Research On Multi-level Battery State Estimation Methods And Reconfigurable Battery Management System Technogoly

Global warming,petroleum crisis and the legislation pushing for higher fuel economy and lower emissions are leading to the development of electric vehicle?EV?.As the key component of any EV,the energy storage system attracts more and more attentions.A variety of electrochemical energy storage devices are currently used in EV applications,such as lithium-ion?Li-ion?battery,nickel metal hydride?NiMH?battery,lead acid?LA?battery.The battery system is a complex electrochemical system with strong nonlinear characteristic,and the performance is affected by many factors such as current,environment temperature and battery age degree.Therefore,there will be a great meaning for improving the power,economy and safety of EVs by ensuring the safety and efficient operation of traction battery under complex working conditions.The main research topics of this dissertation focus on the state of charge?SOC?and state of health?SOH?estimation methods and the design of battery management system,The main contents of this dissertation are as follows:To solve the‘cell-level'state of charge estimation problem,the nonlinear predictive filter?NPF?based SOC estimation method is proposed by combining the Thevenin equivalent circuit battery model and NPF algorithm,which realizes the accurate SOC estimation of battery cell with unknown process noise.A designed battery cell test plan is carried out to validate the proposed method under different application environments:the estimation results with different automotive driving cycles show that the proposed method is robust to different types of battery loading profiles;The estimation results of LiCoO₂?LCO?and LiFePO₄?LFP?battery cells indicate that the proposed method can be applied to different batterie types with satisfactory estimation performance;The estimation results under different battery working temperatures show that the application of the proposed method can be extended to different working temperature ranges by employing the relationship between battery model parameter and battery working temperature;The analysis of the SOC estimation accuracy under different model parameter errors reveals that the estimation accuracy is linear to the molde parameter error.In order to show the advantages of the proposed method,the comparison between the classical extended Kalman filter?EKF?based SOC estimation method and NPF based SOC estimation method is performed,the results show that the NPF based method has superior performance over EKF based method with respect to estimation accuracy and convergence rate.To solve the‘cell-level'state estimation problem of aged battery,the joint battery SOC and SOH estimation method based on dual nonlinear predictive filter?DNPF?is proposed,which takes account of the interrelationship between battery SOC and SOH.With respect to the SOH estimation,the key model parameters of battery cell are online estimated to accurately predict the capacity SOH and power SOH.With respect to SOC estimation,the estimation accuracy under total battery life cycle is effectively improved by online updating the values of key model parameters.The cycle life test of a prismatic LFP battery cell is performed to validate the joint estimation method,the results show that the proposed method has achieved accurate SOC and SOH estimation under total battery life cycle.To solve the‘pack-level'SOC and SOH estimation problem,the battery state definition is extended from‘cell-level'to‘pack-level'according to different pack topologies and different balance control strategies,then a multi time-scale SOC and SOH estimation method for battery pack is proposed with respect to the multi time-scale nature of SOC and SOH.In this method,the SOH estimation is performed with macro time-scale and the SOC estimation is performed with micro time-scale.The experiments conducted on a series-connected lithium-ion battery pack with passive balance control strategies are applied to validate the performance of the proposed method.The validation results indicate that the SOC and SOH of the battery pack can be accurately estimated by proposed method and the computation of the estimation algorithm can be effectively reduced.The battery modules with different characteristics cann't be well managed by the traditional battery management system?BMS?,and the structure of the traditional BMS is not flexible.In order to solve these problems,the concept of reconfigurable battery management system?RBMS?is proposed and the prototype system of RBMS is developed.In RBMS,the user is able to determine the number of‘in-pack'battery module due to the specific usage requirement,which reduces the cost of battery pack and minimizes the weight of the vehicle,and thus improving the power and economy of EVs.Meanwhile,based on the Master-Slave topology,a battery module switching system is added in RBMS to realize the flexible management of different battery modules,which increases the compatibility to the battery modules with different characteristics.In this dissertation,in order to solve the battery SOC and SOH estimation problem in both‘cell-level'and‘pack-level',an integral multi-level state estimation method for battery system is established,which provides a theoretical support for the state estimation algorithm in BMS.Besides,the proposed design concept of RBMS is able to realize the flexible structure of battery pack and can effectively improve the compatibility performance to the battery modules with different characteristics,it has a good prospect for engineering application.