The SOC Estimation of NIMH Battery Pack for HEV Based on BP Neural Network

Abstract

The state of charge (SOC) is the most important parameter of power battery using in electric vehicles (EVs) and is the most difficult parameter to estimate. Considering of the nonlinear character of the power battery system, the back propagation (BP) neural network method is proposed in this paper. At first, divide the working range of SOC (25%-70%) into 3 parts, the low range (25%-40%), the medium range (40%-55%) and the high range (55%-70%). Then, aiming at the SOC estimation of the 3 parts, 3 models of BP neural network are established and trained by the typical battery data of charge and discharge. Before the estimation of BP neural network, primarily determine the range of the SOC value by the relationship between the open circuit voltage (OCV) and SOC based on the 4 situations of the battery, charging, discharging, laying aside after charging or laying aside after discharging. In the high range of SOC, the simulation results show that, the precision of SOC estimation can meet the requirement of HEV.