Research On SOC Estimation Of Lithium-ion Battery Based On VFFRLS-EKPF Algorithm

Taking lithium-ion battery as the research object,the following studies are carried out in this paper:(1)The second-order RC equivalent model of lithium ion battery is established.Then,the pulse discharge experiment is carried out based on Samsung INR18650-30 Q lithium-ion battery.Based on the experimental data,the OCV-SOC relationship curve is determined,and the parameters in the battery model are identified offline.Finally,the MATLAB simulation results show that the voltage error at the output of the second-order RC equivalent model is within-0.026V～0.057 V,which verifies the validity of the model parameter identification results.(2)The particle filter(PF)SOC estimation algorithm is analyzed in detail,and the EKPF algorithm combining extended Kalman filter(EKF)and particle filter(PF)is proposed for the particle degradation problem.This method uses the EKF algorithm and the current observation to update the particles of the PF algorithm,and achieves the effect of suppressing particle degradation.The MATLAB simulation results show that compared with the PF algorithm under the fixed battery model parameters,the EKPF fusion algorithm improves the accuracy of SOC estimation by 23%,and the convergence speed is increased by 117.2s.(3)Aiming at the problem that offline parameter identification cannot reflect the dynamic changes of model parameters in actual battery use,This paper first introduces the data windowing theory into the traditional variable forgetting factor method,and then proposes a variable forgetting factor recursive least squares(VFFRLS)online parameter identification method based on this,which can realize the real-time update of model parameters.Then combined VFFRLS and EKPF algorithm,and proposed a VFFRLS-EKPF battery model parameter and SOC joint estimation method.This method can realize the mutual update and correction of model parameters and SOC estimation results.Finally,the SOC estimation results of the VFFRLS-EKPF joint algorithm are compared with the SOC estimation results of the EKPF algorithm under fixed battery model parameters.The results show that the SOC estimation accuracy based on the joint algorithm is improved by 49%,and the robustness is better,which is suitable for engineering applications.(4)Based on the SOC estimation algorithm in this paper,a battery management system that complements photovoltaic energy storage and mains power is designed for water conservancy testing terminal equipment in a remote water source area.The data communication between the lower computer and the upper computer based on the RS485 bus is completed,and the accurate estimation of the battery pack SOC is realized in the upper computer software.The VFFRLS-EKPF joint SOC estimation algorithm and the designed battery management monitoring system proposed in this paper have certain practical application value and reference significance for the development of battery management systems in remote areas.