Optimal sizing of AC railway power conditioner in autotransformer-fed railway power supply system

Abstract

This paper proposes a procedure to determine the optimal railway power conditioner (RPC) size in an autotransformer-fed traction power supply system, using the current-based Newton-Raphson power flow calculation in combination with the particle swarm optimisation (PSO) method. The PSO is chosen because of searching capability for highly non-linear objective functions, and having better performance in this application comparing to the Genetic Algorithm (GA). The optimisation-based RPC steady-state model is also introduced in the power flow calculation. The optimal capacity is designed to accommodate an overall range of traction load, regardless of overload tolerance and a period of overload; moreover, it is designed on the basis of partial compensation. The East Corridor commuter rail line in Denver is used as a case study for the proposed procedure. The power supply of the line is modified to have more severe power quality problems than the original ones, and two different load levels are simulated for optimal size searching. The results show that the optimal power capacities were successfully obtained with the power factor and voltage unbalance factor within the permissible levels; the optimal values were obviously smaller in comparison with the full compensated ones.