Methodology for Optimal Sizing of Hybrid Power System Usingparticle Swarm Optimization and Dynamic Programming

Abstract A methodology for optimal sizing of hybrid battery-ultracapacitor power system (HPS) is presented. The purpose of the proposed methodology is to locate the optimal voltage levelfor HPS used in a plug-in hybrid electric vehicle (PHEV). A combined optimization framework for a HPS is proposed and the optimization problem is solved in a bi-level manner. The framework contains two nested optimization loops. The outer loop evaluates the selected parameters throughparticle swarm optimization (PSO) algorithm, while the inner loop generates the optimal control strategy and calculates the costs through dynamic programming (DP) algorithm. The Chinese Typical City Bus Drive Cycle (CTCBDC) has beenused to verify and evaluate the performance of the proposed methodology. The optimization result shows that higher voltage degree usually means better performance and the battery tends to provide a constant power for the HPS. It is noted that the constant powercloses to the high efficiency district of the battery and DC/DC convertor. After that the optimal result is further analyzed undervarious optimization goals andbattery charge/discharge current constrains.