Оptimization of planned and preventive maintenance of rolling stock taking into account the risk of failures

Abstract

The paper considers the problem of increasing the efficiency of planned preventive maintenance (PPTO) of rolling stock by moving from rigidly fixed scheduled intervals to risk-based planning (Risk-Based Maintenance, RBM). The relevance of the topic is due to the simultaneous increase in requirements for safety and punctuality of transportation, aging of the fleet and increasing the cost of downtime, while the throughput capacity of the depot (posts, personnel, spare parts) remains limited. In such conditions, the same maintenance intervals for all units of equipment are a compromise: for some of the rolling stock they are overly conservative and cause unnecessary costs, for others they create an increased probability of failures in motion. The key idea of RBM is to coordinate the periodicity and priorities of PPTO with the probability of failure and the consequences of failure. The consequences in railway systems are multi-component and include safety effects, operational losses (delays, schedule disruption, reduced fleet availability) and economic costs (repairs, downtime, fines). The purpose of the study is to propose and justify an approach to choosing maintenance intervals that minimizes the expected total losses while ensuring an acceptable level of risk. The methodological basis of the work combines a review of modern concepts of maintenance, RCM, CBM and PdM/PHM with the formalization of the “cost + risk” problem based on parametric reliability models. For a quantitative description of degradation, a time-to-failure model with a Weibull distribution was used, which allows us to relate the maintenance interval T to the failure probability F (T ) and the expected costs for planned and corrective intervention. A demonstration example was used to calculate a rational maintenance interval and conduct a sensitivity analysis to the ratio of the costs of planned and emergency work. The existence of a region of optimal (rational) intervals is shown, in which a compromise is achieved between the frequency of planned work and losses from failures. The results obtained can be used as a basis for further integration with condition monitoring data (CBM/PdM) and for expanding the statement by introducing explicit risk constraints and depot resource constraints.