铁路论文；LP-ENGLISH（英文）.pdfVIP

PERFORMANCE OF A REAL TIME L/V FORCE PREDICTION MODEL G.W. English, T.W. Moynihan, TranSys Research Ltd. 682 Milford Drive Kingston, ON K7M 6B4 Canada english@transys.ca moynihan@transys.ca KEYWORDS: performance measures, track geometry, vehicle dynamics, wheel climb, derailment risk ABSTRACT: This paper presents the results of the latest phase of a multiyear test and analysis research program with funding contributions from Canadian Pacific Railway, Transport Canada and Canadian National Railway. An engineering-based L/V predictor model has been developed, which can be run in real time on board the geometry car to predict where the occurrences of undesirable L/V will occur. In instrumented wheel set testing, the predictor performed well in isolating track conditions that will produce high L/Vs, while the existing track standards identified very few sites. The model can predict L/V ratios in real time, for all axles, of multiple vehicles, at multiple speeds, for both directions of travel. The focus of this paper is on a case study evaluation of the model’s performance in identifying geometry conditions that later contributed to wheel climb derailments. 1 INTRODUCTION The principal means deployed by all major railways in measuring the geometric quality of track is a track geometry car or track evaluation vehicle. While these vehicles have the capability of measuring track parameters of interest, the interpretation of the measurements in assessing track condition and in deciding what to do about it varies widely and is in continuous evolution. Regulatory defects (or exceptions) are also derived from the track evaluation cars. These are local geometric conditions with threshol