Managing Risk on the Railway Infrastructure

1 Managing Risk on the Railway Infrastructure Allan M. Zarembski, , , FASME1, Joseph W. Palese, PE, MSCE, MBA1 ZETA-TECH Associates, Inc., Cherry Hill, New Jersey, USA, Abstract Risk management and together with its associated management of safety is a fundamental focus of Railway officers worldwide. This includes the issue of derailment prevention as well as the related issue of risk management for those categories of track failure or degradation that can result in a derailment. Recently, a new generation of track safety management tools has been developed that quantify and analyze the risk associated with key track failure modes.

2 This paper will describe three such new generation risk management tools that deal directly with track safety in key track failure (and derailment) areas. These models have been implemented on railways in the US and worldwide and have been shown to be effective tools in controlling and reducing the risk of derailments in their respective track risk areas. The specific models to be discussed in this article are: Broken rail risk model, which quantifies the risk of occurrence of a broken rail (and associated broken rail derailment) and allows railways to adjust their UT test schedules to reduce that risk. Track buckling risk model which identifies and prioritizes locations of high potential buckling risk and directs Railway engineers to them for appropriate action Vehicle/track geometry risk model, which identifies and prioritizes locations of high potential for vehicle/track geometry related derailments.

3 All three of these models rely on research that has been developed over the last several decades, and the resulting body of knowledge on the theory behind these track failure modes. All of these models are designed to be applicable in large-scale applications, to identify potential failure sites across an entire route, division or Railway system. Applications up to 25,000 miles (40,000 km) have been made using these risk analysis tools across an entire rail network. Results, in full-scale applications on major railways in the US and Europe have shown significant and quantifiable reductions in derailments themselves and in corresponding derailment causing conditions.

4 In several large-scale applications of the broken rail risk model, significant benefits in terms of reduced occurrence of broken rails and reduced broken rail derailments have been reported. In one such application, a reduction in broken rails (service defects) on the order of 28% was reported over a three-year application period. Concurrently, the same Railway reported a reduction in rail caused derailments of 33%. Similar results were reported on another large-scale railroad application this past year. Likewise, full-scale applications of the buckling risk assessment model resulted in a high percentage of high risk sites identified for immediate action by local forces.

5 On one major Railway application, between 30 and 50% of the identified sites were the subject of immediate corrective action by local forces. This Railway likewise reported a dramatic reduction in both track buckling related derailments and track buckling incidents. WCRR2006 2 Introduction In recent years, railways have turned to the discipline of risk management to improve safety and reduce the potential risk of accidents or derailments. Since accidents and derailments are very low probability of occurrence events, it is necessary to focus on the derailment causes themselves and develop risk management tools that quantify and analyze the risk for each key derailment or accident area.

6 One such high risk area, for which a risk management approach has been developed, is broken rail safety, a major track-caused derailment category. In the United States alone, an average of over 280 derailments a year have occurred over the last 8 years with a total average annual cost of over $100 Million. Increased testing has a direct effect on this behavior, where both the service defect rate and the number of derailments decrease with increased testing. However, increased testing is expensive and should be judiciously applied. Simply increasing testing across the board is not the best approach.

7 Rather, assessment of the rail condition, and the risk of broken rails (service defects) offers the most efficient and cost effective approach to increasing rail testing. A second such high-risk area is that of track buckling. Track buckling, the sudden lateral movement of the track due to thermally generated longitudinal rail forces, remains a major track failure mode both in North America and worldwide. Over the last five years, there have been, on average, 34 derailments a year in the US alone, with a FRA reported cost of over $11 Million annually. These derailments, coupled with high maintenance costs associated with buckling prevention, have forced Railway systems to pay careful attention to those factors that can influence the potential for track buckling.

8 Since techniques to monitor the neutral temperature of the rail, and thus predict the onset of buckling by measurement, are very limited and difficult to employ, railroads generally rely on preventive measures to control buckling risk. Yet a third such high-risk area is the vehicle/track interaction area, where a combination of geometric parameters or repeated geometric anomalies can lead to unsafe vehicle dynamic response. Current track recording car based measurement of geometric parameters and comparison to pre-set safety limits is insufficient. This is evidenced by the more than 100 such vehicle-track interaction related derailments reported in the US alone each year.

9 The need fro a more focused approach to identify these high-risk areas is apparent. This paper will address these three high derailment areas and show how a risk based approach has been used to help identify high risk areas and effectively and economically focus resources on these high risk areas to control and reduce the number of derailments. Broken Rail Risk Management Control of broken rail derailments is a serious concern to maintenance of way officers worldwide. Research studies have shown a relationship between rail defect occurrence and broken rail derailments [1, 2]. Furthermore, breaking rail defects into two detection categories - detected defects1 and service defects2- have shown a strong correlation between service defects and derailments.

10 Early US studies found that for fatigue related defects, a broken rail derailment occurred every 133 service defects for a derailment per defect rate of [2]. Subsequent research, focusing on control of rail service defects by improved inspection efficiencies, showed the potential for controlling the risk of broken rail derailments by reducing the percentage of service defects and the associated service defect rate [3, 4]. Examination of data from over 50,000 km (30,000 miles) of railroad date for the period 1995 though 2003 supports the findings of a relationship between rail defects, particularly service defects, and derailments [5].