A Study of the Friction of the Power Cylinder System in Internal
Combustion Engines using a Floating Liner Engine
ARCHNES
by
NAASSAC'"USETTS INSTITUTE
OF EcHNOLOLGY
Zach Westerfield
Bachelor of Science in Mechanical Engineering
The United States Air Force Academy (2013)
[JUL 3 0 2015
LIBRARIES
Submitted to the Department of Mechanical Engineering in Partial Fulfillment of the
Requirements of the Degree of
Masters of Science in Mechanical Engineering
at the
Massachusetts Institute of Technology
June 2015
D 2015 Massachusetts Institute of Technology. All rights reserved.
Signature redacted
Author:
Department of MechanicaEngineering
Certified by:
Signature redacted
May 9, 2015
Tian Tian
Principle R
a ch Engineer, Deartment of Mechanical Engineering
Signature redacted
Thesis Supervisor
Certified by:
David E. Hardt
Professor of Mechanical Engineering
Chairman, Department Committee on Graduate Theses
MITLibraries
77 Massachusetts Avenue
Cambridge, MA 02139
http://Iibraries.mit.edu/ask
DISCLAIMER NOTICE
Due to the condition of the original material, there are unavoidable
flaws in this reproduction. We have made every effort possible to
provide you with the best copy available.
Thank you.
The images contained in this document are of the
best quality available.
2
A Study of the Friction of the Power Cylinder System in Internal
Combustion Engines using a Floating Liner Engine
By: Zach Westerfield
Submitted to the Department of Mechanical Engineering in Partial Fulfillment of the
Requirements of the Degree of Masters of Science in Mechanical Engineering
Abstract
The recent worldwide quest to increase the fuel efficiency of internal combustion engines has led
to great effects to reduce friction of many of the components in these engines. One area of major
concern pertaining to this area is the friction of the piston/ring pack assembly. Because of the
importance and necessity of this component to the internal combustion engine, any
improvements can have relatively large implications for friction reduction. The purpose of this
paper is to investigate several key components of the piston/ring pack assembly and how they
influence friction levels. Specifically, experimental friction trends related to the oil control ring,
piston skirt, liner surface and lubricant will be discussed. The Floating Liner Engine is used in
this study in both the motored and fired configuration to isolate results from components and
provide data for comparative analysis.
The oil control ring (OCR) controls the supply of lubricating oil to the top two rings of the ring
pack and has a significant contribution to friction of the system. This study investigates the two
most prevalent types of OCR in the automotive market: the twin land oil control ring (TLOCR)
and three piece oil control ring (TPOCR). The effect of changing the land width and spring
tension on different liner surfaces for the TLOCR is investigated, and distinct trends are
identified. A comparison is then done between the TLOCR and TPOCR on different liner
surfaces. Results showed the TPOCR displayed different patterns of friction compared the
TLOCR in certain cases.
The piston skirt is also an important contributor of friction in the piston assembly. This thesis
discusses the investigation into low friction coatings on the piston skirt. A brief study of piston
skirt patterns is presented, with little gains being made by applying patterns the piston skirt
coating. Next the roughness of the piston skirt coating is analyzed, and results show that
reducing piston skirt roughness can have positive effects on friction reduction. Finally, an
introductory study into the profile of the piston skirt is presented, with the outcome being that
friction reduction is possible by optimizing the skirt profile.
The final section of this thesis discusses the effects of lubricants pertaining to friction in the
piston assembly. The effects of changing lubricant viscosity through both temperature and
formulation are presented, as well as results from testing the effects of select anti-wear additives
in the oil. The results identify new developments related to lubricant/additive effects on the liner
surface, and how these effects can influence friction.
Thesis Supervisor: Dr. Tian Tian, Department of Mechanical Engineering
3
4
Acknowledgement
My time at MIT has gone by in what seems like an instant. I have had a wonderful experience
here and have learned more than I could ever imagine. The extent of education and experience I
have gained here would not be possible without several people whom I would like to personally
thank.
First and foremost, I would like to thank my research advisor, Dr. Tian Tian. I cannot begin to
list the knowledge and experience, both academic and otherwise, he has shared with me.
Working with him was the highlight of my time at MIT.
I would also like to thank the colorful and unique people that made each at MIT day fun. My
peers at MIT with which I shared my time are some of the greatest people I have ever met. I
must start first with my office mates, who never hesitated to help me when I had a question:
Camille Baelden, Mathieu Pickard, Yang Liu, and Tianshi Fang. I must also thank the other
students in our research group who provided countless hours of help with modeling and
experimental knowledge: Eric Zanghi, Pasquale Totaro, and Renze Wang. I must also thank Dr.
Dallwoo Kim, who served as my mentor by teaching me all about the Floating Liner Engine.
Finally, I would like to thank a few students of the Sloan Automotive Lab who also provided
assistance: Felipe Rodriguez, Morgen Sullivan, and Jacob McKenzie.
No research project is possible without backing from sponsors, and I was fortunate enough to be
a part of a wonderful group of people in the automotive industry. This group is comprised of the
members of the Consortium on Lubrication in IC Engines and includes: Argonne National
Laboratories, Daimler, Mahle, MTU, PSA, Renault, Shell, Toyota, US DOE, Volkswagen and
Volvo. Much appreciation is giving to N. Demas, 0. Ajayi, and C. Lorenzo-Martin at the
tribology department of Argonne National Laboratory for their collaboration in analyzing the
liner surfaces. Thanks to Brian Papke at Shell for his recommendations and information on
tribofilms.
In addition to sponsorship, I also had the pleasure of working at what I consider one of the best
laboratories at MIT: the Sloan Automotive Laboratory. One of the things that makes this lab so
great is the people who make it run on a daily basis. Janet Maslow was an administrative miracle
worker and coordinator for everything. Thane Dewitt and Raymond Phan were there whenever I
needed assistance in the test cells.
Finally, I would like to thank my family, whose support has been undying throughout my
academic journey.
5
6
Contents
Abstract ...........................................................................................................................................
3
Acknowledgem ent ..........................................................................................................................
5
List of Figures ...............................................................................................................................
10
List of Tables ................................................................................................................................
15
Nom enclature........................................................................................................................-....
17
Introduction...........................................................................................................................
19
1.
2.
3.
1.1.
Background ....................................................................................................................
19
1.2.
Piston A ssem bly .............................................................................................................
20
1.3.
Lubrication Theory in the Piston A ssembly................................................................
21
1.4.
Development of Experimental Study on the Engine Power Cylinder System............ 24
1.5.
Objective of Thesis Work ...........................................................................................
25
1.6.
Organization of the Thesis ...........................................................................................
25
The Floating Liner Engine (FLE) System ........................................................................
27
2. 1.
Purpose ...........................................................................................................................
27
2.2.
Test Capabilities.............................................................................................................
27
2.3.
Floating Liner Engine System Validation......................................................................
31
Friction of the Oil Control Rings ......................................................................................
32
3.1.
Test M ethods and Rationale.........................................................................................
32
3.2.
Liner Surfaces ................................................................................................................
32
3.2.1.
Liner W ear ........................................................................................................
33
3.2.2.
Honing.....................................................................................................................
36
3.3.
TLOCR Param eter Com parison..................................................................................
41
3.3.1.
Ring Param eters ..................................................................................................
42
3.3.2.
TLOCR Experim ental Results .............................................................................
43
7
3.3.3.
3.4.
TPOCR Com parison ....................................................................................................
3.4.1.
4.
6.
Ring Type Com parison and Liner Finish..............................................................
Effects of Piston Skirt Designs ...........................................................................................
56
57
57
73
4.1.
Skirt Patterns ..................................................................................................................
73
4.2.
Skirt Roughness..............................................................................................................
75
4.2.1.
RA02.......................................................................................................................
75
4.2.2.
RS5..........................................................................................................................
77
4.2.3.
RS12........................................................................................................................
78
Skirt Profiles...................................................................................................................
85
4.3.
5.
TLOCR Ring Param eter Conclusions..................................................................
4.3.1.
Piston Skirt Profile Results ..................................................................................
86
4.3.2.
Conclusions.............................................................................................................
91
Effects of Lubricants.............................................................................................................
92
5.1.
Effects of Lubricants under Fired Conditions .............................................................
92
5.2.
Effects of Lubricant Additives ....................................................................................
95
5.2.1.
Introduction to Tribofilm s.....................................................................................
96
5.2.2.
Test Procedures....................................................................................................
96
5.2.3.
Friction Results and Hypothesis ...........................................................................
98
5.2.4.
Liner Exam ination Results....................................................................................
110
Conclusion ..........................................................................................................................
112
6.1.
Summ ary ......................................................................................................................
112
6.2.
General Conclusions ....................................................................................................
113
6.3.
Future W ork .................................................................................................................
115
8
9
List of Figures
Figure 1: Breakdown of total engine energy consumption; mechanical friction loss; pistons, rings
19
and rod friction; and ring pack friction [3] .................................................................................
Figure 2: Piston Rings on the Piston Assembly [4]...................................................................
20
Figure 3: Confocal Microscopy Measurement Showing Characteristic Roughness of Typical
........---...................... 21
C ylinder Liner [8]... ...........................................................................
Figure 4: Instantaneous Stribeck Curve Indicating Lubrication Regimes [10] ........................
22
Figure 5: Floating Liner Engine..................................................................................................
28
Figure 6: Cross-Sectional View of FLE Cylinder....................................................................
29
31
Figure 7: Example of the FLE in Motored Test Configuration .................................................
Figure 8: FMEP Comparison Showing Smooth (GG22) Liner Wear Over Time (0.15mm 19.5N
O C R at 80C ) ......................................................................
.......... .
-.... ---------------.......................
34
Figure 9: FMEP Comparison for Rough (GG07) Liner Wear in Mixed Region under Motored
35
Conditions on GG07 Liner IOOC with 0.15mm 19.5N TLOCR...................................................
Figure 10: FMEP Comparison for Rough (GG07) Liner Wear in Hydrodynamic Region under
35
Motored Conditions on GG07 Liner 40C with 0.15mm 19.5N TLOCR ..................................
Figure 11: FMEP Comparison for Rough (GG07) Liner Wear under Fired Conditions GG07 with
36
0.15m m 19.5N O C R .....................................................................................................................
Figure 12: FMEP Comparison of Smooth and Rough Liners under Fired Conditions at 2 bar
37
-.............
...........................................................................................
IM EP
Figure 13: Comparison of Smooth and Rough Liners at 1500 RPM 2 bar IMEP........... 37
Figure 14: Average Stribeck Curve Comparing Liner Roughness in Mixed Friction Regime .... 38
Figure 15: Average Stribeck Curve Comparing Liner Roughness in Hydrodynamic Friction
39
..... . . ------------.............................................
R egim e .................................................................
Figure 16: Modeled Relationship between Hydrodynamic Pressure and Film Thickness for
40
TLO C R s [33] ........................................................................................-----------------....................
Figure 17: Twin Land Oil Control Ring Cross-Section.............................................................
41
Figure 18: Friction Traces Comparing OCR Land Widths; GG22, 1500 RPM, 10.5N TLOCR, 2
......................... 44
bar IME P .......................................................................................................
10
Figure 19: Comparison of FMEP between OCR Land Widths; GG22, 10.5N TLOCR, 2 bar
IM E P .......................................................................................................................................----
44
Figure 20: Comparison of FMEP between OCR Land Widths; GG07, 19.5N spring, 2 bar IMEP
.......................................................................................................................................................
45
Figure 21: Comparison of FMEP between TLOCR Land Widths in Hydrodynamic Region (40C)
under Motored Conditions with Smooth (GG22) Liner and 10.5N Spring .........
46
Figure 22: Comparison of Friction Traces between TLOCR Land Widths in Hydrodynamic
Region (40C) under Motored Conditions (800 RPM) with Smooth (GG22) Liner and 10.5N
S p ring ............................................................................................................................................
46
Figure 23: Comparison of FMEP between TLOCR Land Widths in Mixed Region (100C) under
Motored Conditions with Smooth (GG22) Liner and 10.5N Spring ............
47
Figure 24: Comparison of Friction Traces between TLOCR Land Widths in Mixed Region
(IOOC) under Motored Conditions (200 RPM) with Smooth (GG22) Liner and 10.5N Spring... 47
Figure 25: Comparison of Friction Traces between TLOCR Land Widths in Mixed Region
(100C) under Motored Conditions (500 RPM) with Smooth (GG22) Liner and 10.5N Spring... 48
Figure 26: Land Measurements after Break-In for 0.06 mm TLOCR ......................................
49
Figure 27: Land Measurements after Break-In for 0.15 mm TLOCR ...........
49
Figure 28: Land Measurements after Break-In for 0.23 mm TLOCR ...........
50
Figure 29: Effect of Increasing Spring Tension on Smooth Liners (GG22) at 1000 RPM .......... 51
Figure 30: FMEP Comparison between OCR Spring Tensions for Smooth Liners (GG22) ....... 51
Figure 31: Effect of Increasing Spring Tension on Rough Liners (GG07) at 1500 RPM............ 52
Figure 32: FMEP Comparison between OCR Spring Tensions for Rough (GG07) Liners ......... 52
Figure 33: Comparison of FMEP between TLOCR Tension in Mixed Region (40C) under
Motored Conditions with Rough (GG07) Liner and 0.15mm Land Width ..............................
53
Figure 34: Comparison of FMEP between TLOCR Tension in Boundary Region (100C) under
Motored Conditions with Rough (GG07) Liner and 0.15mm Land Width .........
53
Figure 35: Comparison of Friction Traces between TLOCR Tension in Mixed Region (100C)
under Motored Conditions (100 RPM) with Rough (GG07) Liner and 0.15mm Land Width..... 54
Figure 36: Comparison of FMEP between TLOCR Tension in Hydrodynamic Region (40C)
under Motored Conditions with Smooth (GG22) Liner and 0.15mm Land Width ......
11
54
Figure 37: Comparison of FMEP between TLOCR Tension in Mixed Region (100C) under
Motored Conditions with Smooth (GG22) Liner and 0.15mm Land Width .............................
55
Figure 38: TLOCR Hydrodynamic Friction Trace Comparison (40C) under Motored Conditions
(1000 RPM) with Smooth (GG22) Liner and 0.15mm Land Width....................
55
Figure 39: TLOCR Mixed Friction Trace Comparison (100C) under Motored Conditions (100
RPM) with Smooth (GG22) Liner and 0.15mm Land Width.......................................................
56
Figure 40: Cross Sectional View of Three Piece OCR.............................................................
57
Figure 41: Friction Coefficient Comparison for Different OCR's on GG22 .............................
59
Figure 42: FMEP Comparison for Hydrodynamic Regime GG08 40C ...................................
60
Figure 43: FMEP Comparison for Hydrodynamic Regime GG22 40C ....................................
60
Figure 44: Hydrodynamic Friction Trace Comparing OCRs on Smooth Liner at 40C and 1000
R P M ..............................................................................................................................................
61
Figure 45: Hydrodynamic Friction Trace Comparing OCRs on Rough Liner at 40C and 1000
R PM ..............................................................................................................................................
61
Figure 46: Friction Coefficient Comparison for Different OCR's on GG08 .............................
62
Figure 47: FMEP Comparison for Mixed Regime GG08 80C ..................................................
63
Figure 48: FMEP Comparison for Mixed Regime GG22 80C ..................................................
63
Figure 49: Mixed/Boundary Friction Trace Comparing OCRs on Smooth Liner at 100C and 200
R P M ..............................................................................................................................................
64
Figure 50: Mixed/Boundary Friction Trace Comparing OCRs on Rough Liner at 100C and 200
R P M ..............................................................................................................................................
64
Figure 51: FMEP Comparison between Liner Roughnesses for Hydrodynamic Region (40C) for
T PO C R ..........................................................................................................................................
65
Figure 52: FMEP Comparison between Liner Roughnesses for Mixed Region (100C) for TPOCR
.......................................................................................................................................................
66
Figure 53: Average Stribeck Curve Comparison Liner Roughness for TPOCR ......................
66
Figure 54: Modeled vs. Experimental FMEP Data for TPOCR on GG22 at 40C.................... 68
Figure 55: Modeled vs. Experimental FMEP Data for TPOCR on GG22 at 60C.................... 68
Figure 56: Modeled vs. Experimental FMEP Data for TPOCR on GG22 at 100C.................. 69
Figure 57: Modeled vs. Experimental FMEP Data for TPOCR on GG08 at 40C.................... 69
Figure 58: Modeled vs. Experimental FMEP Data for TPOCR on GG08 at 100C.................. 70
12
Figure 59: FMEP Comparison with Standard TPOCR Configuration and TPOCR with Bottom
R ail Flipped at 40C .......................................................................................................................
71
Figure 60: FMEP Comparison with Standard TPOCR Configuration and TPOCR with Bottom
R ail Flipped at 60C .......................................................................................................................
71
Figure 61: Assumed and Theorized Contact Profiles for TPOCR.............................................
72
Figure 62: Piston Skirt Patterns Tested......................................................................................
74
Figure 63: Comparison of FMEP from Various Piston Skirts at 2 bar Cylinder Pressure .....
74
Figure 64: Comparison of FMEP from Various Piston Skirts at 4 bar Cylinder Pressure .....
75
Figure 65: Measure Roughness Profile for RA02 1002 after Coating.......................................
76
Figure 66: RA02 Piston Break-In FMEP for 2 bar IMEP ........................................................
76
Figure 67: RA02 Piston Break-In FMEP for 4 bar IMEP ........................................................
77
Figure 68: Measure Roughness Profile for RS5 1007 after Coating ........................................
77
Figure 69: RS5 Piston Break-In FMEP for 2 bar IMEP ...........................................................
78
Figure 70: RS5 Piston Break-In FMEP for 4 bar IMEP ...........................................................
78
Figure 71: Measured Roughness Profile for RS 12 1009 after Coating ....................................
79
Figure 72: RS12 Piston Break-In FMEP for 2 bar IMEP ........................................................
79
Figure 73: RS12 Piston Break-In FMEP for 4 bar IMEP ........................................................
80
Figure 74: FMEP Comparison between Skirt Roughness for Broken-In Pistons...................... 81
Figure 75: Friction Comparison between Skirt Roughness at 1000 RPM 2 bar............ 82
Figure 76: Friction Comparison between Skirt Roughness at 1500 RPM 2 bar............ 82
Figure 77: Friction Comparison between Skirt Roughness at 2000 RPM 2 bar............ 83
Figure 78: Friction Comparison between Skirt Roughness at 1000 RPM 4 bar............ 83
Figure 79: Friction Comparison between Skirt Roughness at 1500 RPM 4 bar............ 84
Figure 80: Friction Comparison between Skirt Roughness at 2000 RPM 4 bar............ 84
Figure 81: Typical FLE Piston with GRAFAL Skirt Coating ..................................................
85
Figure 82: Optim ized Piston Skirt Profile .................................................................................
86
Figure 83: FMEP Comparison for Optimized Profile Piston....................................................
88
Figure 84: Friction Comparison between Skirt Profile at 1000 RPM 2 bar ...............
88
Figure 85: Friction Comparison between Skirt Profile at 1500 RPM 2 bar ...............
89
Figure 86: Friction Comparison between Skirt Profile at 2000 RPM 2 bar ...............
89
Figure 87: Friction Comparison between Skirt Profile at 1000 RPM 4 bar ...............
90
13
Figure 88: Friction Comparison between Skirt Profile at 1500 RPM 4 bar ...............
90
Figure 89: Friction Comparison between Skirt Profile at 2000 RPM 4 bar ...............
91
Figure 90: Comparison of FMEP at 2 bar IMEP due to Changing Viscosity with GG22 and
0.23m m 10.5N T L O C R ................................................................................................................
93
Figure 91: Friction Trace Showing Increased Boundary Contact for Lower Viscosity Oil under
Fired Conditions with GG22 and 0.23mm 10.5N TLOCR......................................................
93
Figure 92: FMEP Effect from Varying Oil Temperature under Fired Conditions with HTHS 2.9
Oil and 0.15mm 19.5N TLOCR at 2 bar IMEP........................................................................
94
Figure 93: Friction Change due to Changing Oil Viscosity with Temperature under Fired
Conditions with HTHS 2.9 Oil and 0.15mm 19.5N TLOCR ...................................................
95
Figure 94: Theorized Effect of ZDDP Film on Liner Surface ..................................................
96
Figure 95: Tribofilm Investigation Test Sequence .......................................................................
97
Figure 96: Boundary/Mixed Friction Region for Firing Effect with HA Oil at 100C.............. 98
Figure 97: Hydrodynamic Friction Region for Firing Effect with HA Oil................................
99
Figure 98: Friction Trace Showing Firing Effect on GG07 at 100 RPM 40C.............
99
Figure 99: Average Stribeck Curve Showing Firing Effect with Fully Formulated HA Oil on
GGO7 Liner.................................................................................................................................
100
Figure 100: Hydrodynamic Region Friction Tribofilm Investigation with HK Oil on GG07 at
40 C ..............................................................................................................................................
10 1
Figure 101: Friction Trace Showing Change in Hydrodynamic Friction ...................................
101
Figure 102: Mixed Region Friction for Tribofilm Investigation with HK Oil...........................
102
Figure 103: Change in Boundary Friction Due to Firing with HK Oil 300 RPM 80C .......
103
Figure 104: Mixed/Boundary Region Friction for Tribofilm Investigation with HK Oil .......... 103
Figure 105: Average Stribeck Curve Showing Firing Effect with HK Oil on GG07 Liner with
0.15m m 19.5N OC R ...................................................................................................................
104
Figure 106: Hydrodynamic Region Friction for Tri-Solvent Wash ...........................................
105
Figure 107: Mixed Region Friction for Tri-Solvent Wash .........................................................
106
Figure 108: Change in Mixed Region Friction Due to Firing with Tri-Solvent Wash............... 106
Figure 109: Mixed/Boundary Region Friction for Tri-Solvent Wash ........................................
107
Figure 110: Stribeck Curve Showing Friction Coefficient for HA Oil with Tri-Solvent Wash on
GG07 w ith 0.15m m 19.5N TLOCR ...........................................................................................
14
108
Figure 111: Smoothing Effect Theory Due to Firing..................................................................
109
Figure 112: Pressure Relationship for Washed vs Fired Liners in the Hydrodynamic Regime. 109
Figure 113: Optical Image of Liner Surface at TDC ..................................................................
110
Figure 114: Profilometry Measurement of Liner at TDC...........................................................
111
Figure 115: Optical Image of Liner Surface at BDC..................................................................
111
List of Tables
Table 1: Floating Liner Engine Specifications ..........................................................................
28
Table 2: Roughness of Daimler Liners Tested with Unit Pressure Notation [33]....................
33
Table 3: O C R Specifications .....................................................................................................
58
Table 4: Pistons Available for Testing......................................................................................
85
Table 5: Oil Formulations Used in Testing...............................................................................
95
15
16
Nomenclature
CA
crank angle
CAD
crank angle degree
RPM
revolutions per minute
TDC
top dead center
BDC
bottom dead center
FMEP
friction mean effective pressure
IMEP
indicated mean effective pressure
FLE
floating liner engine
TLOCR
twin land oil control ring
TPOCR
three piece oil control ring
OD
outer diameter
HTHS
high temperature high shear
17
18
1. Introduction
1.1. Background
The worldwide focus on energy consumption and emissions control has led to efforts by the
automotive industry to further investigate ways to increase efficiency while reducing emissions
in internal combustion engines. This quest led to a focus on mechanical friction, which can
consume up to 15% of the total energy input from fuel combustion [1]. Figure 1 shows that this
mechanical friction can be broken into several areas, of which approximately 50% comes from
the piston assembly [2, 3].
Mechanical friction breakdown
Total energy breakdown
Mechanical
Friction
Pistons,
(4-15%
Rings,
Rod
40-55%)
Other
Losses
(47-58%)
Output
(36-41%)
Other
(40--0%)
Ring pack friction breakdown
Pistons, rings and rod friction breakdown
Top Ring
(13-40%)
Rods
(18-33%)
Rings
(8-450%)
Oil Ring
(50-75%)
Second Ring
(10-22%)
Piston
(25-47%)
Figure 1: Breakdown of total engine energy consumption; mechanical friction loss; pistons, rings
and rod friction; and ring pack friction [3]
19
1.2. Piston Assembly
The piston assembly serves as the mechanism that converts the chemical energy of combustion
into mechanical energy output by the engine. In order to accomplish this, the assembly has two
main functions: first, reciprocate in the cylinder and second, effectively seal the combustion
chamber. The assembly features several design characteristics which help accomplish this. The
standard assembly features a series of three rings, shown in Figure 2, which are set in
circumferential grooves around the piston.
Combustion Chamber
Top Ring
Second Ring
Twin Land Oil
/
Control Ring
Figure 2: Piston Rings on the Piston Assembly [4]
The top ring serves to seal off the combustion gasses, while the second ring provides secondary
control of oil flows on the liner. Primary control of oil flow is accomplished by the third ring,
known as the oil control ring or OCR. This ring controls the amount of oil that is available to the
top two rings. The dilemma in designing the system lies in the need for the rings to seal off the
combustion chamber and minimize the amount of oil that passes by into the combustion chamber
(a process known as oil consumption), all while minimizing the friction of the system as much as
possible. Increasing ring tension may decrease oil consumption, but is also increases friction.
Any oil that passes by can burn during combustion, which increases oil consumption and
emission levels, as well as poisons the exhaust after-treatment with inorganic components of
additives in the oil [5, 6].
20
In addition to the piston and rings, the cylinder liner has certain characteristics which play a
critical role in the quandary between oil consumption and friction. The surface is machined with
a characteristic roughness that typically takes the form of smoother plateau areas and deep
valleys, in a cross-hatched pattern [4]. In recent years, manufacturers have developed processes
to hone the cylinder liners in such a way as to achieve a desired characteristic roughness [7].
Figure 3: Confocal Microscopy Measurement Showing Characteristic Roughness of Typical
Cylinder Liner [8]
1.3. Lubrication Theory in the Piston Assembly
When discussing the contribution of friction from the piston assembly, there are some
fundamentals which must first be understood. The friction between the piston rings and the
cylinder liner is governed by the principles of lubrication theory, which were originally
developed for bearing applications [9]. The friction coefficient is typically plotted on a Stribeck
Curve as a function of the oil dynamic viscosity and rotational speed divided by the load applied.
In the piston assembly application, the friction coefficient is plotted as a function of the dynamic
viscosity of the oil, p; the instantaneous piston linear speed, U; the ring tension in tangential
direction, Ft and the diameter of the engine cylinder bore [10]. The Y axis of the curve is
calculated from the friction data at each crank angle.
21
Bown&yl*
Mixed
S
Hydrodynamic
AtU Bom /fft
Figure 4: Instantaneous Stribeck Curve Indicating Lubrication Regimes [10]
When discussing friction between two sliding surfaces, the friction coefficient is the sum of the
sliding contact coefficient between two dry, solid surfaces (fb) and the hydrodynamic coefficient
of friction, fhFO = af + (1 -
H*fh
d(1)
In this case, a represents the solid-to-solid contact constant and is a value ranging from 0 tol,
with I being complete solid-to-solid contact [2]. There are three regimes that describe the nature
of the contact, as shown in Figure 4. The first regime, called the boundary region, occurs when a
approaches 1, and there is significant contact between the asperities of the two sliding surfaces.
This region is characterized by a high overall coefficient of friction and high material wear.
The second region, on the right side of the chart, is the hydrodynamic region. In this region,
where a approaches 0, hydrodynamic pressure in the lubricant trapped between the two sliding
surfaces causes them to separate, so that no asperity contact occurs. The friction that arises in this
region comes from the shearing of the fluid, and is much lower than in the boundary region [2].
The shear stress can be written as a function of the dynamic viscosity of the lubricant, p, and the
velocity gradient in the y direction.
22
dv;
y)
T
(2)
The hydrodynamic friction coefficient, fl, comes from the non-dimensional form of the shear
stress. In this term, U is the sliding speed of the surfaces, and h is the clearance between them
[11].
Typically, little to no wear occurs in the hydrodynamic regime, as long as the lubricant maintains
its integrity.
The third (and middle) region is the mixed region, named for the fact that it includes a mixture of
the other two regimes. While some asperity contact still occurs, there is also a degree of
separation caused by hydrodynamic pressure in the lubricant [2]. As seen on the Stribeck curve,
this region can exhibit a rapid change in friction coefficient with only a relatively small
adjustment to input parameters. Equation 4 shows the breakdown of the friction coefficient in
this region.
Fc~
A
+-(4)
h
In addition to the instantaneous Stribeck curve, there is also an average Stribeck curve. The
difference in this approach is that the friction coefficient is calculated from the measured FMEP
for a cycle, rather than the frictional force at each crank angle. This term, known as the average
friction coefficient, makes up the Y axis and is a function of the FMEP per cycle, the displaced
volume of the cylinder, Vd, the distance traveled by the piston in one cycle (4 times the stroke
length, 1), and the tangential force of the OCR ring, Ft.
Avg.Fc
FMEP
* Vd
4=1 2wF
(5)
4 * 1 * 27rFt
The X axis of the average Stribeck curve differs from the instantaneous version only by the
velocity term. Since the Y axis is a function of the average friction work per cycle (FMEP), only
a characteristic velocity is needed. Due to its simplicity of calculation, the maximum piston
speed, U,,, was chosen. The complete X axis term is shown in Equation 6.
23
*
Umax * Bore
2Ft
(6)
A unique feature of the average Stribeck curve is that ring tension appears in both the X and Y
axis, meaning it will cancel out and the resulting graph is normalized for ring tension, allowing
for comparisons of different OCR tensions at the same time.
1.4. Development of Experimental Study on the Engine Power Cylinder
System
Because of the implications on fuel efficiency and oil consumption, the piston/cylinder system
has been an area of study for quite some time. The basis for information comes from bench tests
to study friction and wear behavior of piston/cylinder materials, such as rotary tests [12, 13, 14,
15, 16] and reciprocating tests [17, 18, 19]. These tests were used to determine properties of wear,
scuffing and the friction coefficients for the materials and oils typically used in the engine power
cylinder. In addition, a modified Cameron-Plint high frequency machine can be used to study the
effects of oil properties [17, 20], new materials and coatings [12, 21] on engine friction and wear.
The next step in understanding the Engine Power Cylinder comes from a tool known as the
Floating Liner Engine (FLE). The FLE is a specialized internal combustion engine that is used in
the traditional fired configuration to measure friction and wear related to the power cylinder
system [12, 22, 23, 24, 25, 26, 27, 28, 29]. It allows for very accurate and repeatable
experimental measurements that have been invaluable in developing the understanding of
friction and wear as well as aiding the development of computer modeling tools.
The FLE system has been used by many people and undergone multiple modifications to study
specific areas of the engine power cylinder system. In order to study the effects of piston skirt
design, piston skirt profile, lubricant film thickness distribution, oil viscosity, and cylinder
clearance on the piston assembly during engine operation, Takiguchi et al. made various
modifications of the system [22, 23, 24, 25, 30].
Expounding on that work, Cho et al. used motoring tests to study the lubrication behavior of a
barrel-shaped OCR under fully flooded conditions, and was successful in isolating the friction
from the piston rings, without the effect of the piston skirt [31].
24
In his PhD thesis work, K. Liao demonstrated the repeatability and self-consistency of the FLE
used by MIT. He developed methods and test procedures to acquire data under motoring
conditions, to include identifying the engine speeds and liner temperatures that display all
lubrication regimes: boundary, mixed, and hydrodynamic. Liao further discussed proper methods
to measure and select components such as piston rings, liners, and oil. Another main focus of the
research work was documenting and analyzing the friction change of cylinder liners throughout
the break-in process. Finally, Liao developed and discussed ways to match existing models to
experimental results, with great success [10].
1.5. Objective of Thesis Work
While many of the foundation blocks have been laid in the realm of friction in the piston power
system, there has been little work bringing these principles together in order to create a greater
understanding of how each of the components interacts to contribute to the friction of the overall
system. Bench tests have been well documented showing the friction and wear of various
material and lubricant types, but the general understanding of the many parameters of the piston
power system and how they contribute to friction is still undeveloped.
The purpose of this thesis is to identify key trends and principles of friction contributions from
the various parts of the piston assembly. Experimental results will be presented that build upon
this foundation, with the intent of expanding the understanding of the various components of the
system. Specifically, results showing friction trends of different OCR types, piston skirt types,
and lubricant types will be presented. These results will be pivotal in establishing a broader
understanding of friction and wear in the piston power system, which in turn will aid in the
world-wide quest to improve efficiency and reduce oil consumption.
1.6. Organization of the Thesis
This thesis starts in the second chapter by introducing the FLE system and its capabilities. A
brief history of the development of the FLE is given and the procedures for testing are discussed.
The main results of this thesis are split into three sections. While separate in effect, each of these
sections remains interrelated, just as the piston assembly is an interrelated system of parts. The
first of these sections, located in the third chapter, discusses findings related to the friction
25
contributions of oil control rings. It details several specific designs and identifies friction trends
related to key parameters in OCR design. Finally, this section discusses the effect of changing
the liner roughness on friction.
The fourth chapter deals with friction contributions from the piston skirt. Results are presented
on different patterns as well as skirt roughnesses. Finally, new piston skirt profiles are introduced
and discussed.
The fifth chapter investigates the effects of lubricants in the piston assembly. The results show
how changing lubricant viscosity can affect friction in this realm. Next, some effects of lubricant
additives are introduced and explored.
The sixth and final chapter summarizes the results presented and concludes the thesis. It also
details areas of future study that are related to the topics previously presented.
26
2. The Floating Liner Engine (FLE) System
2.1. Purpose
The purpose of the Floating Liner Engine (FLE) is to measure the friction contributions from the
piston/liner assembly. This includes friction from any part that comes in contact with the
cylinder liner, namely, the piston itself, oil control ring, oil scraper ring, and compression ring.
This does not include the friction contributions from the piston wrist pin or connecting rod. The
advantage of this system is it allows for very accurate measurements of friction for a system that
is quite complicated and unsteady. The system is most useful when used to identify trends by
isolating certain components, such as the effect of only increasing OCR tension. While these
trends can help to better understand the broad picture of friction from the piston assembly, they
should not be taken as an ultimatum, such as "a 10 N OCR is the best".
2.2. Test Capabilities
This particular FLE was provided by Professor Masaki Takiguchi at Tokyo City University
(formerly Musashi Institute of Technology) in 2009. The FLE is a specialized internal
combustion engine in which the cylinder sleeve is essentially "free floating", and connected to
the cylinder via two load sensors (Figure 6). This setup allows accurate measurement of friction
contributions from the piston assembly while the engine is in operation.
27
ICylindE He
Cylinder Block
(Floating Liner)
~-F~ase
Secondary Balaer
Figure 5: Floating Liner Engine
Table 1: Floating Liner Engine Specifications
0.496L
92.8 mm
82.51 mm
10:1
0.7MPa
3000RPM
Displaced volume
Stroke
Bore
Compression ratio
Maximum BMEP
Maximum engine speed
28
Lateral
Then
P iston
Pre-Load Bolt
Piston Cooling Jet
Cylinder Sleeve
Figure 6: Cross-Sectional View of FLE Cylinder
The uniqueness of the FLE is that a number of parameters in the piston assembly can be easily
altered and isolated for testing. A list of these parameters includes:
" Engine speed
*
Liner finish
" Compression (top) ring
" Oil scraper (second) ring
29
*
OCR
* Piston
* Liner temperature
" Oil temperature
*
Oil type
*
Cylinder pressure
In addition to these parameters, there are two types of standard test procedures that can be
accomplished with the FLE.
Operating the engine under fired conditions (referred to as a Fired or Firing Test) involves
operating the engine in the conventional sense, using standard 87 octane gasoline. Cylinder
pressures are controlled at both 10 and 20 bar peak pressures, which correlates to 2 bar and 4 bar
IMEP respectively. The thrust side of the liner is maintained at 100 degrees C
the oil is kept steady at 85 degrees C
1 degree while
1 degree. This procedure involves first warming up the
engine, then collecting data "points" at 1000, 1500, and 2000 RPM at 2 bar IMEP cylinder
pressure. For fired tests, each data collection "point" represents the average of data collected for
90 engine cycles. Next, the cylinder pressure is increased to 4 bar IMEP, and the data is again
recorded at the same engine speeds. This complete procedure, referred to as a "sweep", is
repeated 5 times to complete a test "run".
The FLE can also collect data in an alternate method referred to as motored testing (referred to as
motored tests). This procedure, developed for this specific engine by K. Liao [10], utilizes a
dynamometer to power the engine, rather than combustion. The cylinder head is removed, and
the crank case is opened to alleviate pressure imbalances. In addition, a smaller diameter piston
is used, so as to minimize friction contributions from the piston skirt and further isolate OCR
friction. Using the FLE in this manner allows consistent analysis of factors with small friction
traces (such as the OCR), without the unsteady effects of combustion. The standard procedure
for testing under motored conditions is to collect a data point (average of 30 engine cycles) at
100 RPM intervals from 100 to 1000 RPM. This "sweep" is conducted five times with both the
oil and thrust side of the liner temperature being held at 40 degrees C
30
0.5 degrees. The
procedure is then repeated at oil and liner temperatures of 60, 80, and 100 degrees C
0.5
degrees C to complete the entire test "run" [10].
Figure 7: Example of the FLE in Motored Test Configuration
2.3. Floating Liner Engine System Validation
In his Ph.D. research K. Liao demonstrated the repeatability of the FLE under motored
conditions. The FLE was demonstrated to record data with a standard deviation under 3 percent,
and a relative standard error under 2 percent [10].
31
3. Friction of the Oil Control Rings
3.1. Test Methods and Rationale
Since the tension of the OCR can surpass the tension of the other two rings combined, there is a
great interest in focusing on friction trends associated with different parameters and types of
OCRs. Furthermore, the OCR controls the amount of oil supplied to the other two rings, and
therefore has a direct effect on their friction as well. In addition to standard fired tests, standard
motoring tests were conducted with a high clearance piston and only the OCR installed. This
piston, referred to as the "small" piston, has an outer diameter of 82.382mm plus an approximate
15 pm coating on each side of the skirt (for an additional 30 pm in diameter). The "large" piston
used under fired conditions has an outer diameter of 82.455mm plus an approximate 15 Im
coating on each side of the skirt (for an additional 30 pm in diameter). Testing in the motored
configuration allows for the isolation of the effects of the OCR without the contributions of the
top two rings and only minimal contributions from the piston skirt. This procedure was
developed and discussed by K. Liao at MIT [10]. Throughout the investigation, conclusions were
drawn by changing only one parameter at a time and comparing the trends in friction results.
3.2. Liner Surfaces
The single most important surface when discussing the piston assembly is that of the liner
surface. Every component discussed in this study interacts in some way with the liner surface.
Just a few functions of the liner surface are: provide a friction coefficient due to asperity contact,
promote oil retention, and transport of worn debris. When referring to friction, the liner surface is
critical to achieving mixed friction (and the lowest friction coefficient on the Stribeck curve) as
the piston assembly slides back and forth. The exact nature of this roughness is an area of
concern which is easy tested in the FLE.
A variety of liner roughnesses in the classical cross-hatched honing were provided by Daimler
for testing in the FLE. These liners are made of grey cast iron, and are classified according to
structure height and plateau character, shown below in Table 2. Smoother liners are located to
the left of the graph, such as the GG21 and GG22, while rougher liners such as the GG07 fall to
the right.
32
Table 2: Roughness of Daimler Liners Tested with Unit Pressure Notation [32]
3
1GG21
GG30
*GG22
+L
A3
2
B3
C-
0
C3
*GG28
GG28
CO)
OGG09
4-
A2
1
B2
GG22
GG21
C2
AGG07
U
0
C9)
AG
GG09
AGG08
EU
A1
0
0
B1
C1
1
2
Structure Height R3k = Rpk + Rk + Rvk [ m]
*GG30
3
3.2.1. Liner Wear
The original roughness of the liner surface is not a static condition. Just like any sliding solid
contact surface, there is wear which continuously changes the nature of the surface. For starters,
the liner undergoes a break-in period in which this wear is significant at first, and then stabilizes
for a time to a nominal rate. During this period, only the outstanding features of the asperities are
removed, but the main geometrical structure of the liner finish is unchanged. It is the hope of
designers that the liner surface remains in this broken-in condition for the life of the engine.
However, excessive wear may cause vertical scratches to form on the liner surface reducing the
flow resistance and in turn the generation of hydrodynamic pressure [4]. If the wear is allowed to
continue, all of the asperities will be removed from the liner surface, a condition known as
polishing. Complete failure of the system is likely to follow this case.
Break-In
The break-in period of the Daimler liners used in this investigation were studied by K. Liao [10].
In this study, he identified a period of stabilization in friction that indicated the end of break-in.
33
He then repeated the procedure for all liners used in this investigation, so that all were
sufficiently broken in for further use.
Long Term Wear Analysis
In addition to the initial break-in analysis, tests were conducted to determine how much the liner
surface was wearing with time. A comparison of FMEP under motored tests for the smooth
(GG22) liner is shown below in Figure 8. As shown, the liner shows a change in friction
predominantly in the mixed/boundary region, indicating the asperities may be changing to some
degree but the hydrodynamic performance stays the same.
-- A-6/28/13 (~100 hrs)
-++-3/6/2013 (~50 hrs)
9000
8000
6000
5000
0.
4000
u- 3000
2000
1000
n
________
________
100
________
200
________
________
300
400
500
600
700
1 ~ I
________
800
________
900
1000
RPM
Figure 8: FMEP Comparison Showing Smooth (GG22) Liner Wear Over Time (0.15mm 19.5N
OCR at 80C)
For the rough (GG07) liner, both fired and motored tests were conducted under the same
conditions as tests conducted approximately 9 test hours apart in December and January. Figure
9 shows the comparison of FMEP under motored conditions for 100C in the mixed region, while
Figure 10 shows the hydrodynamic region. Figure 11 shows the FMEP comparison under fired
conditions. The period between 10 December and 7 January represents about 9 hours of mixed
motored and fired testing, while the period between 7 January and 27 May represents about 63
hours of mixed testing. The data labeled break-in comes from the original break-in data for the
GG07 back in 2012, which represents the original surface before initial break-in and several
hundred hours of testing [10].
34
--
Break-In (Ohrs)
-r-12/10/2014
-4-1/7/2014 (+9 hrs)
(~300 hrs)
-3--5/27/2014 (+63 hrs)
14000
13000
12000
11000
10000
'U
a-
9000
8000
7000
6000
5000
4000
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 9: FMEP Comparison for Rough (GG07) Liner Wear in Mixed Region under Motored
Conditions on GG07 Liner lOOC with 0.15mm 19.5N TLOCR
--- 12/10/2014 (~300 hrs) -+-1/7/2014 (+9 hrs) -dr-5/27/2014 (+63 hrs)
14000
12000
-
11000
-
13000
-
10000
9000
-
z
8000
7000
6000
5000
4000
100
200
300
400
500
600
700
800
900
1000
RPM
Figure 10: FMEP Comparison for Rough (GG07) Liner Wear in Hydrodynamic Region under
Motored Conditions on GG07 Liner 40C with 0.15mm 19.5N TLOCR
35
17000
-
0 HTHS 2.9HA 1/8 (~309 hrs) S HTHS 2.9HA 5/28 (+66 hrs) U HTHS 2.9HA 5/29 (+3 hrs)
16000
u 12000
11000
10000
9000
1000 RPM 2
bar
1500 RPM 2
bar
2000 RPM 2
bar
1000 RPM 4
bar
1500 RPM 4
bar
2000 RPM 4
bar
RPM
Figure 11: FMEP Comparison for Rough (GG07) Liner Wear under Fired Conditions GG07 with
0.15mm 19.5N OCR
The figures show that motored tests reveal there is a degree of wear that has occurred on the liner.
While one may think that a wearing liner should act like a smoother liner (and therefore display
the pattern of increasing friction in the hydrodynamic region and decreasing in the boundary) the
presence of scoring marks can allowed oil to escape at boundary contact regions, decreasing
hydrodynamic pressure. Ideally, the liner surfaces should be re-measured from time to time to
verify the characteristic roughness. It is important to consider long term wear on the liners when
comparing data spanning a large number of test hours.
3.2.2. Honing
Tests show that friction change from changing honing, and therefore the roughness of the contact
surface, is dependent on the speed of the engine. Although several liners were tested, it is only
necessary to analyze two of these, one from either end of the roughness spectrum. The GG22
represents a smooth liner while the GG07 represents a rough one. The FMEP comparison under
fired conditions, displayed in Figure 12 below, shows that FMEP at low speeds where the
friction is predominantly contact friction is lower for smoother liners. As the speed of the engine
increases, and the regime becomes more and more hydrodynamic, the rougher liner shows
increasingly lower FMEP when compared to the smooth.
36
-
20000
15000
2 10000
*GG22
LU
* GG07
5000
0
1500
RPM
1000
2000
Figure 12: FMEP Comparison of Smooth and Rough Liners under Fired Conditions at 2 bar IMEP
This effect is also shown by a comparison of the friction traces, as seen in Figure 13. At the same
engine speed, the rougher GG07 liner exhibits higher friction in the boundary friction areas of
TDC and BDC, while showing lower friction at the hydrodynamic mid-stroke.
30
20
10
C
0
0
U- -10
-
GG22
-
GG07
-20
-30
I
-40
-360
-180
0
180
360
Crank Angle [deg]
Figure 13: Comparison of Smooth and Rough Liners at 1500 RPM 2 bar IMEP
In order to advance the understanding of the effect of changing the characteristic roughness of
the liner surface, data collected under motored conditions for different temperature, lubricants,
37
and tension was plotted on an average Stribeck curve. The friction regime could be changed by
changing oil viscosity, through either temperature or actual oil formulation. Figure 14 shows
three different roughness liners compared using the lower (HTHS 1.4) viscosity oil, in which the
friction regime is dominated by mixed and boundary friction. In this regime the smoother liners
(GG22 and GG28) show a lower friction coefficient. Figure 15 increases viscosity to the HTHS
2.9 oil, so that the friction regime stays mainly hydrodynamic. This leads to the rougher liners
having a progressively lower friction coefficient.
---
-GG28
GG09 -).-GG22
0.16
0.14
0.12
Rough
0.1
U
w
0.08
ba
<
0.06
0.04
Smooth
0.02
n
0.E+00
2.E-05
4.E-05
6.E-05
8.E-05
1.E-04
1.E-04
V*Umax/(2Ft/Bore)
Figure 14: Average Stribeck Curve Comparing Liner Roughness in Mixed Friction Regime
38
GG09 -- O-GG22
---
-GG28
0.18
0.16
Smooth
0.14
0.12
U
U..
W
bU
Rough
0.1
aJ
0.08
0.06
-AO"
0.04
A n2
0.E+00
5.E-05
1.E-04
2.E-04
2.E-04
3.E-04
3.E-04
p*Umax/(2Ft/Bore)
Figure 15: Average Stribeck Curve Comparing Liner Roughness in Hydrodynamic Friction
Regime
The overall conclusion is when the regime is dominated by mixed or boundary friction, the
smoother liner surfaces perform better. This principle can be explained as follows. Equation
7 shows how the friction between two sliding surfaces (in this case the liner and piston
assembly) is dependent on the viscosity of the oil, the speed of the sliding contact, and the
separation between the surfaces. However, this relationship is only true in purely
hydrodynamic situations, when the separation of the surfaces is large enough so that no
contact is made. In this case friction comes from the shear stress in the lubricant.
Pure Hydrodynamic: %- fh-
pU
(7)
When either speed or viscosity is decreased enough so that the separation is comparable to
the size of the asperities, solid on solid contact begins to occur. At this point, another term is
introduced into the friction equation, which accounts for this contact. This new equation,
which represents the friction in the mixed/boundary region, is shown below as Equation 8.
39
Mixed: Fe-
fb
+ fh~
IU
-+
h
boundaryfriction
(8)
When in the mixed/boundary friction regime, the additional term due to increased solid on solid
contact between sliding surfaces can influence overall friction. For smoother liners, smaller
asperities mean that contact stops sooner, so that at the same speed a smoother liner has less
contact that a rougher liner and therefore lower overall friction. This is particularly true at TDC
and BDC when slower sliding speeds decrease hydrodynamic pressure.
When friction is predominantly hydrodynamic, the rougher liners show less friction than the
smoother liners. The reasons for this can be explained by graphing the hydrodynamic pressure
between the ring and liner
(Ph)
versus the clearance, h, on a log-log scale, with a constant pVU
value of 0.015 N/m.
*GG07
1.OOE+07
AGG21 XGG28
*GG09
j
_
T
z
t
2*(Yp
$ 1.OOE+06
>
ket
E 1.OOE+05
6*(Yp
0
1.OOE+04
E
.
1
5.OOE-08
5.OOE-07
Clearance (h)
5.OOE-06
Figure 16: Modeled Relationship between Hydrodynamic Pressure and Film Thickness for
TLOCRs [33]
The key to the different liners graphed in Figure 16 is their position along the x and y-axis.
Notice how each graphed line contains five data points each. For each line, the upper-left most
point represents a clearance of 2 *up. ap is the standard deviation of the characteristic heights of
the peaks and valleys that make up the plateau liner roughness. Therefore, based on the
principles of a Gaussian distribution, 99.7% of the asperity heights lie within three standard
deviations or 3*ap. This means that any position on the graphed lines above and left of the
40
second data point (3*ap) will have some degree of asperity contact, as the clearance is within the
height of the asperities. Any point below or to the right of the second data point with have
virtually no asperity contact, and will be purely hydrodynamic.
For a specific unit pressure, say 1 bar (shown on the graph by the blue dashed line), the different
liner finishes have different levels of clearance. For example, the GG07 has a clearance just
below its 3cyp point, meaning it will still be close to the mixed region, and closer to the minimum
friction coefficient point on the Stribeck Curve. Furthermore, notice that this point lies further
along the x-axis, which means the GGO7 will have a higher film thickness to support the same
load. Since film thickness is inversely proportional to hydrodynamic friction coefficient, the
rougher GG07 liner will exhibit lower friction in the hydrodynamic regime.
It is important to note that all data collected in the previous section was with the TLOCR. The
trends discussed should be limited to TLOCR applications. Similar trends observed for the
TPOCR have not been fully explored at this time, and could possibly occur for other reasons.
3.3. TLOCR Parameter Comparison
The Twin Land Oil Control Ring, or TLOCR, is a design that utilizes two pieces to perform the
functions needed: a twin land ring and tension spring.
Tension
spring
Ring lands
Figure 17: Twin Land Oil Control Ring Cross-Section
The two ring lands, highlighted in Figure 17, provide the unit pressure radially outward against
the cylinder liner to control the amount of oil on down-stroke. Also, the ring features slots in the
horizontal direction (which can be observed in Figure 2) which allow oil to flow from the contact
area, and thereby reduce pressure buildup. The two parameters of TLOCR design that were
studied were the width of the two lands and the tension of the spring exerting a tangential force.
41
3.3.1. Ring Parameters
The outward radial pressure exerted by the OCR controls the thickness of the oil film on the liner
during the down-stroke of the piston. By altering the width of the ring's normal surface to the
liner (increasing surface area) and/or the normal force applied by the ring (by changing the
spring tension), the unit pressure can be altered and in turn change the amount of oil supplied.
This means the friction of the top two rings is highly reliant on the OCR [34]. Equation 9 shows
how changing the land width changes the unit pressure of the OCR on the liner, with F, being the
ring tension, L, being the land width, and B being the cylinder bore diameter.
Unit Pressure =
(9)
LwB
However, changing tension and changing land width do not affect friction at the same rate. The
reason lies in the derivation for friction in the hydrodynamic regime. First, Equation 10 shows
the hydrodynamic pressure, Ph, between the ring and the liner surface where p is the dynamic
viscosity of the lubricant, U is the sliding speed, h is the film thickness separating the two
surfaces and C, is a constant incorporating roughness characteristics and initial conditions [4].
Ph= Cp Uh-a typically: a > 2
(10)
In order for forces to balance, the unit pressure from the ring must equal the hydrodynamic
pressure. When Equation 9 and Equation 10 are set equal and rearranged, the result an
expression for the oil film thickness, h, in terms of the ring parameters [4].
1
h= [CpBptULwf
Ft
L
(11)
Now, recall from earlier in Equation 3 that the hydrodynamic coefficient of friction, fl, is equal
to a constant multiplied by the viscosity times the sliding speed over the film thickness.
Furthermore, the friction from the hydrodynamic shear stress of an oil control ring is shown
below in Equation 12 [4].
Friction= 2lrBLwfh
42
(12)
By substituting Equations 3 and 11 into Equation 12, the final result is an expression for TLOCR
ring friction in terms of the TLOCR land width and tension [4].
1
(ULw)13)
Friction= 2rcf ()B
Notice how in this equation (13), the TLOCR spring tension and land width scale differently. For
simplicity, these relationships are shown below in Equations 14 and 15.
1
Friction~ Ft
(14)
Friction~ Lwl-a
(15)
The final conclusion is that although both TLOCR land width and tension affect unit pressure,
they affect friction at different rates for the same viscosity and speed. The experimental results
presented next give further insight on this reasoning.
3.3.2. TLOCR Experimental Results
The effects of changing TLOCR land width and tension were explored using both fired tests and
motored tests. For motored tests, the top two rings were removed and a larger clearance piston
was used to isolate the effects of the TLOCR.
Land Width Effect under FiredConditions
The three land widths tested under fired conditions were 0.06mm, 0.15mm and 0.23mm. Various
results from changing OCR land width on a smooth (GG22) liner are shown below, with a
friction trace comparison shown in Figure 18 and an FMEP comparison shown in Figure 19. The
same pattern is observed in Figure 20 for a comparison of the 0.06mm and 0.15mm land widths
on the rough (GGO7) liner. Notice the minimal change in friction due to changing land width in
both cases. Other land width combinations and variations in engine speed produced similar
results.
43
30
20
10
0
z
0
-10
LL
U-
-0.23
mm
-0.15
mm
-20
-30
-40
-360
-180
180
0
360
Crank Angle [deg]
Figure 18: Friction Traces Comparing OCR Land Widths; GG22, 1500 RPM, 10.5N TLOCR, 2 bar
IMEP
E0.23mm
M0.15mm
16000
14000
12000
CU
10000
a. 8000
E6000
4000
2000
0
1000
1500
2000
RPM
Figure 19: Comparison of FMEP between OCR Land Widths; GG22, 10.5N TLOCR, 2 bar IMEP
44
n 0.06mm
0.15mm
20000
15000
10000
LU
5000
0
1000
1500
2000
RPM
Figure 20: Comparison of FMEP between OCR Land Widths; GG07, 19.5N spring, 2 bar IMEP
As seen, there is little change in friction from changing the land width of the TLOCR regardless
of load, roughness, ring tension, etc. Reasons for this are presented later in the conclusion of this
section.
Land Width Effect under Motored Conditions
In an effort to further explore the friction contributions of the TLOCR, data was collected under
motored conditions. Three different land widths were compared, all with the same 10.5N tension
spring. The three land widths were: 0.06mm, 0.15mm, and 0.23mm. All tests were conducted on
the smooth (GG22) liner. Figure 21 shows the FMEP comparison from the hydrodynamic region.
The data trends, reinforced by the friction trace in Figure 22, show that friction can increase with
increasing land width in the hydrodynamic region. However, this relationship is not a linear
relationship, and may have a limit as seen by the minimal change in friction from 0.15mm to
0.23mm. Recall from Equation 15 that friction does not in fact vary linearly with land width.
When the region becomes mixed, as shown by the FMEP graph in Figure 23, the developed
equations no longer hold true. At this point the mixed hydrodynamic and boundary effects taking
place, as shown by the friction trace in Figure 24, make the friction much less predictable.
45
-4--0.06 mm
-U-0.15 mm
-*-0.23 mm
12000
10000
8000
(U
c-1.
6000
0.
LU
4000
2000
0
100
200
400
300
500
RPM
600
700
800
900
1000
Figure 21: Comparison of FMEP between TLOCR Land Widths in Hydrodynamic Region (40C)
under Motored Conditions with Smooth (GG22) Liner and 10.5N Spring
-0.06
mm
mm
-0.15
-0.23
mm
20
15
10
5
0
0
U.
-5
-10
-15
-20
-360
-180
0
Crank Angle [deg]
180
360
Figure 22: Comparison of Friction Traces between TLOCR Land Widths in Hydrodynamic Region
(40C) under Motored Conditions (800 RPM) with Smooth (GG22) Liner and 10.5N Spring
46
mm
-$--0.06
-*-0.15 mm
-*-.0.23 mm
6000
5000
I
4000
cc
:.
3000
LU
L. 2000
Unpredictability in mixed region
1000
0
100
200
300
400
500
600
700
800
900
1000
RPM
Figure 23: Comparison of FMEP between TLOCR Land Widths in Mixed Region (100C) under
Motored Conditions with Smooth (GG22) Liner and 10.5N Spring
-0.06
mm
-0.15
mm
-0.23
mm
10
8
------
6
-------
4
2
0
0
-2
-4
-
-6
-
LL
-8
-10
-360
-180
0
Crank Angle [deg)
180
360
Figure 24: Comparison of Friction Traces between TLOCR Land Widths in Mixed Region (100C)
under Motored Conditions (200 RPM) with Smooth (GG22) Liner and 10.5N Spring
47
-0.06
mm
-
0.15 mm
-
0.23 mm
10
8
6
- - - - --
_ _--_
_
4
0
U-
0
-2
-4
-6
-8
-10
-360
-180
0
Crank Angle [deg]
180
360
Figure 25: Comparison of Friction Traces between TLOCR Land Widths in Mixed Region (100C)
under Motored Conditions (500 RPM) with Smooth (GG22) Liner and 10.5N Spring
The data showed that the larger the land width, the lower the speed at which the minimum
friction coefficient, or transition point, is reached when all other things are held equal. It also
showed that the smallest land width has the lowest friction in the hydrodynamic regime.
However, there is some inconsistency in the results, such as variability in friction levels in the
mixed regime, and no friction increase from 0.15 mm to 0.23 mm land widths in the
hydrodynamic regime. This uncertainty could stem from several sources. First, there is the
possibility of measurement uncertainty, which may be indicated by the step up in friction near
TDC for the 0.23 mm ring in Figure 25. Second, the large piston clearance used in motored tests
and low TLOCR spring tension may not make both lands of the ring conform to the liner as well
as it should, meaning one land may be in contact while the other is not. Finally, all three of the
rings are not made the same. The 0.23 mm ring is cast iron and was broken in during a durability
test. The 0.15 mm and 0.06 mm rings are steel and have different land designs as well as land
widths which are not completely uniform. The inspections showing this variability are shown
below in Figure 26 through Figure 28. The dimensions shown at specific intervals around the
ring in black numbers are the measured widths of the upper and lower lands, respectively.
48
Profile measurement (OD face contact after test)
4.
a
60pm
64pm
47pm
63pm
200
3400
62pm
55pm 315*
M
IHLE
64pm
60pm
450
Upper Land
Lower Land
60pm 2 7 0*-
.-
90 pm
.- .
LD 1 -------
45pm
63p
Tension Lt (N): 28.5N
Gap: 0.40
62pm 2250
1356opm
69pm
42pm
1800
67pm
56pm
Figure 26: Land Measurements after Break-In for 0.06 mm TLOCR
MNHLE
Profile measurement (OD face contact after test)
0.16 mm
0.10 mm
20*
0.18 mm
0.11 mm
340*
0.18 mm
0.15mm
0.17 mm
3135*
450
Upper Land
Lower Land
017mm270
0.10 mm 2 7
- - - - --
B2
- - -
.002m
----
0.15 mm
Tension Lt (N): 10.5N
Gap: 0.42
22511
0.18 mm
0.13 mm
135"
0.19 mm
0.15 mm
1800
0.21 mm
0.17 mm
Figure 27: Land Measurements after Break-In for 0.15 mm TLOCR
49
TOP
0.24
0.25
0.23
0.23,
0-2
0-23
Figure 28: Land Measurements after Break-In for 0.23 mm TLOCR
Spring Tension Effect under Fired Conditions
Another aspect of the OCR that was tested was changing the spring tension. Again, the TLOCR
tension spring came in three varieties, ION, 19.5N, and 28.5N. The change in unit pressure due
to altering spring tension is exemplified in Equation 14. The procedure was conducted with
standard HTHS 2.9 oil, a 0.15mm land width ring, and both the rough (GG07) and smooth
(GG22) liners. Figure 29 shows how increasing spring tension increases friction in the
hydrodynamic regime for smooth liners, leading to an overall increase in FMEP (Figure 30). The
friction trace in Figure 29 also shows this increase can also occur when boundary contact occurs
at TDC and BDC. Figure 31 and Figure 32 shows that the same tension change increases friction
at nearly the same rate for rough liners as it did for smooth liners.
50
30
20
10
0
LL
0
7--- - - -
-
----- - - - -
-10.5
-10
28.5 N
N
-20
-30
-40
-360
-180
0
Crank Angle [deg]
180
360
Figure 29: Effect of Increasing Spring Tension on Smooth Liners (GG22) at 1000 RPM
20000
18000
16000
14000
12000
10000
N
CL
8000
28.5 N
0 10.5 N
U-
6000
4000
2000
0
1000
1500
2000
RPM
Figure 30: FMEP Comparison between OCR Spring Tensions for Smooth Liners (GG22)
51
40
30
20
10
0
0
-10
4-J
LA-
-20
-30
-28.5
N
-10.5
N
-__
-40
-50
-360
-180
0
180
360
Crank Angle [deg]
Figure 31: Effect of Increasing Spring Tension on Rough Liners (GG07) at 1500 RPM
18000
16000
14000
12000
10000
LU
LA.
N
8000
28.5 N
* 10.5 N
6000
4000
2000
0
1000
1500
2000
RPM
Figure 32: FMEP Comparison between OCR Spring Tensions for Rough (GG07) Liners
Spring Tension Effect under Motored Conditions
A look at the change in spring tension under motored conditions without the top two rings shows
the same trends as the fired tests. Figure 33 through Figure 39 show the FMEP graphs and
52
friction traces, all of which show that friction increases steadily with spring tension across all
regimes regardless of roughness, just as in the fired tests.
-U--19.5N
-0-10.5N
14000
12000
10000
'U
8000
LU
6000
aa-
U-
4000
2000
0
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 33: Comparison of FMEP between TLOCR Tension in Mixed Region (40C) under Motored
Conditions with Rough (GG07) Liner and 0.15mm Land Width
-'--10.5N
-4--19.5N
14000
12000
10000
I
8000
LU
6000
LA-
4000
2000
0
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 34: Comparison of FMEP between TLOCR Tension in Boundary Region (100C) under
Motored Conditions with Rough (GG07) Liner and 0.15mm Land Width
53
--
10.5N
-
19.5N
25
20
15
10
-
......
..
..
.
z 5
0
.9 -5
U-
-10
-15
-20
-25
-180
-360
0
Crank Angle [deg]
360
180
Figure 35: Comparison of Friction Traces between TLOCR Tension in Mixed Region (100C) under
Motored Conditions (100 RPM) with Rough (GG07) Liner and 0.15mm Land Width
- r29.5N
-- I-19.5N
-+--10.5N
18000
16000
14000
12000
a.10000
u
8000
A
6000
4000
2000
0
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 36: Comparison of FMEP between TLOCR Tension in Hydrodynamic Region (40C) under
Motored Conditions with Smooth (GG22) Liner and 0.15mm Land Width
54
---
-r-29.5N
-4--19.5N
10.5N
12000
10000
8000
a.
a-
6000
4000
2000
0
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 37: Comparison of FMEP between TLOCR Tension in Mixed Region (100C) under Motored
Conditions with Smooth (GG22) Liner and 0.15mm Land Width
-10.5N
-19.5N
-29.5N
30
20
----
00%rj-
--
10
0
LA-10
-20
-30
-360
-180
0
Crank Angle [deg]
180
360
Figure 38: TLOCR Hydrodynamic Friction Trace Comparison (40C) under Motored Conditions
(1000 RPM) with Smooth (GG22) Liner and 0.15nun Land Width
55
10.5N
-
-
19.5N
-
29.5N
30
20
10
0
0
m--10
-20
-,:M
-360
-180
0
Crank Angle [deg]
180
360
Figure 39: TLOCR Mixed Friction Trace Comparison (100C) under Motored Conditions (100
RPM) with Smooth (GG22) Liner and 0.15mm Land Width
The results of motored testing indicate that increasing spring tension increases friction,
regardless of the contributions of the other two rings.
3.3.3. TLOCR Ring Parameter Conclusions
In general, the experimental results provide great insight into the effects of changing TLOCR
land width and spring tension, which in turn provide reinforcement for the model equations
developed. In regard to land width, any change showed a minimal effect on friction under fired
conditions. Using the equations put forth in section 3.3.1 give rise to a theory called the
"canceling" effect. This theory suggests that reducing the land width, while maintaining the
same ring tension, will lead to a lower oil film thickness and also lower friction for the oil ring in
the hydrodynamic regime. However, a smaller oil film thickness will increase the friction of the
top two rings.
Therefore the total friction may not change much with changing TLOCR land
width under fired conditions. The motored tests reinforced this when they showed reducing land
width can in fact reduce friction when the TLOCR is isolated.
On the other hand, increasing TLOCR spring tension shows clear increases in friction. This
relationship holds true for both fired and motored conditions, smooth and rough liners, as well as
mixed and hydrodynamic friction regimes. Equation 14 supports these finding by showing the
relationship between increasing TLOCR spring tension and increasing friction. Furthermore,
56
Equation 11 shows that when the TLOCR tension is increased, oil film thickness of the OCR is
decreased. This decrease in film thickness reduces the oil supply to the top two rings, which in
turn will increase their friction. Therefore, the friction of all three rings is increased when OCR
tension is increased, leading to a more noticeable effect.
3.4.TPOCR Comparison
The three piece oil control ring (TPOCR) is an alternate ring design shown below in Figure 40.
The design features two rings separated by an expander. The three parts are separate entities,
with the radial tension provided mainly by the two flat rings. Due to the manufacturing process,
the contact surfaces are initially rounded, but may wear to a flatter profile with break-in. This
design has been modeled by Tian's OCR Model [35], which was used in conjunction with
experimental results to further explore the friction trends for the TPOCR.
Figure 40: Cross Sectional View of Three Piece OCR
3.4.1. Ring Type Comparison and Liner Finish
The initial investigation of the TPOCR focused on comparing trends of the two types of OCR:
TPOCR and TLOCR. For this study, two different liner roughnesses (Table 2) were used, the
GG22 (smooth) and GG08 (rough). Table 3 shows the specifications of both of the two types of
rings tested, which are all of similar tension and contact area. Although manufacturing dictates
that the TPOCR does not have a flat land area like the TLOCR, wear after break-in leaves a worn
surface area that was measured. These specific rings were broken-in in a separate endurance test
engine for approximately 100 hours. These measurements, which are comparable to land widths
of the TLOCR, were averaged and are denoted in Table 3 as the average contact surface width. It
57
should also be noted that the tensions used for the TPOCR are approximated 15N, while the
TLOCRs used were 19.5N and 10.5N. The range of tensions for the TLOCR should provide
adequate comparisons of unit pressure to the other two types of OCR. Furthermore, the TLOCRs
have been used in the FLE for several years, and are adequately broken in.
Table 3: OCR Specifications
Identification Number
Tension (N14.2
Avg. Contact Surface 'Width (MM
0.115
2
14.7
0.107
Identification Number
10.5
19.5
10.5
19.5
0.15
0.15
Tension,
LandWidt (mm
1
Ring Type Comparison Experimental Results
Figure 41 shows the comparison between the two types of rings on an average Stribeck curve
normalized for unit pressure, using the smoother GG22 liner and HTHS 2.9 lubricant. This graph
shows mainly the hydrodynamic region, and it is clear that the TLOCR shows a higher friction
coefficient in the hydrodynamic region, while the TPOCR shows a lower friction coefficient.
58
--
TL 0.15mm 19.5N
0.0001
0.00015
-41-3p #2
0.14
-
0.16
0.12
0.1
U
U
w
0.08
0.06
0.04
0.02
0
0
0.00005
0.0002
0.00025
0.0003
0.00035
p*Umax/(2Ft/Bore)
Figure 41: Friction Coefficient Comparison for Different OCR's on GG22
A look at the FMEP comparison shows this trend is repeated on the hydrodynamic region of the
rougher liner, which is displayed in Figure 42 above 400 RPM. Figure 44 and Figure 45 show
some select friction traces exhibiting hydrodynamic behavior for both the smooth (GG22) and
rough (GG08) liners, respectively. While the shapes are slightly different between the two liners,
the ordering of the friction levels of the OCRs is the same. The data reinforces the development
that the friction trends between various OCR types with similar unit pressure seem to be
independent of liner finish when the regime is strictly hydrodynamic.
59
---
-- 3p #1
-
TL0.15mm 19.5N
TL 0.15mm 10.5N
13000
12000
11000
- 10000
I
9000
__________________
7000 -
_________
_________
________
_________
_________
-
8000
6000
200
100
300
400
500
RPM
600
700
800
900
1000
Figure 42: FMEP Comparison for Hydrodynamic Regime GG08 40C
-3p
#2
-
TL 0.15mm 10.5N
-
TL 0.15mm 19.5N
14000
12000
10000
8000
a.
aL
j 6000
Ul-
4000
2000
0
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 43: FMEEP Comparison for Hydrodynamic Regime GG22 40C
60
-
3p #2
-TL
0.15mm 10.5N
-
TL 0.15mm 19.5N
25
20
15
10
5
0
0
L-
-5
-10
-15
-20
-25
180
0
-180
360
360
Crank Angle [deg)
Figure 44: Hydrodynamic Friction Trace Comparing OCRs on Smooth Liner at 40C and 1000
RPM
-3p
#2
-TL
0.15mm 10.5N
-
TL 0.15mm 19.5N
20
15
10
5
0
-
z
-5
-10
-15
-20
360
-180
0
Crank Angle [deg]
180
360
Figure 45: Hydrodynamic Friction Trace Comparing OCRs on Rough Liner at 40C and 1000 RPM
The independency of OCR type and liner roughness does not hold true, however, when in the
mixed or boundary region of the friction regime. Figure 46 shows how the average friction
coefficient varies for the OCR types on the rougher (GG08) liner. Notice how the TLOCR now
shows the lowest average friction coefficient in the mixed region, but is sharply increasing in the
boundary region. In contrast, the TPOCR maintains a "flatter" shape.
61
-*--TL
0.15mm 19.5N
-f--3p #1
0.15
0.14
0.13
0.12
LL
0.11
-o
4)
ba
0.1
0.09
-L-600
0.08
0.07
0.06
0
0.00005
0.0001
0.00015
0.0002
0.00025
0.0003
0.00035
p*Umax/(2Ft/Bore)
Figure 46: Friction Coefficient Comparison for Different OCR's on GG08
A look at the FMEP comparison for the rougher GG08 liner in Figure 47 shows this same trend,
with the TLOCRs showing a lower dip in the mixed region followed by a sharper increased as
the region shifts to the boundary regime (in the leftward direction) when compared to the
TPOCR. In contrast, the FMEP comparison in the mixed regime shown on the smoother GG22
liner (Figure 48) shows that moving leftward toward the boundary region yields increasing
friction at nearly the same slope for all OCR types. Investigating further, the friction traces show
more evidence of a correlation between OCR type and liner roughness in the mixed/boundary
friction regime. For the smooth liner (Figure 49) at the same temperature and speed, the TPOCR
exhibits a friction trace characteristic of the mixed region that resembles a step function, with a
flat top. However, both of the TLOCRs show a different shape, with spikes at TDC and BDC.
This contrast does not hold true when analyzing the traces on the rougher liner (Figure 50). In
this case, all OCR types show the same friction trace shape, one characteristic of the boundary
region.
62
-0.15mm
-3p#1
--- TL
19.5N
0.15mm 10.5N
12000
11000
10000
'U
9000
100
8000
200
3 40 0 0 7 0 9
U.'
LL
7000
6000
5000
4000
200
100
300
400
500
RPM
600
700
1000
900
800
Figure 47: FMEP Comparison for Mixed Regime GG08 80C
-3p
#2
-
TL 0.15mm 10.5N
-
TL 0.15mm 19.5N
7000
6000
-
-
-
----------------
5000
'U
a.
LL
4000
3000
2000
1000
0
100
200
300
400
500
RPM
600
700
800
900
Figure 48: FMEP Comparison for Mixed Regime GG22 80C
63
1000
3p #2
-
TL 0.15mm 10.5N
-
-
TL 0.15mm 19.5N
15
10
-r
5
0
-- - --------
0
4.UL
'U-
-5
____________________________
_________________________
- ____________________________
____________________________
-10
-15
180
0
-180
-360
360
Crank Angle [deg]
Figure 49: Mixed/Boundary Friction Trace Comparing OCRs on Smooth Liner at 100C and 200
RPM
-
-
3p #1
-
0.15mm 19.5N
TL 0.15mm 10.5N
25
20
15
10
5
z
0
L-
0
-5
-10
-__
-15
_
-20
-25
-360
0
-180
180
360
Crank Angle [deg]
Figure 50: Mixed/Boundary Friction Trace Comparing OCRs on Rough Liner at 100C and 200
RPM
TPOCR LinerRoughness Effects
The final step in the experimental investigation of the TPOCR that was completed was a
comparison of friction on different liner roughnesses. While the effects of liner roughness are
well understood for the TLOCR, they are still largely un-documented for the TPOCR. This
64
investigation is an initial look into trends that will serve as a basis for future modeling and
experimental pursuits.
Figure 51 shows the FMEP comparison between a rough (GG08) and smooth (GG22) liner in the
hydrodynamic regime for the TPOCR. It is interesting to note that the previously defined
relationships for the TLOCR differ for the TPOCR. That is, the swap to lower friction for
rougher liners in the hydrodynamic region seems to occur at a much later point. In fact, the
testing procedure in the figures does not even identity when this point occurs. Moving into the
mixed region (Figure 52) shows the separation between friction increases, with the rougher liner
remaining at a higher friction level when compared to the smoother liner. A look at the average
Stribeck curve, shown in Figure 53, indicates that the rougher liner does in fact show a "flatter"
friction curve by decreasing at a much lower rate when compared to the smoother liner in the
mixed region. In contrast, the smooth liner shows a large dip in the friction coefficient. The
friction coefficient in the hydrodynamic region appears to be converging, just as with the
TLOCR, but at a much higher speed.
-1-GG22
-
-+-GG08
12000
10000
8000
:.
6000
IL4000
0.
2000
0
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 51: FMEP Comparison between Liner Roughnesses for Hydrodynamic Region (40C) for
TPOCR
65
-U-GG22
-4--GG08
9000
8000
7000
6000
5000
LU
4000
LL.
3000
...
......
.
2000
1000
0
100
200
400
300
500
RPM
600
700
800
900
1000
Figure 52: FMEP Comparison between Liner Roughnesses for Mixed Region (100C) for TPOCR
---
GG08 -1-GG22
0.16
0.14
0.12
0.1
0.08
'
0.06
0.04
- - ---0.02
0
0
0.00005
0.0001
0.00015
0.0002
0.00025
0.0003
0.00035
p*Umax/(2Ft/Bore)
Figure 53: Average Stribeck Curve Comparison Liner Roughness for TPOCR
Experimental Results Conclusions
It is important to remember the data presented does not allow for absolute conclusions such as
"the TPOCR has higher friction." The important take away is the trends that are identified, which
66
can be used to form a foundation for understanding OCR behavior. There are several such trends
identified in the previous section. First, OCR design can influence friction levels. Second, there
seems to be an insignificant correlation between OCR design and liner roughness in the
hydrodynamic friction regime, as similar trends are observed on both rough and smooth liners.
However, when contact between surfaces occurs in the mixed and boundary region, it seems that
different OCR types do behave differently based on liner roughness. Finally, the results indicate
that the principles identified for the roughness effects with the TLOCR differ for the TPOCR.
The advantage of rough liners with the TLOCR in the hydrodynamic region appears to happen at
a much later point for the TPOCR. The reasons for this are believed that the TPOCR relies on its
profile to generate hydrodynamic pressure while the TLOCR relies on the liner roughness
geometry. Since the TPOCR rails are flexible and axial clearance is in the order of 100 microns,
the rails of the three piece experience relatively large dynamic twist - in the order of a couple of
degrees. As a result, the contact surface (shown in Figure 61) is not large and flat like the
TLOCR, but instead rounded. The shape of this profile can become quite sharp, with a barrel
drop of a few microns. The ratio of this barrel drop compared to the minimum oil film thickness
(MOFT) can become large, in which case the scale of the contact surface profile is comparable to
the scale of the liner roughness. This means the contact area is very small, like a roller pin on a
flat surface. This increases the effective unit pressure and keeps the ring in contact through much
higher speeds when compared to the TLOCR, which delays pure hydrodynamic behavior. In
contrast, the wide flat lands of the TLOCR allow it to generate hydrodynamic pressure quickly,
and therefore quickly leave the mixed region. Once the hydrodynamic behavior is achieved for
the TPOCR, the profile becomes useful as the wide lands of the TLOCR are unable to create
adequate film thickness in this region.
Model Results
When it comes to modeling TPOCR friction, Tian et al.'s OCR model can be used to help
explain some of the results found in experimentation for the TPOCR [35]. This model has
recently since been improved to consider the dynamics of the expander.
Additionally, the
original Reynolds equation is used in the current model instead of the average Reynolds equation
in the original model. D. Kim had previously obtained data on the FLE of friction levels of the
piston with no rings under motored conditions. This data was subtracted from the measured data
for the TPOCR, in order to compare it to the model results, which do not account for piston
67
friction. The comparison between the model and experimental data on the smooth (GG22) liner
can be found in Figure 54 through Figure 56, while the comparison on the rough (GG08) liner is
shown on Figure 57 and Figure 58.
A model data
-
3p #2 -p
6000
------
5000
4000
CO
3000
2000
1000
100
200
300
400
600
500
RPM
700
800
900
1000
Figure 54: Modeled vs. Experimental FMEP Data for TPOCR on GG22 at 40C
A Model Data
-
3p #2 -p
5000
4500
4000
,.-'''
2
L"-
2000
-
3500
1500
-
1000
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 55: Modeled vs. Experimental FMEP Data for TPOCR on GG22 at 60C
68
-
3p #2 -p
-
Model Data
7000
6000
5000
-4000
U5 3000
2000
1000
0
100
200
300
400
500
RPM
600
800
700
900
1000
Figure 56: Modeled vs. Experimental FMEP Data for TPOCR on GG22 at 100C
A model data
-3p
#2 -p
8000
7500
7000
A
I-
,6500
0.
u- 6000
5500
5000
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 57: Modeled vs. Experimental FMEP Data for TPOCR on GG08 at 40C
69
A
Model Data
-
3p #2 -p
9000
8500
8000
LU
7500
7000
6500
~ann
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 58: Modeled vs. Experimental FMEP Data for TPOCR on GG08 at 100C
As shown by the data, the model results match the trends of the experimental results quite well.
However, the magnitudes are different, with the experimental FMEP being higher than the model
in every instance. At this time, there is no solid explanation for this anomaly. In order to
investigate what caused this, further experiments were conducted with the TPOCR. In these tests,
the bottom rail of the TPOCR was flipped, with the reasoning being the profile of the worn
surface does not wear completely flat, but instead is worn to an asymmetrical rounded profile as
shown in Figure 61. The estimation of this profile can easily be a source of error. The FMEP data,
displayed in Figure 59 and Figure 60, shows how flipping the bottom rail increases friction
nearly uniformly across the regime, just as seem in the modeled versus experimental
comparisons.
70
-+--3p BF -0-3p
9000
8000
7000
6000
=- 5000
CL
4000
3000
2000
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 59: FMEP Comparison with Standard TPOCR Configuration and TPOCR with Bottom
Rail Flipped at 40C
-4-3p BF -0-3p
6000
5500
5000
4500
cc
a.
4000
2 3500
3000
2500
2000
100
200
300
400
500
RPM
600
700
800
900
1000
Figure 60: FMEP Comparison with Standard TPOCR Configuration and TPOCR with Bottom
Rail Flipped at 60C
Currently, the model assumes a parabolic profile, as seen in Figure 61. The results seem to
indicate this assumption may not be accurate, and suggest the contact surface may wear more at
an angle, in which case flipping the bottom ring left less oil for the top ring, and effectively
increased overall FMEP. Also, the extent of the barrel drop is unknown. Future re-measurements
71
of the TPOCR will be acquired to attempt a more accurate representation of the ring contact
surface.
Assumed Profile
: 0.06mm
0.05mm
0.05mm
s.
LA
..............................
Upper rail
0.06mm
EE
r-
ao
Lower rail
MeI sured P rofile
0,2
0 08 mm
m
Figure 61: Assumed and Theorized Contact Profiles for TPOCR
Another reason for the discrepancy in magnitudes may be that the model assumes both rails of
the OCR are fully flooded at the rail/liner interface. In reality, one rail may be starved of
lubrication by the rail moving ahead of it.
72
4. Effects of Piston Skirt Designs
The main contributor to friction from the piston itself is the piston skirt. While some designs
exist that do allow contact of the piston lands between the rings, the majority of contact comes
from the skirt. Because of this, manufacturers have developed methods to coat the skirt with a
low-friction coating. The specific coating on the pistons used in the study is known as the
GRAFAL coating. This is a sliding lacquer coating with fine graphite particles embedded in a
polymer matrix, which is designed to not wear under normal operating conditions. Under
extreme loads, it can wear locally, resulting in greater clearance and therefore resistance to
seizure [36]. Since the manufacturing of this coating is very controllable, there has been an
interest in recent years to explore the possibility of reducing friction by exploring different
coating patterns and roughnesses, as well as skirt profiles.
In order to explore the friction trends associated with the piston skirt profile as well as coating
pattern and roughness, a variety of pistons were provided. These pistons were essentially the
same diameter, having an outer diameter of 82.455mm plus an approximate 30 Pm of coating.
All comparisons were done under fired test conditions.
4.1. Skirt Patterns
One of the capabilities of the spray coat manufacturing process is the development of patterns.
For different coating patterns, there were four configurations, shown below in Figure 62. Of the
four patterns tested, the "rows" piston had the best results, as shown in Figure 63. The friction
results of the base piston were close behind the rows, while both the dots and voids showed
significantly high FMEP. Although the rows piston displayed the best results, it is important to
consider production cost for this complex design when considering actual application. The main
conclusion is that none of the alternate piston patterns tested shows a significant improvement in
friction over the base pattern. In addition to the friction data presented here, E. Zanghi studied oil
transport for the voids piston [37] and P. Totaro discussed the comparison between the
experimental and model results for the piston skirt patterns [38].
73
Figure 62: Piston Skirt Patterns Tested
18000
12000
" Base
U
a-
" voids
" Dots
6000
0
" Rows
L_-
1000
1500
RPM
2000
Figure 63: Comparison of FMEP from Various Piston Skirts at 2 bar Cylinder Pressure
74
-
18000
12000
E Base
-
E voids
U Dots
6000
U Rows
0
1000
1500
2000
RPM
Figure 64: Comparison of FMEP from Various Piston Skirts at 4 bar Cylinder Pressure
4.2. Skirt Roughness
In addition to the patterns, the effect of changing the actual roughness of the coating was
analyzed. A number of pistons with the solid base pattern were obtained in three different
roughnesses. The three different types of pistons were essentially identical in all aspects except
skirt roughness. The different designators are RA02, RS5, and RS 12 in ascending order of
roughness, and the measurements of this roughness are shown in Figure 65, Figure 68, and
Figure 71. The numbers in these designators refer to the nominal peak to valley height in
microns. An initial baseline was established by conducting a fired test with a piston that has been
used at MIT for several years. This piston initially had a RS12 skirt roughness, and is of the same
dimensions as the new pistons. The only difference is this piston has several hundred hours of
testing under fired conditions in the FLE. This baseline was conducted with the GG22 liner with
the 0.06mm, 10.5N TLOCR. After the baseline was established, a break-in analysis was
conducted for each of the new pistons, in order to determine any patterns or trends in friction
during break-in.
4.2.1. RA02
The first break-in analysis was conducted on the RA02 roughness piston. This is the smoothest
of the three different roughness pistons, and an example of the measured roughness after coating
75
is shown in Figure 65. In order to determine the break-in pattern, the first RA02 piston (piston
number 1002) was subjected to three test runs, or 15 RPM sweeps. After this, the FMEP was
analyzed to determine the pattern of stabilization. Figure 66 shows the FMEP during 2 bar IMEP
stabilizes around the seventh or eighth RPM sweep, or approximately four to five hours of
engine testing. Figure 67 shows the same pattern holds true for 4 bar IMEP. While there is still
variability after this point, the initial decline in FMEP associated with break-in has leveled off.
PF PrfW ausgerichtet LcJLs= AUS
5.0
0,0
Tastr TKU300 LU 40 mm Lc = 0A00 mm A =0.50 mms
40
Figure 65: Measure Roughness Profile for RA02 1002 after Coating
-+-1000 RPM
-- h2000 RPM
-13-1500 RPM
14000
13000
12000
11000
LU
10000
9000
8000
1
2
3
4
5
8
6
7
RPM Sweep Number
9
10
11
12
Figure 66: RA02 Piston Break-In FMEP for 2 bar IMEP
76
13
14
15
---
10004bar
---
15004bar
-*- 20004bar
15000
14000
13000
'U
12000
L.
11000
--
o " rl .
o
10000
9000
8000
1
2
3
5
4
6
10
8
9
7
RPM Sweep Number
11
12
13
14
15
Figure 67: RA02 Piston Break-In FMEP for 4 bar IMEP
4.2.2. RS5
The same break-in procedure was conducted for the RS5 piston. The roughness measurement for
the RS5 1007 piston is shown below in Figure 68. Figure 69 shows the 2 bar IMEP data, while
Figure 70 shows the 4 bar IMEP data. This time 6 test runs were conducted or, 30 RPM sweeps.
The 2 bar IMEP data shows a stabilizing trend around the eighth or ninth RPM sweep, only
slightly higher than the RA02 piston. At 4 bar IMEP, FMEP stabilized around the same point for
the 2000 RPM FMEP line, but interestingly enough continues to decrease for 1500 and 1000
RPM. Based on previous experience with the break-in of rough surfaces, the trend is not
unexpected, and it is not uncommon for FMEP to continue to slowly decrease over the effective
life of the surface.
P.
PMdN aasgedcd
LCILs= AUS
5.0
TosterTKU0
LI=4.80
mm Lc=0800 mm
-
A
0,0
Vt:= 0.50mi4
Figure 68: Measure Roughness Profile for RS5 1007 after Coating
77
-4--1000 RPM
-4-1500 RPM
-*-2000 RPM
14000
13000
12000
11000
LU
U-
10000
9000
8000
2
4
8
6
10
12
20
18
16
14
RPM Sweep Number
22
24
26
28
3C
Figure 69: RS5 Piston Break-In FMEP for 2 bar IMEP
1000 RPM
--
-4.-1500 RPM
-*-2000 RPM
14000
13000
12000
2LU 11000
U-
1uuuu
9000
8000
2
4
6
8
10
12
20
18
16
14
RPM Sweep Number
22
24
26
28
30
Figure 70: RS5 Piston Break-In FMEP for 4 bar IMEP
4.2.3. RS12
An example profile measurement of the roughest piston skirt tested, the RS12, is shown in
Figure 71. The roughness for this piston is approximately twice that of the other two pistons. For
this roughness, both the 2 bar IMEP data (Figure 72) and 4 bar IMEP data (Figure 73) show a
FMEP stabilization around the eighth or ninth RPM sweep (approximately 6 hours of testing).
78
P- PmQ ausgevcktf tLzs =VJS
10,0
-10.0
Tastr TKU3O
L
= 4,80 mn
4,0
Lc =0.800 mm V = OM noo
Figure 71: Measured Roughness Profile for RS12 1009 after Coating
-4-10002bar
---
15002bar
---
2000 2 bar
16000
15000
14000
(U
0~
13000
0~
LU
12000
U-
11000
~ ~ 1~ ~ ~~ ~ ~~ ~ wim o2K 01
2 31 51 71 92
10000
9000
8000
1
2
3
4
5
6
7
8
9 10 11 12 13
RPM Sweep Number
14
15
16
Figure 72: RS12 Piston Break-In FMEP for 2 bar ITMEP
79
17
18
19
20
rIArI
---
10004bar
-*-20004bar
-U-15004bar
17000
16000
15000
-
14000
. 13000
au 12000
LL_
11000
1
10000
4
5
6
7
4
5
6
7
1
1
12
1
4
5
16
1
8
14
15
16
17
18
92
9000
Q800
1
2
3
8
9 10 11 12 13
RPM Sweep Number
19
20
Figure 73: RS12 Piston Break-In FMEP for 4 bar IMEP
After the break-in procedure was completed, the data for each of the broken-in pistons was
compared to analyze the friction effects arising from different skirt roughness. For this analysis,
the results from the baseline piston test described in the previous sections are included under the
label "base". The FMEP comparison between the different roughnesses is shown in Figure 74. It
is important to point out that data from a second RS12 piston (piston number 1005) was included
as well. The RS12 piston shows substantial increase (approximately 10 percent) in FMEP over
the smoother piston skirts. This increase is consistent for both of the new RS12 pistons tested.
Furthermore, the same pattern is shown for the baseline piston, which is also of the RS12
roughness. As for the comparison between the RA02 and RS5 pistons, there is minimal
difference in FMEP.
80
a RS5 1007
0 RS12 1005
14000 -
--
-
12000 -
-
- -
-
-
--
-
- --
-
-
-
-
U RS12 1009
-
16000 -
U base (RS12)
-
E RA02 1002
10000
cc
LU
6000
4000
2000
0
1000 2 bar
1500 2 bar
2000 2 bar
RPM
1000 4 bar
1500 4 bar
2000 4 bar
Figure 74: FMEP Comparison between Skirt Roughness for Broken-In Pistons
In order to determine where the differences in FMEP come from, various friction traces are
compared. Figure 75 through Figure 80 show these comparisons. An analysis of all of the graphs
shows the RS 12 pistons consistently display a different form than the other two pistons,
particularly in the region at and just after TDC. The shape of the trace indicates the rougher
pistons are experiencing significant boundary contact after TDC, while the RA02 and RS5
pistons remain largely in the hydrodynamic region. While the base piston does not always follow
the same pattern as the newer RS12 pistons, its shape closely reassembles them. Any variation
can be attributed to the more worn surface of the base piston skirt. It is also evident that the
disparities in friction are magnified when load is increased on the engine. The 4 bar IMEP graphs
clearly show this trend.
When comparing just the RA02 and RS5 piston, there is minimal difference. While the smoother
RA02 consistently shows an advantage in friction, this advantage is small, especially when
compared to the difference in the RS12 friction. Another trend to note is that while a large
magnitude of difference occurs around the TDC region, there are still some differences in the
81
mid-stroke as well. While these differences are small, they do show that the smoother pistons
consistently have lower friction in this area as well.
base -
-
RA02-3 1002 -
RS-6 1007 -
RS12-4 1009 -
RS12-4 1005
25
15
5
0
--------
-
-5
LL
-15
-25
-35
-360
-180
180
0
Crank Angle [deg]
360
Figure 75: Friction Comparison between Skirt Roughness at 1000 RPM 2 bar
-
base -
RA02-3 1002 -
RS5-6 1007 -
RS12-4 1009 -
RS12-4 1005
20
15
10
5
0
r-5
-10
-15
-20
-25
-30
-35
-360
-180
0
Crank Angle [deg]
180
360
Figure 76: Friction Comparison between Skirt Roughness at 1500 RPM 2 bar
82
-
base -
RA02-3 1002 -
RS5-6 1007 -
RS12-4 1009 -
RS12-4 1005
20
15
10
5
z 0
o
-
--------
-5
-10
-15
-20
-25
-30
-360
-180
360
180
0
Crank Angle [deg]
Figure 77: Friction Comparison between Skirt Roughness at 2000 RPM 2 bar
-
base -
RA02-3 1002 -
RS-6 1007 -
RS12-4 1009 -
RS12-4 1005
35
25
15
5
-5
0
U-
-15
-25
U.
-- -- --- ---
-
-35
-45
I
-
-55
-65
-360
-180
0
180
360
Crank Angle [deg]
Figure 78: Friction Comparison between Skirt Roughness at 1000 RPM 4 bar
83
RA02-3 1002 -
base 30
RS5-6 1007 -
RS12-4 1005
RS12-4 1009 --
-
-
10 -
-
-
-
-
20
-
-
-30 --
-
-
.0 -20 ---
-
-
--
z -10 ---
-
0 ------
-40
-50
-60
-180
-360
0
Crank Angle [deg]
180
360
Figure 79: Friction Comparison between Skirt Roughness at 1500 RPM 4 bar
RA02-3 1002 -
base -
RS12-4 1009 -
RS12-4 1005
15 - - -
-
-
-
25
RS5-6 1007 -
-
-
15
z-5
U
LL . -2I
-45
-360
-180
0
Crank Angle [deg]
180
360
Figure 80: Friction Comparison between Skirt Roughness at 2000 RPM 4 bar
In conclusion, the data shows a consistent trend that the smoother RA02 and RS5 pistons
produce significantly lower friction that the rougher RS12 piston. While this large difference
may initially be suspect, further verification by testing additional pistons of the same roughness
has reinforced the conclusions. The fact that the RA02 and RS5 pistons show little variation in
friction can be explained by comparing the surface measurements in Figure 65 and Figure 68.
84
While the micro-scale variation (roughness) of the RA02 is around 3 pm, the overall variation is
more on the order of 5 pm. This is because the desired scaled for the RA02 is below the
machining capabilities of the process used. The smallest overall roughness the process can
produce is on the order of 5 pm. At this level, the RA02 asperity height nearly matches that of
the RS5, which explains while their friction levels are so similar.
4.3. Skirt Profiles
After the analysis of piston skirt roughness was completed, a new investigation was started
involving the profile of the piston skirt. This investigation is the result of collaborations between
the manufacturer and MIT using modeling tools to optimize the machined profile of the piston
skirt. The results of this work yielded an optimized profile, shown below in Figure 82. In order to
test the new piston, it was first broken in, and then compared to tests conducted on both smooth
(RA02) and rough (RS12) pistons. It is important to note these tests were redone from the
previous skirt analysis, as recalibration of the engine dictated previous tests could no longer be
compared to new tests. Therefore, the pistons were tested again in order to have an accurate
comparison to the new profiles.
Figure 81: Typical FLE Piston with GRAFAL Skirt Coating
Table 4: Pistons Available for Testing
Piston Profiles
Piston
Piston Number
Roughness
85
base
RS12
1002
1006
RA02
1011
1012
1001
1003
1004
RS5
1007
1005
1009
RS12
Optimized
Profile
Optimized Profile
Ipm]
Thrust
Side
2
Anti-Thrust
Side
Figure 82: Optimized Piston Skirt Profile
4.3.1. Piston Skirt Profile Results
After analyzing the effects of skirt roughness, the next step was to analyze the effect of actually
changing the piston skirt profile during the manufacturing process. Previously, work had been
86
done by P. Totaro at MIT using his model to develop an optimized profile for oil distribution that
will minimize friction (optimized profile). He also calculated the results using several other
profiles [38]. In order to verify the results of the model simulation, these profiles were
manufactured with a RS12 roughness (previously shown in Figure 82). The minimum clearance
for these piston skirts was the same as the standard base pistons.
In order to analyze the effects of piston profile, both a RA02 and RS 12 from the previous
analysis were retested to use for comparison. A comparison of the FMEP from these tests is
shown in Figure 83. The results show that the optimized profile piston has lower FMEP at all
points, except when compared to the RA02 at 1000 RPM, 4 bar IMEP. However, considering the
roughness of the optimized profile piston is the same as the roughest, or RS 12 piston, it is
remarkable to see the gains that were made by changing piston skirt profile. Figure 84 through
Figure 89 show the friction trace comparisons between the smooth, rough, and optimized profile
pistons. The results show that the optimized profile design achieves lower overall FMEP by both
reducing the amount of boundary contact around the TDC area as well as reducing the length of
the contact when it does occur. Figure 85 and Figure 86 are examples of when the boundary
contact is reduced, and the piston skirt remains more hydrodynamic around the TDC area.
Interestingly enough, they show reductions in the mid-stroke region as well. Figure 84, Figure 87,
and Figure 88 are examples of the optimized profile design reducing the amount of time that
boundary contact occurs after TDC. The magnitude of the spike after TDC that insinuates
boundary contact is comparable to the other two pistons, but is narrower in size. Finally, Figure
89 is example where both of these phenomena occur simultaneously.
87
N RS 12 1009 (rough)
0 RA02 1006 (smooth)
U RS12B-4 1004 (optimized)
16000
14000
12000
10000
'0
8000
6000
4000
2000
0
1000 2 bar 1500 2 bar 2000 2 bar 1000 4 bar 1500 4 bar 2000 4 bar
RPM
Figure 83: FMEP Comparison for Optimized Profile Piston
20
RA02 1006 (Smooth) -
RS12B-4 1004 (Optimized)
-
RS12 1009 (rough) -
-
10
-10
2-20
L
-30
-360
-180
0
Crank Angle [deg]
180
360
Figure 84: Friction Comparison between Skirt Profile at 1000 RPM 2 bar
88
RS12 1009 (rough) -
-
RA02 1006 (Smooth) -
RS12B-4 1004 (optimized)
15
5
-01
Ww"
-5
-15
AF
0
LL
-25
-35
-45
-180
-360
180
0
Crank Angle [deg]
360
Figure 85: Friction Comparison between Skirt Profile at 1500 RPM 2 bar
-
RS12 1009 (rough) 20
10
RS12B-4 1004 (optimized)
RA02 1006 (Smooth) -
'WWI
0
z -10
r
Ai
-Ardop,
CL
A/
AWW
ALAVY
.49
-20
-30
-40
-50
-360
-180
0
Crank Angle [deg]
180
360
Figure 86: Friction Comparison between Skirt Profile at 2000 RPM 2 bar
89
-
RS12 1009 (rough) -
RA02 1006 (Smooth)
-
RS12B-4 1004 (optimized)
25
10
-5
-
---------
-20
0
4'
-35
-50
.U
-65
-80
-95
-360
0
-180
180
360
Crank Angle [deg]
Figure 87: Friction Comparison between Skirt Profile at 1000 RPM 4 bar
-
RS12 1009 (rough) -
RA02 1006 (Smooth) -
RS12B-4 1004 (optimized)
20
0
-20
0
-40
U
LL
-60
-80
-100
-360
-180
0
180
360
Crank Angle [deg]
Figure 88: Friction Comparison between Skirt Profile at 1500 RPM 4 bar
90
-
RS12 1009 (rough) -
RA02 1006 (Smooth) -
RS12B-4 1004 (optimized)
0
20
- -
- --
-
- -
1-0
-
r-30
'Z; -40 - -S
-
-
-
-
-
-
-
.
-
-10
-
0
-60
-70
-80
-360
-180
0
180
360
Crank Angle [deg]
Figure 89: Friction Comparison between Skirt Profile at 2000 RPM 4 bar
4.3.2. Conclusions
The initial investigation into the effects of piston profiles has shown that substantial gains can be
made by utilizing different piston skirt profiles. Furthermore, the results verify the findings of the
computer model, and open the door for further optimization. Additionally, the tests indicate that
a combination of shirt roughness and optimal profile may reduce the contributions to friction
from the skirt even further. It is important to note that this is just an initial investigation, and
further tests will be conducted in the future to reinforce the findings.
91
5. Effects of Lubricants
5.1. Effects of Lubricants under Fired Conditions
A major factor to consider when discussing friction is not just the surfaces themselves but also
the lubricant between them. The oil which serves this purpose in automotive engines has
undergone decades of development in order to perform a host of tasks. A difficulty in designing
engine oil is the range of environments in which it must perform. The same oil which lubricates
the journals and bearings in the valve train must also lubricate the piston assembly, in which the
temperatures can differ drastically. Furthermore, consideration must be given to the small
amounts of oil which inevitably is burned in combustion, and escapes to the exhaust system.
Several aspects of the engine oil were investigated in order to provide further insight into its
interactions with the piston assembly. The focus is limited to effects of viscosity and certain oil
additives in the piston assembly.
Two different oils were used in order to test the effects of changing oil viscosity on liner friction.
One had a High Temperature High Shear (HTHS) rate of 1.4 mPa-s, while the other was 2.9
mPa-s. The HTHS 2.9 is a common production oil under the label of OW-30.
In addition,
viscosity for individual oils can be further modified by changing the temperature of the oil:
increasing temperature decreases viscosity and vice-versa. Tests were conducted by varying both
oil type and temperature and holding all other factors constant. This allowed the isolation of
viscosity effects. Results from these tests are shown below. In general, the lower the viscosity of
the oil, the lower the friction was in the hydrodynamic regime, due to the reduction in shear
stress. However, this gain for lower viscosity came at the cost of allowing boundary friction to
occur at higher speeds, at which point FMEP was increased. This effect is exemplified in Figure
90, where tests shows that at lower speeds (1000 RPM) the FMEP is actually higher than at 1500
RPM for the lower viscosity oil (HTHS 1.4). The increased boundary contact that causes this
effect is shown in the friction traces displayed in Figure 91. However, FMEP at all speeds for the
lower (HTHS 1.4) viscosity oil is still lower than the higher (HTHS 2.9) viscosity oil.
92
8 HTHS 2.9
N HTHS 1.4
16000
14000
12000
10000
8000
LU
6000
LL
4000
2000
0
1000
2000
1500
RPM
Figure 90: Comparison of FMEP at 2 bar IMEP due to Changing Viscosity with GG22 and
0.23mm 10.5N TLOCR
30
20
10
0
U
0
-
HTHS 2.9
-
HTHS 1.4
-10
More
-20
contact
for HTHS 1.4
-30
-360
-180
1 0
0
Crank Angle [deg]
36 0
Figure 91: Friction Trace Showing Increased Boundary Contact for Lower Viscosity Oil under
Fired Conditions with GG22 and 0.23mm 10.5N TLOCR
Figure 92 shows how overall FMEP decreases with increasing temperature. In order to test the
effect of changing oil temperature, the oil, coolant, and thrust side cylinder liner temperatures
were maintained at the desired level with a margin of one degree Celsius. The oil temperature is
measured in-line right before flowing into the engine. As expected, the change in friction due to
changing oil temperature is directly related to the oil viscosity change with temperature. Since
93
viscosity decreases with increasing temperature, the change in FMEP is attributed to the change
in viscosity. This is reinforced in Figure 93 where the friction trace shows the difference in
FMEP comes mainly from changes in hydrodynamic friction, which corresponds to the
previously discussed viscosity effect. It is important to note, however, that decreasing viscosity
also increases the amount of boundary contact friction on the liner, which can have negative
effects such as increased wear which lead to higher maintenance costs in automotive applications.
20000
15000
mj
* 60C
10000
" 80C
0.
* 100C
5000
0
1000
1500
2000
RPM
Figure 92: FMEP Effect from Varying Oil Temperature under Fired Conditions with HTHS 2.9
Oil and 0.15mm 19.5N TLOCR at 2 bar IMEP
94
30
-------- -
20
10
0
-Z;-10
-60
C
-80
C
100 C
-20
-30
-40
-180
-360
0
Crank Angle [deg]
180
360
Figure 93: Friction Change due to Changing Oil Viscosity with Temperature under Fired
Conditions with HTHS 2.9 Oil and 0.15mm 19.5N TLOCR
5.2. Effects of Lubricant Additives
In addition to the honed surface of the liner, other factors can contribute to the roughness of the
liner surface. These factors include additives in the oil which form hardened films under specific
environments, as well as by-products of combustion which introduce fine particulates to the liner.
In order to analyze some of the effects these factors may have on friction, different oils of
various formulations were provided, shown below in Table 5. Results indicate that liner
roughness is a dynamic parameter influenced not only by wear, but also by other factors which
may or may not be permanent.
Table 5: Oil Formulations Used in Testing
2.9
Fully Formulated
1.4
2.9
Fully Formulated adjusted to 1.4
Fully Formulated without Anti-Wear Additives
Fully Formulated with alternate Viscosity
2.9
Modifiers
95
5.2.1. Introduction to Tribofilms
The initial investigation began with the desire to analyze the effect tribofilms may have on liner
friction. Tribofilms are films that form under certain temperatures and pressures from additives
in the engine oil. These films form a hard, protective layer to prevent wear. For this study,
specifically the zinc dialkyl dithiophosphate (ZDDP) film formed from the presence of zinc and
phosphorous was targeted. ZDDP films form under the presence of direct, rubbing, solid contact
of solid surfaces and form a film approximately 150 nm thick. Topolovec-Miklozic et al.
demonstrated that the formation of these films in bench tests takes about 30 minutes [39]. While
these films are used to prevent wear for many contact surfaces in the engine (such as valve train
and crankshaft), they have been shown to have a higher friction coefficient when compared to
surfaces without the film [39]. Because of this tradeoff, it was previously theorized that the
presence of ZDDP films on the liner surface will increase the friction of the piston assembly. The
explanation for this put forth by Topolovec-Miklozic et al. is that the film forms only on the
peaks of the liner surface (shown below in Figure 94). This increases the height of these peaks,
which in turn deepens the valleys, allowing them to serve as drainage paths in the contact inlet
and prevent the build-up of fluid pressure. The net effect is the prevention of hydrodynamic
effects (similar to having a rougher liner honing). In order to investigate this theory, a test
sequence was developed with the FLE in order to shed more light on the subject [39].
ZDDP
Liner Surface
Peaks
Valleys
Figure 94: Theorized Effect of ZDDP Film on Liner Surface
5.2.2. Test Procedures
The investigation into tribofilms on the liner surface requires a unique testing procedure due to
the fact that the film buildup is expected to occur in fired conditions. However, the small
differences in friction from the films dictate the repeatability of motored tests is necessary.
Therefore a sequence incorporating both fired and motored tests was needed.
96
In order to test the possible effects of the ZDDP film, a washing procedure was recommended
which was proven to remove the film in lab tests. This procedure consisted of soaking the liner
and rings in a 0.05M solution of Ethylenediaminetetraacetic acid (EDTA) for five minutes, and
then rinsing well with water. The final step was to rinse the parts with acetone and dry them with
a clean paper towel. This procedure is referred to as the EDTA wash [40].
An outline of the developed test sequence is shown below in Figure 95, along with labels that
identify different motored tests used for the comparison. For all tests, the 0.15mm 19.5N TLOCR
was used.
Test Sequence:
1. Liner wash
2. Motored test
3. 3 Fired tests (Approx. 9 hrs)
4. Motored test
5. Liner wash
6. Motored test
Test Label
> Washed
>
Post Fired
Post Fired Washed
Figure 95: Tribofilm Investigation Test Sequence
The first step of the test sequence was to wash the liner and the rings with the EDTA wash to
remove any existing tribofilms. After this, a motored test was conducted to serve as a baseline
point for friction comparisons. This test is referred to as the washed test. The next step was to
fire the engine for a series of three fired tests, which effectively exposed the liner to the
combustion environment for approximately nine hours. After this, a second motored test was
conducted (referred to as the post fired test) in order to compare with the initial baseline, and see
if any effects could be seen from the formation of tribofilms. The final step was to wash the liner
and rings again, and conduct a third motored test, labeled the post fired washed test. The purpose
of this test was to serve as a comparison to the initial baseline, to see if any results from the post
fired test were permanent or removable by washing. All data presented is from the motored tests
in the sequence.
97
5.2.3. Friction Results and Hypothesis
Firing Effect with Standard Oil
The results from the test sequence using the fully formulate HA oil were quite astounding. Figure
96 shows that the post fired test showed lower friction in the boundary region, but higher friction
in the hydrodynamic region (Figure 97). This effect was opposite of previous convention, and is
the same effect one would expect with changing liner honing from a rougher to smoother liner.
In essence, exposing the liner to the combustion environment had the effect of "smoothing" the
liner. This effect was dubbed the "firing effect".
-U-washed
---
post fired
10000
9000
8000
-
7000
_
I
LU
U-
6000
I
5000
4000
100
200
300
400
500
600
700
800
900
1000
RPM
Figure 96: Boundary/Mixed Friction Region for Firing Effect with HA Oil at 100C
98
-4-Post Fired
-U--washed
13000
12000
11000
10000
'U
9000
8000
LU
7000
6000
IV mw
5000
4000
100
200
300
500
400
600
700
800
900
1000
RPM
Figure 97: Hydrodynamic Friction Region for Firing Effect with HA Oil
Figure 99 shows the comparisons for the tests compiled on an average Stribeck curve. This curve
also includes data from a motored test conducted just before the first wash, and is labeled prewashed. The friction trace in Figure 98 shows that just as in the liner roughness study, this
difference comes from decreased boundary contact for the post fired test (particularly at TDC),
and increased hydrodynamic friction at the mid-stroke.
-
Washed
-
Post-Fired
20
15
10
5
0
0
I-
-5
-10
-15
-20
360
0
-180
180
360
Crank Angle [deg]
Figure 98: Friction Trace Showing Firing Effect on GG07 at 100 RPM 40C
99
-4-Washed
-U-Post-Fired
-
Post-Fired Washed
-+0-Pre-Washed
0.15
0.14
0.13
0.12
0.11
U
0.1
T 0.09
4 0.08
0.07
0.06
0.05
0.04
0.E+00
5.E-05
2.E-04
1.E-04
2.E-04
3.E-04
3.E-04
p*Umax/(2Ft/Bore)
Figure 99: Average Stribeck Curve Showing Firing Effect with Fully Formulated HA Oil on GG07
Liner
FiringEffect with Oil without Anti-Wear Additives
To advance the study, and oil designated HTHS 2.9HK was used, which is a version of the fully
formulated HTHS 2.9HA that lacks the anti-wear additives Zn and P. The previously developed
firing test sequence was conducted using this oil in order to compare results to the same test
conducted with the fully formulated oil and determine the effect, if any, that anti-wear additives
have on liner friction. The result was that the same pattern observed with the fully formulated
HA oil was repeated with the HK oil. A graph of FMEP versus RPM at 40C is shown in Figure
100 and shows how firing has the effect of increasing friction in the hydrodynamic region, which
is the same effect observed with the HA oil.
100
-0--Post Fired Washed
-W--Post Fired
-*-Washed
-
14000
13000
12000
11000
10000
9000
Firing caused friction to
increase in hydrodynamic
LU
8000
U-
7000
6000
___n_
___mi_
_1
5000
100
200
300
400
600
500
700
800
900
1000
RPM
Figure 100: Hydrodynamic Region Friction Tribofihm Investigation with HK Oil on GG07 at 40C
A closer look at the friction trace at 800 RPM, shown in Figure 101, verifies that hydrodynamic
friction does indeed increase after the liner is exposed to the combustion environment, and effect
which is reversible with liner washing.
-Washed
-
Post Fired
-Post
Fired Washed
25
20
------------ - - -
15
10
5
z
0
0
-5
-.
-10
-15
-20
-25
-360
-180
0
180
Crank Angle [deg]
Figure 101: Friction Trace Showing Change in Hydrodynamic Friction
101
360
When the friction regime transitions from hydrodynamic to mixed, as shown in tests conducted
at 80C, it is evident that the opposite effect occurs: liner friction is decreased after exposure to
combustion. This effect is shown in Figure 102.
-U--Washed
-4-Post Fired
*&
di-Post Fired Washed
10000
Firing decreases friction
iin mixed region
9500
9000
Lines merge as region
transitions to
hydrodynamic
8500
8000
(U
a.
7500
7000
6500
6000
5500
5000
100
200
300
400
500
600
700
800
900
1000
RPM
Figure 102: Mixed Region Friction for Tribofiln Investigation with HK Oil
A further analysis in of the individual friction traces at 200 RPM shown in Figure 103 reinforces
that boundary friction is decreased by firing, especially at TDC (which has the most exposure to
combustion effects).
102
-
Washed
-
Post Fired
-
Post Fired Washed
20
15
--------
-
10
-'5
-25
-20
-360
180
0
-180
360
Crank Angle [deg]
Figure 103: Change in Boundary Friction Due to Firing with HK Oil 300 RPM 80C
FMEP graphs at 100C show a predominantly mixed friction regime, and it is clear that exposing
the liner to combustion effects decreases friction. Figure 104 displays these results.
-E-Washed
-4--Post Fired
-*-
Post Fired Washed
11000
10000
i
I
9000
8000
U.'
LL
7000
6000
.9
777-
5000
4000
100
200
300
400
500
600
700
800
900
1000
RPM
Figure 104: Mixed/Boundary Region Friction for Tribofilm Investigation with HK Oil
Finally, an analysis of the average friction coefficient (Fc) verifies what the FMEP and friction
traces indicate. Exposing the liner to combustion effects changes the friction coefficient of the
liner surface. Friction is increased in hydrodynamic regions and decreased in mixed regions, a
103
process that is reversible with washing therefore ruling out the possibility of permanent liner
alteration. Figure 105 shows a plot of the average friction coefficient versus maximum piston
speed (normalized for unit pressure) for the entire firing test sequence.
--
0.15
Washed
-U--Post-Fired
-
Post-Fired Washed
-_
0.14
0.13
0.12
0.11
U
U-
a
0.1
0.09
0.08
0.07
0.06
0.05
0.04
G.E +00
5.E-05
1.E-04
2.E-04
2.E-04
3.E-04
3.E-04
p*Umax/(2Ft/Bore)
Figure 105: Average Stribeck Curve Showing Firing Effect with HK Oil on GG07 Liner with
0.15mm 19.5N OCR
Firing Effect with the Tr-Solvent Wash
In order to isolate the effects of the tribofilm, the firing test sequence was conducted with a trisolvent wash instead of the normal EDTA wash, and the fully formulated HA oil. The tri-solvent
wash was recommended as a wash that does not remove the tribofilm. For this wash, the liner
and rings are soaked in a solution of equal parts toluene, heptane, and acetone for five minutes. It
is then dried with a clean paper towel. The basic theory is that is the same trend is observed from
the firing tests with the tri-solvent wash, then it can be concluded that the tribofilm has little to
no effect on liner friction.
104
An FMEP analysis at 40C shows the hydrodynamic regime, in which the same pattern as
previous firing tests occurs; namely, exposing the liner to combustion effects increases friction in
the hydrodynamic regime. This is shown in Figure 106.
-$-Post Fired
-U-Washed
-&- Post Fired Washed
-4uuu
I
13000
12000
____________
____________
____________
--
___________
____________
_____________
____________
____________
.11000
a-
10000
2 9000
Uj-
8000
_____
-
-____________
-
7000
6000
5000
100
200
300
400
500
600
700
800
900
1000
RPM
Figure 106: Hydrodynamic Region Friction for Tri-Solvent Wash
Tests at 80C show the transition from the mixed to hydrodynamic regimes, and Figure 107
shows a very nice example of the firing effect. In the mixed region, friction is reduced by firing.
However in the hydrodynamic region, it is increased.
105
-0-Washed
-- *-Post Fired
-*-
Post Fired Washed
10000
9000
8000
U
0.
7000
6000
INL
N
5000
4000
100
200
300
400
500
600
700
800
900
1000
RPM
Figure 107: Mixed Region Friction for Tri-Solvent Wash
Analyzing the friction trace at 400 RPM and 80C shows exactly how this transition occurs.
Exposing the liner to combustion effects causes the trace to act more hydrodynamic versus the
same conditions for washed liners. This is shown in Figure 108.
-Post
-Washed
Fired
-
Post Fired Washed
20
15
10
5
0
0
4-
-5
-10
-15
-360
-180
0
Crank Angle [deg]
180
360
Figure 108: Change in Mixed Region Friction Due to Firing with Tri-Solvent Wash
Figure 109 shows an FMEP graph that is almost completely in the mixed region, and how the
effect is reinforced.
106
-U-Washed
---
Post Fired
-i--Post Fired Washed
11000
10000
9000
cL 8000
7000
- -- M--I--S-
6000
5000
4000
100
200
300
400
500
600
700
800
900
1000
RPM
Figure 109: Mixed/Boundary Region Friction for Tri-Solvent Wash
Finally, just as with the HA and HA EDTA washed tests, plotting the average friction coefficient
of the tri-solvent wash fired test sequence shows that firing has the same effect, which is
changing the friction coefficient of the liner. This result is shown below in Figure 110.
-4-Washed
-*-Post-Fired
-
Post-Fired Washed
0.16
0.14
0.12
U
LL
a'
(U
a'
0.1
0.08
0.06
0.04
0
0.00005
0.0001
0.00015
p*Umax/(2Ft/Bore)
107
0.0002
0.00025
0.0003
Figure 110: Stribeck Curve Showing Friction Coefficient for HA Oil with Tri-Solvent Wash on
GG07 with 0.15mm 19.5N TLOCR
Firing Effect Conclusions
It is important to note the original motivation for the firing effect investigation, that is, to
determine the possible formation and effects of the anti-wear tribofilm on liner friction. Bench
tests have shown that under the right conditions the HTHS 2.9HA oil will from a tribofilm, and
this film increases the friction coefficient of the contact surface [39]. However, in these tests the
films form under unit pressures several orders of magnitude higher than the ring/liner interface.
It is unknown to what extent these films form on the liner surface, and if they do what effects
they have on friction. Convention would lead to the theory that anti-wear additives in the oil
should form a tribofilm in the heat and pressure of the combustion environment that increases
friction between the piston and liner assembly, especially at areas where there is more contact
such at TDC and in mixed/boundary friction regimes.
The results from the investigation proved this theory to be incorrect. In fact, all tests using two
different oils and two different washing procedures showed the opposite: that exposing the liner
to the combustion environment reduced friction in the mixed/boundary region where contact
occurs. However, friction was increased in the hydrodynamic region. The fact that this effect
could be reversed by washing the liner again (and therefore removing whatever effects
combustion had on the surface) proved that the effect was not due to a permanent change in the
liner surface, but by some other phenomenon.
A closer analysis of the results from the firing effect tests shows that exposing the liner to the
combustion environment has the same effect as swapping the liner for one with a smoother
characteristic roughness, as shown in the previous liner roughness study. Firing the liner has the
effect of "smoothing" the surface, an occurrence that was dubbed the "smoothing effect". While
the reasons for this are not yet clear, a proposed theory is that byproducts of combustion such as
ash or unknown additives are filling the crevices in the liner surface, which in essence makes it a
smoother liner.
108
Liner Surface
Ash/Unknown
Tribo FilmAdivs
Increases friction
Changed
coefficient of
boundary region
hdoyai
behavior
Figure 111: Smoothing Effect Theory Due to Firing
If there is a tribofilm forming, it will most likely only form on areas that are exposed to higher
contact pressures, such as the peaks of the liner surface. Whatever contribution this tribofilm has
on liner friction (if it is formed at all) are overridden by the effect of the valleys of the liner being
filled and smoothing the surface.
Washed Liner: Unfilled crevices
c)
CL
Filled Crevices
h (clearance)
Figure 112: Pressure Relationship for Washed vs Fired Liners in the Hydrodynamic Regime
The reasoning behind the phenomenon of washed liners showing friction trends comparable to a
rougher surface is synonymous to the explanation for hydrodynamic relationships on liner
surfaces. Figure 112 shows the theory of how washed surfaces (which have deeper valleys) cause
hydrodynamic pressure to decrease at slower rate than an unwashed liner for the same load. A
good analogy to understand this is water that flows over smooth ground travels faster than water
over rough ground. The washed liner acts like a rougher surface, in which the oil takes longer to
109
escape, so as hydrodynamic pressure decreases the clearance between the surfaces stays larger.
Since the clearance is inversely proportional to the hydrodynamic friction coefficient, a rougher
liner exhibits a lower friction coefficient in hydrodynamic regions.
5.2.4. Liner Examination Results
In an effort to further investigate the fundamentals of the firing effect, the liner used in the study
was analyzed with a microscope. While detailed measurement results are still currently being
compiled and analyzed, an optical microscope image of the surface shows how the EDTA wash
does in fact change the nature of the surface. In Figure 113 below, the circled region on the left
shows more interference in the form of various colors, while the same area lacks these colors
after washing on the right.
After EDTA wash
Fired (before EDTA)
Figure 113: Optical Image of Liner Surface at TDC
The profilometry measurement of the liner at TDC before and after the EDTA wash shed even
more light on the effect. The liner after the wash in Figure 114 shows many more grooves than
the liner that has not been washed. These grooves represent liner valleys that were filled due to
firing.
110
After EDTA wash
Fired (before EDTA)
20
0.64
04
2.0
0,60
0.60
1.0
050
1.5
1.0
0.50
0.5
0
.0s
040
0.30
-0.5
010
-0
0.20
-1.5
0.10
0.00
0.0
--2.0
l0
-215
0.2
0.6
0.4
-2.0
0,10
0.8 0.9
.....
..
.I.
..
.
..
...
.....
..
...
....
.
...
...
..
...
..
..
......
..
..
....
.~
...
04
06
08 09
-2.5
.
Im
02
Figure 114: Profilometry Measurement of Liner at TDC
At BDC, the same results do not occur. Figure 115 shows how images of the liner surface at
BDC look largely identical. This reinforces what the experimental data suggested, that the effect
is stronger at TDC, possibly due to its proximity to the combustion environment.
After EDTA wash
Fired (before EDTA)
Figure 115: Optical Image of Liner Surface at BDC
111
6. Conclusion
6.1. Summary
This thesis used experimental methods to explore different components of the piston/ring pack
systems such as the oil control ring, piston skirt, liner surface, and lubricant. By making use of
the unique testing capabilities of the Floating Liner Engine, each of these components was able
to be isolated in order to analyze its effect separately. The testing methods developed by past
studies were used and expounded upon in order to record reliable, usable data. This data was
then analyzed in order to identify trends in friction related to each of the components studied.
Two types of oil control ring were analyzed, the Twin Land Oil Control Ring (TLOCR) and
Three Piece Oil Control Ring (TPOCR). The effects of changing specific parameters of the
TLOCR and their relationship to different liner honings were studied in detail and several key
trends were identified. Modeling efforts were mentioned that helped to explain some of these
trends. The level of analysis completed on the TLOCR gives a detailed understanding of its
performance in the piston/ring pack system.
The main focus of study for the TPOCR was an initial experimental investigation to compare
trends to those identified for the TLOCR. Several differences were observed which will provide
a basis for future exploration. The trends observed were described by modeling effects with
some degree of success, but more work is needed to fine tune the efforts and produce a clearer
picture of TPOCR friction trends in the future.
The piston skirt also plays a critical role in the friction contributions of the piston assembly. This
thesis completed a brief exploration of the effects of producing patterns in the low friction
coating on the piston skirt. In addition, the effects of changing the roughness of the piston skirt
coating were identified and conclusive evidence was obtained. Finally, the profile of the piston
skirt itself was considered based on past efforts to optimize the profile through a modeling
approach. This area of study proved to show some promise in the area of friction reduction, and
further investigation into this area is encouraged.
Finally, the effects lubricants can play in the friction of the piston assembly were explored. This
study started by detailing the outcome on friction by changing lubricant viscosity through both
112
temperature and formulation. Next, specific anti-wear additives were considered in order to
determine their effect, if any, on friction in the piston/liner assembly. These testing efforts led to
new observations about the changing nature of the liner surface due to combustion by-products.
Again, this was just an initial investigation which highlighted areas of future concern.
6.2. General Conclusions
The general conclusions for this document will match the format in that it will consist of three
parts. In general, the FLE proved to be a vital tool for obtaining experimental data on the
piston/liner assembly. By studying the data collected as well as working with existing modeling
tools, many trends were identified which will serve as a basis for future designs and
investigations.
OCR Conclusions
The TLOCR is now well documented and most friction trends associated with it are supported by
both experimentation and modeling. The efforts of this thesis have led to several key conclusions
regarding the TLOCR.
1. When used in conjunction with the TLOCR, smoother liners (those with a lower structure
height) show lower friction levels when significant contact occurs in the boundary and
mixed region. Rougher liners perform better in the hydrodynamic region.
2. For a given oil viscosity and unit pressure, rough liners exhibit a minimum friction
coefficient at larger sliding speeds than smoother liners. They also show a minimum
friction coefficient that is larger in magnitude. It is important to remember that these
results are for one type of liner from a single manufacturer.
3. In regards to TLOCR land width, observations under motored conditions show that the
larger the land width, the lower the speed at which the minimum friction coefficient
(transition point) is achieved, when all other things are equal. Under these conditions the
smallest land width displayed the lowest friction in the hydrodynamic regime.
4. Under fired conditions, changing TLOCR land width has a minimal effect on friction,
due to the "canceling effect" without changing tension. However, common industry
practice is to reduce the land width and the tension proportionally to maintain the same
unit pressure. Once can expect a ring pack friction reduction by doing so.
113
5. Increasing TLOCR spring tension increase friction across all regimes for both smooth
and rough liners under both motored and fired conditions. Modeling showed that this was
due to the increasing tension reducing the film thickness the TLOCR. This not only
increased the friction of the TLOCR but also starved the top two rings, in turn increasing
their friction as well.
Unlike the TLOCR, the TPOCR is still in the beginning phases of study, and experimental
efforts here focused more on general comparisons and identification of trends, as well as
collaboration with modeling efforts.
6. The trends identified for the TLOCR differ for the TPOCR. The TPOCR seems to delay
contact separation, meaning friction levels stay higher for rougher liners for much higher
speeds (or viscosities/unit pressures) than the TLOCR. The pure hydrodynamic behavior
observed with the TLOCR on rough liners was not fully analyzed for the TPOCR due to
limitations in the testing window.
7. Once hydrodynamic behavior is developed, the TPOCR shows an advantage over the
TLOCR, possibly due to its small contact areas developing high film thicknesses.
8. When compared to experimental results, TPOCR modeled data shows the same shape
and form, but a significantly lower magnitude. The reasons for this are still under
exploration but are believed to be related to the assumed ring profile.
Piston Skirt Conclusions
Several key conclusions were drawn in regard to piston skirt friction.
1. The patterns of low friction coating on the piston skirt yielded no conclusive advantage
over the base solid coating. In fact, some of the patterns showed higher levels of friction.
2. In general, the lower the roughness of the piston skirt coating, the lower the friction
contributions. The smoother piston skirt coatings show a significant improvement in the
friction spike just after TDC during the expansion stroke.
3. Optimizing the piston skirt profile can have an advantage in friction. While the results
remain preliminary, the data shows promise in the area of piston skirt profile optimization.
The optimized profile used in this study showed an advantage by decreasing both the
114
magnitude of the friction that occurs right after TDC (during the expansion stroke) and
the duration.
Lubricant Conclusions
The investigation into lubrication effects on friction yielded some surprising results that
definitely warrant future exploration.
1. Decreasing oil viscosity can decrease friction in the hydrodynamic regime, but is also
causes boundary contact to occur at higher speeds. This contact can increase overall
friction compared to a higher viscosity oil at low engine speeds.
2. In regards to the ZDDP anti-wear films stemming from additives in the oil, the expected
result of increasing friction in the piston assembly was not observed. In fact, the under
the conditions and procedures tested in this study the ZDDP film was found to have a
minimal effect.
3. A new effect, dubbed the "firing effect", was observed. This phenomenon occurs when
the liner surface is exposed to the fired conditions in the combustion environment. The
result as a change in the performance of the liner surface, which makes it act like a
"smoother" surface than before. The effect is non-permanent as washing the liner
returned friction levels to pre-fired conditions. It is believed that byproducts of
combustion are filling the crevices in the liner, but further investigation is recommended.
6.3. Future Work
This thesis developed many of the building blocks needed to further understand the friction
contributions from the piston/liner assembly. However, this is just one piece in a larger puzzle
needed to extend the knowledge necessary to ultimately increase overall engine efficiency. As a
result, there are a number of areas the warrant future work.
As previously stated, the investigation in to the TPOCR is in the early stages. This means there is
still much more work to be done, both on the experimental and modeling side. In reference to
modeling, methods to develop a more accurate contact profile are necessary which will hopefully
alleviate some of the discrepancies between modeled and experimental data. On the experimental
115
side, further investigation into TPOCR parameters is needed, as well as a comparison of TPOCR
and TLOCR friction under fired conditions.
More testing of the piston skirt is needed as well. At this point, only one of the new profiles was
tested in the FLE. In the future other profiles provided need testing for comparative purposes, as
well as further collaboration between MIT modeling and manufacturers to develop and test
optimized designs.
In regards to the firing effect, the identification of this phenomenon means much more inquiry is
needed. To start with, repeating the investigation with a smooth liner to see if the same pattern
exists may be helpful. Furthermore, future collaboration with manufacturers on this topic may
lead to alternative lubricant formulations with which to isolate specific factors in the results.
116
References
[1] J. A. McGeehan, "Literature Review of the Effects of Piston Ring Friction and Lubricating
Oil Viscosity on Fuel Economy," SAE Paper, 780673, 1978.
[2] J. B. Heywood, in Internal Combustion Engine Fundamentals, McGraw-Hill Book
Company, 1988, pp. 725-730.
[3] D. E. Richardson, "Review of Power Cylinder Friction for Diesel Engines," ASME Journal
of Engineeringfor Gas Turbines and Power, pp. 506-519, 2000.
[4] H. Chen, "Modeling the Lubrication of the Piston Ring Pack in Internal Combustion
Engines Using the Deterministic Method," PhD Thesis, Massachusetts Institute of
Technology, 2011.
[5] C. Drury and S. Withehouse, "The Effect of Lubricant Phosphorus Level on Exhaust
Emissions in a Field Trial of Gasoline Engine Vehicles," SAE Paper 940233, 1994.
[6] F. Ueda, S. Sugiyama, K. Arimura, S. Hamaguchi and K. Akiyama, "Engine Additive
Effects on Deactivation of Monolithic Three-Way Catalysts and Oxygen Sensors," SAE
Paper 940746, 1994.
[7] P. Pawlus, T. Cieslakb and T. Mathia, "The Study of Cylinder Liner Plateau Honing
Process," Journal of MaterialsProcessing Technology, no. 209, pp. 6078-6086, 2009.
[8] T. Heroke, "Comprehensive 3D Evaluation of Honed Surfaced for Combustion Engines".
[9] R. Stribeck, in The Basic Propertiesof Sliding and Rolling Bearings, Zeitschrift des Vereins
Deutscher Ingenieure, 2002, pp. 1341-1348; 1432-1438; 1463-1470.
[10] K. Liao, "Factors Affecting Piston Ring Friction," PhD Thesis, Massachusetts Institute of
Technology, 2013.
117
[11] R. Panton, Incompressible Flow, John Wiley and Sons, 2013.
[12] M. Noonnan, "Overview of Techniques for Measuring Friction Using Bench Tests and
Fired Engines," SAE Paper 2000-01-1780, 2000.
[13] R. Benzing, M. Peterson and e. al, "Friction and Wear Devices," ASLE Publication, Park
Ridge, Illinois, 1976.
[14] "ASTM G99 Standard: Test Method for Ultra Testing with a Pin-On-Disk Apparatus,"
ASTM, PA, 1990.
[15] D. Patterson, S. Hill and S. Tung, "Bench Wear Testing of Engine Power Cylinder
Components," in Presentedat the ASME Fall Technical Conference, Muskegon, Michigan,
1991.
[16] "ASTM G77 Standard: Test Method for Ranking Materials to Sliding Wear Using BlockOn-Ring Wear Tester," ASTM, PA, 1983.
[17] S. Tung and e. al, "Engine Oil Effects on Friction and Wear Using 2.2L Direct Injection
Diesel Engine Components for Bench Testing Part 2 - Tribology Bench Test Results and
Surface Analyses," SAE Technical Paper 2004-01-2005, 2004.
[18] S. Hartfield-Wunsch, S. Tung and C. Rivard, "Development of a Bench Test for the
Evaluation of Engine Cylinder Components and the Correlation with Engine Test Results,"
SAE 1993 Transactions, vol. Section 3, no. Paper No. 932693, pp. 1131-1138, 1993.
[19] S. Hartfield-Wunsch and S. Tung, "The Effect of Microstructures on the Wear Behavior of
Thermal Sprayed Coatings," in Reprint from the 1994 7th Thermal Spray Conference
Proceedings,Boston, Massachusetts, 1994.
[20] S. C. Tung and S. I. Tseregounis, "An Investigation of Tribological Characteristics of
Energy-Conserving Engine Oils Using a Reciprocating Bench Test," SAE Technical Paper
2000-01-1781, 2000.
118
[21] S. Hartfield-Wunsch and e. al., "Development if a Bench Wear Test for the Evaluation of
Engine Cylinder Components and the Correlation with Engine Test Results," SAE
Technical Paper No. 20075101, 1993.
[22] M. Takiguchi, H. Ando, T. Takimoto and A. Uratsauka, "Characteristics of friction and
lubrication of two-ring piston," JSAE Review, no. 17, pp. 11-16, 1996.
[23] D. Madden and e. al., "Part 1: Piston Friction and Noise Study of Three Different Piston
Architectures for an Automotive Gasoline Engine," SAE Paper 2006-01-0427, 2006.
[24] K. Kim, T. Godward, M. Takiguchi and S. Aoki, "Part 2: The Effects of Lubricating Oil
Film Thickness Distribution on Gasoline Engine Piston Friction," SAE Paper 2007-01-1247,
2007.
[25] K. Kim and e. al., "Part 3: A Study of Friction and Lubrication Behavior for Gasoline Piston
Skirt Profile Concepts," SAE International, 09PFL- 1163, 2009.
[26] R. Wakabayashi and e. al., "The Effects of Piston Rings and Liner Break-in on Lubricating
Condition," SAE Paper 2007-01-1250, 2007.
[27] S. Furuhama and M. Takiguchi, "Measurement of Piston Frictional Force in Actual
Operating Diesel Engine," SAE Paper No. 790855, 1979.
[28] S. Furuhama and e. al., "Effect of Piston and Piston Ring Designs on the Piston Friction
Force in Diesel Engines," SAE Paper No. 810977, 1981.
[29] I. Sherrington and E. H. Smith, "The Measurement of Piston-Ring Friction by the 'FloatingLiner' Method," SAE Paper No. 884707, 1988.
[30] M. Takiguchi and S. Aoki, "Friction Analysis of a Piston for Gasoline Engines," JSAE
Technical Paper No. 20075101, 2007.
[31] S. Cho and e. al., "Frictional Modes of Barrel Shaped Piston Rings Under Flooded
119
Lubrication," Tribology International,no. 33, pp. 545-551, 2000.
[32] Mercedez-Benz, "Honen von Zylinder-Laufflachen," 2008.
[33] Y. Liu, "Comparison Between the Modeling and Experimental Results," in Consortium on
Lubrication in IC Engines, July 2013.
[34] T. Tian, "Modeling the Performance of the Piston Ring Pack in Internal Combustion
Engines," PhD Thesis, Massachusetts Institute of Technology, 1997.
[35] T. Tian, V. Wong and J. Heywood, "Modeling the Dynamics and Lubricaiton of Three
Piece Oil Control Rings in Internal Combustion Engines," SAE Paper No. 982657, 1998.
[36] MAHLE GmbH, Pistons and Engine Testing, Stuttgart: AZ Druck and Datentechnik, 2012.
[37] E. Zanghi, "Analysis of Oil Flow Mechanism in Internal Combustion Engines via High
Speed Laser Induced Fluorescence (LIF) Spectroscopy," M. S. Thesis, Massachusetts
Institute of Technology, 2014.
[38] P. Totaro, "Modeling Piston Secondary Motion and Skirt Lubrication with Applications,"
M. S. Thesis, Massachusetts Institute of Technology, 2014.
[39] K. Topolovec-Miklozic, T. Reg Forbus and H. Spikes, "Flim Thickness and Roughness of
ZDDP Antiwear Films," Tribology Letters, vol. 26, pp. 161-171, May 2007.
[401 Shell Inc, "Lubricant Antiwear Flim Formation and Removal," in Consortium on
Lubricationin IC Engines, 2014.
120