Concrete Tie Degradation Study -...
Transcript of Concrete Tie Degradation Study -...
Federal Railroad Administration Office of Research and Development
Concrete Tie Degradation Study Tie Condition and Crack Growth Rate Assessment
Final Results
Funded By Federal Railroad Administration
Office of Research and Development Under Broad Agency Announcement BAA-2010-1
June 5, 2014 Presented by: Jeff Henderson
ENSCO, Inc.
Federal Railroad Administration Office of Research and Development
Study Overview • Main Objective
Using machine vision technology, develop a better understanding of concrete tie degradation
• Project Team – FRA Office of Research and
Development – AMTRAK – ENSCO
Sponsored by the Federal Railroad Administration under BAA-2010-1
Federal Railroad Administration Office of Research and Development
Technical Approach Overview
1) Image 100 miles of track 3 times over ~1-year 2) Use software to manually align track bed images 3) Grade ties and determine crack growth rates 4) Assess concrete tie degradation over time
Federal Railroad Administration Office of Research and Development
Technology Used Hi-rail-based Machine Vision Imaging System
• High Data Rate Line Scan Machine Vision Cameras • High Data Rate Solid State Hard Drives • High-intensity LED Lighting
Federal Railroad Administration Office of Research and Development
Track Number
Mile Post Start
Mile Post End
Total Miles
1 187 158 29 1 218 189 29 2 143 158 15 2 187 159 28 2 189 218 29
Total Number of Miles 130
Test Zone Location (Amtrak’s AB Line, Northeast Corridor)
Federal Railroad Administration Office of Research and Development
Summary of Test Schedule • Survey 1 – August 16 to 19, 2012 • Survey 2 – April 15 to 19, 2013 • Survey 3 –September 23 to 25, 2013
Federal Railroad Administration Office of Research and Development
Concrete Tie Degradation Study Summary of Analysis Approaches
Subjective Analysis • Assess general tie conditions • Evaluate tie degradation based on tie grades
• Grading scale: 1-good to 5-ineffective • Assess a large population of ties ~26,000
Objective Analysis • Use machine vision images to measure crack area • Derive growth rate from measured crack area • Conduct parametric analysis of crack growth rates
• From track charts: Speed, Grade, Curvature, • From other sources: Manufacturer, Crack Size, and
Crack Location
Federal Railroad Administration Office of Research and Development
Tie Condition Analysis Types of Conditions Observed
Cracking
Crumbling Erosion Missing Fastener
Chipping
Federal Railroad Administration Office of Research and Development
Tie Grading Approach Use software to visually review and grade tie images
Grade 1 – Good Tie
Grade 3
Grade 2
Grade 4 Grade 5 – Ineffective Tie
Federal Railroad Administration Office of Research and Development
Crack Growth Rate Determination Use Software to Measure Crack Growth Rate
Unannotated Crack
Survey 1 Annotation
Survey 3 Annotation
Δ Area = 7 mm2
Area 1 = 928 mm2 Area 2 = 935 mm2
Growth Rate = 6.5 mm2 /yr
Annualize the result:
Federal Railroad Administration Office of Research and Development
Parametric Analysis Approach Use Statistics to Assess Growth Rates Across
Multiple Crack Populations
0%
20%
40%
60%
80%
100%
120%
0.00
0.03
0.05
0.08
0.10
0.13
0.15
3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 78 81 84 87 90 93 96 99 More
Cum
ulat
ive
Freq
uenc
y
Nor
mal
ized
Freq
uenc
y
Annualized Growth Rate (mm^2/yr)
Frequency distributioinCumulative percent curve
• Cumulative percentage is the integral (running total) of normalized frequency • Combines crack growth rate within a population of cracks into a single curve • Provides a means of quantifying crack growth rate over a population of cracks
Federal Railroad Administration Office of Research and Development
Tie Condition Results Distribution of Tie Grades
Grade 1
Grade 2
Grade 3
Grade 4
Grade 5
Grade 1 – 97%Grade 2 – 2%
Grade 3 – 0.8%
Grade 4 – 0.2%
Grade 5 – 0.2%
•Distribution based on ~26,000 ties •Typical tie age 33 years
12
Federal Railroad Administration Office of Research and Development 13
Tie Condition Results Distribution of Condition Types
46%
44%
4%6%
CrackChipMissing FastenerCrumbling
Federal Railroad Administration Office of Research and Development
Tie Condition Results Percent Change in Number of Tie Grades Over One Year
• Most ties did not advance in grade
• Most ties that did advance in grade did so by 1 grade level
• Relatively few ties advanced by 2 grade levels
• No ties advanced by more than 2 grade levels
-0.23%
0.22% 0.01% 0.00% 0.00%
0.0%
-6.5%
6.2% 0.3% 0.0%
0.0% 0.0%
-3.8%
3.0% 0.8%
0.0% 0.0% 0.0%
-5.8%
5.8%
Grade 1 Grade 2 Grade 3 Grade 4 Grade 5
Gra
de 1
Gra
de 2
Gra
de 3
Gra
de 4
55 Ties 3 Ties
32 Ties 2 Ties
6 Ties 2 Ties
3 Ties
58 Ties
34 Ties
8 Ties
Federal Railroad Administration Office of Research and Development
Crack Growth Rates for Total Assessed Crack Population
50%
60%
70%
80%
90%
100%
0102030405060708090
100110
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 More
Cum
ulat
ive
Perc
enta
ge
Freq
uenc
y
Annualized Crack Growth Rate ( mm^2/year)
1386
Measured Growth Rate
Half Normal Distribution
Cauchy Distribution
Cumulative Percentage
• Number of cracks assessed: 2139 • Number that did not grow: 1386 • Number that grew: 748 • Median growth rate: 6.5 mm2/year
Federal Railroad Administration Office of Research and Development
Parametric Analysis Results Parameters that Influenced Crack Growth Rate
Parameter Growth Rate of a Typical Crack (mm2/Yr)
Growth Rate Spread (mm2/Yr)
2-Sigma Uncertainty (mm2/Yr)
Residual (mm2/Yr)
Combined Factors Worst case Best case
13 3.8
9.2
2.4
6.8
Crack Size Small crack Medium crack Large crack
4
8.3 7.9
4.3
2.0
2.3
Field VS Gauge Gauge side Field side
8.9 5
3.9
2.1
1.8
Curvature Non-tangent track Tangent track
8.9 5.3
3.6
2.1
1.5
Tie Manufacturer Manufacturer “B” Manufacturer “A”
8.1 5.9
2.2
2.1
0.1
Federal Railroad Administration Office of Research and Development
Parametric Analysis Results Parameters that did not Influence Crack Growth Rate
Parameter Growth Rate of a Typical Crack
(mm2/Yr)
Growth Rate Spread
(mm2/Yr)
2-Sigma Uncertainty
(mm2/Yr)
Residual (mm2/Yr)
Posted Speed High range (130 - 150 mph) Low range (30 - 125 mph)
6.5 6
0.5
2.0
-1.5
Grade Neutral Grade (|Grade|≤ .116) Non-neutral Grade (|Grade|> .116)
6.4 6.2
0.2
1.9
-1.7
Federal Railroad Administration Office of Research and Development
Impact of Measurement Uncertainty
0.00
0.20
0.40
0.60
0.80
1.00
1.20
0 1 2 3 4 5 6
Stan
dard
Dev
iatio
n of
Med
ian
RSS Measurement Uncertainty
Federal Railroad Administration Office of Research and Development
Tie Manufacturer Comparison
0.0%
0.2%
0.4%
0.6%
0.8%
1.0%
1.2%
1.4%
1.6%
1.8%
2.0%
2.2%
-52-48-44-40-36-32-28-24-20-16-12-8 -4 0 4 8 12 16 20 24 28 32 36 40 44 48 52
Perc
enta
ge o
f Cra
cked
Tie
s
Approximate Distance from Tie Ceter (inches)
Manufacturer "A"
Manufacturer "B"
Federal Railroad Administration Office of Research and Development
Conclusions • The observed crack growth rates were not alarming (median
growth rate = 6.5 mm2/year • Tie grades progressed at a non-alarming rate • Curvature, crack size, crack location (field or gauge side of the
tie), and tie manufacturer had a statistically significant impact on crack growth rates
• Posted speed (track class) and track grade (the degree of incline) did not have a statistically significant impact on crack growth rates
• Tie manufacturer and/or manufacturing lot appears to play an important role in crack prevalence
Federal Railroad Administration Office of Research and Development
Remaining Unknowns • The progression rate of failure modes other than cracks (e.g.,
crumbling and chipping) • The impact of tonnage on crack growth rates • The impact of seasonal variations on crack growth rates • The impact of different geographical regions on tie
degradation • Potential variations among additional tie manufacturers
Federal Railroad Administration Office of Research and Development
Acknowledgements Federal Railroad Administration Office of Research and Development • Mr. Hugh Thompson – FRA Project Manager • Mr. Cameron Stuart – BAA Program Manager
National Railroad Passenger Corporation (Amtrak) • Mr. Mike Trosino • Mr. Steven Sawadasavi • Mr. Joe Mascara Computer Vision Laboratory at the University of Maryland, College Park • Professor Rama Chellapa • Mr. Xavier Gibert-Serra