Ground Damage Database
David Anderson
Nancy Rockbrune
26th IGHC Conference
Ground Damage Database
Introduction
26th IGHC Conference
Introductions
 David Anderson
• GDDB Chair
• Head of Operational Safety ~ British Airways
 Nancy Rockbrune
• GDDB Secretary
• Assistant Director SMS and Operational Data Management ~ IATA
3
26th IGHC Conference
Ground Damage Database
Overview
26th IGHC Conference
History
 Ground damage has been long reported to cost airlines $4b annually
•
•
Estimate
Unplanned costs
 Significant changes need to be made
•
Thin margins
 Focus on Ground Operations has been made
 Collaboration to make a difference
5
26th IGHC Conference
History
 Launched 2011, with limited membership
 Shift in IATA’s data management and analysis approach
 During 2011 and 2012 new reporting protocols and requirements
were developed
•
•
•
•
Updated Contract
Launched Q1 2012
First useable data received from 10 members (with some manipulation)
Consistently receiving useable data from 14 participants
 2013 focus on expanding participation
•
Any airline, ground service provider, and / or airport which provide ground services
are eligible to participate in the program
6
26th IGHC Conference
GDDB Coverage ~ As of December 2012
7
26th IGHC Conference
GDDB Coverage ~ As of May 1, 2013
8
26th IGHC Conference
GDDB Coverage ~ As of May 1, 2013
9
26th IGHC Conference
Efforts to Support Growth
 Introductory letter to all Stakeholders
 Updated contract with reporting criteria
 IT development
•
•
•
•
•
Automated quality check
Web Form
Discussions with strategic partners to develop GDDB submission extract from
existing reporting systems
Common dimension tables
Participant query tool
10
26th IGHC Conference
Efforts to Support Growth
 Introduction of IOSA provision ~ ISM Ed. 7
GRH 1.11.6
The Operator should have a process to ensure aircraft ground damages
are reported to IATA for inclusion in the Ground Damage Database (GDDB). Such
reports should be submitted in accordance with the formal IATA ground damage
reporting structure. (GM)
11
26th IGHC Conference
Ground Damage Database
Purpose
26th IGHC Conference
Purpose
 Facilitate data driven improvements to effectively improve
performance
 Gather and analyze global data with Industry partnership
•
•
•
Provide information not otherwise possible
Identify trends and contributing factors allowing for the
development and assessment of effective mitigation actions
Establish a baseline of ground damage performance in
which future comparisons can be made
13
26th IGHC Conference
Use of Data
 Conduct statistical analysis on clean defensible data
•
•
•
•
•
•
Statistical analysis produces more tangible information
Measures process performance
Identify and prioritize contributing factors to process performance
Measure and predict process performance improvements
Provides confidence interval
Measures the quality of the data
 Communicate findings to applicable WGs and TFs
14
26th IGHC Conference
What Data?
 Data (even seemingly benign) is critical
•
•
•
Identifying issues
Determine the effectiveness of any mitigation actions
Demonstrating effectiveness of the program as a whole
 Data driven decisions
 Not only improves safety / performance but also makes for a more
efficient organization as a whole
15
26th IGHC Conference
Types of Data?
 Reactive ~ wait for incidents to happen and try to understand why
 Proactive ~ analyze identified risks to mitigate before they turn into an
accident / incident
 Predictive ~ mature system which conducts predictive analytics
(statistical modeling) to identify and mitigate unknown risks
16
26th IGHC Conference
Confidential Reporting
 Confidential reporting can be used for the following safety concerns:
•
•
•
•
•
Unsafe behaviors
Inadvertent errors and mistakes
Near miss occurrences (incidents that did not occur but could have easily resulted
in a serious event)
Inadvertent errors or violations of aircraft handling or servicing systems
Procedures or processes that could be improved
17
26th IGHC Conference
Data Collection Process







Identify hazards
Report on hazards and occurrences
Collect and risk assess reported hazards
Trend and analyze information
Identify mitigation action
Monitor for effectiveness
Review and monitoring for continuous improvement
18
26th IGHC Conference
Challenge ~ Data Quality
 Any airline, ground service provider, and / or airport which provide
ground services are eligible to participate in the program
 Variance in data received
 Data integrity the utmost of importance
 Confidence in analysis and decisions derived from
it is equal to the confidence in the data itself
19
26th IGHC Conference
Data Management Principles
 Information producers and knowledge workers alike must know the
meaning of information; otherwise they cannot perform their work
properly
 Information producers must also know the business rules, valid
values, and formats to create information correctly
 Information definition is to data (content) what manufacturing
product specifications are to the manufactured product
 Quality “Information Product Specifications” are necessary for the
consistent production of quality information
20
26th IGHC Conference
Data Management Principles
 Managing by averages leads to flawed decision making as you are
not accounting for process variation
 If measurement system variation is too large there is an increased
risk of:
 Rejecting good data
 Accepting bad data
 Important to know how much of the observed variation of a process
is due to the actual process itself (normal) and how much is due to
the measurement system
21
26th IGHC Conference
Data Management Principles
 Operational definitions (includes taxonomies) help reduce
subjectivity and variance in your measurement system (data)
 Operational definitions can be:
• A written statement
• Templates
• Display of comparisons (colour chart)
 Operational definitions should be:
• Something people can really use
• Enables different people to reach the same conclusion (repeatability)
• Enables the same person to reach the same correct conclusion at
different times (reproducibility)
22
26th IGHC Conference
Taxonomy / Operational Definitions




Controls data inputs
Reduce subjectivity
Reduce variation
Means for integration (internal and external)
23
26th IGHC Conference
Solution ~ Defined Fields
 GDDB TF developed reporting criteria
•
Representatives from Operators, GSP’s, Manufacturers and Industry groups
 Identified data to be consistently reported amongst ALL members
•
•
Includes definitions / assumptions
Minimize data variance
 Identify means in which data and analysis will feed ground operations
working groups and vice versa
24
26th IGHC Conference
Ground Damage Database
Reporting Criteria
26th IGHC Conference
Definition
“Any occurrence / event associated with ground operations that
results in aircraft damage”
Out of Scope
In Scope
 While parked at Gate / Stand
or other parked area
 During Marshaling or using
Stand Guidance
 During Deicing
 While being Towed
 Near Miss
 Slide Deployments




26
FOD
Wildlife Damage
Lightning Strikes
Environmental
26th IGHC Conference
Field Categories
 Mandatory
•
•
•
•
•
•
•
•
•
•
 Optional
•
•
•
Incident details
Location details
Aircraft details
Ramp conditions
Phase of operation ~ definitions
included
Activities
Type of damage
Damage to aircraft
Ground equipment
Severity ~ definitions included
27
Causal factors
Corrective actions
Free text
26th IGHC Conference
Field Categories
 Mandatory
•
•
•
•
•
•
•
•
•
•
Incident details
Location details
Aircraft details
Ramp conditions
Phase of operation
Activities
Type of damage
Damage to aircraft
Ground equipment
Severity
28
26th IGHC Conference
Field Categories
Incident
Details
Type of
Report
Damage by
3rd Party
Date
Reporting
Station
Time
Previous
Station
Stand / Gate
Type of
Stand
Found
Yes
Cargo
Reported
No
Passenger
Terminal
Passenger
Remote
N/A
29
26th IGHC Conference
Field Categories
Location
Of Incident
Apron
Manoeuvring
Area
De-Icing
Pad
Taxiway
30
Stand / Gate
26th IGHC Conference
Aircraft
Details
Field Categories
Flight
Number*
Aircraft
Registration*
Date of
Flight*
Aircraft
Model
ETA
ATA
ETD
ATD
Manufacturer
Model
Series
Configuration
Freighter
Passenger
31
26th IGHC Conference
Ramp
Conditions
Field Categories
Weather
Surface
Conditions
Phase of
Day
Clear
Dry
Dawn
Rain
Wet
Day
Fog
Ice
Dusk
Freezing
Rain
Snow
Night
Hail
Contaminated
Snow
Sand / Dust
Storm
Wind
Electrical
Storm
32
26th IGHC Conference
Field Categories
Arrival
Phase of
Operation /
Activities
Towing
Servicing
Departure
 Arrival ~ Time period from when the aircraft nose wheel crosses onto the stand until
the anti-collision light is off
 Towing ~ Time period when an aircraft is being towed from one location to another
 Servicing ~ Time period an aircraft is being serviced at a gate / stand
 Departure ~ Time period when the anti-collision light is turned on and the brakes
are off until control is handed over to Flight Operations
33
26th IGHC Conference
Phase of
Operation /
Activities
Field Categories
Arrival
Towing
Servicing
Marshalling
Stand to Stand
Deicing
Guidance
To
Maintenance
Fueling
From
Maintenance
Catering
All Activities broken
down into three periods:
1. Positioning
2. Performing
3. Removal
Departure
Pushback
Loading
Passenger
Baggage
Cargo
Etc
34
26th IGHC Conference
Field Categories
Type of
Damage
Scratch /
Dent / Scuff
Puncture
Tear or
Crack
Delamination
35
Detached
Slide
Deployment
26th IGHC Conference
Area of
Damage
Field Categories
Engine
Cargo Hold
Fuselage
Lights
Doors
Landing
Gear
Wings
Engine
Intake
Fwd Cargo
Hold
Radar Nose
Cone
Cabin Door
Wing Surface
Main
Undercarriage
Door
Engine
Exhaust
Aft Cargo
Hold
Pitot Head /
Probe
Cabin Door
Surround /
Frame
Wing
Leading Edge
Undercarriage /
Wheels /
Tires
Engine
Cowling
Bulk Cargo
Hold
Antenna
Cabin Door
Seal
Wing
Trailing Edge
Nose
Undercarriage
Door
Propeller
Loading
System
Skin /
Structure
Cargo Door
Wing Tip
Nose Gear /
Wheels /
Tires
Locks / Stops /
Guides
Panels /
Hatches
Cargo Door
Surround /
Frame
Flaps
Tail Wheel
Panel /
Side Walls
Tail
Cargo Door
Seal
Aileron
Scuff Plates
36
26th IGHC Conference
Field Categories
Ground
Equipment
Pax Stairs /
Steps
Lav Truck
Belt Loader
Pushback
Tractor
ULD
Loading
Bridge
Catering
Vehicle
Etc.
Conventional
Pallet
Towbarless
37
Container
26th IGHC Conference
Field Categories





Minor ~ No operational effect
Low ~ Aircraft inop < 60 mins
Moderate ~ Aircraft inop ≥60 minutes <24 hours
High ~ Aircraft inop ≥ 24 hours
Catastrophic ~ Hull loss
38
26th IGHC Conference
Field Categories
Time Out of
Service
 Hours out of service ~ rounded up to the nearest hour
Note: this field is for calculation purposes only, and should not be confused with
severity
39
26th IGHC Conference
Field Categories
 Optional
•
•
•
Causal factors
Corrective actions
Free text
40
26th IGHC Conference
Causal Factors
Field Categories
Behaviour
Organizational
Equipment
Regs / SOPs
Not Followed
Physical
Cirumstances
Excess Speed
Lack of SOPs
Defective
Equipment
SOP
Weather
Conditions
Spatial
Misjudgment
Inadequate
Time...
Incorrect Use
Safety
Regulations
Surface
Conditions
Distraction
Inadequate
Supervision
Unsuitable
for Task
Traffic
Regulations
Inadequate
Lighting
Incapacitation
Insufficient
Personnel
Unsafe
for Task
PPE
Noise
Etc.
Inadequate
Training
Design
Problem
Validity of
License
Congestion
Left in Wrong
Place
Etc.
Incorrect ULD /
Pallet Contour
41
26th IGHC Conference
Ground Damage Database
Statistical Analysis
26th IGHC Conference
Statistical Analysis
 Summarizes data
 Measures the quality of the data
 Assess the degree of variation
•
•
Critical to understanding the process performance
Measure process performance
 Calculate separate descriptive statistics for each group, allowing to
see how the groups differ
• Identify “critical x’s”
•
Predict process performance with improvements
43
26th IGHC Conference
Sample ~ Box Plot
 Quickly compare distributions
 Highlights the variability of the
data
 Displays data from different
categories
 Can compare several groups
of data at once
44
26th IGHC Conference
Sample ~ Control Charts
 Displays the control of a process
•
•
In control process shows random variation
Out of control process shows unusual variation due to special causes
 Help to determine where to focus problem-solving efforts by
distinguishing between common and special-cause variation
45
26th IGHC Conference
Sample ~ Pareto Charts
 Displays defects from largest to
smallest
 Prioritize issues and focus
improvement efforts on areas
where largest gains can be
made
 Separates the "vital few"
problems from the "trivial many”
46
26th IGHC Conference
Sample ~ Probability Charts
 Measure process
improvements
 If distributions are normal can
estimate the performance if
new procedures are put in
place
47
26th IGHC Conference
Hypothesis Test ~ P-Value
Help us determine whether observed differences are:
Statistically significant
Or
Due to chance
(random variation)
48
26th IGHC Conference
Hypothesis Test ~ P-Value
 Null Hypothesis = No statistical difference
•
•
Things are “on Target”, “same”, difference is due to “random variation”
Not likely to be a critical x or may require more data
 Alternative Hypothesis = statistical difference
•
•
Things are NOT “on target”, “the same”, due to “random variation”
Data supports this x as a likely cause for further investigation
49
26th IGHC Conference
Hypothesis Test ~ P-Value
 P-Value represents the risk that we are wrong if we conclude that the
null hypothesis is false
•
That we claim there is a difference and there isn’t one
 Numerically
•
•
P-Value < 0.05 we can conclude that we found a statistical difference
We may say we have a 5% chance of being wrong when we conclude that
something is “off-target”
50
26th IGHC Conference
Hypothesis Test ~ P-Value
We have found a statistical difference if the
p-value is < 0.05
51
26th IGHC Conference
Initial Analysis ~ Some Comparisons
 Type of Report (found vs. reported) and Severity
Correlations: Num report, Num Sev
Pearson correlation of Num report and Num Sev = -0.334
P-Value = 0.000
Statistically Significant
52
26th IGHC Conference
Initial Analysis ~ Some Comparisons
 Type of Report (found vs. reported) and Organization
Correlations: Num Org, Num report
Pearson correlation of Num Org and Num report = 0.124
P-Value = 0.040
Statistically Significant
53
26th IGHC Conference
Initial Analysis ~ Some Comparisons
 Month of Report and Severity
Correlations: Mnth, Num Sev
Pearson correlation of Mnth and Num Sev = 0.065
P-Value = 0.292
Not Statistically Significant
54
26th IGHC Conference
Initial Analysis ~ Some Comparisons
 Type of Report (found vs. reported) and Reporting Region
Correlations: Num report, Num Region
Pearson correlation of Num report and Num Region = - 0.232
P-Value = 0.000
Statistically Significant
55
26th IGHC Conference
Initial Analysis ~ Some Comparisons
 Reporting Region and Severity
Correlations: Num Region, Num Sev
Pearson correlation of Num Region and Num Sev = 0.221
P-Value = 0.000
Statistically Significant
56
26th IGHC Conference
Initial Analysis ~ Some Comparisons





Type of Report (found vs. reported) and Severity
Type of Report (found vs. reported) and Organization
Month of Report and Severity
Type of Report (found vs. reported) and Reporting Region
Reporting Region and Severity
57
26th IGHC Conference
Initial Analysis ~ Comparison of Means
 Severity and Organization
•
Does the organization affect the severity of reports
One-way ANOVA: Num Sev versus Num Org
Source DF
SS MS
F
P
Num Org 8 68.510 8.564 18.30 0.000
Error 256 119.830 0.468
Total 264 188.340
At least one
organization’s
mean is different
S = 0.6842 R-Sq = 36.38% R-Sq(adj) = 34.39%
58
26th IGHC Conference
Initial Analysis ~ Comparison of Means
 Severity and Organization
•
Does the organization affect the severity of reports
One-way ANOVA: Num Sev versus Num Org
Source DF
SS MS
F
P
Num Org 8 68.510 8.564 18.30 0.000
Error 256 119.830 0.468
Total 264 188.340
At least one
organization’s
mean is different
But how much
variation does
this explain?
S = 0.6842 R-Sq = 36.38% R-Sq(adj) = 34.39%
59
26th IGHC Conference
Initial Analysis ~ Comparison of Means
 Severity and Organization
•
Does the organization affect the severity of reports
One-way ANOVA: Num Sev versus Num Org
Source DF
SS MS
F
P
Num Org 8 68.510 8.564 18.30 0.000
Error 256 119.830 0.468
Total 264 188.340
S = 0.6842 R-Sq = 36.38%
But how much
variation does
this explain?
36%
R-Sq(adj) = 34.39%
60
26th IGHC Conference
Initial Analysis ~ Comparison of Means
61
26th IGHC Conference
Statistical Analysis ~ In Summary






Statistical analysis produces more tangible information
Measures process performance
Identify and prioritize contributing factors to process performance
Measure and predict process performance improvements
Provides confidence interval
Measures the quality of the data
» Need trained resources
» Need consistent data
» Dependent on controlled inputs
62
26th IGHC Conference
Statistical Analysis ~ In Summary
 Must know restrictions of data
•
•
•
Sample size
MSA
Practical importance
 GDDB data
•
•
•
Small sample size
Much variance in data but significantly better
Practical importance will be vetted by WGs and TFs
63
26th IGHC Conference
Ground Damage Database
Sample Observations to Date
26th IGHC Conference
Observations ~ Type of Report
43%
57%
Source: IATA Q4 2012 GDDB
65
26th IGHC Conference
Observations ~ Type of Report by Severity
Source: IATA Q4 2012 GDDB
66
26th IGHC Conference
Observations ~ Areas of Damage
Engine
7
Fuselage
46
Wings
13
Lights
0
Doors
81
Cargo Hold
58
Landing Gear
9
67
Source: IATA Q4 2012 GDDB
26th IGHC Conference
Observations ~ High Severity Reports
Engine
0
Fuselage
3
13 High events reported
Note: 1 report indicated damage
to both Fuselage and Wings
Wings
3
Lights
0
Doors
4
Cargo Hold
1
Landing Gear
3
69
Source: IATA Q4 2012 GDDB
26th IGHC Conference
Observations ~ High Severity Reports
Location of Damage Type of Damage
Door
Tear / Crack
Scratch / Dent / Scuff
Tear / Crack
Scratch / Dent / Scuff
Equipment
Food Truck
Bridge
Unknown
Cargo Loader
Wing
Scratch / Dent / Scuff
Puncture
* Scratch / Dent / Scuff
Pax Stairs
Catering Truck
Pax Stairs
Fuselage
Scratch / Dent / Scuff
Tear / Crack
* Scratch / Dent / Scuff
Unknown
Pushback Tractor ~ Conventional
Pax Stairs
Cargo Hold
Tear / Crack
Main Loader
Landing Gear
Scratch / Dent / Scuff
Tear / Crack
Puncture
Pushback Tractor ~ Unknown
Pushback Tractor ~ Towbarless
Pushback Tractor ~ Unknown
70
Source: IATA Q4 2012 GDDB
26th IGHC Conference
Ground Damage Database
Ongoing Development
26th IGHC Conference
GDDB Development
 Numerous requests to expand the database to include entire scope of
AHM and ISAGO
 TF meetings planned for May (DOH) and Sep (YUL)
•
•
•
Establishment of a costing model
Identify reporting challenges / solutions
Development of a phased plan to expand scope
72
26th IGHC Conference
Strategy ~ Continuous Improvement Circle
Regulatory
Environment and
Framework
• Standards
• Polices
Operational
Procedures
• Procedures
• Job cards
Framework
Implementation
Database
Verification
Hazard
Identification
• Data analysis
• Feedback
Verification
• Audit
73
26th IGHC Conference
Strategy ~ Continuous Improvement Circle
Regulatory
Environment and
Framework
• Standards
• Polices
Operational
Procedures
• Procedures
• Job cards
AHM
IGOM
GDDB
ISAGO
Hazard
Identification
• Data analysis
• Feedback
Verification
• Audit
74
26th IGHC Conference
GSIC
ODM
GADM
75
26th IGHC Conference
Participant Support
 IT development
•
•
•
•
•
Automated quality check
Web Form
Discussions with strategic partners to develop GDDB submission extract from
existing reporting systems
Common dimension tables
Participant query tool
76
26th IGHC Conference
77
26th IGHC Conference
Ground Damage Database
Thank You
26th IGHC Conference