Research On The Joint Estimation Method Of The Lithium-ion Battery SOC And Capacity In The Whole Life Cycle

Accurate battery state of charge（SOC）estimation is always the most important and urgent problem for electric vehicles,which is also very important for the normal and safe use of Lithium-ion batteries.Meanwhile,it can improve the battery energy utilization efficiency,reduce the use costs and prolong the cycle life of Lithium-ion batteries.However,due to the strong nonlinearity of the battery and the influence of self-discharge,temperature and noise,it is difficult to guarantee the accuracy of the estimated battery SOC.In addition,the accurate estimation of battery SOC is highly dependent on the accurate maximum available capacity(C_max)of the battery.However,the battery C_max will decay continuously with battery aging,and it is difficult to measure directly in the working process.Therefore,this paper mainly studies the joint estimation of battery SOC and capacity in the whole life cycle,some works have been done and summarized as follows:（1）Battery modeling and parameter identification.A series of battery tests were implemented to study the charging-discharging characteristics of lithium-ion batteries under different aging cycles,current ratios and temperatures.To reflect the dynamic behavior of batteries in a real situation,a hybrid battery model considering the dynamic characteristics of battery capacity was established on the basis of the second-order RC model.Besides,an improved exponential decay particle swarm optimization（EDPSO）was proposed to identify the open-circuit voltage and electrical impedance of the circuit model online.（2）Prediction of the battery future working conditions.Most of the existing researches are executed under an assumption that the future conditions of the battery are priori known.But in the practical applications,the loading current profiles are extremely uncertain.Therefore,a Markov chain model（MCM）is established to predict the future driving conditions of vehicle,and the battery future load is then obtained based on a validated EV system model.（3）Joint estimation of the battery SOC and capacity in the whole life cycle.A discrete nonlinear proportional integral observer（PIO）was established based on the proposed hybrid battery model to estimate the battery SOC.In addition,considering the slow time-varying characteristics of the battery capacity,it is online estimated using the accumulated charge and the variation of battery open-circuit voltage（OCV）and unusable capacity(C_not)under a fixed time segment,which helps to update the SOC estimation at different aging cycles.Finally,complex urban conditions under different aging cycles were carried out to verify the accuracy and robustness of the EDPSO-PIO method.The results show that the proposed method can well adapt to the situations with different temperature and different aging cycles,the error of the estimated battery SOC and maximum available capacity(C_max)can be controlled within 2%in the whole life cycle.Moreover,the efficiency of the proposed EDPSO-PIO method is more than 30%higher than that of the Kalman filter and its extensions,which indicates that the proposed EDPSO-PIO method has better real-time performance and is more suitable for the real-time battery SOC estimation via the embedded system.