SOC Estimation Algorithm Study For Lithium-Ion Battery

In response to the concerns of the energy depleting crisis and environment protection,many governments are actively promoting changes in the drive system of vehicles.Among them,the electrical vehicle is an effective method.In electric vehicles,power battery is one of the most important factors that can affect the performance of electric vehicles.The storage and release of battery energy is accomplished through the internal chemical reaction.The battery management system(BMS),which is particularly important for monitoring the working status of the battery,mainly undertakes the task of monitoring and managing the power battery.The main task of BMS for electric vehicles contains the estimation of State of Charge(SOC)of the battery.Accurate SOC estimation is prerequisite for a safe and effective journey of electric vehicles.This dissertation considers the characteristics of the battery equivalent circuit model,relies on the fractional calculus theory,establishes a new fractional-order battery model,based on fractional calculus theory,and explores fractional-order SOC estimation algorithms to improve the accuracy of SOC estimation.This dissertation takes a ternary lithium battery as the research object and carries out the following work:A detailed analysis of the research background and research status of BMS and SOC estimation for electric vehicles is presented.The research status of fractional calculus is introduced as well.The electrochemical reaction principle of ternary lithium batteries and the multiple performance indicators of lithium batteries are analyzed as well as the dynamic characteristics of the battery.The charge and discharge experiments of the selected Sony US18650GR G7 ternary lithium battery were performed.The experiments included pulse power characteristics test,static condition experiment with constant discharge current of 1C,and dynamic condition experiment of UDDS with irregular current magnitudes and cycles.Experimental data of lithium battery load current and battery terminal voltage in three experiments is recorded.A fractional second-order RC equivalent circuit model is established based on the fractional-order capacitance model.The lithium battery is described as a state-space equation that can express the characteristics of the battery.The least-squares method is used,combined with the pulse power characteristic test curve,to identify the parameters of the model,including fractional orders.The parameter identification results show that the model estimation of terminal voltage approximates the actual terminal voltage in the pulse power characteristic test well.In this dissertation,the characteristics and deficiencies of the standard Kalman Filter and the Extended Kalman Filter are analyzed in detail.Then the Unscented Kalman Filter is introduced and the derivation of the Unscented Kalman Filter,which is used for the fractional-order circuit model,is proposed as fractional Unscented Kalman Filter.After that,the SOC value of the lithium battery is reckoned with load current data and terminal voltage data,which is recorded under static and dynamic conditions.The comparison of estimation value with the actual measurement value is performed.The result shows that the proposed fractional Unscented Kalman filter has a more accurate estimation outcome than integer Unscented Kalman Filter.H_? observer is also introduced in this dissertation.With the parameters of the integer second-order RC model,an H_?observer is designed.The SOC of the lithium battery is estimated under two operating conditions.The results show that the H_?observer also estimate the battery SOC accurately and it has good robustness.