Dynamic Performance Analysis and SOC Predication of Battery for Electric Vehicle

Real-time prediction of power batteries'residual capacity accurately is one of the key technologies of battery management system(BMS).In order to predict the residual capacity of battery accurately,charge and dis- charge performance analysis of MH-Ni battery for EV-3 developed by Hunan University has been made,and cur- rent,voltage,capacity characteristic curves have also been presented.The experiments show that the MH-Ni battery is suitable for EV.Based on the experiment data and the theory that the open circuit voltage is correlated with SOC, GRNN neural network of voltage descent model has been proposed.The BP neural network indicates the correlation between open circuit voltage and SOC,and the difference between the net prediction result and the experiment data is small.Tests and simulation results show that the neural network can build an effective prediction system of SOC for batteries.