SOC Prediction Of Ternary Lithium Lon Power Battery Based On Composite Algorithm

The new energy electric vehicle industry is one of the ten emerging industries in our country.The development and application of electric vehicles can effectively alleviate the global energy crisis and curb environmental deterioration,which is the main way to implement "environmental protection" in our country.However,in recent years,the development of electric vehicle industry has been hindered and battery technology has been slow.The main reason is that the development of electric vehicle power batteries has encountered new difficulties.The key problems hindering the development of electric vehicle power batteries include the energy density,cycle life and estimation of battery(state of charge,SOC).In order to solve the problem of battery state of charge estimation,with the support of the schoolenterprise cooperation project,the following work is carried out on how to establish an accurate battery model,design a reasonable SOC estimation strategy and realize offline and online simulation prediction.(1)Analyze the original data of charge and discharge test of ternary lithium battery and import them into excel.The abnormal data in the table are eliminated,incomplete data are supplemented,and the same data are classified.Finally,a new battery parameter information table is obtained,which is ready for subsequent identification of battery model parameters and verification of model accuracy,offline simulation of coupling estimation strategy and online prediction.The data processing software is used to generate various pictures related to battery SOC from battery parameter data,and the influencing factors of battery SOC estimation abnormality are separated from the pictures.(2)Considering the precision requirement of battery model,external environment current drift,noise interference and factors affecting battery SOC estimation,an electrochemical noise model with current,SOC and temperature as inputs is established.Recursive least square method with forgetting factor is adopted to import the corresponding data in the battery parameter information table to identify the unknown parameters in the battery model.The accuracy of the battery model is verified by the experimental data of ternary lithium battery and lithium iron phosphate battery under different working conditions and at multiple temperature points.The results show that the electrochemical noise model can better adapt to the lithium battery used in the experiment.(3)Considering that the battery is a complex,time-varying nonlinear system,the traditional estimation method of ampere-hour integration and open circuit voltage combination is difficult to accurately estimate the battery state of charge,so this paper adds particle filter algorithm to the traditional estimation method for coupling.In order to make the SOC estimation more accurate,the factors affecting the SOC estimation,such as battery capacity,differential pressure,temperature difference,etc.,are proposed to be corrected and compensated in the coupling algorithm.The coupling algorithm and the traditional estimation method are verified by off-line simulation through three periods of discharge data.The comparison results show that the estimation accuracy of the coupling algorithm is better than the traditional estimation method.(4)Based on the simulation and prediction model of coupling algorithm built in MATLAB/simulink module,the online simulation and prediction of SOC of ternary lithium battery are completed for different working conditions and different discharge rates.The results show that the prediction error of coupling estimation strategy is less than 5% under different working conditions,which proves that coupling algorithm has better applicability and higher estimation accuracy.