SOC estimation algorithm based on improved PNGV model

Based on the improved PNGV equivalent circuit model for LiFePO₄ battery, a method of SOC estimation combining Kalman filtering and AH integration is put forward. Considering the effects of temperature and self-discharge on model parameters, a simulation model is set up with Matlab/Simulink. The results of comparison between battery SOC values estimated with the combination of Kalman filtering and AH integration and sole AH integration show that improved PNGV model can actually reflect the characterizations of battery and accurately estimate battery SOC in allowable error range.