Fusion of VMD and SABO-LSSVM for Lithium-ion Battery State of Health Prediction
The state of health(SOH)of lithium-ion batteries is an important indicator reflecting the health management of lithium-ion batteries.To solve the problems of inaccurate SOH prediction of lithium-ion batteries and that the parameters of the least squares support vector machine(LSSVM)model are prone to fall into local optimums,the paper proposes an LSSVM lithium-ion SOH prediction method by combining variational mode decomposition(VMD)and the subtraction average based optimizer(SABO)algorithm.It extracts the potential health indicator(HI)from the charging and discharging processes that contain battery degradation information firstly,analyzes the correlation between HI and capacity using the grey relation analysis(GRA)method secondly,decomposes the HI into a series of modal components by VMD and inputs each modal component regarded as a separate subsequence into the LSSVM of the SABO optimization respectively,and superimposes&reconstructs each subsequence of the predictions and evaluates errors finally.The experimental validation is carried out using four battery data provided by NASA and the battery data of CALCE from University of Maryland is additionally selected to verify the adaptability of the method.The results show that the method has high prediction accuracy and RMSE increases by 69.8%,86.9%,and 78.1%respectively compared with the VMD-LSSVM,LSSVM,and VMD-SABO-SVM models.
Lithium-ion batteryVariational modal decompositionLeast squares support vector machineSubtraction average based optimizerState of health
万方数据
维普
