You are on page 1of 17

Introduction to Broken Rails

• Broken rails are caused by inherent flaws (internal


defects) in the rail and/or by fatigue on the surface of
the rail.
Broken Rail Risk Analysis
– A broken rail derailment is the occurrence of a train
accident due to the rail failure
Xiang Liu, Ph.D.
Assistant
– A service failureProfessor in Railway
is a detected Engineering
broken rail event without the
Department
occurrence of aoftrain
Civilderailment.
and Environmental Engineering
Rutgers, The State University of New Jersey
• Usually detected by the signal system
Number of cars derailed by accident cause, Class I mainline
FRA-reportable freight-train derailments, 2001 to 2010
Broken
Number of Cars

Rails or
Welds

Percentage
Cumulative
Track Geometry Defects
Buckled Track

Accident Cause
Reference: Liu, X., Saat, M.R., Barkan, C.P.L. (2013). Safety effectiveness of integrated risk reduction strategies for rail transport
of hazardous materials. Transportation Research Record 2374, 102-110.
Frequency and severity of Class I mainline
freight-train derailments: 2001 to 2010
25 Average frequency = 89
Average Number of Cars Derailed

Other Rail and Joint Defects


20
Rail Defects at Broken Rails
Bolted Joints
or Welds
15 Joint Bar
Defects Broken Rails or Welds
Buckled Track

Obstructions

Wide Gauge
10
Mainline Brake Operation Average severity = 8.6
Train Handling (excl. Brakes)
Bearing Failure (Car)

Broken Wheels Track Geometry


5 (Car) (excl. Wide Gauge)

0
0 100 200 300 400 500 600 700

Reference: Liu, X., Saat, M.R., Barkan, C.P.L. (2012). Analysis of causes of
Number of Derailments
major train derailment and their effect on accident rates. Transportation
Research Record 2289, 154-163.
Broken rail causes
•  Transverse/compound fissure (FRA cause code T220), detail
fracture (T207) and vertical split head (T212) collectively
caused over 70% of FRA-reportable broken-rail-caused freight-
train derailments on Class I mainlines from 2001 to 2010
•  These rail defects may be identified and removed during
ultrasonic rail defect inspection
Transverse fissure Compound fissure Detail fracture
Ultrasonic rail defect inspection
•  Ultrasonic technology is used to identify rail defects
•  The undetected rail defects may lead to broken rails
•  A small proportion of broken rails may cause derailments
(1 derailment per 100 to 200 broken rails)

Rail Defects Broken Rails Derailment


The scale of rail defects versus broken rails is for illustration
only. It may not fully represent real-word conditions

Undetected Undetected

Identified by
Detected by track inspection
Reference: Davis, D.D., Joerms, M.J., rail inspection
Orringer, O., Steele, R.K. (1987). The or track circuits
economic consequences of rail integrity.
Report No. R-656, Association of
American Railroads, Chicago, IL.
Rail defect formation
•  The model for rail defect formation is based on
field data from the Test Track by the Association
of American Railroads (Besunner et al. 1978)
Test Center in Pueblo, Colorado and on several se
revenue track studied by the Association of
Railroads [4]. Moreover, theRaildata can be characteri
age (cumulative
Cumulative fraction of rails that have
Weibull
developed a defect distribution
by rail age T [5]: tonnage on the rail)

3
T
F (T ) 1 exp

References: where T is the rail age in cumulative MGT, is


characteristic life, and F(T) is the cumulative fracti
Besunner, P.M., Stone, D.H., DeHerrera, M.A., Schoeneberg, K.W. (1978). AAR Chicago Technical Center.
simulation model, defect growth per MGT is projected from a
seasonally adjusted rate,Rail defect
based on one of growth
the curves shown in
thedefect
The figure, for each
growth calendaronmonth.
rate depends several factors such as rail properties,
axle load, weather, and other service conditions (Jeong et al. 2009)
100

80
Defect Size (%HA)

60

40

20

0
0 20 40 60 80 100

Tonnage (MGT)
Jeong, D.Y., Gordon, G.E. (2009).Evaluation of rail test
frequencies using risk analysis. Joint Rail Conference
identify the defects missed by the
Detection of rail defect
rail samples containing the defec
true sizes.
•  Detection performance depends Under these
on the equipment, the circum
location and size of a defectapproach is to infer p(s), via a tria
•  Although larger defects areavailable
more likelystatistics for overall
to be detected, they sy
still can be missed during the inspection
prior research, process
national statistics
curve corresponding to older rail
RESULTS
In this paper, the
derived curve is given byis[3]
detection used to st
axle load on the occurr
concept of high-speed
s 5 rail testing s
alternative
p ( s ) 1 exp Results from the
14 defect rate as
service
Here the defect rate is t
(Jeong et al.year.
2009) Moreover, risk (o
where s is greaterthethanincreaseor(orequal
decreas
minimum detectableThesize). duration ofTh
th
number of years varied
detection performancedepending ofonultrason
the parti
single 70º sensor perlevel probe
of safetywheel
in terms
Removal of rail defects
•  Detected rail defects are removed or remediated,
according to the current FRA regulations

•  Alternatively, certain “non-critical” small defects might


be left for repair within a grace period (aka. risk-
based repair) (Orringer et al. 1990)

Orringer, O., Tang, Y.H., Jeong, D.Y., Perlman, A.B. (1999). Risk/benefit assessment
of delayed action concept for rail inspection. Volpe Center Final Report.
Simulator of rail defect
occurrence, growth, detection and repair
•  A simulator is being developed that Preliminary results for illustration only
incorporates rail defect occurrence,
growth, detection and repair
processes
•  Sample inputs include
–  Route mileage
–  Rail age
–  Annual inspection frequency
–  Rail type
–  Threshold for delayed repair
–  Grace period for delayed repair
•  Sample outputs include
–  Detected rail defect rate
–  Broken rail rate
–  Detection car utilization rate
(mile/day)
Broken-rail-caused number of railcars
derailed per freight-train derailment
Freight-train derailment severity distribution by major
accident cause,
X. Liu U.S.Analysis
et al. / Accident Class andIPrevention
mainlines, 2001–2010
59 (2013) 87–93

Broken Rails or Welds Bearing Failures

120 100%
(a) (b)
120 100%

Cumulative Percentage
96 80% 96 80%
Sample size: 197
Frequency

Sample size: 458


Mean: 7.0
72 Mean: 14.2 60% 72 60%
Standard deviation: 9.6
Standard deviation: 9.5
25% quantile 1
25% quantile: 7
50% quantile (median): 1
48 50% quantile (median): 12 40% 48 40%
75% quantile: 10
75% quantile: 19

24 20% 24 20%

0 0% 0 0%
1 5 9 13 17 21 25 29 33 37 41 45 49 1 5 9 13 17 21 25 29 33 37 41 45 49

Number of Cars Derailed Number of Cars Derailed

Fig. 2. Freight-train derailment severity distribution by major accident cause, U.S. Class I mainlines, 2001–2010.
Factors affecting derailment severity
•  Accident cause
(Saccomanno et al. 1989, 1991; Liu et al. 2013)
•  Accident speed
(Nayak et al. 1983; Liu et al. 2013)
•  Train length
(Saccomanno et al. 1989; Bagheri et al. 2011; Liu et al. 2013)
•  Point of derailment (the position of the first car derailed)
(Saccomanno et al. 1989, 1991; Liu et al. 2013)
•  Proportion of loaded cars
(Liu et al. 2013)
Statistical modeling of number of cars derailed

•  Modeling the mean number of cars derailed


–  Zero-truncated negative binomial (Liu et al. 2013)

•  Modeling the 25 percentile, median, 75 percentile or 95


percentile
–  Quantile regression (Liu et al. 2013)

Liu, X., Saat, M.R., Qin, X., Barkan, C.P.L. (2013). Analysis of U.S. freight-
train derailment severity using zero-truncated negative binomial regression
and quantile regression. Accident Analysis and Prevention
Broken-rail-caused
crude oil train accident risk
Number of hazmat cars releasing in freight-train derailments on
U.S. railroads by accident cause from 2002 to 2011
Number of Hazmat Cars Releasing
0 10 20 30 40 50 60

Broken Rail - Transverse/compound fissure 56

Broken Rail - Head and web separation 37

Other rail and joint bar defects 25

Broken Rail - Weld (field) 20

Journal (roller bearing) failure from overheating 16

Broken Rail - Detail fracture 16

Washout/rain/slide/flood/snow/ice damage to track 15

Joint bar broken (noninsulated) 13

Broken rim 9

Cause under active investigation 9

Liu, X., Dick. C.T. (2016). Risk-Based Optimization of Rail Defect Inspection Frequency for
Petroleum Crude Oil Transportation. Transportation Research Record
Ongoing research
•  The final deliverable will be a
simulation-based rail defect engineering risk
management framework that can address questions
such as:
–  How to schedule ultrasonic rail tests?
–  How to arrange the repair or replacement of
defective rails?
–  How to reduce crude oil transportation risk by
improving rail conditions?
–  How to evaluate new technologies (e.g., advanced
rail testing methods or new rail design)?
Acknowledgement
•  Dr. David Jeong (Volpe Center of US Department of
Transportation)

•  This research was funded by Rutgers University

16
Thanks very much for your attention!
Your questions are welcome and can be directed to

Xiang Liu, Ph.D.


Assistant Professor in Railway Engineering
Department of Civil and Environmental Engineering
Rutgers, The State University of New Jersey

Email: xiang.liu@rutgers.edu
Phone: (848)-445-2868

You might also like