Kesearch On Estimation For SOH Of PEV Li-ion Battery Pack

The deterioration of the environment and energy crisis bring double pressures to the development of the traditional vehicle, therefore electric vehicle (EV) has become the main research direction of the automobile development in the future. As the energy source of EV, power battery packs should be managed and controlled essentially in order to ensure the security and stability of EV operation. Battery state of Health (SOH) which is one of the most important parameters in battery management system (BMS), is accurately estimated to timely know the battery healthy status, provide information for the detection and diagnosis of the battery, and contribute to replacement of aging battery cell in time. Furthermore, this can improve the overall life of the battery and the power performance of EV. Therefore, it has very important and practical significance to estimate SOH accurately.At present, there are a few people to do research on battery SOH estimation. And the studies are mainly focused on the battery cell. However, studies using the whole battery pack as the study object are less. In addition, the research collected data from the test of the laboratory instead of real-time operational EV. Therefore, different from the previous studies, this paper estimates battery SOH based on real-time data of pure electric taxi running in Beijing.Firstly, according to the definitions of battery SOH, this paper uses battery internal resistance as the evaluation index of battery SOH. Secondly, based on the real-time data of electric taxis in operation, this paper discusses the characteristics of the charge and discharge processes in detail, and shows the reasonability of a simple equivalent circuit model of battery cell representing the properties of the whole battery pack. Next, according to the equivalent circuit model of battery cell, the structure of the battery pack model is analyzed based on the circuit theory. Then the two-order RC equivalent circuit model of battery pack is established. At last, based on the real-time charging data, the model parameters are identified by using the forgetting factor recursive least-squares estimation and the battery pack model is confirmed.The battery pack is a complex nonlinear system, and the particle filter (PF) algorithm has good performance in solving nonlinear problems, so this paper explores the estimation of battery pack’s internal resistance using particle filter algorithm. Combining particle filter algorithm and the equivalent circuit model of battery pack, based on the real-time charging data, the estimation experiments about battery pack’s internal resistance are carried out. The results show that, the particle filter algorithm can predict battery pack’s internal resistance accurately in charge process, and the proposed algorithm has good applicability.The degradation problem in the PF algorithm influences the accuracy of estimation directly. Therefore in order to better solve the degradation problem of particle and improve the estimation accuracy of battery pack’s internal resistance, this paper tries to apply genetic algorithm in the PF algorithm and design the steps of the algorithm in detail to estimate the battery pack’s internal resistance. Through the analysis on the experimental results of comparing to standard PF and improved PF, the conclusions show that:the estimation curves of battery pack’s internal resistance using genetic particle filter algorithm have better stability. Therefore, the genetic particle filter algorithm has more superior performance on estimation of battery pack’s internal resistance.