Reliability-based model for optimizing resources in the railway vehicles maintenance

Railway vehicles, due to a large number of hours spent in service operation, have a high frequency of breakdowns, thus, maintenance planning can be complex and expensive. The concept of maintenance is certainly connected to the concept of reliability, and the consequences of neglect can cause enormous resources losses. This research develop tool to optimize the required stocks of spare parts / materials and required number of employees for a given level of vehicle reliability. In the first phase of the research, data for the total fleet of 216 passenger wagons divided into 36 different technical groups were structured and analyzed according to the criteria defined by domestic regulations and international standards. In the second phase, reliability dependences on resource costs for all groups were modelled with using of regression analysis and enhanced by neural networks. Furthermore, goal and constraint functions were set and optimization algorithms were used in order to obtain optimal resources depending on the given system reliability. All the results of the applied algorithms were compared. The new model in practice serve as a tool to support the management of passenger car maintenance in making strategic decisions about resource planning with regard to available financial resources or the required reliability of the entire rolling stock.