Research On Modeling And State Of Charge Estimation For Power Battery

The power battery is the energy source of the new energy electric vehicle.The battery management system manages the power battery safely and effectively,and is a key technology for the development of the electric vehicle industry.In order to make the power battery meet the requirements of electric vehicles for the cruising range and safe driving under the actual complicated conditions,it is necessary to study the model and critical SOC estimation of power battery.The paper studies the model and SOC estimation of lithium-iron phosphate battery for electric vehicles.The main work is as follows:(1)The power battery test platform is built for the power battery characteristic test,in which Arbin EVTS was used to collect test data for the lithium-iron phosphate power battery.The terminal voltage characteristics,current rate characteristics,open circuit voltage characteristics,ohmic internal resistance characteristics and working condition characteristics of lithium-iron phosphate battery are analyzed,and the experimental data is provided for the power battery modeling,parameter identification and state estimation.(2)The battery equivalent circuit model is the foundation of SOC estimation.Aiming at the problem that the battery equivalent circuit model does not consider the measurement current and voltage error,the first-order Thevenin equivalent circuit model based on the controllable autoregressive moving average(CARMA)structure is proposed to effectively reduce the influence of the external measurement error to improve the accuracy of the model.Then,according to the time-varying characteristics of the model parameters,the variable forgetting factor is added to identify the model parameters based on the recursive least squares algorithm.The adjustment of the adaptive forgetting factor makes the model parameters match the dynamic changes characteristics,and thus improving the stability of the model parameters.The simulation results show that the CARMA model and the variable forgetting factor recursive least squares algorithm can obtain the model parameters more accurately.(3)The characteristic of the open circuit voltage-state of charge(OCV-SOC)for the lithium-iron phosphate battery is used for describing the relationship between the SOC and the OCV in the battery equivalent circuit model.However,the OCV-SOC curve of the lithium-iron phosphate battery has a local flat characteristic,which leads to an amplification of the influence of the OCV error on the SOC estimation.To solve this problem,an open circuit voltage(OCV)feedforward compensation strategy is proposed to improve the OCV-based SOC estimation accuracy.The OCV is feedforward compensated by adjusting the terminal voltage deviation based on the trimming factor of the slope of the OCV-SOC curve.(4)The battery equivalent circuit model has the nonlinear characteristics and is subject to the external random interference.In the state estimation by Kalman filter,the nonlinear observation function approximation,the probability density approximation of the nonlinear observation function and the interference suppression are studied,respectively.Aiming at the nonlinear observation function approximation problem of single-innovation Kalman filter in the measurement update,a battery SOC estimation method based on the multi-innovation extended Kalman filter is proposed.Aiming at the approximation of the probability density of the nonlinear observation functions,a battery SOC estimation method based on Gauss-Newton iterative cubature Kalman filter is proposed.For the interference suppression,the observation variance matrix and gain matrix are adjusted by selecting the robustness factor to attenuate the influence of external noise on the battery SOC estimation.(5)Due to the difference of the lithium-iron phosphate battery under the same process conditions,the low-order equivalent circuit model of the battery cannot cover the complete dynamic and static characteristics of the battery,and the uncertainty of the equivalent circuit model is unavoidable.Aiming at the model error,the H-infinity filter method based on the minimum maximum error criterion is introduced to estimate the battery SOC.In order to reduce the influence of the statistical characteristics of unknown noise on the Hinfinity filter method,the state noise variance and the measured noise variance are updated by using the adaptive criterion based on the maximum likelihood estimation.The strong tracking filter is introduced into the adaptive H-infinity filter.The suboptimal attenuation factor is added to the a priori estimation variance.The correlation quadrature information in the estimated residual sequence is used to update the state variance to enhance the tracking ability of H-infinity filter on the battery SOC.Considering the correlation between OCV and capacity,a capacity error compensation strategy is introduced to improve the accuracy of capacity estimation.In summary,based on the first-order Thevenin model,the controllable autoregressive moving average battery model and the variable forgetting factor recursive least squares algorithm are used to identify the model parameters.It solves the problems of non-synchronous changes and identification fluctuations of power battery model parameters.According to the flatness characteristic of OCV-SOC curve,the SOC estimation accuracy based on the multi-innovation extended Kalman filter algorithm is improved by using the open circuit voltage(OCV)feedforward compensation strategy.The Gauss-Newton iterative method is used to iteratively update the SOC and state variance in the measurement correction phase of cubature Kalman filter,and adaptively select the robust factor to suppress external noise interference.For the low-order model error,the adaptive strong tracking H-infinity filter algorithm is used to estimate the battery SOC.For the view of various influencing factors,the proposed three methods effectively improve the accuracy and anti-interference for SOC estimation.The results obtained provide theoretical and technical reference for the efficient and safe operation of the power battery system.