Track Buckling Hazard Detection and Rail Stress Management

Challenge C: Increasing Freight capacity and services
Track Buckling Hazard Detection and Rail Stress Management
Andrew Kish, Ph.D.
President- Kandrew Inc. Consulting Services
Peabody, MA, USA
Phone: 978-535-3342
Email: kandrewinc@aol.com
Ryan S. McWilliams
Vice President-Technology & Business Development
Salient Systems, Dublin, OH, USA
Phone: 719-337-5852
Email: rmcwilliams@salientsystems.com
Harold Harrison
Salient Systems - Technology Consultant
Friday Harbor, WA, USA
Phone: 360-317-8505
Email: h_sqd@salientsystems.com
ABSTRACT
A key aspect of managing the stress state of the railroad involves controlling longitudinal rail stresses to
safe levels. Rail networks are therefore focused on reducing the risk of both buckled track and broken rail
for safety and security enhancement. The efficient management of longitudinal stress induced thermal
forces in continuous welded rail (CWR) is an important aspect of railroad safety and maintenance.
Balancing such forces is a complex issue which requires intimate knowledge of many parameters which
include track strength as well as environmental operating conditions.
CWR track is prone to lateral buckling when the rails are under high compressive forces. Conversely,
when high tensile forces dominate under cold temperature conditions, rail defect growth can cause rail
fractures, weld failures and rail joints to pull apart. The magnitude of longitudinal thermal forces is
governed by the rail neutral temperature (RNT) which is the rail temperature at which the net force is
zero. It is well known that RNT is a highly variable parameter and that its measurement has proven to be
both difficult and costly.
This paper presents evolving and proven measurement technologies as well as new safety tools for
longitudinal force management. The safety aspects include track buckling prevention through hazard
identification diagnostics in terms of a “yellow”, “orange” and “red” buckling margin of safety (BMS)
warning, and rail break detection through sudden stress discontinuity (interruption) measurements. The
maintenance aspects include providing and applying data for rail defect/break repairs when readjusting
RNT to safe values in various operation conditions, and providing data for curve realignment and stability
improvement through more efficient RNT management. The applications are illustrated through case
study examples.
1.0 Introduction
Prevention of excessive longitudinal thermal forces in continuous welded rail (CWR) is an important track
maintenance and safety issue. Thermal forces are generated when the rail temperature varies from the
rail’s neutral temperature (RNT) i.e. the rail temperature at which the longitudinal force is zero, which is
typically the rail’s installation or fastening temperature. (Note: common terminology also uses SFT and T N
to designate RNT). Compressive forces are produced when the rail temperature is higher than the RNT,
and tensile forces are produced when the rail temperatures are below the RNT. High compressive forces
can cause track buckling, while high tensile forces can accelerate rail defect growth and cause rail joints
to fail or pull apart in cold weather, with both being potential derailment causes.
Longitudinal forces are controlled by installing and maintaining the rail at RNT levels within specified
limits. However, it is a well known fact that the RNT will vary over time and that it typically shifts to a lower
value that can increase the risk of track buckling. These lowered RNT values can result from rail repair or
1

Challenge C: Increasing Freight capacity and services
other maintenance work undertaken during cold weather and from excessive rail movement. A key
impediment to knowing these low and potentially dangerous RNTs is a lack of adequate measurement
capability which has eluded researchers for the past three decades. There have been several
measurement concepts/techniques proposed and evaluated over the years with mixed results. These
included techniques based on various physics concepts as illustrated in Figure 1.
RNT Measurement Concepts/Issues
Concepts Researched
• Mechanical/electrical
resistance strain gage
• Rail uplift
• Rail vibration
• Vibrating wire/filament
• Ultrasonic wave
• Acoustic wave
• Electromagnetic/acoustic
wave (EMAT)
• Magnetic permeability:
(Barkhausen noise)
(Magnetostriction)
• X-ray diffraction
Measurement Difficulties
• Accuracy Issues:
sensitivity to rail
microstructure, residual
stresses, track parameters;
hard to get force/RNT from
stress; “zeroes” required
• Don’t provide real-time,
continuous data output
• Not nondestructive and
not easily deployable
• No direct applicability
to railroad maintenance
and safety practices
Figure 1 – Summary of RNT Measurement Concepts
The key compromising aspects being (1) measurement accuracy, (2) ease of deployment, (3) providing
continuous data on a real time basis, and (4) having direct applicability to railroad maintenance and
safety. Although several concepts partially fulfill these needs, the US Salient System’s Rail Stress Monitor
(RSM) and its supportive StressNet TM data base system have evolved to be the most responsive to the
above needs. The RSM and its StressNet TM technology have been developed and tested over twenty-five
years of R&D, and have been recently expanded to support specific industry needs to not only measure
RNT but interpret it for practical application to enhance overall CWR safety and performance. More
specifically the RSM and StressNet TM technology:
Measures longitudinal force and RNT
Detects rail breaks and track buckles
Alerts on potential buckling hazards
Enables more effective rail break/defect repairs
Monitors rail joint condition
In this paper several of these applications will be discussed with a particular emphasis on track buckling
prevention through potential hazard identification diagnostics in terms of a “green”, “yellow”, “orange” or
“red” buckling hazard warning.
2.0 The RSM-StressNet TM System (The RSM Technology)
The rail-mounted hardware represents continuing evolving technology developments starting with a
stress measurement circuit first developed at Battelle Columbus Laboratories in the early eighties [1, 2,
2

Challenge C: Increasing Freight capacity and services
3], and culminating in today’s expedient and railroad friendly iPod/iPhone based data acquisition system.
The system is based on current radio technology and novel antenna designs that allow for a seamless
collection process in a hardened design capable of surviving normal freight railroad operations and
maintenance with a power management system that allows 10+ year useful life, and the ability to collect
the data with a variety of intermediate devices, including:
Hand held iPod/iPhone reader carried by a track inspector
End of Train (EOT) device
High rail inspection vehicle
Wayside reader
Figure 2 below shows the various uses and deployment options of the RSM-StressNet TM system.
RSM Deployment Options: When and Where?
• When installing new rail
• When making rail break/defect repairs
• When destressing “hot” rail
• In curves to manage lateral stability
• At locations where fastener conditions are substandard
• At locations of high stress build-up (approaches to
bridges, special track work, tunnels, etc.)
• In high tonnage lines to manage speed restrictions
• In high-speed passenger lines for RNT safety monitoring
• At joints for performance evaluation
3.0 RNT Monitoring and Rail Break Detection
Figure 2 – Deployment Options for the RSM
Figure 3 is an example of a StressNet TM output of 5 years of RSM data at one location of Union Pacific test
segment of newly laid CWR on a 2 degree (875 m radius) concrete tie track curve experiencing 230 MGTs
of annual tonnage where both rails were instrumented with RSMs. The data show the RNT variation with
time for the two rails (red for the low rail and green for the high rail) exhibiting the key features of: the overall
RNT behavior trends, the daily variations (the shaded data and tooth-like spikes), the detection and
influences of three rail breaks on the high rail (as shown by the sudden drop in RNT as denoted by RB1,
RB2 and RB3), and the influence of track maintenance involving surfacing and lining. To get a more detailed
look at a rail break scenario, Figure 4 shows an expanded view of RB1 and RB2 data taken by the RSM
closest to the breaks. Further inspection of Figure 4 shows that the low rail (red data) is relatively stable in
terms of RNT variation (ignoring the daily changes - tooth-like spikes). The high rail (green) data, however,
show the dramatic influences of rail breaks as indicated by the sudden discontinuity in RNT. The influence
of interim joint repairs is also evident in terms of large daily RNT shifts due to joint-gap movement with
temperature as shown by A and B. Finally, the green dashed arrow shows the effects of joint welding and
RNT restressing to the target RNT, and its effectiveness. Thus RNT knowledge and RNT monitoring during
rail repair and readjustment is highly essential in CWR repair and maintenance, hence to more effectively
managing the repair of broken and defective rail. Some inherent difficulties with these repair procedures are
briefly outlined below in context of rail break mechanics.
Daily RNT
Variation
2°(875m) Curve
Low Rail
50
Ne utra l Tempe ra ture (°F)
3
High Rail
Cell Phone Outage
&
RSM Refurbishment
40
30
Ne utra l Temperature (

Challenge C: Increasing Freight capacity and services
Figure 3 – Five Years of RSM Data on the Union Pacific
50
40
Neutral Temperature (°F)
RB1
A
RB2
B
30
20
Neutral Temperature (°C)
2°(875m) Curve
Low Rail
10
High Rail
0
Calendar Days
Figure 4 – RNT Behavior in Rail Breaks
4.0 Rail Break Mechanics, Repair and Restressing
A major reason for the reduced and possibly unsafe neutral temperatures is the difficulty in resetting the
RNT to the desired target value after a rail break or after a rail defect removal. The fundamentals of rail
break mechanics are discussed in [4, 5], while Figure 5 illustrates the basics. It is a well known that when
the rail is broken or cut for defect removal the RNT drops to the temperature of the rail when the break/cut
occurs. For example, in Figure 5 when the rail breaks at 40°F(4°C) its RNT is 40°F (4°C) at that location.
Additionally in this case, the rail break/cut influence is experienced for a track length exceeding 635 ft
(194m) on either side of the break/cut as indicated by L d . As shown in [4, 5], these L d ’s can be on the
4

Challenge C: Increasing Freight capacity and services
order of 200–1200 ft (60–365 m) depending on the tensile force in the rail and on the rail/fastener’s
longitudinal resistance.
Rail Break/Cut RNT Profile
T Break =40°F(4°C); SFT PreBreak =100°F (38°C); Gap=3in (7.6cm)
RNT(°F)
Pre-Break RNT = 100°F (38°C)
RNT Locked in After
Joint Closes
100
80
Add
3 in
Rail
60
T BR = 40°F
3 in
136#Rail
Timber Ties
40
20
RNT Profile
Just After
The Break/Cut
L d =635 ft (194m)
900 800
700
600
500
400
300
200
100
0
Distance from break (ft)
100 200 300 400 500 600 700 800 900
Figure 5 – Rail Break/Cut RNT Profiles
The resulting rail gaps require closure by “adding rail” through a rail-plug insertion. Note that as shown in
[4] closing the gap by pulling rail together is generally not an effective procedure since it might only
restore the RNT to what it was in the rail prior to the break/cut, which may be much lower than the
required target RNT. Adding rail through a plug insertion, however, does result in locking in the reduced
RNT as indicated by the green RNT profile. It is this reduced RNT profile that is required to be readjusted
before the onset of warm temperatures to avoid buckling prone conditions. The usual readjustment is
made by cutting rail out and unfastening a prescribed length of rail at warmer temperatures. As indicated
in [4, 5], the basic difficulty with this procedure is that the “correct” amount of rail to cut out and the
“correct” unfastening lengths required are not known because the reduced RNT profile requiring
adjustment is not known. The Salient RSM-StressNet TM System can be applied to provide this knowledge,
thus promoting a more effective management of rail break/defect removal repairs in terms of RNT
readjustment. Specifically, the RSM-StressNet TM System:
Determines and monitors the RNT after plug/joint interim repair
Alerts on when to return for the final repair/RNT readjustment to avoid a buckling prone condition
(see Buckling Hazard description in next Section)
Provides information on how to readjust to the desired RNT
Provides information on effectiveness of the RNT readjustment
One such application is shown in Figure 6 where the RSM data shows the specifics of rail break (RB3
shown in Figure 3) repair on the Union Pacific test site. The RSM measured RNT data clearly shows the
interim joint’s RNT of 60°F (16°C) to be readjusted. Without RSM/RNT measurement this would not be
known, hence the amount of rail to cut out and unfasten would also be not known. RSM data also shows
the effectiveness of the readjustment procedure and tracks the subsequent RNT condition.
UP South Morrill Sub Rail Break on 2/8/06 , Joint Repair and RNT
Readjustment As Monitored By Rail Stress Module (RSM)
5
RNT
Readjustment
50
40
Rail Break
30
emp (° F)
Joint Repair RNT
40°F (4°C)
Joint RNT
60°F (16°C)
20
RNT and

Challenge C: Increasing Freight capacity and services
Figure 6 – RNT Monitoring and Readjustment during a Rail Break and Repair
5.0 Track Buckling Hazard Detection
With the annual buckle derailments in the US ranging from 20 - 40 at a cost of over $10M per year, the
prevention of track buckle caused derailments remains a key industry goal. Railroad efforts are
continuing to focus on improving buckling prevention practices and on to better identifying buckling prone
conditions. To assist in these efforts Salient Systems has recently developed a “Buckling Hazard Index”
which provides a warning on potential buckles in terms of a yellow, orange, or red alert. The yelloworange-red
alerts are based on the track having a prescribed buckling margin of safety (BMS). The BMS
is determined by the StressNet TM system and is based on the RSM measured rail and neutral
temperature data coupled with the track’s buckling strength evaluations. The latter is based on the
Volpe/FRA “CWR-SAFE” model [6], a part of which has been incorporated into the StressNet TM system,
referred to as BUCKLE. The BMS is a numerical temperature index on how close the track is to a “critical”
condition in accordance with theoretical developments in [7, 8]. For completeness, a brief review of
buckling safety concepts and BMS are provided below.
5.1 Buckling Safety Concepts for BMS Applications
Figure 7 shows the fundamental safety concept and criterion for buckling prevention as detailed in [9], as
well as applied by the UIC in line with [10, 11].
T Bmax T Bmin
Temperature increase
above neutral
Buckle
Temperature increase
above neutral
Step 1 Step 2
Buckle
o
Deflection
o
Deflection
6
Temperature increase
above neutral
T Bmax T Bmin
Step 3
Buckle
Buckling
regime
Temperature increase
above neutral
T Bmax T Bmin
Step 4
Buckle
Buckling
regime
Safe temperature
increase

Challenge C: Increasing Freight capacity and services
Figure 7- Track Buckling Stability Response and Safe Temperature Concept
Step 1 shows the track lateral deflection from an initial line defect with amplitude o with temperature
increase. Step 2 exhibits the theoretically determined buckling equilibrium curve (shown in dashed)
identifying the upper and lower critical temperatures ΔT Bmax and ΔT Bmin . Step 3 shows that the actual
buckling occurs at a temperature in a buckling regime bounded by the upper and lower critical
temperatures, and Step 4 identifies the lowest point on the buckling response curve as the safe allowable
temperature increase. According to theoretical concepts [5], at temperatures below this value buckling
should not occur, and the track’s deflections are confined to the growth of the initial line defects with
temperature. The safety criterion on track buckling prevention is based on this safe temperature value i.e.
on the ΔT Bmin on the buckling response curve.
Buckling prevention safety criterion then requires that rail temperature increase above neutral to be
confined to values less than ΔT Bmin. Since ΔT Bmin is an analytically determined quantity based on theory
and on track condition/parameter inputs, for additional safety considerations it is customary to add a
safety factor on ΔT Bmin . For most heavy-haul applications this safety factor is typically 10°F (6°C) less
than ΔT Bmin [9]. For the European rail network’s safety factors refer to [10, 11]. It is important to note that
the safe temperature increase value, ΔT Bmin , and the buckling regime’s upper bound, ΔT Bmax , are highly
track parameter/type/condition dependent, and are referenced to the track’s neutral temperature. Hence
for buckling hazard detection both the safe allowable temperature (as modified by any additional safety
factors) and the track’s RNT are required in accordance with the safety criterion that: thermal load <
buckling strength, where the thermal load is based on the RNT, and the track strength is based on the
safe allowable temperature.
In line with the above, the buckling margin of safety (BMS) is defined as the reserve “buckling strength” in
terms of the additional temperature or thermal load required to produce a buckling prone condition i.e.
reaching the track’s safety limit. This is graphically illustrated in Figure 8.
T
T Bmax T Bmin
T ALL
T actual
Safety Factor
BMS = (T ALL – T actual )
Buckle Amplitude
7

Challenge C: Increasing Freight capacity and services
5.2 Buckling Hazard Index
Figure 8- Buckling Margin of Safety (BMS) Definition
Salient’s StressNet TM System has been recently upgraded to provide a buckling hazard warning based on
the RSM’s RNT output. The warning is based on a “Buckling Hazard Index” which provides a color
indexed alert in terms of a green, yellow, orange, or red conditions. The green-yellow-orange-red alerts
are based on having a prescribed buckling margin of safety (BMS) where the BMS is determined by the
StressNet TM . This determination is based on the RSM provided rail and neutral temperature data, and on
the track’s buckling strength as computed by BUCKLE. The BUCKLE algorithm based on the Volpe/FRA
“CWR-SAFE” model (3), and it is customized for easy railroad use by requiring only simple input
parameters readily available to the railroad engineer. The program determines the “allowable
temperature increase”, and based on the RSM measured rail temperature and RNT it computes a
buckling margin of safety (BSM). Based on the calculated BMSs falling into prescribed “safe”, marginally
safe” or “unsafe” regimes in accordance with Table 1, alerts can be issued on a potential for an incipient
buckle. Note that in Table 1 the green-yellow-orange-red BMS regimes are illustrative and provisional in
line with current US practice, and are based on the RSM’s real time output and StressNet’s real time
application for BMS determination. Automated “triggers” can be developed for specific railroad property or
network needs.
Table 1 - Provisional Criterion for BMS Application to Hazard
Identification and Advisories
BMS
BMS
Larger than 25°F
(14°C)
ACTION
ACTION
OK; no action
Between 10 and 25°F
( 6 and 14°C)
Caution; advisory
Between 0 and 10°F
(0 and 6°C) Slow order; readjust RNT
Less than 0
Red flag; immediate attention
5.3 Buckling Hazard Warning Example and Case Study
Based on RSM/StressNet TM data as outlined in Figures 2 and 3, buckling hazard evaluations were
conducted for the Union Pacific at two RNT conditions for a maximum summer rail temperature of 140°F
(60°C) application in line the green yellow-orange-red index definition. The results are shown in Figure 9
where the left-hand side shows the RNT profile, while the right-hand side the StressNet’s BUCKLE output
providing the BMS and the hazard index. At RNT avg=103°F(40°C) a high BMS=50°F (28°C) is calculated
showing the “green” condition at a rail temperature of 140°F (60°C) i.e. that at the track is not buckling
prone till 140 + 50 =190°F (88°C). At a lower RNT avg = 75°F (24°C) BUCKLE output indicates a BMS=22°F
(12°C) or a “yellow” alert i.e. that the track is not vulnerable to buckling till 140 + 22 =162°F (72°C). Based
on such outputs the railroad can determine and apply remedial action as needed. For the Union Pacific,
this test provided information on not needing summer slow orders. For additional hot-weather speed
restriction concepts and applications refer to [12].
8
Is there a buckling hazard at T RAIL = 140°F (60° C)?
RNT avg =103°F(40°C)

Challenge C: Increasing Freight capacity and services
Figure 9 – Example of Buckling Hazard Warning
6.0 Conclusions
Track buckling hazard detection and rail stress management largely depend on the track’s RNT condition.
Although techniques to measure and monitor RNT with sufficient accuracy have been lacking, a
promising evolving technology based on continuous and real time RNT measurement system has been
developed. The applications of this system’s measurements have been extended to cover several rail
longitudinal stress management aspects, including rail break detection, improved rail break/defect
removal repairs, and to a track buckling hazard detection and warning capability, all aimed at providing for
more efficient CWR safety and maintenance strategies.
7.0 References
[1] H. D. Harrison, “Evaluation of Methods for Measurement of Longitudinal Rail Force in Unloaded
Track”, Battelle Columbus Laboratories, Technical Memo, Contract No. DOT-FRA-9162, Task 2, 1981
[2] H. D. Harrison, “Strain Gage Based Longitudinal Rail Stress Measurements at FAST”, Battelle
Columbus Laboratories, Technical Note, Contract No. DOT-FRA-9162, October 12, 1983
[3] D. R. Ahlbeck, H. D. Harrison, “Monitoring Rail Neutral Temperature and the Long Term Stability
of Track”, presented at the American Society of Mechanical Engineers, Annual Winter Meeting, San
Francisco, California, 1989
9

Challenge C: Increasing Freight capacity and services
[4] A. Kish, “Destressing/Restressing for Improved CWR Neutral Temperature Management”, in
“Guidelines to Best Practices for Heavy Haul Railway Operations: Track”, International Heavy Haul
Association publication, 2009, pp. 6-58 to 6-73
[5] A. Kish, G. Samavedam, “Improvements in CWR Destressing for Better Management of Rail
Neutral Temperature”, Paper presented at 2005 Transportation Research Board Annual Conference -
January 2005, and in Journal of the Transportation Research Board #1916, pp 56-65, 2005
[6] A. Kish, W. Mui, “CWR-SAFE”, Software and User’s Guide, February 2001, US DOT/Volpe Center
[7] A. Kish, “Track Lateral Stability”, in “Guidelines to Best Practices for Heavy Haul Railway Operations:
Track”, International Heavy Haul Association publication, 2009 pp. 1-71 to 1-84
[8] A. Kish, G. Samavedam, “Dynamic Buckling of Continuous Welded Rail Track: Theory, Tests,
and Safety Concepts,” Transportation Research Record No. 1289, “Rail: Lateral Track Stability 1991”
[9] A. Kish, G. Samavedam, “Track Buckling Prevention: Theory, Safety Concepts and Applications”,
DOT/FRA/ORD Final Report, October 2004, in FRA publication
[10] A. Kish, et al, “Laying and Maintenance of CWR Track”, Union of International Railways (UIC) Code
720, ERRI D202/RP#10, Chapter 6, April 1999, Utrecht, The Netherlands
[11] Union of International Railways (UIC) Code 720R, “Laying and Maintenance of CWR Track”, 2nd
Edition, December 2002
[12] A. Kish, D. Clark, “Track Buckling Derailment Prevention through Risk-Based Train Speed
Reductions”, 2009 AREMA Annual Conference, Chicago, September 2009
10