Research On Co-Estimation Strategy Of SOC And SOH For Lithium Batteries Based On BP Neural Network Modification And Optimization

To deal with environmental deterioration and energy crises,developing clean and sustainable new energy has become the strategic goal of all countries in the world.Lithiumion batteries are currently the best solution for power and energy storage in the new energy industry and are also the main power source for new energy vehicles.State-of-charge(SOC)and state-of-health(SOH)are important indicators to measure the safety and efficiency of a battery management system.Therefore,an LMRLS-BP-DEKF synergistic prognostic model is proposed for the SOC and SOH of power lithium-ion batteries,and the following research is carried out.(1)Aiming at the problem of poor interference rejection ability and easy data saturation when there is too much data for the recursive least square algorithm and forgetting factor recursive least square algorithm,a limited memory recursive least square(LMRLS)algorithm in this study that removes the old data and uses new data with only the restricted length is designed to effectively improve the accuracy of on-line parameter identification.(2)Considering the coupling effect between SOC and SOH,a collaborative prediction system of the dual extended Kalman filter(DEKF)is constructed with ampere hour integral as the bridge.Where the first EKF realizes the estimation of SOC,and the estimation result is taken as the input of the second EKF to realize the progressive estimation of SOH.Through the mutual correction feedback of the two,a closed loop is formed to achieve the collaborative estimation of the SOC and SOH.(3)To address the problem that the DEKF algorithm neglects the higher order terms of the Taylor expansion when dealing with non-linear systems and there is a model error,a back propagation(BP)neural network is introduced in this study.With the powerful self-learning ability and nonlinear processing ability of neural network,the errors of DEKF model can be compensated and corrected to further improve the accuracy of SOC and SOH estimation results.(4)To verify the effectiveness and adaptability of the LMRLS-BP-DEKF synergistic estimation algorithm model,a collaborative SOC and SOH prediction experimental verification scheme is designed.The proposed algorithm model is verified by simulation experiments under different temperatures and different initial values of estimation state,and the estimation results are compared and analyzed with the traditional algorithm.Experimental results show that the LMRLS-BP-DEKF synergistic prognostic algorithm model proposed in this study can effectively improve the accuracy and robustness of SOC and SOH co-estimation of power lithium-ion batteries.Moreover,it has certain adaptability and correction ability for complex working conditions under different temperatures and different initial values of estimation state conditions,which not only lays a theoretical foundation for battery condition monitoring but also provides a guarantee for the safe application of lithium batteries and the safe driving of electric vehicles.