Equivalent Safeguard for Broken Rail Detection in Positive Train Control Territory
Federal requirements for Positive Train Control (PTC) systems
are specified in 49 CFR 236.1005. Paragraph
(a)(5) of this regulation requires that the speed of passenger and freight
trains be limited to 59 miles per hour and 49 miles per hour, respectively, in
areas without broken rail detection or equivalent safeguards. Also, 49 CFR 236.0
paragraph (c)(2) requires that a block signal system be installed for such
speeds, unless a railroad has applied under 49 CFR 236.0 paragraph (e) for
approval of discontinuance or material modification of a signal or train
control system. This can be done in conjunction with a request for approval of
a Positive Train Control Development Plan (PTCDP) or Positive Train Control
Safety Plan (PTCSP).
However, the US railroads also have many miles of
conventional signaling systems where trains operate slower than these
regulatory speed limits. Under existing regulations, track circuits and block
signal systems would not be required for such lines, unless they already
existed. If track circuits already existed, a railroad that wished to replace
its existing signal system with a stand-alone PTC solution would have to apply either
to discontinue the existing signal system under 49 CFR 235, or for a discontinuance/material
modification to the signal system under 49 CFR 236.0 paragraph (e). However, no
railroads are known to have made any such applications.
Rather, the industry has chosen to implement PTC as an
“overlay” to existing signal systems. While this choice has limited the
potential for benefits from PTC implementation, the industry’s position has
always been that this choice was a short term expediency that was necessitated
by the requirement to conform to an extremely aggressive deadline for PTC
system implementation. It was never suggested that this approach would be the
long-term optimum. The rail industry has always acknowledged the high probability
that a stand-alone PTC system could over the long term, produce greater
business benefits than the current overlay approaches are able to do.
The current Interoperable Electronic Train Management
System (I-ETMS) developed by Wabtec does not yet have vital certification. However,
as I-ETMS moves towards vitality, railroads will be able to obtain business
benefits from it. Once full vital certification is attained, it is likely that
railroads will start to show interest in stand-alone PTC implementations that
can fully replace the functionality of existing signaling systems. Currently, most
PTC products have anchored their architecture to the existing signal system. The
proposed research is needed for allowing PTC developers to evolve their
products in new directions that can provide better functionality at a lower
cost to the railroad industry.
In conventional block signaling
systems, track-circuits are the de facto
means for establishing train occupancy.
However, they also provide a measure of broken rail protection at no
additional cost. As railroads overlay existing rail lines with PTC to comply
with the federal requirements, they will retain the track-circuit systems which
provide broken rail detection at little to no additional cost. However, to satisfy the “equivalent safeguard”
requirement, developers of new rail lines will be obligated to two signal
systems; one for PTC vitality and one for broken rail detection. Given the lack
of alternative broken rail detection technologies, developers will employ track
circuits. This will impose an additional cost on developers of intercity passenger
rail lines and
new freight lines, as they will be responsible for the installation and
maintenance of two signal systems or at least track circuits.1
In signalized territories, a “defense in depth” strategy
that uses multi-layered protections is employed. The intention is to create a
reliable system using multiple layers, rather than making any one layer
perfectly reliable. All safety activities, whether organizational, behavioral
or equipment related, are subject to layers of overlapping provisions, so that
if a failure should occur it would be compensated for or corrected without
causing harm to individuals or the public at large. The problem is that even with the existing
two layers of broken rail defense in signalized territories: periodic rail
inspection coupled with track circuit-based broken rail detection, broken rails
a leading cause of train derailments in the United States.2 In many
of these cases, the rail break will occur under a train, thus making it
undetectable by either traditional track circuits or by visual observation. It
should be noted that broken rails detected by the signal system, train crews,
and track inspectors are classified as service
failures and are strongly correlated with derailments. Rail defects
detected during non-destructive testing (NDT) (ultrasonic or other), known as detected defects, exhibit a less direct
cause and effect relationship with derailments.
Track-circuit-based broken rail detection is not perfect
because rails often break under trains, and there are many kinds of rail
defects that can’t be detected by signal system track circuits. While improved rail maintenance methods have
led to a sharp decline in accidents over the past several years, other accident
causes have declined as well. This leads
to a concern that the current “defense in depth” strategy may not be sufficient
to address the full range of broken rail derailments. Numerous investigations
and studies conclude that track circuits in no way guarantee reliable
identification of rail breaks. Estimates of 20% to a maximum of 60% broken rail
detection rate have been given.3
NDT is a much more reliable technique for capturing rail
defects as they form and propagate. In
fact, numerous research studies have shown that ultrasonic rail testing finds
between 85% and 90% of all rail defects, greatly reducing the risk of broken
rail derailments.4 This has been further enhanced by the current “risk” based approach to rail
test scheduling which, together with an increase in the rate of installation of
new replacement rail and other factors, has resulted in significant further
reduction in broken rail caused derailments.5,6 While track circuit broken rail detection is
more or less associated with a fixed cost (aside from maintenance), increasing
the frequency of NDT has a one-to-one correlation with cost.
Therefore, it is desirable to consider alternatives to the
continued regulatory requirement for track circuits for broken rail detection.
Clearly, it is better to use improved inspection techniques to identify rail
defects before the rail breaks, rather than to use track circuits to
identify rail defects after the rail breaks.
Because of recent improvements in rail inspection
technology, achievement of this vision has become a distinct possibility.
However, in the US, the railroads are still installing track circuits for
meeting the requirements of 49 CFR 236.1005. paragraph (a)(5). No one really
questions this cost or knows how effective it is, in terms of reducing the
actual derailment rate, or for improving safety. Rather, this is done simply as
a matter of regulatory compliance. Internationally of course, this requirement for
using track circuits is not universally accepted.
1) As noted
in FRA’s final rule on PTC, BNSF successfully demonstrated a Track Integrity
Warning System in dark territory, employing basic track circuits of about 5
miles in length and communications capabilities. This was viewed as a simpler, less costly means
of detecting service failures than maintaining a full-blown signal system (with
redundant logic controllers, signal masts, etc.). 75 FR 2598, 2601 (Jan. 15, 2010).
2) Liu, X., Saat, M.,
& Barkan, C. Analysis of causes of major train derailment and their effect
on accident rates. Transportation
Research Record: Journal of the Transportation Research Board 2289, 2012,
3) Gifford, J. ARTC & FIRSE, Trevor Moore ARTC
& FIRSE and Joseph Borg SELECTRIX: Axle counters for heavy rail traffic applications
(23 March 2007, Case study in Australia).
4) Zarembski, A.M., Palese, J.W., “Characterization of
Broken Rail Risk for Freight and Passenger Railway Operations”, 2005 AREMA Annual Conference, Chicago,
IL, September 25-28, 2005.
5) Zarembski, A.M., Palese, J.W., "Use of Risk
Management in Improving Track Safety", AREMA 2007 Annual Conference &
Exposition, Chicago, IL, September 2007.
by FRA-sponsored research in the 1990’s, the performance-based approach has
been embodied in the track safety regulations. See: 49 CFR §§ 213.113,
213.237, 213.238, and 213.241, as revised at 79 FR 4234 (Jan. 24, 2014).
Since track circuits fail
to detect a significant number of rail breaks, they are a weak “safety net” to
fall into. One may imagine the equivalent plight of a trapeze artist who knows
that their safety net will only work 50% of the time. Such an artist would not
be expected to live very long! In that context, a risk reduction rate of only
50% would not be effective. If the objective is actually to reduce the
frequency of broken rail accidents, then either:
A more effective rail monitoring system than conventional track circuits needs to be developed, and demonstrated cost-effective, or
An alternative approach would increase the frequency and effectiveness of rail inspection, while seeking to reduce the cost of rail inspection, so developing defects can be confidently and economically identified before they turn into actual rail breaks.
In fact, if it were possible to confidently identify all
rail defects before they could develop into actual broken rails, this would
remove both the economic justification and safety need for using track circuits
to detect broken rails. The more resources are put into rail inspection and
maintenance, the less likely it is that a defect resulting in a broken rail will
be detected by track circuit monitoring systems. Eventually the frequency of broken
rail failures is reduced to such a low level that there is no longer any
economic or safety case for rail circuit monitoring.
The proposed research will take a fresh look at the
business and safety case for track circuits, specifically in regard to rail
safety and the value of broken rail detection. This will allow the industry to
better understand the best way to spend its limited safety dollars. Most certainly, axle loadings will influence
the choice; some Australian heavy-haul operators, for example have voluntarily
decided to include track circuits, because they feel that the broken rail
protection is valuable and useful in the context of their own operations. Other
operators who run light tonnage on new rails, even at high speeds may find that
the risk of broken rails is extremely low which does not warrant the cost of
installing track circuits.
As a result, the central focus of the proposed research
will be on identifying the best way for the rail industry to reduce its
currently high rate of broken rail accidents. The proposed research would assess
the full continuum of rail operations including representative low speed, high
tonnage lines all the way up to representative high speed, low tonnage lines,
for assessing the value proposition that is actually associated with the
installation of track circuits. It will answer the question, for example, if resources
were redirected away from track circuits (high capital and maintenance costs)
towards better rail inspection and maintenance, would it result in a better
The objective of this research would be to develop a
current and comprehensive assessment of the economic and safety case for all promising
means available to reduce the rate of derailments associated with broken rails:
- The research would assess the full range of available rail monitoring and rail inspection technologies, with a view towards finding the most cost effective combination for any given application and risk safety profile. For example, it is possible that a higher level of accident risk may be economically tolerated on a rail line that carries only coal, as compared to one that carries hazardous materials and/or passengers.7 It is likely that different approaches may be optimal for different lines that carry different combinations of tonnage and traffic mix3.
- The proposed research would develop a methodology for identifying the risks and economic tradeoffs, so the optimum combination of maintenance and monitoring for reducing risk on any given line segment can be identified.
- In terms of how this may impact the requirements for PTC implementation, 49 CFR 236.1005 does not define what an equivalent safeguard is, or how any specific proposed equivalent safeguard may be evaluated or approved. It is possible that a risk-based approach based on an approved rail inspection regimen might be able to substitute for track circuits as an equivalent safeguard under 49 CFR 236.1005 (a)(5).
Given the known ineffectiveness of track circuits as a
safety net for detecting broken rails, it should not be difficult to show that
an increased inspection regimen and more aggressive maintenance practices could
provide an equivalent level of risk reduction to what track circuits currently provide.
However, this may not be good enough to gain FRA’s approval. To qualify as an “equivalent
safeguard” under 49 CFR 236.1005 it may need to be shown that a regimen of
increased inspection and rail maintenance would reduce rail failure risks to “de minimis” levels, eliminating the need
for rail monitoring. The residual risks would then be so small that there is no
economic or safety case for continuing the requirement that signal systems must
provide broken rail detection.
This research would consider available promising options
for preventing train accidents caused by broken rails, and would recommend the
best combination of systems approaches under various circumstances for reducing
rail caused accidents. It would specifically address the contribution made by
track circuits to rail risk reduction, and would identify alternatives to the
use of track circuits that may reduce accident risks by even more than track
7) Liu, X.,
Optimizing rail defect inspection frequency to reduce the risk of hazardous
materials transportation by rail. Journal of Loss Prevention in the Process
Industries, 48, 2017, 151-161.
The proposed research would further shift the focus away
from detection of service failures, which account for around 15% of all defects
and are five times8 more likely to result in a derailment, to detection of defects by testing and accident
prevention by maintenance, including the laying of new rail when actions such
as inspection and profiling have reached their practical limits. As such, it
would point the way to further reductions in rail-caused train accidents, and would
lessen the large economic and social costs associated with such accidents. To
the degree that the proposed research focuses on Economic Assessment, it will help
ensure that resources are directed to the areas of greatest potential benefit
and risk reduction – thereby leading to the greatest possible accident risk
reduction for any given level of safety expenditure by the railroad industry.
Furthermore, by clarifying system requirements in regard
to railway signaling, PTC developers could take a more flexible and cost
effective approach in regard to the design and deployment of their train
control systems. By reducing the cost of PTC, additional safety benefits could
be had if railroads were encouraged to further expand the scope of implementation
of such systems.
A. M. Management of Broken Rail Risk for High Speed Passenger Rail. In 2010
Joint Rail Conference (pp. 15-22), American Society of Mechanical Engineers,
The Federal Railroad
Administration has supported research into rail integrity and encouraged
innovation. However, FRA has also noted
the practical considerations which predominate in many circumstances, at least
with existing technology. For example, as
FRA has noted to its personnel responsible for rail integrity,
- FRA Track Safety Standards Part 213.237 designates responsibility to the rail flaw detector car operator to properly identify the types of rail head surface conditions that can result in an improper or invalid test of the rail section in which the condition is contained. When rail head surface conditions are encountered that may influence test results extra care should be taken in the data interpretation process. The operator should also be aware and diligent when encountering other conditions that may result in an invalid test.9
Although advances in artificial
intelligence and other software-based solutions may narrow the discretion
presently available to those who conduct internal rail flaw inspections, it is
likely that human factor issues, bearing on day-to-day safety management, will
continue to present concerns.
In the past, the regulator has
frowned on rail integrity practices that relied too heavily on personnel in the
field to make decisions regarding actionable findings during the inspection
process, without a framework of quality assurance to create boundaries on, or
even estimate, the degree of risk tolerated.
In the view of the regulator, rail inspection contractors have sometimes
been subject to heavy influence from transportation department personnel eager
to get inspection cars over the territory.
Further, the regulator may be concerned that contractors employ closely
held proprietary technology which has not been validated against accepted
New approaches to rail
integrity may succeed in cabining unwarranted discretion and adding quality
assurance elements not traditionally available.
However, to ensure this is the case, research into technologies and
practices must be accompanied by a strong emphasis on safety management,
including human factors (such as test car operator training, qualification and
oversight) and institutional considerations.
There is a large body of knowledge of rail safety in the
area of rail inspection and maintenance practices. However past over reliance on the ability of
signal systems to actually detect rail breaks may have led to a false confidence
of safety, which has not been justified by the actual rail accident
history. In other words, the existing
body of knowledge is “stovepiped.” The
proposed research would take a more holistic view of rail safety, taking into
account the effectiveness of both rail inspection and monitoring approaches. From this systems perspective, it may be
possible to obtain a more complete and balanced perspective of rail safety, so
that overall system approach to preventing rail-caused train accidents can be
changed for obtaining a better result.
9) Track Inspector
Rail Defect Reference Manual at 53 (FRA Office of Safety August 2011).
a full range of available rail monitoring and rail inspection technologies,
with a view towards finding the most cost effective combination for any given
application and risk safety profile. The research should identify the best way(s)
for the rail industry to reduce its currently high rate of broken rail
accidents, which continue to occur even in spite of the presence of track
circuits on many lines.
a methodology for identifying the risks and economic tradeoffs, so the optimum
combination of maintenance and monitoring for reducing risk on any given line
segment can be identified.
the risks associated with elimination of an alternate broken rail detection
technology, such as track circuits, and reliance on increased frequency of
conventional inspection technology.
“acceptable” level of broken rail risk for different traffic types, such as:
speed passenger routes
material freight routes
Priority conventional freight routes
and heavy rail transit systems
route configurations as appropriate
existing or potentially new technology that can be used to further improve
broken rail safety without the major cost of a separate signaling system. This
may include identification of supplemental rail break monitoring technologies.
requirements for integrating new approaches into safety management systems that
include attention to human factors and conflicting demands within the
institutions providing the rail service.
strategies for quality assurance to strengthen confidence in program
a cost benefit analysis of the different ranges of rail inspection and
monitoring levels to include as a minimum:
rail testing alone based on “acceptable” level of broken rail risk
rail testing supplemented by conventional inspection
rail testing supplemented by other (non-track circuit) technologies
of the role of rail replacement in optimization of the system, particularly
where mixed passenger and freight traffic drives high cumulative tonnage.
o Other combinations as appropriate
o Can a
maintenance approach be defined as an “Equivalent Safeguard” to 49 CFR
236.1005.paragraph (a)(5) and which it may argued would have a valid safety
case that the FRA could accept?
This research can assist the Federal Railroad
Administration, the Federal Transit Administration, and U.S. railroad industry
in their continued quest for better understanding and managing broken rail
derailment risk. This topic is particularly timely and important in light of
nationwide implementation of PTC technologies. The project team should
demonstrate experience in working with industry collaborators in acquiring
necessary data and transforming analytical results into actionable decisions.
This research is anticipated to be completed within 48 months, with a final report detailing the analyses and suggestions regarding the assurance of equivalent safe guard for broken rail detection in PTC territory, accounting for various combinations of rail inspection, maintenance and monitoring technologies. The estimated funding requirement for the requested research is approximately $500,000. KEYWORDS: Rail, safety, risk, positive train control, broken rail
|Sponsoring Committee:||AR030, Railroad Operating Technologies
|Research Period:||Longer than 36 months|
|Source Info:||• AR030 Railroad Operating Technologies Committee (primary sponsor)|
• AR070 Railroad Operational Safety Committee (co-sponsor)
|Index Terms:||Rail (Railroads), Railroad safety, Positive train control, Nondestructive tests, Broken rail detection, |
|Cosponsoring Committees:||AR070, Rail Safety|
Operations and Traffic Management
Safety and Human Factors