基于红外的TPA和IAOA BiLSTM电路芯片故障诊断  

Fault diagnosis of TPA and IAOA BILSTM circuit chips based on infrared

作　　者：王力 朱猛 马江燕 WANG Li;ZHU Meng;MA Jang-yan(Airborne Electronic Systems Deep Maintenance Laboratory,College of Vocational Technology,Civil Aviation University of China,Tianjin 300300,China)

机构地区：[1]中国民航大学电子信息与自动化学院机载电子系统深度维修实验室,天津300300

出　　处：《激光与红外》2024年第4期574-583,共10页Laser & Infrared

基　　金：民航安全能力建设基金项目(No.[2023]50,复杂环境下“黑匣子”搜寻探测与深度数据提取装备研制)资助。

摘　　要：为了提高电路芯片故障诊断准确率,超参数设置的效率以及特征提取效率,提出一种基于时间模式注意力机制(TPA)的改进算数优化算法(IAOA)优化双向长短期记忆网络(BiLSTM)的电路故障诊断方法。首先,利用IAOA搜寻BiLSTM的最优超参数组合,提高模型诊断精度;然后使用TPA提取重要特征并分配权重,改善模型特征提取能力;最后,将红外摄像仪采集的红外温度数据输入到最优诊断模型中,实现电路芯片故障诊断。实验采用0~30 V可调稳压电源电路进行验证。结果表明,该模型对电路芯片故障诊断准确率高达9827,可实现对电路芯片的高准确率故障诊断。To improve the accuracy of circuit chip fault diagnosis,the efficiency of hyperparameter setting and the efficiency of feature extraction,an improved arithmetic optimization algorithm(IAOA)based on temporal pattern attention mechanism(TPA)is proposed to optimize the bi directional long and short term memory network(BiLSTM)for circuit fault diagnosis.Firstly,IAOA is employed to search for the optimal hyperparameter combinations of BiLSTM to improve the diagnostic accuracy of the model.Then TPA is used to extract important features and assign weights to enhance the model feature extraction capability.Finally,the infrared temperature data collected by the infrared camera is inputted into the optimal diagnostic model to achieve circuit board chip fault diagnosis.The experiments are verified by using 0~30 V adjustable regulated power supply circuit board.The results show that the model for circuit chip fault diagnosis is as high as 98.27%,which can achieve high accuracy fault diagnosis for circuit board chips.

关 键 词：红外技术 芯片故障诊断 双向长短期记忆网络 算数优化算法 时间模式注意力机制 

分 类 号：TP219[自动化与计算机技术—检测技术与自动化装置] TP274[自动化与计算机技术—控制科学与工程]