Multiobjective optimization of preventive maintenance schedule on traction power system in high-speed railway

5000 km high-speed railway with 350 km/h speed will be constructed in China in the next five years. Traction substations convert the high voltage electricity from the public power system to 25 kV single phase electricity, which are implemented to support the high-speed locomotives through the catenary system. These substations and the catenary system form the traction power supply system (TPSS). Researches show that the reliability of TPSS is mainly dependent upon the catenary system which is essentially a multi-component mechanical system. Three types of maintenance actions have been categorized: (la)-maintenance (mechanical service), (lb)-maintenance (repair) and (2p)- maintenance (replacement). Different maintenance actions result in different system reliability and maintenance costs. Therefore dynamic reliability model and the maintenance cost model on TPSS using different maintenance actions have been established in this paper. The objective is to optimize the maintenance actions so that the maximum reliability and minimum maintenance cost can be achieved. For such a multiobjective optimization problem, methods such as the pareto-optimal set can be applied. In this paper, an advanced evolutionary algorithm (EA), chaos self-adaptive EA, called CSEA, has been proposed. This approach is an enhancement from the nondominated sorting genetic algorithm II (NSGA-II). The chaotic initial population helps to improve the initial diversity. The grouping selection strategy has been suggested to give the dominated solutions more chance to enter the mate pool. Finally a self-adaptive genetic operator is adopted to give the dominated solutions higher possibility to cross and variation. All of the above help to avoid prematurity and enhance the global searching ability of the algorithm. Simulation results show that, CSEA outperforms NSGA-II in terms of diversity-preservation and in converging close to the Pareto-optimal set.