Study On Lithium-ion Battery Temperature Dependent Modeling And State Of Charge Estimation In Electric Vehicles

At present,electric vehicles(EVs)have become the development mainstream of the world's automobile industry because of their advantages,such as energy clean and environmental friendly.Among them,hybrid vehicles have been commercial,and plug-in hybrid vehicles and pure EVs have also occupied the automotive market gradually.Power battery is one of the core component and main power source of EVs and it is neccessary to monitor and manage the battery reasonably and perfectly.In the battery management system(BMS),the state of charge(SOC)of power battery represents the percentage of the remaining capacity to the rated capacity,and it is a critical parameter in battery monitoring and management.The accurate SOC estimation is the basis of energy optimization algorithm of EVs.Power battery is a complex nonlinear time-varying system due to the internal complex chemical reactions,and the battery external characteristics are changing,such as current,voltage,capacity and SOC.In the application of EVs,the battery life,safety and performance are affected by operating conditions,ambient temperature and other factors.However,it is difficult to ensure that the batteries are always working in the good temperature range.Therefore,the study of the battery modeling and the SOC estimation algorithm affected by temperature has theoretical and practical value.For Lithium-ion phosphate battery used in EVs,the following research is carried out in this paper.Firstly,the development of EVs and power battery is introduced,and the functional structure of BMS is expounded.The difficulties and significance of SOC estimation are analyzed.This paper summarizes the research status and development trend at home and abroad of the SOC definition,SOC estimation method,the factors influencing SOC estimation accuracy,battery model and SOC estimation algorithm simulation platform.The temperature dependent battery models and SOC estimation methods are summarized.Most of the existing research about battery model and SOC estimation is discussed under constant temperature condition,and the battery model parameters are fixed.And some researches that considered temperature effects are using look-up table or fuzzy logic method,but the accuracy needs to be improved.Next,the 26650 LiFePO4 battery,which is commonly used in EVs,is considered as a single particle in this paper.A temperature controllable battery test platform is set,a series of calibration test is carried out to measure battery capacity,charging efficiency,open circuit voltage(OCV),ohmic resistance,polarization capacitance and polarization resistance under different temperature.According to the experimental data,the battery external characteristics along with temperature are analyzed.The results show that the battery parameters(available capacity,charging efficiency,battery resistance)obviously change with the temperature.And OCV-SOC characteristic curve has significant hysteresis in the charge and discharge process.Therefore,based on the electrochemical principle of the battery and the analysis of the classical battery model,this paper focuses on the ambient temperature and hysteresis influence of Lithium-ion battery parameters.The dual polarization(DP)model of Lithium-ion battery with hysteresis and parameter changing with temperature is established.According to the experimental data,the model parameters are identified at each temperature by fitting the relation between the parameters and the temperature,which can offset the scatter of data and reduce the error caused by measured data.In the Matlab/Simulink environment,the DP model considering the hysteresis and the parameter changing with temperature is built,and it is compared with the DP model without considering the hysteresis and the parameter temperature changing respectively.The accuracy of the model under different ambient temperature and charge/discharge switching conditions is verified,which solves the problem of poor adaptability and inadequate accuracy caused by the fixed parameters.Then the classical SOC estimation methods are analyzed synthetically.For the nonlinearity of the model in this paper,the SOC estimator based on H? observer method is designed.But it is not ideal in dealing with the details of the current mutations.And the adaptive sliding mode observer(ASMO)based SOC estimator is built with the updated observer gain matrix in real time.ASMO realizes the accurate SOC estimation of Lithium-ion battery under different ambient temperature and charge/discharge switching conditions.The SOC estimation model is established in the Matlab/Simulink environment,to compared the robustness and effectiveness of SOC estimation method based on ASMO and H? observer at different ambient temperature and charge/discharge switching conditions.Finally,the SOC estimation is co-simulated by electric vehicle model in AMESim and battery SOC estimation model in Simulink.Combined with the model of SOC estimation algorithm based on ASMO established in Simulink,the effectiveness and accuracy of the SOC estimation algorithm under actual driving conditions are verified by different actual vehicle conditions.xPC Target based semi-physical simulation is applicated to collect the real battery data in real time and estimate SOC.The validity,accuracy and reliability of the proposed model and algorithm are verified in the hardware-in-the-loop simulation condition,and the application of SOC algorithms realizing in real EVs is promoted.