Defective battery screening method, equipment and medium based on machine learning

Abstract

The invention provides a machine learning-based defect battery screening method, equipment and medium, including obtaining the open circuit voltage, AC internal resistance and self-discharge rate of the battery to be tested at high temperature; high-temperature self-discharge abnormal primary screening and re-screening; Screen abnormal battery capacity; obtain open-circuit voltage, AC internal resistance, and self-discharge rate of battery under test at normal temperature; pre-screen and re-screen abnormal self-discharge at room temperature; screen abnormal internal resistance of battery under test; screen abnormal battery under test. The invention adopts a sliding screening mechanism based on a machine learning algorithm to collect, analyze, and screen data on the battery to be tested, and simultaneously sets a re-screening mechanism and a final screening mechanism to avoid missed screening and sieving, and improve the accuracy of battery detection.

Description

technical field [0001] The invention relates to the technical field of battery detection, in particular to a method, equipment and medium for screening defective batteries based on machine learning. Background technique [0002] In order to deal with carbon emissions, environmental pollution, energy crisis and other issues, the new energy industry, especially the battery industry, is being vigorously developed around the world. In the mass production process of lithium-ion batteries, after the battery is formed (the first charge is activated), the self-discharge rate test (commonly known as the K value test), capacity test, and internal resistance test (commonly known as the DCR test) are generally carried out for screening. Produce defective bad batteries to avoid safety accidents caused by entering the market. [0003] At present, various tests for batteries are screened based on traditional mathematical statistics methods, that is, a threshold is set for each test, and w...

AI Technical Summary

Problems solved by technology

[0003] At present, various tests for batteries are screened based on traditional mathematical statistics methods, that is, a threshold is set for each test, and when each test is performed, it is determined whether the battery to be tested is a defective battery according to the corresponding threshold. This method exists The following problems, first of all, the setting of the threshold value can only be based on experience, there will be situations of screening and leakage, and often cannot be well adapted to different production batches of batteries, not only will make the yield rate not high, but also the battery quality is not good, and even It will cause safety problems for batteries in the market; secondly, there is no correlation between each test, because the threshold value is different, it may cause misjudgment for batteries with marginalized test data; and this method often requires a large number of batteries to accumulate. Makes the enterprise inventory pressure is high, and the product circulation efficiency is low

Images