Multifunctional Estimation And Analysis Of Lithium-ion Battery State Based On Data Model Fusion Under Multiple Constraints

Accurate battery state estimation is essential for ensuring the security and efficiency of lithium-ion batteries.Generally,the signal detected by a sensor inevitably has noise and drift.Therefore,the prediction process will indefinitely accumulate this random error,resulting in lower estimation accuracy,which is particularly prominent when the performance of vehicle sensors is poor.To overcome the problems of over-idealised estimation results,low efficiency,and insufficient reliability of traditional battery state estimation methods,this study proposes a multifunctional estimation and analysis model of battery state of charge(SOC),battery capacity,and state of power(SOP),based on data model fusion.The main works of this thesis are as follows:First,a data-driven multi-scale extended Kalman filter(MDEKF)was used to de-noise the original data with random errors observed by the sensor in each cycle.The de-noised data were input to a temporal convolutional network(TCN)as training samples,and the estimation model was obtained through TCN neural network machine learning.Furthermore,a peak power estimation method based on multiple constraints was used.The accurate SOC estimation results obtained through the TCN network were used to describe and enhance the relationship between the SOC,voltage,and peak power.In this way,it avoids the disadvantage that the TCN depends too much on the accuracy of training data and retains the advantages of the high robustness of the MDEKF and strong nonlinear characteristics of the TCN network.The proposed method is compared with an TCN method without pre-processing the input data on the Dynamic stress test(DST),the New European Driving Cycle(NEDC)and the Worldwide Harmonized Light Vehicles Test Cycle(WLTC)datasets.Subsequently,the experimental results are analysed and compared with other algorithms under various initial noise,temperatures,and unknown initial SOC conditions.The results show that the accuracy of SOC estimation is improved a lot in the case of compensating the random input error by the MDEKF in advance in comparison with feeding it into a neural network algorithm directly,converging quickly under inaccurate initial conditions and having strong robustness against noise and other factors.The dangers of over-charging and over-discharging are effectively avoided,and the safety and reliability of lithium-ion batteries are improved.