A versatile method in the space domain to study short-wave rail undulatory wear caused by rail surface defects

Abstract

This paper examines the development of short-wave undulatory wear caused by vehicles running over defects on rail surfaces. The track model used was obtained using a Rational Fraction Polynomial (RFP) method modified using genetic algorithms. This produces an extremely accurate and simple wheel-track model requiring minimum computational times. As corrugation formation simulations in the time domain require such a large number of thousands of wheelset passovers, it was essential to develop such a simple but precise model. A non-linear Hertzian spring was used to study the interaction between wheel and rail. Wear is considered to be proportional to the friction work, as some other authors assume for this type of problems, emerging upon contact. This model has been validated by comparing its results with experimental results of undulatory wear measured on a track. The developed tool enables the evolution of rail wear to be studied for different ride conditions that are not affordable to study experimentally. Some of the factors studied in the paper are: the influence of the frequency at which the partially worn rail surface profile is updated during the wear evolution simulation; the influence of the train speed dispersion that usually exists around the nominal speed; the influence of the position of the rail defect within a span between two sleepers; and the influence of the train׳s nominal speed.