Research On SOC Estimation Of Li Ion Battery Based On H_∞-EKF Hybrid Filtering Algorithm

In recent years,due to the fossil energy crisis and environmental pollution,electric vehicles have developed rapidly.A good battery management system can ensure the safety of the battery and maintain the stable operation of the vehicle.The accurate estimation of the battery’s state of charge(SOC)is critical to the performance of the BMS system.This article focuses on improving the estimation accuracy of lithium-ion battery SOC,and has done the following research.1)The composition structure,reaction principle and related battery technical parameters of lithium-ion batteries are analyzed,and common battery models such as electrochemical model,thermal model,and battery equivalent circuit model are compared and studied.Considering the accuracy and complexity of the model,this paper adopts the Thevenin battery model as the battery equivalent model in this paper.In order to reflect the dynamic changes of model parameters,an online parameter identification method with forgetting factor recursive least squares algorithm is introduced.The accuracy of the battery equivalent model is simulated and verified under the pulse constant current discharge experiment.The terminal voltage output by the battery model is used as the simulation value,and the terminal voltage measured by the multimeter is used as the real value.The simulation results are more consistent with the real value.The maximum error of the voltage is less than 0.05 V,which proves that the accuracy of the model is very high.2)A comparative analysis of common lithium-ion battery SOC estimation methods is carried out,and the extended Kalman filter algorithm is used to estimate the SOC based on the battery-based state space model to achieve the optimal parameter estimation under the interference of white noise.Considering that the extended Kalman filter algorithm does not perform well against colored noise,especially in the environment with huge current changes,it is easy to cause the deviation of the estimation result.As a robust filtering algorithm,the H∞ filter algorithm has conservative estimation accuracy,but it is suitable for various Colored noise and model inaccuracy have better robustness.In actual applications,it is mostly mixed noise.In order to further improve the accuracy of SOC estimation,the paper designs a H∞-EKF weighted mixed filtering algorithm based on the Thevenin battery model.The weighting coefficient is determined according to the principle of minimum error covariance,which is compared with a single algorithm.,The hybrid filtering algorithm can take into account the advantages of both,while maintaining its robustness while improving the accuracy of SOC estimation.3)Build a SOC estimation platform for the SOC estimation function of the battery,design related hardware and software,and conduct performance tests on the completed platform.The results show that the designed hardware and software system has high data acquisition accuracy,which meets the requirements for data acquisition accuracy in SOC estimation.On this basis,the SOC estimation method under the condition of the discharge tester simulating constant current discharge has a higher SOC.Compared with the EKF algorithm and the H∞ algorithm for SOC estimation under the same conditions,the H∞used in this platform-EKF hybrid filtering algorithm has higher SOC estimation accuracy and better robustness,meets the requirements of electric vehicles for SOC,and has better application value.