Application of the Bayesian method in assessing the reliability of load-bearing structures of locomotives

Abstract

Diesel locomotives of the ČKD ČME3 class are the basis of the fleet of shunting locomotives of Ukraine. Today, the average service life of these machines is 35–38 years, while the established standard is 25 years. Accordingly, if the operation is too long, there arise some problems which are associated not only with equipment failures, but also with failures of the supporting structures of the corresponding machines. The main failures caused by metal fatigue are the occurrence of cracks along the contour of the vertical sheet of the bogie frame along the welded joints and the occurrence of cracks in the area of fastening of the traction motor brackets. Determination of such damages requires significant costs and downtime of the rolling stock, because quality control of the objective condition and suitability is possible only when lifting the locomotive body, disconnecting the bogies and completely cleaning them from dirt. Now such control is carried out every 3 years, regardless of the circumstances. In this regard, the probabilistic approach to assessing the state of supporting structures and predicting both the period of condition monitoring and the service life have its advantages. It requires much lower costs and can be used to obtain recommendations for the selection of samples specifically for timely control, including the probability of failure, which is based on taking into account various kinds of indicators. The article proposes to use the Bayesian method which has found wide application in the theory of reliability. It has been demonstrated that it allows one to take into account the probabilistic characteristics of processes and operating conditions, on the basis of which it is possible to judge the reliability of load-bearing structures of diesel locomotives most reliably. Applying the Bayesian method together with the probabilistic-physical one, which uses a diffusion-monotonic distribution, it becomes possible to calculate the average operating time before failure and predict the time of safe operation of the equipment under study.