
temperature history obtained by the experiment.
4 Results and discussions
To verify the accuracy of the model, the analytical results were compared with the
experimental results. The results of temperature histories and out-of-plane
deformations were shown in Figures 6 and 7, respectively. Since the temperature
histories of the specimens were similar, only the temperature histories of the
specimen 2 is shown. The analytical results of temperature histories are consistent
with the experimental results. The correction amount of deformation at the center
of the specimen for the two specimens after heating correction were about 1.1 mm
and 0.6 mm, respectively. The correction amount of deformation at the center of
the specimen by FE analysis was about 0.9 mm. The analytical result of out-ofplane deformations was between the two experimental results.
Figure 5: Temperature
measurement by infrared camera
Figure 6: Temperature histories
5 Conclusions
In this study, a series of fundamental studies were conducted on how to simulate
linear heating correction on an I-shaped welded steel bridge member by FE
analysis. The main findings are as follows:
(1) By using the moving heat input method based on the Gaussian combined heat
source model, the temperature histories of the heating correction were reproduced
by thermal elastic-plastic analysis. The analysis results were consistent with the
experimental results.
Figure 7: Out-of-plane
deformations
(2) Only the heating correction process was reproduced without considering the
welding process. The tendency of the out-of-plane deformation after the heating correction could be reproduced roughly.
In the future, it is necessary to study a more accurate and detailed heat input model, and to investigate residual stresses
from welding process to heating correction process. However, the proposed heat input model has a possibility for being
utilized to predict the heating correction process effectively and easily.
References
[1] Hirohata M., Nozawa S., Tokumaru Y.: Verification of FEM simulation by using shell elements for fillet welding
process. International Journal on Interactive Design and Manufacturing (IJIDeM), 2022, 16(4): 1601-1613.
[2] Kumada M., Nakatogawa T., Hirata K.: Heat and mass transfer by Impinging Jet. Journal of Japan Society of
Mechanical Engineers.1973, 76, 822-830 (in Japanese).
[3] Friedman E.: Thermomechanical analysis of the welding process using the finite element method. 1975, 206-213.
164
13th - Japanese-German Bridge Symposium, Osaka, Japan
Development of a temperature model for small-sized box girders
Malik Ltaief, M.Sc.*
Prof. Dr.-Ing. Dipl. Wirt. -Ing- (NDS) Martin Mensinger **
* Technische Universität München, Chair of Metal Structures, Germany, m.ltaief@tum.de
** Technische Universität München, Chair of Metal Structures, Germany, mensinger@tum.de
Keywords: Temperature loads, thermal model, small-sized box girders
1 Introduction
Bridge structures are influenced by climatic conditions throughout their service life. Early research in the 1970s found a
direct relationship between shade temperature and bridge temperature, categorized by construction type (concrete, steel
composite, and steel bridges) [1]. Subsequent investigations in [2] involved simulations and measurements to investigate
the random behavior of temperature loads. Part of this research contributed to the current EN 1991-1-5 [3] wording. In
Germany, bridge constructions with non-accessible hollow boxes are welded airtight due to inspection difficulties.
Fluctuations of the internal pressure in the small-sized box girders will occur as a result of the air tightness and temperature
changes. However, this load case is not separately addressed in the current code EN 1991-1-5 [3]. Part of the research
project "Economic dimensioning of fillet welds of tightly welded box girders" investigates this temperature load case. A
numerical model based on weather data is developed for transient temperature field simulations. An initial parameter
study explores the influence of geographical directions, shading effects, and geographical location on the heating behavior
of the internal air in small-sized box girders.
2 Climatic effects on structures
Structures are subject to varying climatic effects based on their geographical location, orientation, and geometry. These
climatic effects include solar radiation, also known as shortwave radiation, which leads to peak temperature values in
structures during the summer months. Additionally, structures interact with their surrounding environment, experiencing
radiation exchange between the ground and the structure, as well as between the atmosphere and the structure. This
radiation exchange is also referred to as longwave radiation interaction. Furthermore, structures undergo temperature
exchange with the environment due to convection. The combined effects of these climatic interactions are summarized in
the following Figure 1 from [4].
Figure 1: Climatic effects on bridges according to [4]
For the description of all heat transfers, pre-processed measured climate data is required, which is fed into a finite element
model. In this modelling process, the calculated heat transfer coefficient and the shade temperature are used to model the
convective heat transfer. Furthermore, the emissivity coefficients and shade temperature determined as part of this preprocessing step are used to describe the longwave radiation exchange with the environment. To consider the heating effect
of the solar radiation, shortwave emissivity coefficients, shading effects and solar irradiation are taken into account.
3 Parameter study and evaluation of preliminary investigation results
The parameter study examines two cross-sections considering different geographic directions and locations. Additionally,
the effect of shading on the heating of the hollow box is investigated by varying the cantilever length. Three locations
with different annual average shade temperature values are analyzed. The geometry of the examined cross-sections and
the investigated geographic directions are illustrated in Figure 2.
165
Figure 2: Numerically investigated cross-sections and investigated geographic directions of the cross-sections
To investigate the influence of geographic directions on the heating of the internal air temperature, simulations are
conducted using climate data of the German city Mannheim. It can be observed that the north-south direction leads to the
highest increase in internal air temperature, while the west-east direction results in the lowest temperature inside the
hollow box. The directions northeast-southwest and southeast-northwest yield similar air temperatures within the hollow
box and the daily maximum temperatures are in between those of the north-south and the west-east directions.
Furthermore, the influence of the geometry of cross-sections 1 and 2 are compared. It is observed that the shading of the
hollow boxes by cantilevers significantly influence the internal air temperature of the hollow box. The hollow box with
less shading from the cantilever leads to higher temperatures. This can be attributed to the longer duration of direct
sunlight on the web surfaces due to reduced shading.
For section 1, simulations over a time period of 15 to 18 years are carried out for the German cities Mannheim, Schleswig
and Stuttgart. The simulated annual maxima of internal air temperature are compared with the annual maxima of the
shade temperature (Figure 3). A correlation between the annual maxima of the internal air temperature and the shade
temperature is observed for small-sized box girders. A similar correlation was found in [1] regarding the uniform bridge
temperature and the shade temperature.
Figure 3: Correlation between internal air temperature of cross-section 1 and shade temperature
4 Conclusion and outlook
Temperature simulations were performed for two different cross-sections by numerical analysis of climate data from
different locations. The parameter study investigated the influences of geographical directions and varying shading
conditions on the internal air temperature of small-sized box girders. It was observed that the north-south direction leads
to the highest internal air temperatures. It was identified that the shading caused by the cantilever has an impact on the
internal air temperature. By comparing the shade temperature and internal air temperature, a correlation between the
temperatures was found. Further numerical investigations of different cross-sectional shapes and thermal parameters will
be carried out in the course of the research project "Economic dimensioning of fillet welds of tightly welded box girders".
The objective is to establish a correlation between internal air temperature and shade temperature using regression
methods. This will determine whether the relationship between shade temperature and uniform bridge temperature given
in EN 1991-1-5 can be applied to internal air temperature for the design of steel and steel-composite bridges. This analysis
is also required for pure steel box girders, such as those used in tied-arch bridges.
5 References
[1] Emerson, M.: Bridge Temperatures Estimated from the Shade Temperature. Transport and Road Research
Laboratory (TRRL). Wokingham, Berkshire United Kingdom, 1976.
[2] Frenzel, B. et al.: Bestimmung von Kombinationsbeiwerten und -regeln für Einwirkungen auf Brücken.
Bundesministerium für Verkehr. Bonn-Bad Godesberg, Germany, 1996
[3] EN 1993-1-5: Eurocode 1: Actions on structures - Part 1-5: General actions; Thermal actions. 2003
[4] Ltaief, M., Mensinger, M., Mangerig, I.: Fillet welds of tightly welded hollow boxes. Stahlbau, Ernst & Sohn.
Forthcoming.
166
13th - Japanese-German Bridge Symposium, Osaka, Japan
Experimental Study of Sound-based Hammer Test on Composite Structure
Yu Yiran *
Goi Yoshinao **
Sugiura Kunitomo ***
* Dept. of Urban Management, Kyoto University, Kyoto, Japan, yu.yiran.58a@st.kyoto-u.ac.jp
** Dept. of Civil and Earth Resources Engineering, Kyoto University, Kyoto, Japan, goi.yoshinao.2r@kyoto-u.ac.jp
*** Dept. of Urban Management, Kyoto University, Kyoto, Japan, sugiura.kunitomo.4n@kyoto-u.ac.jp
1 Introduction
The hammer test is an inspection method to discriminate defects by listening to the sound generated during the hammer
tap. It has considerable application in visual inspection due to its low cost and convenience. However, the judgment
mainly relies on the technician's experience, and a lack of available technicians limits its use in large-area inspections.
In previous studies, researchers investigated its application in tile de-voiding [1] and concrete defect detection [2] and
used acoustic techniques in concrete void detection [3]. This study aims to quantify the sound pressure during
hammering test in the composite deck and to make the testing results interpretable with higher reliability.
2 Test setup
In this study, an investigation of the applicability of the sound base-hammer test and the correlation between the sound
and vibration data is carried out on a specimen with built-in artificial defects. The test target is a 1600mm×1600mm×
252mm concrete block with an 8mm thickness steel plate covered above, as shown in Figure 1. The defects are
distributed on the surface of the concrete blocks in two types: cavities simulated by plastic foam and artificial
honeycomb in concrete. A scalable automatic modal hammer is applied to excite the specimen, a microphone captures
the generated sound, and a non-contact portable laser-Doppler Vibrometer and an attached accelerometer captures the
vibration. The experiment setup is shown in Figure 2. To investigate the effect of defect type, defect size, boundary
conditions, hitting location, and microphone distance for the test, I used the combination shown in Table 1. In each test,
the defect area will be hit by an auto-hammer ten times with variations on microphone distance and hitting location.
Figure 1: Testing specimen
Table 1:
Figure 2: Experiment setup
Table 2:
Parameter combination
Type
Size
Surroundings
Fine
\
\
30mm
\
Hitting
location
Edge,
Center,
Inside
Microphone
distance
Center
\
50mm
Center
Side support
Honeycomb
80mm
\
Center
\
Side support
Edge,
Center,
Inside
\
Center
100mm
80mm
Cavity
10mm,
15mm,
20mm
\
100mm
Center
Side support
Dominant frequency
Type-sizehitting area
Microphone
distance
Acceleration
data (Hz)
Sound data
(Hz)
Fine-c*
10mm
5780
2294
HC30c*
10mm
5371
2392
HC50c
10mm
3710
3662
HC100c
10mm
3515
3613
HC100c
15mm
3564
3613
HC100c
20mm
3466
3564
Cavity80c
10mm
4980
2001
Cavity100c
10mm
5078
2441
*c for center (hitting area), ** HC for honeycomb
167
3 Data processing and analysis
Defining the correlation between the dynamic response and sound is essential to address subjectivity in the
hammer test. Initial screening of the test results is done by plotting the waveforms. Data that lacks consistency
will be screened out at this step. Then peak-picking algorithm is utilized to segment the test series into singlehitting blocks with a time duration of 0.1 seconds. The Fast Fourier Transform is mainly used to obtain the
dominant frequency of the vibration and sound data. Segments with different frequency domain characteristics are
also screened out in this process. Power Spectral Density (PSD) is then used to verify the FFT result. The most
predominant peaks in the PSD curves are summarized in Table 2.
By the mentioned process, the dominant frequency of the acceleration and sound data are compared. For
honeycomb, sound data has almost the same dominant frequency as the accerelation data. For the other scenarios,
the dominant frequencies were not consistent.
Figure 3: Power spectral density of acceleration
Figure 4: Power spectral density of sound pressures
By comparing the result of the same size honeycomb and cavity, the results show that the honeycomb is easier to
find in the hammer inspection. The similar dominant frequency of sound and vibration data explain why it is
sensually easier to detect.
A longer microphone distance will not affect the frequency feature by comparing the result of the 100mm size
honeycomb by hitting at the center. However, appropriate distance will effectively stabilize the test data quality.
The sound amplitude decreases as the microphone distance increases while the background noise remains the
same; thus, the signal-to-noise ratio increases. 15cm has a better signal-to-noise ratio than 20 cm.
By comparing the result from different hitting locations, as the hitting point gets closer to the center of the
damage, the dominant frequeny of sound decreases, but the peak height increases.
4 Summary
The correlation between the sound and acceleration data in the hammer test for a composite deck is studied.
Investigations of the effect of the defect type, defect size, defect location, and microphone distance on the sound test are
conducted. Results indicate that dominant frequencies in sound signal correspond to dominant frequencies of the
acceleration response near the hitting point under several scenarios with honeycomb in concrete. The honeycomb defect
is more easily captured since it has more evident frequency feature variation in the frequency domain.
References
[1] Tong, F. (2008). Evaluation of tile–wall bonding integrity based on impact acoustics and support vector machine.
Sensors and Actuators A: Physical 144.1, 104.
[2] Lim, Yujin. (2018). Comparison of conventional and acoustic impact echo tests for detecting a cavity underneath
a concrete slab track. WIT Transactions on The Built Environment 181, 499-506.
[3] Ju, J. (2023). Detection and Identification for Void of Concrete Structure by Air-Coupled Impact-Echo
Method. Sensors, 23(13), 6018.
168
SESSION 6-B
Fatigue 2 (Composite & Steel Structures)
13th - Japanese-German Bridge Symposium, Osaka, Japan
FATIGUE ANALYSIS OF RC SLAB REPAIRED WITH EARLY-AGE ULTRAHIGH PERFORMANCE FIBER REINFORCED CONCRETE
Amatulhay PRIBADI*1 and Takashi MATSUMOTO2
1Graduate
School of Engineering, Hokkaido University
Faculty of Public Policy, Hokkaido University
2
*contact: pribadi.amatulhay.a0@elms.hokudai.ac.jp
Keywords: Fatigue, RC Slab, UHPFRC, Early Age Strength
Daily and continuous loads on bridges cause the fluctuation of stress and lead to fatigue damage and reduction of
structural performance. Over some decades, new repairing technology called Ultra-High Performance Fiber
Reinforced Concrete (UHPFRC) has been developed and applied in many reinforced concrete bridge deck slabs.
Previous studies about the examination of fatigue reliability on damaged RC slabs repaired with UHPFRC have
been conducted using experimental and numerical analysis which show the improvement of the punching shear and
fatigue performance of the damaged slab. However, there were some cases when the repair construction duration
should be cut short due to the traffic demand. It is possible that the early-age UHPFRC has not developed its full
strength to resist the traffic load, hence leading to crack formation and degradation problems. Therefore, this
fatigue study is conducted in order to analyze the fatigue behavior of the early-age strength of one-day-old
UHPFRC on repairing the RC slab.
This study was conducted on an original RC slab which was designed based on the Specification for Highway
Bridges 2002 and analyzed using finite element analysis software. The fatigue preloading of the original 230 mm
RC slab with the moving wheel load caused bending cracks on the slab which then was removed only 20 mm from
each of the top and bottom layers. This 190 mm thickness of RC Slab became the unrepaired slab as the basis to
investigate the effectiveness of the early age strength UHPFRC repair. As for the repaired RC slab, the damaged
S230 was removed 40 mm and 20 mm from the top layer and the bottom layer, respectively. Afterward, a 20 mm
UHPFRC was overlaid on the top of the RC slab. Both of the slabs were fatigue reloaded again starting from 1
cycle of 120 kN, then increasing stepwise until 445,000 cycles of 200 kN. The strength properties of one-day-old
UHPFRC are reduced from the 28 days strength due to the early age strength. Thus, the compressive strength of
110 MPa, Young’s modulus of 23,400 MPa, and tensile crack strength of 4.2 MPa at 1 day are assumed in this
study. The concept of bridging stress degradation proposed by Li and Matsumoto (1998) is used in order to
reproduce the fatigue behavior of the RC slab under the moving wheel load.
After being loaded for 300,000 cycles, the propagation of cracked elements in the unrepaired slab exhibits a slight
increase in the crack volume due to the propagation in the top slab. On the other hand, there is no crack
propagation of the repaired slab due to the higher stiffness. A similar trend is also experienced for the evaluation
of center displacement where the repaired slab generated smaller values than the unrepaired one. From the
comparative result, it is shown that the displacement and crack development of the repaired slab is still
significantly slower compared to the unrepaired RC slab thus increasing the fatigue durability. This conclusion
quantitatively evaluates the UHPFRC as a solution for the repairment of the RC slab despite the premature
strength of the one-day-old UHPFRC.
References
Deng, P., Kakuma, K., Mitamura, H., and Matsumoto, T. (2020), “Fatigue analysis of partly damaged RC slabs repaired with
overlaid UHPFRC”, Structural Engineering and Mechanics, 75, 19-32
Li, V.C. and Matsumoto, T. (1998), “Fatigue crack growth analysis of f iber reinf orced concrete with ef f ect of interf acial
bond degradation”, Cement Concrete Comp., 20(5), 339-351.
171
172
13th - Japanese-German Bridge Symposium, Osaka, Japan
Wheel running fatigue test for steel plate-concrete composite deck
using peculiar shape ribs with multi-functional projections
Kozo IWATA*
Risa KATSUKI**
Shota NAKAGAWA***
Prof. Dr. Eng. Shigeyuki MATSUI****
Prof. Dr. Eng. Hiroshi HIGASHIYAMA*****
*Kawada Industries, inc., Osaka, Japan, kozo.iwata@kawada.co.jp
**Kawada Industries, inc., Osaka, Japan, risa.katsuki@kawada.co.jp
***Kawada Industries, inc., Osaka, Japan, shota.nakagawa@kawada.co.jp
****Osaka Institute of Technology, Osaka, Japan, Shigeyuki.mstsui@oit.ac.jp
*****Kindai University, Osaka, Japan, h-hirosi@civileng.kindai.ac.jp
Abstract
The steel-concrete composite deck has high durability and has a proven track record as bridge decks. In order to further
rationalize the conventional Robinson-type composite deck, which has simple flat steel transverse ribs and studs welded
on the bottom steel plate, the authors have developed a composite deck using peculiar shape ribs instead of the flat
transverse ribs to eliminate studs.
The purpose of this study is to investigate the rational arrangement of the transverse ribs in actual bridge decks. In this
study, wheel running fatigue test were conducted on composite deck models with the peculiar shape ribs arranged at
intervals of 500mm and 600mm, to confirm the fatigue durability. In addition, a finite element analysis using a threedimensional elastic model was performed to compare with the measured diflection.
Keywords: Steel plate-concrete composite deck, multi-functional projections ribs, peculiar shape ribs
1 Introduction
The steel-concrete composite deck (hereafter abbreviated as composite deck) has high durability and has a proven track
record as bridge decks. In order to further rationalize the conventional Robinson-type composite deck, which has simple
flat transverse steel ribs and studs welded on the bottom steel plate, the authors have developed a new peculiar shape ribs
[1] shown in Figure 1 instead of the original simple flat ribs. The peculiar shape rib has projections with a thickness of
17 mm on the upper and lower edges of the flat steel, and additionally, node-like projections with a height of 3 mm are
arranged between the top and bottom edges on both sides of the rib at a pitch of 3 rows/100 mm. The projections on the
upper and lower edges of the peculiar shape rib restrain the vertical displacement of concrete, enhance the integrity of the
bottom steel plate. And the node-like projections on the side surfaces have a function of restraining the displacement
caused by the horizontal shear force generated in the bridge decks when the wheel loads act.
The purpose of this study is to investigate the rational arrangement of transverse ribs in actual bridge decks. In this study,
wheel running fatigue tests (hereafter abbreviated as the running tests) were conducted on composite deck models with
peculiar shape ribs arranged at intervals of 500mm and 600mm, as shown in Figure 2. In addition, finite element analysis
(hereafter abbreviated as FE analysis) was performed using a three-dimensional elastic model assembled with all elements
above mentioned.
node-like projections
17
7
※2
33.3
4
9
※2
3
※1
4
6
height
※1,※2change depending on slab thickness
(b) Cross
(a) Side view
Figure 2：Wheel running fatigue test
Figure 1：Multi-functional projections rib
173
2 Comparison between Measurement Results and Analytical Results
The measured results from the running test of the transition of the central deflection are presented. These figures include
the analytical values of the FE analysis. he measured values were converted into the essential load of 98kN from the final
maximum load of 392kN after running 520,000 load cycles. The measured values were also converted per 98kN of applied
load.Deflection of the Test Specimen
For the test specimen with a rib spacing of 500mm, the measured values and analytical values of the variation of the
central deflection from the start to the end of the running test is shown in Figure 3(a). The measured deflection is almost
constant and almost coincides with the analytical result of the bottom steel plate adhesion model. The increase in the
elastic deflection is not recognized and it indicate the stiffness with respect to deflection has hardly decreased throughout
the duration of the test.
For the test specimen with a rib spacing of 600mm, the measured values and analytical values of the variation of deflection
from the start to the end of the running test is presented in Figure 3(b). The deflection measurements, as shown in Figure
10, exhibited a slight increase immediately after the start of the test and then gradually increased at a constant gradient
until the end of the test.
The gradual increase in the deflection was likely caused by the influence of concrete cracks that occurred at the detached
regions, leading to a reduction of the stiffness as the transition from the fully effective section to the section with ignoring
tensile resistance.
1.0
1.0
Rib spacing：500mm
Rib spacing：600mm
Analytical values : steel plate in contact
0.8
Deflection（mm）
Deflection（mm）
Analytical values : steel plate adhesion
Measured values(per 98kN load)
0.6
0.4
0.2
0.0
0
4
8
12 16 20 24 28 32 36
Loading cycles（x10,000）
40
Analytical values : steel plate in contact
0.8
44
48
Analytical values : steel plate adhesion
Measured values(per 98kN load)
0.6
0.4
0.2
0.0
52
0
(a) Rib spacing 500mm
4
8
12
16 20 24 28 32 36 40
Loading cycles （×10,000）
44
48
52
(b) Rib spacing 600mm
Figure 3：Variation of deflection from the start to the end of running test
3 Conclusion
The study investigated the fatigue durability of composite deck using the multi-functional projections ribs as the
transverse ribs. The results obtained in this study are shown below.
1) For the test specimen with a rib spacing of 500mm, the elastic deflection is almost coincides with the analytical results
that the bottom steel plate adhesion model until the end of the test. The increase in elastic deflection was minimal,
indicating that the stiffness regarding deflection was sufficiently maintained until the end of the test.
2) For the test specimen with a rib spacing of 600mm, the initial deflection slightly increased due to the detachment
between the bottom steel plate and the concrete even before the test. Nevertheless, the deflection at the end of the test
was slightly smaller than that of the bottom steel plate-rib fully contact model, and the change in stiffness increased
gradually. But the stiffness was sufficiently maintained until the end of the test.
4 Acknowledgments
Finally, we would like to express our sincere gratitude to all the staff members of the National Institute for Land and
Infrastructure Management for their valuable guidance and support in conducting the running tests in this study.
5 References
[1] Japan Patent Office (JP), Design Registration No. 1588530 [Application Date: November 16, 2016, Registration Date:
2017.9.22, Design Rights Holders: Kawada Industries Co., Ltd., Oji Steel Co., Ltd.]
174
13th - Japanese-German Bridge Symposium, Osaka, Japan
Experimental Fatigue Test on historic Railroad Bridge
Fabian Seitz M.Sc. *
Univ. Prof. Dr.-Ing. Max Spannaus **
* University of the Bundeswehr Munich, Institut for Structural Engineering, Germany, fabian.seitz@unibw.de
** University of the Bundeswehr Munich, Institut for Structural Engineering, Germany, max.spannaus@unibw.de
Keywords: Railway bridge, Monitoring, Fatigue, Damage progress, brittle Failure
1 Introduction:
In Germany, there is still a high number of historic railway bridges in service. According to [1] about 45% (state of
2019) were built before 1940. The average service age of steel bridges is 81 years [1]. The majority of these bridges
were constructed using cast iron, wrought iron, or mild steel. From the advent of modern railway bridges until the
1950s, riveting served as the prevailing method for connecting iron girders. However, with the rise in traffic and heavier
axle loads, these bridges now face significantly greater stress than they were originally designed to withstand. Despite
concerns about certain older steel structures being prone to brittle failure, no major failures have been recorded thus far.
Figure 1: year of construction of German railroad bridges [1]
However, these bridges often fail to meet the demands set by modern standards, particularly in terms of fatigue
resistance and projected lifespan. Despite this, numerous historic bridges have endured the past decades and remain in
excellent condition. Not only for economic reasons but also for ecological considerations great efforts are being made to
keep these bridges in operation, as long as safe operation can be ensured. Additionally, some of these historic bridges
hold significant engineering value and should be preserved as important monuments in the field. A test facility was
established at the Universität der Bundeswehr Munich (UniBw M) to explore failure mechanisms and assess the
remaining service life of a historic bridge. This facility enables full-scale testing of authentic historical bridges.
2 Experimental Investigation on historic bridge
2.1 Object of investigation:
The historic bride used for the experimental investigation was built in 1903 and was in service for over 100 years before
it was dismantled in 2020 (Figure 2). The bridge was located in Gau Algesheim between Mainz and Köln, one of the
most frequently used rail tracks in Germany. Several heavy cargo trains as well as high-speed trains passed the
construction over the last decades. The trough bridge is a riveted construction made from casted iron. According to [2]
the single span bridge with a width of 4.8m could no longer be operated safely. As a result, the construction was
dismantled (Figure 2) and shipped to the Universität der Bundeswehr where the test facility shown in Figure 3 was
built. Trough bridges like this one from Gau Algesheim were a very common construction and there are still a lot in
service today. [3].
2.2 The Test Facility:
The test facility at the Universität der Bundeswehr offers an ideal combination of accessibility, loading flexibility, and
well-defined bearing conditions. A weight can be placed exactly in the center of the bridge. This allows you to
175
Figure 2: Dismantling of historic bridge 2020
Figure 3: Test facility at the UniBw M
determine the stiffness of the bridge. For the dynamic test static ballast and a centrally positioned unbalanced vibration
exciter are used to simulate train passages. The addition of static ballast in the form of concrete blocks (Figure 3) to the
bridge effectively lowers its eigenfrequency from approximately 40Hz to 12Hz. This adjustment is crucial as the
maximum rotational speed of the vibration exciter is 1000 rpm. The frequency of excitation from the vibration exciter
can be continuously adjusted to match the bridge's eigenfrequency. When resonance occurs, the amplitude reaches its
maximum level. Regulating the excitation force can be achieved by adding or removing mass from the vibrator. The
bridge is outfitted with several strain gauges, accelerometers, and displacement transducers. In contrast to
measurements conducted on operational bridges, this facility allows for full-scale tests on authentic historical bridges
until damage occurs. As a result, the remaining service life of the bridge can be determined, and failure mechanisms can
be observed [3].
A crucial aspect of this test facility lies in the comparison of gathered measurement data with results from numerical
calculations. This comparison enables the assessment of the monitoring system's validity for this specific type of
structure. The insights gained from these tests can then be applied to similar structures that are still in use.
2.3 Result of Tests:
The test with the ballast placed in the center and accurately weighed showed that the FE models behaved stiffer than the
real bridge. This could be a consequence of the unknown stiffness of the joints, a loss of material due to corrosion, the
bearing stiffness, scattering of material properties or the stiffness of the composite girders. Further testing needs to be
done to verify the differences.
During the fatigue test, a dynamic load was applied 5 x106 times, resulting in a deflection with a 3.1 mm amplitude at
the center of the bridge. The maximum measured stresses reached up to 50 N/mm2 at the middle cross girder. A
measurement campaign conducted on a similar bridge in operation validated that the stresses induced by the described
test setup are within a realistic range. Throughout the entire test, no significant cracks, loose rivets, or other damages
were observed on the bridge. This test serves as evidence that despite experiencing numerous load cycles, historic
bridges are not inherently susceptible to fatigue failure.
3 Conclusion, Outlook
The fatigue test conducted on the historic trough bridge serves as evidence of the resilience of these aged structures.
Despite undergoing 5 x106 load cycles with realistic loads, no damage was incurred. The author intends to develop
monitoring-supported methods to prolong the service life of bridges for both economic and ecological reasons, as long
as safe operation remains feasible. For the subsequent tests, the concrete blocks will be substituted with steel plates to
decrease damping in the system and increase the deflection amplitude. Moreover, artificial damage will be introduced,
such as support settlement or a small cut in the middle cross girder. The extent of damage will be incrementally
increased until observable crack progression occurs, allowing investigation into the following failure mechanisms,
particularly brittle failure. Following the tests on the bridge from Gau Algesheim, other historic bridges of the same
kind will be subjected to similar examinations.
4 Acknowledgements
This contribution from the project RISK.twin is funded by dtec.bw- Center for Digitization and Technology Research of
the Bundeswehr. dtec.bw is funded by the European Union - NextGenerationEU.
5 References
[1] NARANIECKI H., “Zustandsentwicklung und -prognose von Eisenbahnbrücken”, https://doi.org/10.15488/5532
[2] Ril 805, 2012: DB Netz AG: Richtlinie 805 – Tragsicherheit bestehender Eisenbahnbrücken
[3] SEITZ F., SPANNAUS M., „Experimentelle Untersuchungen an Bahnbrücke – Digitaler Zwilling“
176
13th - Japanese-German Bridge Symposium, Osaka, Japan
Lifetime Fatigue Reliability Analysis
Considering Different Distribution Types
Mohamed Zied Mili*
Kunitomo Sugiura **
Yasuo Kitane ***
* Department of Urban Management, Graduate School of Engineering, Kyoto University, Japan
mili.zied.42c@st.kyoto-u.ac.jp
** Department of Urban Management, Graduate School of Engineering, Kyoto University, Japan
sugiura.kunitomo.4n@kyoto-u.ac.jp
*** Department of Civil and Earth Resources Engineering, Graduate School of Engineering, Kyoto University, Japan
kitane.yasuo.2x@kyoto-u.ac.jp
Abstract:
Over the last few decades, Structural Health Monitoring technics and reliability methods have been used to assess and
predict lifetime fatigue performance. SHM data provide real-time data on the structural response and reliability-based
methods permit to consider the uncertainties related to material properties and loads and predict future performance.
Within the reliability analysis process, assumptions about the distribution type of random variables are made. For
instance, the literature review reveals that lognormal and Weibull distributions are used to model the fatigue
resistance, and lognormal, normal and Gumbel distributions were used to model the effective stress range. This paper
assesses the effect of the random variable’s distribution type on the lifetime fatigue reliability index. Different cases
related to distribution types for resistance and loads are considered and the lifetime reliability index is derived
accordingly. The effect of the distribution type on the prediction of future performance and the time to failure is
discussed.
Keywords: Lifetime Reliability, Fatigue, Distribution type.
1 Introduction
Fatigue is one of the most common damage of steel bridges causing the structure's partial or even total failure.
Therefore, it is important to preserve the fatigue performance of the structure during the service life. In this context,
monitoring technics and reliability-based methods are combined to investigate the current and to predict future fatigue
performance [1]. This allows to avoid accidents and to plan lifetime maintenance works [2]. Monitoring allows
collecting structural response data while reliability-based methods improve the accuracy of analysis and predictions
as it accounts for the uncertainties related to the structural response and loads. The limit state for a fatigue
reliability analysis is given by equation 1:
𝑁
𝑚
(1)
(𝑆𝑅𝑒𝑓𝑓 )
𝐶0
where Δ is the Miner’s critical damage accumulation index, N is the number of stress cycles, C 0 is the fatigue detail
coefficient for each category, m is the slope of the S-N curve, and SReff is the effective stress range.Using Hook’s law and
the rain-flow counting method, the monitored strain data provide the load effect, that is SReff and N. The design standards
and Miner’s rule provide the resistance information, that is Δ, C0, and m.
𝑔(𝑋) = 𝛥 −
For the sake of safety, the design S-N curves are lower than the actual ones. A joint may undergo a number of cycles
higher than the one stipulated by the design code. In addition, fatigue experiments to determine the fatigue strength are
known for the scatter of results [2]. Thus, Δ and C0 are considered random variables [2]. The effective stress range is
derived indirectly from the strain data, in other words, from loads. Therefore, SReff is also a random variable to account
for the variability of loads along the service life.
Several researchers have conducted reliability-based fatigue assessments. Assumptions are often made about the random
variables’ type of distributions. For instance, Kwon and Frangopol used a lognormal distribution to model Δ, C0, and SReff
[3] while Leander and Al-Emrani found that the Gumbel distribution presents a good fit for the bridge response (SReff)
[4]. In several studies, it is assumed that the fatigue lifetimes of metallic materials follow either a lognormal distribution
or a Weibull distribution [5]. However, these assumptions are based on intuition, practical experience, and an attempt to
simplify mathematical operations [5].
Our study focuses on evaluating lifetime fatigue reliability by analyzing various distributions of random variables. To
achieve this, we generated strain data and utilized Matlab for analysis. Our findings indicate that the type of distribution
utilized can significantly impact the results of the reliability analysis.
177
Methodology and results
As part of our research, we conducted a simulation of strain monitoring data over a 50-year service life. Throughout our
analysis period, we considered the consistent evolution of annual traffic volume and strain amplitude, with a fixed value
of 3%.
Table 1: Distribution types of random variables.
Cases
1
2
3
4
Δ
Lognormal
Lognormal
Gumbell
Weibull
Type of distribution
C0
SReff
Lognormal
Lognormal
Lognormal
Normal
Lognormal
Normal
Lognormal
Normal
In the first scenario, all random variables are assumed to be lognormally distributed. Taking into account the properties
of the logarithmic function and the limit state function, the reliability index can be expressed as shown in
[3].
𝛽=
𝜇𝛥 + 𝜇𝐶0 −(𝑚 .𝜇𝑆𝑅𝑒𝑓𝑓 +𝐿𝑜𝑔 𝑁)
(2)
2
2 +𝜎 2 +(𝑚.𝜎
√𝜎𝛥
𝑆𝑅𝑒𝑓𝑓 )
𝐶0
where 𝜇𝛥 , 𝜎𝛥 , 𝜇𝐶0 , 𝜎𝐶0 , 𝜇𝑆𝑅𝑒𝑓𝑓 and 𝜎𝑆𝑅𝑒𝑓𝑓 are the statistical parameters of Δ, C0 and SReff respectively.
For the rest of the cases, we used the Rackwitz- Fiessler procedure to derive the lifetime reliability index. The results are
presented in figure 1.
Lifetime reliabilty index
14
12
Scenario
1
Scenario
2
Scenario
3
10
8
6
4
2
0
5
10
15
20
25
30
35
40
45
50
Figure 1: Lifetime reliability index in the 4 cases
2 Conclusion
It was observed that all cases tended to converge towards the end of the service life. The curves of case 2 and case 3
nearly overlap after 20 years. Case 4 had the largest discrepancy compared to the other curve, but eventually converged
towards the end of the service life.
According to this study, the distribution type of a random variable can impact the reliability index for a product's lifetime
and the estimation of its time to failure. If a limit fatigue reliability index (β lim) is established, the intersection with the
curves representing time to failure would yield varied outcomes. Therefore, it is crucial to conduct a thorough
investigation into the appropriate distribution type for random variables.
3 References
[1] M. Susoy and F. Necati Catbas, Dan M. Frangopol: Evaluation of Time-Variant Bridge Reliability Using
Structural Health Monitoring,
[2] A A Wormsen, G Härkegård: Astatistical investigation of fatigue behaviour according to weibull’s weakest-link
theory, ECF15, 2004.
[3] Kwon, Frangopol: Bridge fatigue assessment and management using reliability-based crack growth
and probability of detection models. Probabilistic Engineering Mechanics 26 (2011) 471–480.
[4]
Leander, Al-Emrani (2016). Reliability-based fatigue assessment of steel bridges using LEFM – A sensitivity
analysis. International Journal of Fatigue 93 (2016), 82–91.
[5]
Li, Wen, Lu, Wang, Deng (2016). Identifying the Probability Distribution of Fatigue Life Using the Maximum
Entropy Principle. Entropy 2016, 18, 111.
178