Research On Degradation And Remaining Useful Life Prediction Of Lithium-ion Battery For Electric Vehicles

Due to high energy density and low self-discharge rate,lithium-ion batteries are widely used in electric vehicles,electrical products,wind and solar energy storage devices,military communications,and other equipments.With the development and progress of science and technology,lithium-ion batteries are required to have better durability,reliable safety and relatively long service life.The degradation and failure of battery will cause fire,explosion and other safety incidents.With the increase of charging and discharging cycles as time goes on,the capacity and internal resistance of batteries will deteriorate,as a result,batteries can not meet the required standby time.Therefore,it is necessary to monitor the actual use of battery at real time,estimate the charged and healthy state of the battery residual power,ensure the safety of the battery and predict the remaining lifetime of the battery.In this paper,the performance degradation and residual lifetime prediction of lithium-ion batteries are studied by using Samsung 18650 nickel-cobalt-manganese ternary lithium battery and Xinwei high-precision battery detection system(BTS).Firstly,an equivalent circuit model is constructed to identify the off-line parameters of the second-order RC equivalent circuit model of lithium-ion batteries,and a simulation is used to verify the accuracy of the circuit model.Recursive least squares method with genetic factors is used to identify the parameters of the first-order RC and the second-order RC equivalent circuit models on-line,respectively.The identification results of the model parameters are verified under dynamic stress test conditions.Then,the degradation mechanism of lithium-ion batteries during charging and discharging process is analyzed,and the degradation trend is tested.To hackle the factors affecting the safety and remaining service life of lithium-ion batteries,the internal structure material,electrochemical reaction mechanism and charge-discharge characteristics of lithium-ion batteries are analyzed.Characteristic parameters of charge-discharge characteristic curves are extracted,and the degradation rules of performance parameters,internal resistance characteristics,capacity characteristics and discharge capacity of test battery are given.Thirdly,an extended Kalman filter(EKF)method is proposed to estimate the batteries charging state.The batteries are charged and discharged at different rates,and training data of the charging state,the charging-discharging currents and the charging-discharging voltages under various conditions are obtained.The adaptability and feasibility of Kalman filter in estimating the batteries charging state are verified by experimental data.The dynamic stress power testing conditions are used to verify the batteries performance,and the batteries Charge-discharge characteristics are described by the second-order RC equivalent circuit model.A series of mixed stress tests are carried out on batteries under different charging conditions,and the equivalent circuit model is determined.Based on the multi-scale EKF algorithm,the state space model of batteries is established,and the SOC and SOH of batteries are estimated.Finally,the capacity increment is selected as the degradation amount from the test data of lithium-ion batteries performance degradation test,and the remaining life prediction model of lithium-ion batteries conforming to the Wiener random process is constructed.Combining with Bayesian rule,the test data of cyclic batteries charging-discharging test are preprocessed,and the maximum expectation algorithm is introduced to iteratively update the model parameters.The performance degradation test data are divided into training Samples and test samples to verify the availability of residual life prediction model.