VRLA Battery SOH Estimation Based on WCPSO-LVSVM
The correlation between operation parameters including coup de fouet and SOH was analyzed to choose the input parameters of the battery SOH estimation model, and then the battery SOH estimation model was made based on least square support vector machine (LSSVM). For more prediction accuracy and efficiency, the advanced particle swarm optimization (WCPSO) is used to optimize the parameters of the LS-SVM regression model. The battery SOH was estimated only with measured data (plateau voltage, discharge rate and temperature) of short time discharge, so it is efficient. The verification result shows that the WCPSO-LSSVM model can be used to predict the battery SOH, and the precision is above 93%.
H. M. Zhuang and J. Xiao, "VRLA Battery SOH Estimation Based on WCPSO-LVSVM", Applied Mechanics and Materials, Vol. 628, pp. 396-400, 2014