Lithium Battery SOC Estimation Based On Improved ASRCKF Algorithm

After entering the 21 st century,due to the increasing problem of environmental pollution and energy shortage,the concepts of green environmental protection,energy saving and low carbon are gradually popularized.In this context,new energy electric vehicles will gradually replace fuel vehicles as the mainstream means of transportation in the next few decades.As the core of new energy electric vehicle,battery technology determines its production cost,power performance and mileage.State of charge（SOC）estimation is the key part of battery technology,which affects many aspects such as the working strategy,safety performance and cycle life of battery,and has important research value.In this dissertation,the SOC prediction of ternary lithium batteries with the highest market share is studied and analyzed.The main work is as follows:First of all,the configuration,working principle and several main performance indexes of lithium-ion battery are briefly described.Experiments were designed to study the capacity characteristics,discharge rate characteristics,temperature characteristics and cycle life characteristics of the sample battery,and the effects of several characteristics on SOC estimation were analyzed.Afterwards,the common battery models are compared and analyzed,from which the2RC-Thevenin equivalent circuit model is selected.The fitting and identification of battery model parameters were completed through pulse discharge experiment,and the battery model was built in Matlab/Simulink simulation environment.The accuracy of the battery model is verified by constant current discharge and user-defined dynamic discharge conditions,and the absolute error is less than 0.04 V.In the next part,the principles and characteristics of four kinds of kalman filters are described and analyzed.Based on the CKF,the improved Sage-Husa adaptive filtering algorithm is integrated to improve the filtering accuracy through noise adaptation,and the square-root filter is integrated to improve the stability of the algorithm program.The improved ASRCKF is formed,and the algorithm flow is designed to estimate the SOC of battery.Finally,the convergence performance and filtering accuracy of the improved ASRCKF algorithm are verified by DST condition,UDDS condition,constant current discharge and user-defined dynamic discharge condition in Matlab/Simulink simulation environment and field experiment environment respectively,and compared with EKF,UKF and CKF algorithm.Experimental results show that the improved ASRCKF algorithm has obvious advantages over other algorithms,which not only has good convergence performance,can converge quickly under different initial errors;But also has high filtering accuracy and the prediction error is less than 2.7%.