| In the manufacturing process of lithium-ion batteries,the formation process is one of the most important factors affecting the quality of lithium-ion batteries by charging and activating them with normal electrochemical properties.In practice,the long process and complex mechanism of the formation process leads to a serious lag in the quality testing of lithium-ion batteries.Secondly,the charging and discharging process of lithium-ion batteries has the problem of long distance dependence,which affects the effectiveness of quality inspection.In order to solve these problems,the following research work has been carried out in this paper.(1)In order to solve the problem that the long process flow and complex mechanism lead to the lag of quality detection,a quality classification method based on TS-Res Net(Time Series Bases on Residential Network)is proposed for lithium-ion batteries.At first,the voltage curve of charge and discharge is characterized by residual network,then the quality of lithium-ion battery is classified,and the substandard battery is detected before the charge and discharge is completed,so that the battery can be reworked or scrapped,and the production efficiency and battery quality can be improved to save time and resources.Classified results showed that TS-Res Net had an average accuracy of 91.52%.(2)In view of the long distance dependence of lithium ion battery charging and discharge process and the difficulty of TS-Res Net in extracting global correlation characteristics,a method of product quality classification of lithium ion battery conversion process based on GAFRes Net(Gramian Angular Field Based on Resduial Network)was proposed.Time-series data converted into processes are encoded into images by Gramian Angular Field(GAF)to capture their global correlation characteristics,which can be used as input to residual network models to classify the quality of lithium-ion batteries and more effectively detect substandard batteries.The experimental results show that the method has an average accuracy of 94.16% and can be better applied to the production of lithium-ion batteries. |
