Research On State Collaborative Estimation Of High Energy Density Li-ion Battery System

The global shortage of primary energy and the improvement of human awareness of environmental protection has promoted the development of green and sustainable energy.In recent years,ternary lithium-ion batteries with high specific energy and high-cost performance have become an important part of the new energy field and are recognized as the key to optimizing the existing energy structure.However,due to the unreasonable existing battery management mechanism,safety problems such as fire and the explosion of electric vehicles still occur from time to time,which greatly hinders the development of lithium-ion batteries.With the rapid development of the consumer electronics industry,the market demand for high-performance batteries is constantly expanding,and the potential safety problems of lithium-ion batteries need to be solved urgently.Accurate estimation of state-of-charge and state-of-health is the basis of safety control and efficient management of battery management systems,and also the key to promoting its market application.The following research has been carried out around the co-estimation of SOC and SOH of ternary lithium-ion batteries with high energy density.（1）Research on battery characteristics and construction of equivalent models.Firstly,the internal structure and characteristics of lithium-ion batteries were studied and analyzed by consulting literature,and an experimental platform was set up to conduct experiments on ternary Li-ion batteries to explore the working characteristics of Li-ion batteries.Experiments were conducted by using the control variable method to quantitatively explore the influence of internal and external factors on battery characteristics,and then the battery dynamic equivalent circuit model was built.The sensitivity of each model parameter under different SOC values was tested,and the battery model parameters were estimated by offline identification,online identification,and online identification of some parameters respectively.Compared with the estimation accuracy and time cost,the results showed that the battery parameters with high sensitivity were selected for online identification,and the accuracy of other parameters using offline identification was not significantly reduced compared with the online identification of all parameters,and the calculation cost was greatly reduced.（2）Aiming at the irreversible influence of battery aging on battery internal characteristics and battery capacity,the least square surface fitting is adopted to model the coupling relationship of OCV-SOC-Qmax,and the multi-parameter coupling relationship surface under the principle of minimum error is established.This functional relationship is used in the process of parameter identification and state estimation,which effectively reduces the model representation error and state estimation error caused by battery aging effect on Li-ion batteries.（3）Aiming at the problem of state co-estimation of Li-ion batteries,considering the obvious difference of time interval between battery SOC and SOH,an improved adaptive unscented Kalman filter-unscented particle filter（AUKF-UPF）algorithm based on different update frequencies was proposed,and the state estimation model and adaptive iterative algorithm were constructed respectively to estimate the SOC and SOH of Li-ion batteries in nonlinear systems on-line co-operation.By improving the Sage-Husa algorithm,the estimation error caused by the limitation assumption of the noise of traditional extended Kalman filter algorithm can be reduced,and the closed-loop formed by state co-estimation can make the state quantities feedback and correct each other,thus realizing high-reliability real-time monitoring of various key state quantities of high-energy-density lithium-ion batteries.（4）The validity and robustness verification of SOC-SOH collaborative estimation strategy.To verify the effect of the improved algorithm proposed in the project on the cooperative estimation of the state of high-energy-density lithium-ion batteries,experiments were carried out on batteries under various working conditions based on the battery experimental platform,mainly including constant current discharge,interval pulse discharge,and dynamic test.The input current and output voltage under each working condition are extracted for verification,and different initial values of state estimation and different noise conditions are set to verify the sensitivity of the algorithm to the initial values and noise statistical characteristics.The results show that,compared with other single algorithms,this cooperative estimation strategy greatly improves the estimation accuracy and robustness of SOC and SOH.Aiming at the accurate estimation of SOC and SOH,the core technology of battery management system BMS,this paper explores the battery aging characteristics,constructs the battery OCV-SOC-Q_max coupling surface,and applies it to the battery state collaborative estimation model.According to the battery SOC,SOH definition,and equivalent circuit model,the state space models are established respectively,and the improved AUKF-UPF iterative filtering algorithm is designed to realize the cooperative estimation of battery SOC and SOH,and realize the estimation of remaining available power and life prediction of high-energy-density lithium-ion batteries,effectively reducing potential safety hazards,improving battery energy utilization efficiency and prolonging the service life of lithium-ion batteries.