The flexible DC traction power supply system  has lower losses and higher efficiency compared with the traditional AC railway electrification system. Meanwhile, it has the advantages of easy access to new energy and energy storage equipment. However, increased new energy and energy storage ports will be directly coupled with traction substations and locomotives through the DC traction network, and there are dynamic changes in the position of multiple locomotives. This will significantly increase the complexity of power flow analysis and the computing time. Thus, in this paper, an optimized power flow computation algorithm is proposed. Firstly, the mathematical model of the flexible DC traction power supply system is established for the power flow algorithm. Then, the advantages of the parallel computing method are analyzed. The computation model is validated by comparison with simulation results. Besides, considering the limitation of the parallel computing method, an optimization method for algorithm iteration process based on fuzzy neural network (FNN) is proposed in this paper. According to the calculation results, the computation time is reduced from 3.214 seconds to 2.673 seconds under FNN-based method.

Alternating Iteration, Flexible DC Traction Power Supply System, FNN, Power Flow Computation