Research On State Of Health Estimation And Sorting Method Of Retired Li-Ion Power Batteries

The development of new energy vehicles is a strategic measure to deal with climate change and promote green development,as well as an important way to improve energy shortages and environmental pollution.Affected by the favorable factors of the new energy policy,commercial and civilian new energy vehicles have been launched on a large scale in recent years.The average life of lithium battery systems for electric vehicles is about 3-5 years,and a large number of automotive power battery packs will face retirement in the next few years.The study found that the vehicle battery pack still has about 80%rated capacity after retirement.If it is recycled in a cascade,it can also be safely used in energy storage,household appliances and other fields.The cascade utilization market has huge potential.Due to the‘cask effect’of retired batteries,serious safety problems are often caused during application.The safety status and health status assessment of retired lithium-ion power batteries are the key issues that restrict the smooth development of retired batteries.This thesis aims to evaluate the health status of retired batteries and the consistent sorting and reuse of batteries.The main work and innovations are as follows:First of all,this thesis fully reviews the existing methods of battery health state estimation and sorting into groups.Aiming at the inconsistency of retired lithium-ion power batteries,using retired lithium iron phosphate battery cells as the research object,a battery test platform is built and designed.The full life cycle aging experiment of retired batteries analyzes the changes of different parameters in the aging process of retired batteries in detail,and provides the experimental data basis for subsequent battery health assessment and battery sorting research.Secondly,the characteristic parameters are extracted according to the charging curve of the retired lithium-ion battery.Through the change of the cycle times in the whole life cycle,it is found that some characteristic parameters of the battery have a great correlation with the attenuation of the capacity.The battery health estimation model of the characteristic parameters is extracted from the IC curve and charging curve,and the extracted characteristic parameters are denoised by the method of wavelet domain denoising.The quantitative relationship between parameters and capacity is established through the decision tree model and the Bayesian improved decision tree model,which is intended to realize battery capacity estimation.Thirdly,in view of the problem that the decision tree model is prone to overfitting,which leads to insufficient adaptability,a health assessment model for retired lithium-ion batteries based on ensemble learning is proposed.Multiple decision tree models are selected for model fusion,and the improved XGBoost model based on the boosting algorithm（Boosting）is used.The improved XGBoost model,through the serial iteration of multiple weak learner decision trees,is combined into an XGBoost strong learner to improve the accuracy and adaptability of the health assessment model.Finally,in order to solve the problem that the consistency of each monomer in the retired battery pack is poor and it is difficult to directly recombine and reuse,the retired batteries were sorted based on the Fuzzy C-Means（FCM）clustering algorithm based on genetic simulated annealing.The 75 retired Li Fe PO₄ battery cells in the battery aging experiment were used as the data source for sorting.Various feature parameters and dimensionality reduction processing are carried out,and the principal components are input into the sorting model as the effectiveness index to determine the battery sorting category.The simulation and experimental results show that the retired battery sorting model based on the fuzzy C-means clustering algorithm of genetic simulated annealing has good sorting accuracy,and the battery cells still have good consistency after many cycles.The research in this thesis provides theoretical support for the safe reuse of retired lithium power batteries,and has a certain positive significance for solving important engineering problems such as battery system health assessment and prediction,cascade utilization,sorting and reorganization in the new energy industry.