Prameter Identification And State Of Charge Estimation Of Lithiium Battery Based On Adaptive Kalman Filter

With the popularization of electric vehicles and the promotion of related policies,research on vehicle batteries has gradually become a hot spot of concern.Battery SOC estimation is an important research direction in the field of battery research.Accurately estimating the SOC of the power source is critical to the car.Polymer lithium-ion batteries have the characteristics of high energy density,light weight and easy combination,and the curve of the open circuit voltage is also convenient for mathematical fitting.This paper is based on the research of polymer lithium batteries.Considering the Kalman filtering algorithm's filtering processing,it can gradually approach the real value under noisy interference and does not need much computational complexity.This paper estimates the lithium battery SOC based on the Kalman filtering algorithm optimization process.First,the external characteristic data acquisition experiment is performed on the experimental battery,and then the open circuit voltage curves under different charging and discharging states are analyzed.Based on these experimental data,the Thevenin equivalent circuit model is selected and the model is improved.The factors of charge and discharge state and SOC change are added and the matlab/simulink battery model is built.Under the pulse condition and the variable current condition,the accuracy of the improved model is higher than that of the original Thevenin model,which lays a foundation for subsequent research.Before estimating the battery SOC,it is necessary to identify the parameters of the battery model.Firstly,the offline identification is improved,and the factors of charge and discharge state and SOC change are added on the basis of off-line identification of fixed parameters.Secondly,combining the advantages of offline identification and data-driven online identification,two parameter identification methods combining offline identification and online identification are proposed.The matlab/simulink model of four parameter identification is constructed.Finally,the accuracy of the four methods is verified under the pulse condition and the variable current condition.It is confirmed that the parameter identification method combining offline identification and online identification also has certain significance and its precision is high.Finally,the conclusion is that the offline identification accuracy considering the charge and discharge state and SOC change is the highest.Based on the improved equivalent circuit model,the battery SOC estimation principle is analyzed,and the extended Kalman filter algorithm and the unscented Kalman filter algorithm are used to estimate the SOC model.Based on this,an adaptive algorithm model with noise update module and interaction multi-model are constructed to estimate the battery SOC.Three interactive algorithms,namely interactive multi-model-extended Kalman filter algorithm,adaptive extended Kalman algorithm and adaptive unscented Kalman algorithm,are built.Finally,the experimental results show that the accuracy of the adaptive algorithm is improved compared with the original algorithm in the five algorithms,and the adaptive extended Kalman filter algorithm is the optimal algorithm among all algorithms.Based on the combination of computational accuracy and computational complexity,the adaptive extended Kalman filter algorithm is the three algorithms with the least amount of computation and the highest computational accuracy.