Big-data-AI-based dynamic threshold power battery charging voltage state judgment method

Abstract

The invention discloses a big-data-AI-based dynamic threshold power battery charging voltage state judgment method. The method comprises the steps: extracting a charging data set and training a model;instantiating an RRCF model of a CoDisp value; obtaining a priori Codisp value of a charging vehicle; using the feature data as a priori Codisp value dp; calculating a CoDisp value dn of voltage stream data by using a stream data RRCF model; using a priori CoDisp value to calculate a corrected streaming data abnormal value judgment standard eta'n; taking a time window by taking s as the width; calculating a mean variance sigma wn of the CoDisp in the window, and dividing the distance between the current data CoDisp value dn and the mean value of the time window by the multiple eta n of the variance sigma wn; and comparing the eta n with the eta ', if the eta n is greater than the eta', determining that the voltage data is abnormal, otherwise, determining that the voltage data are normal.Therefore, the self-adaptive adjustment of the abnormal value judgment threshold is realized, so that the early warning accuracy is improved, and the occurrence of false alarms is reduced.

Description

technical field [0001] The invention belongs to the field of battery technology, in particular to a method for judging the charging voltage state of a dynamic threshold power battery based on big data AI. Background technique [0002] As the power source of new energy vehicles, the service life and safety of battery packs are the focus of the new energy industry. The service life and safety of the power battery are not only related to the battery design, material and other factors, but during the use of the power battery, especially during the charging process, due to the intense electrochemical reaction inside the battery, it is particularly important to effectively monitor the battery’s operating status. . Patent application number 201910981877.6 uses the sliding window method to diagnose power battery faults online by calculating basic statistical indicators such as median and variance. The evaluation indicators used in this method are too simple, and some potential fail...

AI Technical Summary

Problems solved by technology

Patent application number 201910981877.6 uses the sliding window method to diagnose power battery faults online by calculating basic statistical indicators such as median and variance. The evaluation indicators used in this method are too simple, and some potential failure modes may not be found, and only the The information of the local data in the window lacks the mechanism to use the overall information

According to the found patents related to cell voltage, most of them belong to hardware circuit-related inventions, and most of the patents related to cell voltage algorithm only focus on a specific problem of the cell, and there are also many data on the selection of model data features. Dimensions are too simple

Images