Author(s)

J. F. Shen, Y. D. Xu, H. F. Li, F. Y. Li, J. X. Qiu, W. Y. Liu

Abstract

Because of the discreteness of ballasted track structure, the randomness of train loads and the periodicity of maintenance, the geometric irregularity tends to be stochastic. The foundation of track management is to select a reasonable probability distribution function and use its characteristics to establish a geometric irregularity prediction model. Nowadays, the improvement in maintenance technologies considerably reduces the overrun of irregularity; however, large values of irregularity that does not overrun lead to a poor fitting effect of two-parameter probability distribution. This study used five kinds of three-parameter probability distribution functions to fit samples from six segments, calculated and compared their P values, selected the best probabilistic distribution function to describe the spatial characteristics of longitudinal irregularity in each segment, and analysed the differences in track quality between segments. The results showed that the three-parameter probability distribution functions fitted effectively to show the characteristics of unilateral thick tail of longitudinal irregularity in ballasted track. The distribution of the longitudinal irregularity in the whole line could not be fitted by a single function.

Keywords

longitudinal irregularity, spatial characteristics, unilateral thick tail, three-parameter probability distribution