Modeling And SOC Estimation Of Lithium-ion Battery Coupling Model Based On Electrochemical Mechanism

Under the dual stimulation of environmental pollution and depletion of oil resources,the world began to vigorously develop new energy electric vehicles.Power battery as the core of electric vehicles,battery performance has a huge impact on the range,maximum power and safety of the car,and lithium-ion batteries are widely used following their excellent performance among many power batteries.Battery management system(BMS)can effectively monitor the status of the battery and manage the energy of the battery.As one of the core functions of BMS,the accurate estimation of battery state of charge(SOC)can effectively avoid the overcharge and overdischarge of the battery,thus prolonging the battery life.SOC estimation relies on accurate battery model,the traditional equivalent circuit model is simple in principle but does not reflect the electrochemical mechanism and has average accuracy,while the electrochemical model has high accuracy but is computationally intensive and cannot be directly applied online.Although the accuracy of electrochemical models is high,they are computationally intensive and cannot be applied directly online.Based on this,this paper aims to investigate the SOC estimation method for Li-ion batteries with the following four main research aspects:(1)Firstly,starting from the mechanism of the battery model,the partial differential equations are simplified by deriving the end-voltage equation of the electrochemical model of the battery and using the Pade approximation polynomial method,and the end-voltage equation of the simplified electrochemical model is derived.Based on this,a coupling model based on the electrochemical mechanism is established by combining the equivalent circuit model and the electrochemical model.(2)The LM(Levenberg-Marquarat)nonlinear least squares algorithm was used to identify the electrochemical parameters involved in the coupled model;and the parameter identification results were input into the coupled model and compared with the equivalent circuit model under the current conditions of HPPC,UDDS,FUDS,and DST,respectively,to verify the results.The results show that the coupled model can significantly improve the accuracy of the model,with the average absolute errors of25.4m V,18.6m V,28.4m V and 24.7m V under several current conditions,respectively;the accuracy of the parameter identification results is also verified.The analysis of parameter sensitivities provides a more in-depth description of the electrochemical mechanistic properties retained in the coupled model.(3)Based on the results of coupled model and parameter identification,the traditional Kalman filter algorithm is improved by using the open-circuit voltage OCV obtained from the electrode Li~+concentration relationship instead of the OCV fitted value obtained by experiment,which largely avoids the influence of OCV on the SOC estimation method and forms a new SOC estimation method.After the simulation verification of two current conditions of HPPC and DST,the improved SOC estimation method can significantly improve the SOC estimation accuracy;secondly,the comparison results of running time demonstrate that the improved algorithm only loses less time and has good real-time performance.(4)Finally,based on the designed experimental DST current condition,the simulation process is experimentally verified and the electrochemical mechanism embedded in the coupling model is analyzed.The results show that the proposed coupling model can effectively retain the electrochemical mechanism information and has higher accuracy compared with the equivalent circuit model,with the 90th percentile error reduced by 30 m V and the maximum error reduced by 60.7 m V,which has important research value for future high-precision battery modeling.On this basis,the improved SOC estimation method well improves the SOC estimation error of Li-ion batteries due to inaccurate SOC-OCV curves,especially the EKF algorithm has the greatest improvement,with the average error reduced by 1%and the SOC estimation error effectively improved.Although the electrochemical mechanism is added,it does not increase too much running time and still converges quickly under different initial values of SOC,with better real-time and robustness,and can estimate the lithium-ion battery SOC well online.