Research On Health Assessment And Fault Diagnosis Of Lithium-ion Battery Based Energy Storage System

As the key component of the new energy electric vehicle(EV)and energy storage grid(ESG),the energy storage system(ESS)plays the role of energy storage and buffer.Lithium-ion battery is the first choice for electrochemical energy storage because of its high energy density,memory-free effect,and other superiorities.However,lithium-ion battery is a complex electrochemical system,and its electrical,thermal and aging dynamic behaviors have obvious nonlinear coupling characteristics.Besides,the grouping and large-scale of lithium-ion batteries make the dynamic behavior of the system more complex,increase the probability of fault,and bring new challenges to the system security,which has become one of the main factors restricting its large-scale application.Therefore,the research on fault diagnosis and health management of lithium battery energy storage system is meaningful for the development and promotion of new energy electric vehicles and energy storage grid.In response to actual application requirements,this work carried out the research on the behavior expression,health assessment,fault diagnosis and health management theory and method of lithium battery energy storage system in complex application environment,aiming to provide theoretical support and technical guidance for ensuring the safe,efficient and long-life operation of the system.The main contributions and innovations of this work are as follows:(1)Aiming at the problem of aging mechanism modeling and health assessment of lithium-ion battery,firstly,according to the characteristics of electrode OCV,an evaluation method of aging mode of lithium-ion battery based on OCV matching model is proposed.Secondly,the semi-empirical models of aging mode attenuation.impedance growth,and state of health(SOH)decay are established.Thirdly,a multi-time scale battery health state assessment method based on particle filter is proposed.Finally,the accuracy of the proposed model and algorithm is verified by the measured aging experimental data.(2)Aiming at the consistency evaluation of lithium-ion battery pack,firstly,based on the measured battery data of EV,the parameters suitable for characterizing the consistency of battery pack are mined.Secondly,the consistency evaluation model of the battery pack based on multi-feature weighting is proposed,and the entropy weight method is used to determine the feature weight to ensure the reliability of consistency evaluation.Thirdly,an improved fuzzy clustering algorithm is proposed for cluster analysis of batteries.Finally,the performance of the proposed model and algorithm is verified by the electric vehicle data of nine months.(3)Aiming at the fault diagnosis of lithium-ion battery pack sensors,first,based on the cell equivalent circuit model and thermal model,combined with the electrical structure of the battery pack,electric and thermal models of the series cylindrical battery pack are constructed.Secondly,voltage and temperature state observers based on the particle filter are constructed for the battery pack,and a sensor fault diagnosis scheme based on characteristic matching is designed for the battery pack.Thirdly,the student residual method is used to preprocess the residual,and a residual evaluation method based on the cumulative sum of the log-likelihood ratio is proposed.Finally,the robustness and accuracy of the proposed method are verified by different fault cases.(4)Aiming at the lithium-ion battery charging management,firstly,based on the charging aging test and aging mode evaluation model,the influence of the charging rate on the three aging modes is explored,and the aging mode growth empirical model is established.Secondly,combined with equivalent thermal and electrical models,a charging optimization model considering charging time,health loss,and temperature rise is constructed.Thirdly,a multi-stage health-conscious charging management scheme based on the NSGA-? algorithm is proposed.Finally,the feasibility of the proposed charging optimization scheme and the charging performance of different solutions are verified by experiments and simulations.