Research On Remaining Useful Life Prediction For Lithium-ion Battery Based On Data-driven

Lithium-ion battery has been widely used in the fields of society,industry,and aerospace because of its high storage energy density,light weight,long service life,and many other advantages.However,with the widespread application of lithium-ion battery in life,the safety and reliability problems caused by their life are gradually exposed.Therefore,effective monitoring of their working status is extremely important.In order to better perform predictive maintenance on the battery,the Remaining Useful Life prediction method has become the focus of social attention.The purpose of this paper is to study a variety of data-driven methods for predicting the remaining useful life of lithium-ion battery.The methods we will use mainly include BP neural networks,Particle Filters,and Relevance Vector Machine,and they predict RUL based on historical data collected from lithium-ion battery experiments.First,this paper describes the internal working principle of the battery,and performance advantages,analyzes the degradation process of lithium-ion battery.Second,three data-driven algorithm principles are introduced in detail,and then we use these algorithms to perform RUL simulation experiments for lithium-ion battery and obtain RUL prediction results.Finally,we compare the experimental results obtained by the three algorithms and draw conclusions: each method has its own advantages,however,relevance vector machine is higher in RUL prediction accuracy,the error is small and uncertainty can also be expressed,it can improve battery reliability,and it has great promotional value.