Accelerated Life Prediction And Life Degradation Of Lithium-Ion Batteries

Lithium-ion batteries have the advantages of low cost,high energy density and long cycle life.It is widely used in electric vehicles,smart grids,aerospace and other fields.In actual use,the performance of the battery will gradually degrade.Obtaining the life of the battery in advance is of great significance for the reliable operation of the system.Nevertheless,Along with the continuous development of batteries with long life,it takes a long time to predict their life at nominal operating conditions using long-term historical battery data.This delays feedback on the battery’s performance and prevents the development and validation of new technologies.Thus,this work starts from accelerating the battery charge and discharge cycle under the accelerated stress working environment.Accelerate battery life prediction by using early cycle data that characterizes battery degradation to reduce the number of battery cycles.Optimized charging strategies are used to reduce the accelerated degradation of battery life under accelerated cycling.The main research content has the following four aspects.Ⅰ.Analysis of working principle and degradation mechanism of lithium ion batteryThe internal structure and charge-discharge working principle of LIBs battery are briefly summarized.The performance parameters such as capacity,voltage and charge-discharge rate of LIBs are introduced.The degradation mechanism of LIBs is analyzed from three aspects:solid electrolyte interface(SEI)growth,electrolyte oxidation and lithium coating caused by degradation of LIBs.The internal and external factors affecting the degradation of LIBs are discussed.Analysis of the reasons why increasing the charge/discharge rate can accelerate battery degradation.Ⅱ.Lithium-ion battery accelerated life prediction based on XGBoost algorithmAiming at the problem of long life prediction time and high cost of LIBs,this work is based on the study of lithium-ion batteries with accelerated cycling at large charging multipliers.The characteristic information closely related to battery life degradation in the early cycle stage of the battery is analyzed from the battery voltage-capacity curve and capacity increment curve.Extract three characteristic parameters that characterize the degradation of battery life are extracted and used as inputs to the prediction model.To establish the mapping relationship between battery characteristic parameters and battery life,build an Extreme Gradient Boosting(XGBoost)prediction model.Continuously improve the learning accuracy of the model by focusing on the residuals between the predicted and true values to effectively simulate the complex nonlinear behavior of lithium-ion batteries.The hyperparameters of the prediction model are optimized by the whale optimization algorithm to improve the prediction accuracy of the model.The prediction results show that the early cycle data of the battery under the accelerated cycle can accurately predict the battery life.The error of the constructed extreme gradient prediction model is within 4%.Ⅲ.Improvement of lithium-ion battery life degradation based on Bayesian algorithmIn response to the accelerated battery life degradation under accelerated stress,based on the physical and chemical reaction characteristics of the battery itself,a coupled electric-aging model of lithium-ion battery is built to reckon the internal state of the battery.Determine the optimization objective of battery life loss in the high rate charging parameter space according to the coupling model,and establish the objective function.Fit the objective function using a probabilistic proxy model with Bayesian algorithm and calculate the mean and variance to continuously evaluate the posterior distribution of the objective function.The optimal value of the objective function is determined by iterative search to find the multi-stage constant-current charging strategy with minimum life loss under accelerated cycles.The optimization results show that the life loss of LIB at the end of charging is 0.8497%under this charging strategy.Compared with the traditional charging strategy,the optimized charging strategy reduces the life degradation of LIB caused by accelerated cycle.Ⅳ.Accelerated life experiment and life prediction model verification of lithium-ion batteriesAccelerated life experiments on a set of Li-ion batteries with a capacity of 1.1 Ah based on the acceleration factor of charge multiplier.Build a battery life test experiment platform.The LIB charging and discharging module was designed according to the experimental requirements.Analyze the principle and device selection of the main circuit,acquisition circuit,charge/discharge circuit and other circuits of the module.Design PCB circuit board and complete charge/discharge test prototype.By connecting the battery and the host computer to charge and discharge the LIB in a constant temperature environment,the life information of the LIB under different large charging rates is obtained.The life prediction of the LIB is carried out in combination with the limit gradient lifting life prediction model used in this paper.The experimental results show that the prediction error is less than 7%.The prediction method used allows for accurate prediction of battery life.