Study On High Power Li-ion Battery Ageing Characterization And State Estimation Algorithms For Vehicles

Energy crisis and environmental issue have driven the electrified transformation in automobile industry.Considered as one of the solutions for further sustainable transportation,the industrial development of electric vehicles?EVs?has attracted consistent attentions.As the main energy sources of the EVs,to ensure batteries are reliable,efficient,and capable of providing power when required,an accurate determination of State of Charge?SOC?and State of Health?SOH?is necessary.Historically,the li-lion battery application for EVs had been along with the discussion of state estimation issue.Although abundant researches were concerning about this topic,still some works were lacking:First,for the study of model-based SOC estimation,the main concern of the existing literatures is focus on the mathematical descriptions of the model,few of them emphasize the effect of parameter identification.In fact,for the automobile applications,the model complexity always need to tradeoff with the computationally lightweight,these are always achieved at the expense of model accuracy.In contrast,accurate model parameters identification under various uncertain factors such as different driving loads and operation environments would be more meaningful for SOC estimation accuracy.Second,for the study of SOH estimation,the issue is usually converted to the estimation problem of some certain parameters of the battery.The systemic evaluation of SOH under multiple constraint indicators are still lacking.In view of the shortages in current study,this dissertation focuses on the LiNi_xCo_yMn_1-x-yO₂?NCM?battery.A series of issues including battery modeling,parameters identification and model-based SOC/SOH evaluation algorithms have been studied,the main contents are as below:?1?Study on control-oriented equivalent circuit model?ECM?and its online parameters identification.First,in view of the hysteresis characterization of NCM batteries,am one-state hysteresis voltage ECM has been proposed.This model can achieve higher estimation accuracy without increasing model complexity too much.Second,in view of the complex and time variable characteristic of EVs'actual operation environment,a mixed optimization method based on particle swarm optimization?PSO?-genetic algorithm?GA?has been proposed.The parameters of the one-state hysteresis ECM is identified online by using the proposed algorithm.Lastly,in view of the nonlinear and mutation state characterization in real-world application,a strong tracking filter?STF?has been proposed for accurate SOC evaluation.Additionally,the accuracy and robustness under various uncertain factors are investigated by simulation.?2?Study on the ageing characterizations and mechanisms of the NCM batteries.First,a set of orthogonal experiments have been designed considering the ageing stress factors of temperature,discharging/charging rate,and end-of-charge voltage.Second,the degradation characterizations of NCM battery performance such as capacity,resistance and voltage are revealed according to the test data.Lastly,the ageing mechanisms under each stress factors are identified and analyzed by employing the constant current charging data.The results lay the data and theoretical foundations for the following SOH model?3?Study on two typical SOH models.On one hand,the classic Pseudo two Dimension?P2D?model is generalized slightly here to allow to model with composite electrodes materials.The transient process of NCM battery is simulated by the model.On the other hand,an ageing mechanism-based semi-empirical model is built up.Consequently,a dual timescale battery key state evaluation algorithm framework has been proposed.The proposed framework consists of the SOC estimation in micro-timescale and model parameters update in macro-timescale.The SOH is evaluated from two perspective including capacity fade and resistance increasing.In the end,the SOC/SOH joint state estimation is achieved.