Research On Model Identification And State Of Charge Estimation For Lith-ium-ion Battery In Electric Vehicles

To cope with the increasing greenhouse effect and oil crisis, many countriesfocus on the research of electric vehicle (EV) for its zero emission. In the auto-motive field, electrochemical batteries are the core of energy source. Among var-ious kinds of batteries, Lithium-ion battery is an ideal type of battery due to itshigh energy ratio and high power ratio. State of Charge (SOC) of batteries is vitalto the performance of EV, and SOC estimation is one of the key techniques of thebatteries management system (BMS). This paper focuses on the battery modeland method of SOC estimation for lithium-ion battery.Firstly, based on the electrochemical impedance spectroscopy (EIS) theory,an impedance-based model was built using the results of lithium-ion battery EIStest. To the problem raised in the impedance-based model, a high-order equiva-lent circuit model (ECM) was introduced. This paper used EIS technique to in-vestigate the influence of battery temperature on the battery. The consequencescan amend the phenomenon that parameters of current model are constant, thusimprove the adaptability to the battery temperature change.Secondly, according to the real-time estimation requirements for the EV, themethod of state estimator was introduced. This paper verified observability of themodel based on the method of nonlinear system observability. The result showedthe battery model is observable, thus a state estimator exists. The overall designprocess includes poles placement and Lyapunov analysis. After pole placement isdone, the LMI condition resulting from the Lyapunov analysis must be checkedto verify the closed loop system is stable. According to the known model parame-ters, a real state estimator was set up. Finally, lithium-ion battery pulse current discharge tests at various tempera-tures implemented using battery test system to verify state estimator method. Re-sults show that even if Li-ion battery was at-20°C, the error using state estima-tor method was within±4%. It means this method has high accuracy. The re-sults of discharge test at-20°C show that at low temperature the resistance of li-ion battery rises significantly, thereafter battery voltage falls below the cut-offvalue and discharge ends.