Prediction Method Of Remaining Useful Life Of Lithium-ion Battery Based On SVR

Lithium-ion batteries are widely utilized in electric vehicles and energy storage systems due to their superior performance.However,with the continuous operation of lithium-ion batteries,aging of lithium-ion batteries has gradually occurred,and the performance of lithium-ion batteries deteriorates over time,which indirectly leads to the degradation of equipment performance or a catastrophic occurrence.Therefore,it is necessary to monitor the state of health and predict remaining useful life(RUL)of the battery in time,to obtain critical information of battery life in advance,maintain and replace the battery in time,and ensure the safe and reliable operation of the battery.This paper mainly studies the RUL prediction of lithium-ion battery,and mainly completes the following work:First of all,the process of the battery charge and discharge cycle is considered,and extract indirect health factors that can characterize the performance degradation characteristics of battery.This paper analyzes the public datasets of NASA,extracts three indirect health factors that can characterize the degradation of lithium-ion battery performance from the constant current stage and the constant voltage stage,and analyzes the correlation between health factors and capacity via using the correlation analysis methods of Pearson.Spearman,and Kendall.To verify the feasibility of the extracted indirect health factors.Secondly,a method based on ant lion optimization(ALO)and support vector regression(SVR)is proposed to improve the accuracy of RUL prediction of lithium-ion batteries.In order to solve the problem of difficulty in selecting the parameters of SVR method,the ALO algorithm is employed to optimize the parameters of SVR method,the ALO-SVR method based on health factors is proposed.The battery datasets of NASA are utilized to verify the ALO-SVR method,compared with the SVR method,the experimental results show that the ALO-SVR method can more accurately predict the RUL of lithium-ion battery.Finally,in view of the problem that the ALO algorithm easily falls into the local optimal solution during the iteration process and cannot provide the optimal parameters for the SVR method,this paper proposes an improved ALO(IALO)algorithm based on the Levy flight algorithm.The introduction of the Levy flight algorithm can solve the shortcoming of the ALO algorithm well,the IALO algorithm can provide optimal parameters for the SVR method.The test functions are utilized to verify the IALO algorithm has better optimization effect and convergence accuracy.The I ALO-SVR method based on health factors is verified via the battery datasets of NASA,the experimental results show that the SVR method based on the IALO optimization algorithm effectively improves the prediction accuracy.To further analyzes the parameter optimization for SVR method,the result shows that the IALO-SVR method can provide smaller and better parameters,and can be provided higher prediction accuracy for RUL of lithium-ion battery.The IALO-SVR method based on indirect health factors proposed in this paper can improve the accuracy of RUL prediction of lithium-ion batteries and has reference value for battery system maintenance.