Study On Li-ion Battery Mechanism Model And State Estimation For Electric Vehicles

The battery system is one of the most important parts in an electric vehicle, and a battery management system is needed to manage the batteries in the vehicle.The battery state estimation is the basic of the other battery management algorithm and the vehicle management algorithm. Thus it is the most critical technology of the battery management system in electric vehicles. The battery state estimation includes the estimation of battery SOC(State of Charge), SOH(State of Health) and SOF(State of Function). The precise battery state estimation algorithm could improve thebattery management algorithm,the battery could be usedsufficiently and the battery life would beextended.The study on the battery mechanism model could help to understand the batteries and improve the state estimation precision.This dissertation focuses on the lithium-ion power battery systems in electric vehicles, the battery mechanism model is studied and based on the battery model the battery state estimation algorithm is studied. In view of the huge computation of the rigorous initial battery mechanism model, the battery mechanism simplification model is studied and the simplified model is used to estimate the battery SOC. The battery aging mechanism identification algorithm of the common lithium-ion batteries is studied, and the battery capacity fade model is built and the battery SOH estimation algorithm is proposed. Considering the loss of the Li ion caused by the side reaction of SEI(Solid Electrolyte Interface) film thickening and lithium deposition is the main aging mechanism of the graphite anode Li ion batteries, the side reaction mechanism model is simplified, and the SOF estimation algorithm is studied.Firstly,the lithium-ion mechanism model is researched. The single particle model is improved and furthermore a novel battery simplified pseudo two dimensional mechanism model is proposed. The simplified model could guarantee enough precision with less computation and it could be used in real battery management system for real time on-line estimation in electric vehicles. Then based on the simplified mechanism model, the corresponding battery SOC estimation algorithm is proposed. The simulation results verify theprecision androbustness of SOC estimation algorithm.Secondly, the battery cycle life tests with changing conditions are conducted in the commonly used batteries in electric vehicles including the C/LFP batteries,C/LMO batteries,and LTO/NCM batteries. The evolution of battery capacity, resistance are investigated. The battery aging mechanism are identified with the method of battery constant current charging curve reconstruction and the main aging mechanisms of the commonly used batteries in electric vehicles are analyzed and discussed. On the other hand,using the genetic algorithm the battery capacity fade model are built according to the changing condition cycle life test and the damage accumulation assumptions. Anovel method of capacity fade model based open-loop capacity estimation with periodical capacity calibration and correction is proposed to estimate the battery SOH.Finally,the side reaction mechanism model simplification of theSEI film thickening and lithium deposition side reactions is studied. The simplified model results shows high precision with less computation and it could be embedded in a microprocessor of battery management system for real time optimization. Considering the principle of theminimum battery side reactions and the battery voltage, current and SOC limitations, the battery SOF could be estimatedbased on the estimation results of battery SOC and SOH.