优化卷积神经网络诊断核辐射探测器电路故障  

Fault Diagnosis of Nuclear Radiation Detector Circuit Based on Optimized Convolution Neural Network

作　　者：杨鹏飞[1] YANG Pengfei(College of Information Engineering,Zhengzhou Tourism College,Zhengzhou 451464,China)

机构地区：[1]郑州旅游职业学院信息工程学院,郑州451464

出　　处：《核电子学与探测技术》2023年第4期772-777,共6页Nuclear Electronics & Detection Technology

基　　金：河南省科技厅重点研发与推广专项项目(212102210321)。

摘　　要：为提高核辐射探测器电路故障诊断精度,提出了一种麻雀搜索算法(SSA)优化卷积神经网络(CNN)的核辐射探测器电路故障诊断新方法。该方法针对CNN网络参数选取主要依靠经验且对诊断精度有较大影响的实际,以SSA算法为优化算法进行CNN网络参数自适应选取,从而提高CNN网络诊断性能。通过核辐射探测器电路故障诊断实例对所提的SSA-CNN诊断方法进行了验证分析,结果表明所提方法提高了诊断精度,且比其他几种方法性能更突出。In order to improve the fault diagnosis accuracy of nuclear radiation detector circuit,a nuclear radiation detector circuit fault diagnosis method based on convolution neural network(CNN)optimized by sparrow search algorithm(SSA)was proposed.In view of the fact that the selection of CNN parameters mainly relies on experience and has a great impact on the diagnosis accuracy,the SSA was used as the optimization algorithm to carry out the adaptive selection of CNN parameters,so as to improve the diagnosis performance of CNN.The proposed SSA-CNN diagnosis method is verified and analyzed by a nuclear radiation detector circuit fault diagnosis example.The results show that the proposed method improves the diagnosis accuracy and has better performance than other methods.

关 键 词：卷积神经网络 麻雀搜索算法 核辐射探测器电路 故障诊断 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]