Study On Multi Stress Accelerated Aging Model Of Ternary Lithium Battery

Ternary lithium battery has the characteristics of high energy density,high voltage platform,poor high temperature resistance and low stability.Studying the accelerated aging model of ternary lithium battery under multi stress and establishing a high-precision and high-level model have theoretical and practical significance for the safe and effective use of battery.In this paper,the accelerated aging model of2.5Ah cylindrical ternary lithium battery is studied.The main research contents include:1 The electrochemical characteristics of ternary lithium battery are analyzed,and the influencing factors of ambient temperature,discharge rate and charging cut-off voltage life are selected as the accelerated stress of battery aging.Determine the accelerated stress intensity range of the battery,and design the accelerated aging experimental process of the battery.2 The accelerated aging model of battery under single stress of ambient temperature,discharge rate and charging cut-off voltage are established.According to the experimental data,the capacity attenuation characteristics of the battery are analyzed,and the accelerated aging model under single stress is established.Through the verification experiment,it is concluded that the accuracy of the model is less than5.0% under the single stress of ambient temperature and discharge rate,and the accuracy of the model is 29.9% under the single stress of charging cut-off voltage.3 Under the condition of single stress,the action level of the model is limited.In order to solve this problem,an acceleration model under coupled stress based on empirical model is established.The influence of coupling another stress on the battery based on the single stress model is proposed.The particle swarm optimization algorithm is used to solve the problem of multi parameter solution of accelerated aging model,and the accuracy of each model is less than 12.0%.4 The accelerated aging model of double stress coupling battery based on datadriven method is established.The accelerated aging model based on empirical model can not get rid of the complexity of the establishment process of specific formulaic model.GA-BP neural network algorithm is used to train the accelerated aging model with experimental data,which is more practical.The BP neural network algorithm takes the battery capacity decline rate and two acceleration stress values as the input and the battery cycle charge and discharge times as the output.The prediction accuracy of each training model is less than 9.0%.