Maintenance Planning for Rail Asset Management—Current Practices (2020)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

8 Broken rails and their associated derailments are of serious concern to railroad and transit systems. A broken rail will often appear as a defect before growing to the point at which an actual break occurs. Thus, identification of these rail defects is critical, as is replacement of the corresponding section of rail before the defect propagates to failure. Studies on major U.S. Class I railroads show a relationship between total defects and derailments (Zarembski, 1988; Zarembski and Palese, 2007). More specifically, these studies, which took place over several decades, show a correlation between rail defects and broken rail derailments—a major class of track-caused derailments and among the most expensive. These studies further note that defects are divided into two primary classes according to how they are found. Detected defects are found by ultrasonic testing of rails, and undetected or service defects are found by track circuits, by walking track inspection, or by train operators. The same research studies have shown a significant correlation of broken rail derailments with service defects, with approximately one derailment per 125 service defects for main line freight railroad track under modern heavy axle loading (Zarembski and Palese, 2005a). Thus, the higher the percentage of detected defects, and corresponding lower percentage of service defects, the lower the risk of derailments. Service defect percentages (service defects divided by total defects) of 10% or lower offer a significantly reduced risk of broken rail derailments than service defect percentages of 15% or higher (Zarembski and Palese, 2005a). Simplistic rail test scheduling approaches, such as those based on specified time inter- vals, annual tonnage levels, or both, do not account for aging rails and the corresponding increased rate of defect initiation (Zarembski and Palese, 2005a; 2007). These traditional test- ing approaches were initially recommended by APTA or included in earlier FRA track safety standards, but they have since been superseded by a risk-based approach (FRA, 2019). These simplistic approaches do not give the railroad the flexibility to adjust test frequency to the actual rail conditions encountered. Likewise, simplified “rules of thumb” for scheduling ultrasonic testing provide limited benefits. They often account for factors including age of rail (usually in cumulative million gross tons [MGT]), level of usage (usually annual traffic density), class of track, type of traffic, and defect counts. However, such rules do not incorporate those factors in a manner that is directly tied to the risk of a derailment. Risk-based scheduling methodologies, however, use site-specific and directly measurable performance parameters that can be related to a defined level of risk, and set ultrasonic test intervals accordingly. In addition, the risk-based approach incorporates inspection effectiveness, as well as defect growth rate based on axle load and other factors. These measures are incorporated to ensure that inspections performed at the determined frequency have a high probability of finding internal rail defects before they turn into a broken rail. This methodology was developed by the U.S. DOT’s Volpe National Trans- portation Systems Center (Orringer, 1990) and further enhanced by subsequent researchers (Palese and Wright, 2000; Zarembski and Palese, 2003). C H A P T E R 2 Literature Review

Literature Review 9 Implementation of risk-based test scheduling techniques has been found to reduce the risk of derailments, specifically for a service defect–based definition of risk—that is, the number of service defects per mile per year (Palese and Wright, 2000; Zarembski and Palese, 2003). These techniques, which are based on more than a decade’s worth of rail integrity research, are used to determine optimum rail test intervals for a defined level of risk and for a defined segment of track. Further, risk-based testing has been shown to reduce the rate of broken rails (service defects) and associated broken rail derailments by 30% or more (Zarembski and Palese, 2003). The use of a service defect–based definition of risk also allows for the identification of the maximum allowable risk for different classes of freight and passenger track. Thus, although the average level of risk for U.S. freight railroads is of the order of 0.1 service defects per mile per year, the risk for passenger-based rail operations such as rail transit systems should be set to much lower levels (e.g., 0.01 to 0.05) (Zarembski and Palese, 2007). Because rail testing is expensive, simply expanding testing is not the best approach. Rather, assessment of rail condition, and the risk of broken rails (service defects) and their associated broken rail derailments, offers the most efficient and cost-effective approach to increasing rail testing (Zarembski and Palese, 2003). A risk-based approach aims to schedule ultrasonic test- ing so that there is a constant low risk of rail failure throughout the service life of the rail, even though rail defects develop more rapidly as the rail ages. This constant low risk is accomplished by increasing the frequency of testing as the rail ages and as the number of defects—both detected and service—increase. Such a methodology takes several factors into consideration, including the age of the rail, the likelihood that a defect will escape detection and grow to failure before the next inspection, and the level of risk itself, defined as the number of service defects (broken rails) per mile per year. This technique has been successfully applied on both passenger and freight railroads. For example, application of this method of scheduling to a Class I railroad led to a significant drop in both broken rails and broken rail derailments over a 9-year period (Palese and Wright, 2000; Zarembski and Palese, 2003). As noted previously, risk-based scheduling intervals have recently been introduced into the FRA track safety standards. Thus, it is the required inspection scheduling approach for those railroads under FRA oversight: Amtrak, all U.S. freight railroads, most commuter railroads, and those rail systems that operate interstate. Internationally, risk-based scheduling has been applied in the United Kingdom, on both Network Rail (Zarembski and Palese, 2003; Bin Osman et al., 2018) and the London Underground; in Sweden, by the Swedish infrastructure owner Banverket, recently renamed Trafikverket (Zarembski and Palese, 2003); in Norway (Vatn and Svee, 2002; Podofillini et al., 2006); and in Australia (Zhang and Wu, 2019). It should be noted that the 2010 study on the London Underground was extensive and included the system’s BCV (Bakerloo, Central, and Victoria) line and SSL (Sub Surface Lines). Other risk-based applications are discussed in Zarembski and Palese (2005b); Podofillini et al. (2006); Liu et al. (2014); Bin Osman et al. (2018); and Zarembski and Palese (2006). The risk-based test scheduling methodology determines the required frequency of ultra- sonic testing on a segment-by-segment basis. The approach uses previous defect history, traffic levels, and acceptable levels of risk as determined by traffic type and speed (e.g., high-speed passenger, passenger, hazardous materials, conventional freight, and so forth). As noted, risk-based scheduling was recently incorporated into the FRA (2019) track safety standards, where it specifies service failures per mile per year as a measure of risk (Orringer, 1990; Palese and Wright, 2000; Zarembski and Palese, 2003). Thus, the allowable level of service defects (risk) varies with type of traffic and class of track (speed), and the inspection frequency is geared toward maintaining that service defect rate as specified in the FRA (2019; see Section 213.237[a]) track safety standards (see Appendix D). The risk-based approach sets a defined level of risk (service defects per mile per year) for each segment; that level of risk must be held constant, even as the rail ages (Zarembski and Palese,

10 Maintenance Planning for Rail Asset Management—Current Practices 2007). In setting that risk level, the risk-based approach increases the percentage of defects found by detector cars, reduces the corresponding level of service failures (service defect–to– detected defect ratio), and thus reduces the risk of derailment. The analytical approach that forms the basis for this risk-based scheduling methodology incorporates three primary phenomena that affect the occurrence of a service defect: 1. Defect Initiation: How frequently do defects initiate? 2. Defect Growth: How quickly does a defect grow from initiation to failure size? 3. Detection Reliability: How probable is it that a defect of a certain size will be found? Defect initiation refers to the rate of development of rail defects, which is directly related to the age of the rail as defined by cumulative tonnage or MGT. Defect growth refers to the rate of growth of an individual defect from initiation to full size, usually defined as the point at which the defect will break under a passing wheel. The key in risk-based rail testing is to find the defect between the point at which it grows to detectable size (i.e., the minimum detection threshold) and the point when actual failure is imminent (i.e., the maximum allowable defect size). This interval is of the order of 10 to 50 MGT, depending on a number of track, traffic, and environmental factors. This interval is thus defined as the defect growth life and represents the maximum amount of traffic that should be permitted to pass over the defect between UT inspections. Detection reliability represents the probability of finding any given defect on the basis of its size. As a defect grows in size, it is much more likely to be found during an ultrasonic inspection. The testing frequency can then be adjusted to improve the overall probability of finding a defect in a given section of track with known defect history and tonnage. Such an adjust- ment is made by combining the rate of defect initiation and growth with the probability of finding a defect of a given size (i.e., the detection reliability). The details of this approach can be found in Orringer (1990), Palese and Wright (2000), and Zarembski and Palese (2003). Although larger railroads and transits can use this methodology on a track segment or sub- division basis, it can also be adapted for smaller transits at the system level, or for larger transits at the line or route level.