An Integrated Approach for Energy-efficient Train Operation Considering Bidirectional Converter in Urban Rail Transit

With the rapid development of urban rail transit, reducing the net energy consumption of urban rail traction system has received widespread attention. Improving the utilization rate of regenerative energy is one of the effective ways to reduce the net energy consumption. When the regenerative energy cannot be utilized by traction trains, the bidirectional converters can invert the remaining regenerative energy back to the middle voltage loop network for use by AC loads to improve the utilization rate of regenerative energy. Therefore, this paper proposes an integrated model for timetable and speed profile considering bidirectional converter. When an urban rail line is used as the optimization object, there is a problem of low efficiency in solving large-scale cases. To solve this problem, an algorithm combining Q-learning with particle swarm optimization (PSO) is proposed. Finally, the proposed method is validated by simulation using the actual data of Beijing Subway Yanfang Line. The results show that Q-learning particle swarm optimization (QPSO) has better global search characteristics. Compared to the original model without the bidirectional converter, the integrated model considering bidirectional converter increases the regenerative energy utilization from 33.9% to 91.0% and reduces the net energy consumption by 37.8%.