基于神经网络的锂电池极耳焊缝微缺陷判别系统设计  

Design of Discriminant System for the Micro Defects in Pole-ear Weld of Lithium Battery Based on Neural Network

作　　者：李泰民 黄平[1] 岳笑含[1] 白石 张晓丹[1] 刘志尧 LI Taimin;HUANG Ping;YUE Xiaohan;BAI Shi;ZHANG Xiaodan;LIU Zhiyao(School of Information Science and Engineering,Shenyang University of Technology,Shenyang 110870,China)

机构地区：[1]沈阳工业大学信息科学与工程学院,沈阳110870

出　　处：《无损探伤》2023年第6期5-8,共4页Nondestructive Testing Technology

基　　金：国家自然科学基金(62001313);辽宁省自然科学基金(2020-MS-211);辽宁省教育厅面上项目(LJKMZ20220477)。

摘　　要：针对锂电池极耳焊缝微缺陷识别问题,提出基于多频差分涡流传感器阵列测量磁数据的方法,并对获得的检测数据采用神经网络训练模型进行分类。实验结果表明,相较于单通道单一频率的检测方式,多频励磁阵列检测效率更高,且能够同时检测极耳焊缝表面和内部缺陷以实现层析成像。此外,所使用的神经网络训练模型解决了多频涡流阵列数据处理问题,识别准确率达到96%以上。In order to identify the micro defects of lithium battery pole-ear welds,a method based on multi frequency differential eddy current sensor array to measure magnetic data is proposed,and the obtained detection data are classified by neural network training model.The experimental results show that compared with the single-channel and single-frequency detection mode,the multi-frequency excitation array detection is more efficient,and can simultaneously detect the surface and internal defects of the pole-ear weld to achieve tomography.In addition,the neural network training model solves the problem of processing multi-frequency eddy current array data,and the recognition accuracy is more than 96%.

关 键 词：极耳焊缝 微缺陷 多频差分涡流 神经网络 层析成像 

分 类 号：TG115.28[金属学及工艺—物理冶金]