Deterministically Setting Hunting Speed Restrictions
Using this method can lower derailment risks and at the same time potentially improve train velocity. This work is part of the Association of American Railroads Car and Truck Systems Strategic Research Initiative program, but it can find application anywhere equipment characteristics and THD data are maintained in a database.
Hunting is a lateral instability that has an onset train speed that depends on various equipment and track characteristics. Historically, railroads managed hunting risks by speed-restricting entire groups of equipment based on derailment or lading damage experience. Each railroad typically had different experiences and put different speed restrictions in place.
The AAR Manual of Standards and Recommended Practices (MSRP) defines hunting in terms of carbody lateral acceleration. As measured by wayside detectors in regular service, hunting is defined by the Hunting Index (HI), with current alert levels at 0.5 (one event) or 0.35 (two events in 12 months). Previously, TTCI explored the relationships between car type, car condition, lateral accelerations, and HI values.1, 2
TTCI accessed five years of THD pass data comprising the majority of freight cars registered in the Railinc UMLER® database. The equipment restrictions provided by the railroads comprised 236,826 cars. For reference, “Population” was defined as all THD passes by cars that were not listed in any railroad restriction. This study focused on three restricted equipment groups: “E” and “G”—gondolas, empty and loaded; “Flat”—bulkhead, centerbeam, and depressed, empty; and “H,” “J,” and “K”—hoppers and coal gondolas, empty.
Hunting index data are usually shown as a scatterplot in terms of pass speed (see figure above). This format is comprehensive for all passes, but it is difficult to describe the population statistically and to compare one population with another. In this analysis, the THD pass data for each population were grouped by speed at 5 mph intervals. The HI data for each speed were then binned into discrete values in intervals of 0.05, and the population was described as a count of passes in each HI bin. The count data were normalized by the total count to produce a percentage. Converting the percentages into exceedances and multiplying by 1 million produced a statistical defect rate in parts per million (ppm).
The defect rate shows the number of truck passes out of 1 million that would be expected to reach a given HI level. The inset table in the figure shows how the HI data for a given car condition can be distilled into a defect rate. The railroad would use a threshold HI and a threshold defect rate to set a speed restriction. Using empty hoppers (yellow cells) as an example, choosing the HI ≤ 0.35 (blue cell) and a maximum defect rate of 1,000 ppm suggests a speed restriction of 50 mph, because the defect rate of 636 ppm (green cell) is less than 1,000. At 55 mph (not shown) the defect rate exceeds 1,000 ppm. The full table would show various types of equipment, in loaded and empty conditions, at speeds ranging from 40 to 60 mph.
A statistical defect rate analysis of HI data provides a means of deterministically setting speed restrictions. The railroad can objectively define the HI and defect rate that will result in an optimized speed restriction. Railroads may still apply restrictions to individual cars as established by current intercept messaging protocols.
1. Tournay, H., S. Chapman, R. Walker. “Evaluation of Cars Registering Salient Hunting Indices at or above 0.25.” June 2015. Technology Digest TD06-025. AAR/TTCI.
2. Tournay, H., S. Chapman, R. Walker. July 2015. “Evaluation of Cars Registering Salient Hunting Indices at or above 0.1.” Technology Digest TD07-012, AAR/TTCI.
The post Deterministically Setting Hunting Speed Restrictions appeared first on Railway Age.