Research On Remaing Useful Life Of Power Lithium-ion Battery Based On Indirect Prediction Method

As an advanced green battery,lithium-ion battery has been widely studied and applied in various fields,with its high specific energy,low self-discharge rate,high safety,long cycle life and other advantages.However,with the increase of the number of cycles,the performance of lithium ion batteries has a decline in performance irreversibly.The electrical system will be affected,even lead to disastrous consequences if the above situation is handled improperly.Therefore,measuring the health factor and predicting the remaining useful life accurately are the basis to ensure the reliability and safety of lithium-ion batteries.In this paper,the selection of indirect health factors,remaining useful life prediction methods and practical applications of lithium-ion batteries are researched furtherly.Firstly,the working principle,life attenuation mechanism and life influencing factors of lithium-ion battery are analyzed.The data sets,including self-measured lithium-ion battery JZ and the battery published by NASA are taken as the research objects,and the measured voltage and discharge time data are normalized during different cycle discharge cycles.According to Matlab fitting and derivation,the selection range of equal discharge voltage is determined.The correlation between time interval of equal discharge voltage and capacity is calculated by using grey correlation analysis method.The validity and rationality of time interval of equal discharge voltage as an indirect health factor is proved.Theoretical basis are provide for the research on the indirect prediction method of the remaining useful life of lithium-ion batteries.Secondly,the relationship between the indirect health factor,voltage drop discharge time series,and the actual capacity of lithium-ion battery is analyzed,and the degradation relationship model based on Elman neural network is established.Based on the training data set,the time interval of equal discharge voltage of the remaining effective cycle times is predicted.The data of the predicted time interval are taken as the input of the neural network model,and the corresponding predicted value is obtained according to the established degradation relation model.And the remaining useful life of lithium ion battery is predicted on the basis of the set failure threshold.The experimental results show that the Elman neural network has the unique advantage of its own receiving layer and can adjust the deviation with feed-forward effect automatically,which can improve the prediction accuracy of lithium ion battery RUL.Finally,it is difficult to obtain the data needed by Elman neural network algorithm in RUL prediction.Two battery packs of NASA and JZ series are selected in the equivalent cycle conditions and different cycle times to research the slope change rule at the equivalent feature point of the first derivative function between discharge time and voltage,and the equivalent cycle life degradation curve of lithium ion battery is established.Furthermore,the method is proved simple and accurate in predicting the remaining useful life of lithium-ion battery.The most important is that the effectiveness and versatility of the method and can be used in practical prediction widely.