Assessment of the reliability of rail fasteners based on expert information using the Monte Carlo method

Abstract

The created software and mathematical support for the evaluation of reliability indicators of rail fasteners of the railway track is given. As a mathematical model of operating time, a three-parameter triangular distribution is adopted, whose parameters are chosen by experts. As reliability indicators, the following factors were selected: the probability of a failure-free operation; average operating time; gamma percent resource. The Visual Studio 2019 development environment was used to create the software. The code is written in the C# programming language. The possibility of obtaining analytical values of reliability indicators made it possible to test their numerical models. Using the Monte Carlo method and the created software, the study of the influence of uncertainty upon the selected reliability indicators was carried out, the average operating time being a random variable. As a probabilistic model for the average value of operating time under uncertainty, a beta distribution with two variants of parameter values is proposed: the beta distribution is a uniform law; the beta distribution is close to the normal law. It is shown that the obtained numerical values of the average operating time, the uncertainty for both variants of parameter values taken into account, fall into the confidence interval. Moreover, the confidence intervals for both variants overlap. Based on these results, it can be concluded that the estimates of the mean values for these variants are statistically indistinguishable. But at the same time, the simulation results showed that the gamma percentage resource for option 1 differs from the analytical value by 2.24%. In absolute terms, due to the uncertainty of the average gamma value, the percentage resource decreased by 5.1 million tkm. And this fact should be taken into account when making management decisions on servicing rail fasteners of a railway track.