基于深度学习的复杂模拟电路故障识别

doi:10.16180/j.cnki.issn1007-7820.2025.02.004

摘要/Abstract

摘要：

复杂的模拟电路故障传递关系复杂,故障类型与故障特征之间存在复杂的非线性关系,导致特征提取困难和故障识别困难。针对该问题,文中提出了一种两测点、自适应噪声完备集合经验模态分解(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise, CEEMDAN)-多尺度伪彩色图像-AlexNet网络结合的模拟电路故障诊断方法。将复杂模拟电路两测点输出信号前后相接构造新的组合故障样本,两测点组合故障样本数据提高了表征复杂模拟电路整体故障状态的能力。将组合的故障样本信号进行多尺度分解,并将多尺度数据矩阵映射为二维伪彩色图,形成信息丰富、特征明显的多尺度故障伪彩色图像。利用深度学习模型AlexNet优异的图像特征挖掘和学习能力,将不同故障类型的多尺度伪彩色图像输入AlexNet中进行模型迁移训练学习并完成故障识别。通过对简单电路和复杂电路的单双故障及混合故障识别的比较分析,证明了所提模拟电路故障诊断方法对复杂模拟电路的不同故障类型能达到更高的识别准确率。

关键词: 模拟电路, 故障诊断, 双故障, CEEMDAN, 伪彩色, AlexNet, 低通滤波器, 特征可视化

Abstract:

For complex analog circuits with complex fault transfer relationships and complex nonlinear relationships between fault types and fault features, which cause difficulties in feature extraction and fault identification. This study presents a fault diagnosis method for analog circuits based on two measuring points-CEEMDAN(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise)-multi-scale false-color image-ALEXNet network. A new combined fault sample is constructed by connecting the output signals of two measuring points in a complex analog circuit. The combined fault sample data of two measuring points improves the ability to characterize the whole fault state of the complex analog circuit. The combined fault sample signals are decomposed in multi-scale, and the multi-scale data matrix is mapped to a two-dimensional false-color graph to form a multi-scale false-color image with abundant information and obvious features. Based on AlexNet's excellent image feature mining and learning ability, multi-scale false-color images of different fault types were input into AlexNet for model transfer training and fault identification. By comparing and analyzing the single and double faults and mixed faults of simple circuit and complex circuit, it is proved that the proposed method can achieve higher recognition accuracy for different fault types of complex analog circuit.

Key words: analog circuit, fault diagnosis, double fault, CEEMDAN, false color, AlexNet, low pass filter, feature-based visualization

中图分类号:

TN710

黄泽华, 毕贵红, 张梓睿. 基于深度学习的复杂模拟电路故障识别[J]. 电子科技, 2025, 38(2): 23-34.

HUANG Zehua, BI Guihong, ZHANG Zirui. Fault Identification of Complex Analog Circuit Based on Deep Learning[J]. Electronic Science and Technology, 2025, 38(2): 23-34.

图/表 16

图1

图2

图3

图4

表1

表2

图5

图6

图7

图8

图9

图10

图11

图12

图13

表3

参考文献 22

[1]	Binu D, Kariyappa B S. A survey on fault diagnosis of analog circuits: Taxonomy and state of the art[J]. International Journal of Electronics and Communications, 2017, 73(1):68-83.
[2]	熊建. 基于时频分析和复数域的模拟电路故障诊断研究[D]. 成都: 电子科技大学,2017:1-62.
	Xiong Jian. Research on fault diagnostic methods for analog circuits based on time-frequency analysis and complex field analysis[D]. Chengdu: University of Electronic Science and Technology of China,2017:1-62.
[3]	刘雪梅. 基于经验小波变换和残差网络的模拟电路故障诊断方法研究[D]. 成都: 电子科技大学,2022:1-74.
	Liu Xuemei. Research on fault diagnosis method of analog circuit based on empirical wavelet transform and residual Network[D]. Chengdu: University of Electronic Science and Technology of China,2022:1-74.
[4]	王月海, 吕晓雯, 程冉, 等. 基于FCM-LLE的模拟电路故障诊断[J]. 北方工业大学学报, 2020, 32(2):85-89.
	Wang Yuexhai, Lü Xiaowen, Cheng Ran, et al. A FCM-LLE-based fault diagnosis algorithm for analog circuits[J]. Journal of North China University of Technology, 2020, 32(2):85-89.
[5]	王力, 贾欣雨. 基于PSO-PF-SVM的模拟电路故障诊断[J]. 江苏大学学报(自然科学版), 2023, 44(2):221-228,234.
	Wang Li, Jia Xinyu. Fault diagnosis of simulation circuit based on PSO-PF-SVM[J]. Journal of Jiangsu University (Natural Science Edition), 2023, 44(2):221-228,234.
[6]	刘沛霖, 刘美容, 何怡刚, 等. 基于改进的VMD和SVM的模拟电路故障诊断方法的研究[J]. 微电子学与计算机, 2022, 39(11):85-94.
	Liu Peilin, Liu Meirong, He Yigang, et al. Research on fault diagnosis method of analog circuit based on improved VMD and SVM[J]. Microelectronics and Computers, 2022, 39(11):85-94.
[7]	张朝龙, 何怡刚, 杜博伦. 基于DBN特征提取的模拟电路早期故障诊断方法[J]. 仪器仪表学报, 2019, 40(10):112-119.
	Zhang Chaolong, He Yigang, Du Bolun. Analog circuit incipient fault diagnosis method based on DBN feature extraction[J]. Chinese Journal of Scientific Instrument, 2019, 40(10):112-119.
[8]	易灵芝, 肖伟红, 于文新, 等. 基于深度学习的模拟电路故障诊断算法[J]. 计算机工程与应用, 2018, 54(24):143-148. doi: 10.3778/j.issn.1002-8331.1708-0334
	Yi Lingzhi, Xiao Weihong, Yu Wenxin, et al. Fault diagnosis algorithm for analog circuits based on deep learning.[J]. Computer Engineering and Applications, 2018, 54(24):143-148. doi: 10.3778/j.issn.1002-8331.1708-0334
[9]	傅晨琦, 季利鹏, 孙伟卿, 等. 基于深度残差网络的模拟电路软故障诊断方法[J]. 飞控与探测, 2021, 4(4):74-81.
	Fu Chenqi, Ji Lipeng, Sun Weiqing, et al. Deep residual learning-based soft fault diagnosis method for analog circuits[J]. Flight Control and Detection, 2021, 4(4):74-81.
[10]	杜先君, 巩彬, 余萍, 等. 基于CBAM-CNN的模拟电路故障诊断[J]. 控制与决策, 2022, 37(10):2609-2618.
	Du Xianjun, Gong Bin, Yu Ping, et al. CBAM-CNN based analog circuit fault diagnosis[J]. Control and Decision, 2022, 37(10):2609-2618.
[11]	张逸, 欧杰宇, 金涛, 等. 基于特征图像组合与改进ResNet-18的电能质量扰动识别方法[J]. 中国电机工程学报, 2024, 44(7):2531-2544.
	Zhang Yi, Ou Jieyu, Jin Tao, et al. Power quality disturbance recognition method based on feature image combination and modified ResNet-18[J]. Proceedings of the CSEE, 2024, 44(7):2531-2544.
[12]	侯思祖, 郭威, 王子奇, 等. 基于小波AlexNet网络的配电网故障区段定位方法[J]. 电测与仪表, 2022, 59(3):46-57.
	Hou Sizu, Guo Wei, Wang Ziqi, et al. Fault segment location method for distribution network based on wavelet AlexNet network[J]. Electrical Measurement and Instrumentation, 2022, 59(3):46-57.
[13]	张永超. 模拟电路的测点选择及故障诊断方法研究[D]. 北京: 北京化工大学,2022:3-5.
	Zhang Yongchao. Research on test point selection and fault diagnosis method of analog circuit[D]. Beijing: Beijing University of Chemical Technology,2022:3-5.
[14]	潘强, 孙必伟. 模拟电路故障诊断中的特征信息提取[J]. 电子科技, 2013, 26(8):116-119.
	Pan Qiang, Sun Biwei. Feature information extraction in fault diagnosis of analog circuits[J]. Electronic Science and Technology, 2013, 26(8):116-119.
[15]	Torres M E, Colominas M A, Schlotthauer G, et al. A complete ensemble empirical mode decomposition with adaptive noise[C]. Prague: IEEE International Conference on Acoustics, Speech and Signal Processing,2011:4144-4147.
[16]	李瑞, 范玉刚. 基于CEEMDAN多尺度排列熵和SO-RELM的高压隔膜泵单向阀故障诊断[J]. 振动与冲击, 2023, 42(5):127-135.
	Li Rui, Fan Yugang. Fault diagnosis of one-way valve of high-pressure diaphragm pump based on CEEMDAN multi-scale permutation entropy and SO-RELM[J]. Journal of Vibration and Shock, 2023, 42(5):127-135.
[17]	周小波, 邹任玲, 卢旭华, 等. 基于人工神经网络的表面肌电信号分类器研究进展[J]. 电子科技, 2021, 34(2):62-67.
	Zhou Xiaobo, Zou Renling, Lu Xuhua, et al. Research progress of surface electromyography signal classifier based on artificial neural network[J]. Electronic Science and Technology, 2021, 34(2):62-67.
[18]	陈军锋, 王雪, 张效天. 非侵入式负荷识别边缘计算颜色编码研究[J]. 仪器仪表学报, 2020, 41(9):12-19.
	Chen Junfeng, Wang Xue, Zhang Xiaotian. Non-intrusive load recognition using color encoding in edge computing[J]. Chinese Journal of Scientific Instrument, 2020, 41(9):12-19.
[19]	王力, 张露露. 基于IHHO-BP神经网络的模拟电路故障诊断[J]. 电子测量与仪器学报, 2024, 38(5):238-248.
	Wang Li, Zhang Lulu. Fault diagnosis of analog circuit based on IHHO-BP neural network[J]. Journal of Electronic Measurement and Instrumentation, 2024, 38(5):238-248.
[20]	颜学龙, 马润平. 基于深度极限学习机的模拟电路故障诊断[J]. 计算机工程与科学, 2019, 41(11):1911-1918.
	Yan Xuelong, Ma Runping. Fault diagnosis of analog circuits based on deep extreme learning machine[J]. Computer Engineering and Science, 2019, 41(11):1911-1918.
[21]	Yang J L, Gao T Y, Jiang S D. A dual-input fault diagnosis model based on SE-MSCNN for analog circuits[J]. Applied Intelligence, 2022, 53(6):128-136..
[22]	张朝龙, 何怡刚, 袁莉芬, 等. 基于GMKL-SVM的模拟电路故障诊断方法[J]. 仪器仪表学报, 2016, 37(9):1989-1995.
	Zhang Chaolong, He Yigang, Yuan Lifen, et al. Analog circuit fault diagnosis based on GMKL-SVM method[J]. Chinese Journal of Scientific Instrument, 2016, 37(9):1989-1995.

故障编号	故障类型	元件标称值	容差范围/%	元件故障值
F₁	NF
F₂	C₁↑	0.01 nF	10	0.020 nF
F₃	C₂↑	0.01 nF	10	0.020 nF
F₄	C₃↓	0.01 nF	10	0.005 nF
F₅	C₄↑	0.01 nF	10	0.020 nF
F₆	R₁↑	1 570.00 kΩ	5	3 140.000 kΩ
F₇	R₇↓	2.64 kΩ	5	1.320 kΩ
F₈	R₁₃↑	1.80 kΩ	5	3.600 kΩ
F₉	R₁₄↓	4.84 kΩ	5	2.420 kΩ
F₁₀	R₁₆↓	3.00 kΩ	5	1.500 kΩ
F₁₁	R₁₈↑	10.00 kΩ	5	20.000 kΩ
F₁₂	R₁₉↓	10.00 kΩ	5	5.000 kΩ

故障编号	故障类型	元件标称值	元件故障值
F₁	R₁↓C₁↑	1 570.00 kΩ,0.01 nF	785.00 kΩ,0.020 nF
F₂	R₁↓C₄↓	1 570.00 kΩ,0.01 nF	785.00 kΩ,0.005 nF
F₃	R₁↓R₂↓	1 570.00 kΩ,2.20 kΩ	785.00 kΩ,1.100 kΩ
F₄	R₁↓R₁₈↓	1 570.00 kΩ,10.00 kΩ	785.00 kΩ,5.000 kΩ
F₅	R₁₃↓C₃↑	1.80 kΩ,0.01 nF	0.90 kΩ,0.020 nF
F₆	R₁₃↓C₄↑	1.80 kΩ,0.01 nF	0.90 kΩ,0.020 nF
F₇	R₁₃↓R₁₄↓	1.80 kΩ,4.84 kΩ	0.90 kΩ,2.420 kΩ
F₈	R₁₃↓R₁₉↑	1.80 kΩ,10.00 kΩ	0.90 kΩ,20.000 kΩ
F₉	R₁₈↓C₃↑	10.00 kΩ,0.01 nF	5.00 kΩ,0.020 nF
F₁₀	R₁₈↓R₇↑	10.00 kΩ,2.64 kΩ	5.00 kΩ,5.280 kΩ
F₁₁	R₁₈↑C₁↓	10.00 kΩ,0.01 nF	20.00 kΩ,0.005 nF
F₁₂	R₁₈↑C₂↑	10.00 kΩ,0.01 nF	20.00 kΩ,0.020 nF
F₁₃	C₂↑C₃↓	0.01 nF,0.01 nF	0.02 nF,0.005 nF
F₁₄	C₂↑C₄↓	0.01 nF,0.01 nF	0.02 nF,0.005 nF
F₁₅	C₂↑R₁₆↓	0.01 nF,3.00 kΩ	0.02 nF,1.500 kΩ
F₁₆	C₂↑R₂₀↑	0.01 nF,2.32 kΩ	0.02 nF,4.640 kΩ

故障诊断方法	电路	准确率
故障诊断方法	电路	单故障/%	双故障/%	混合故障/%
语频谱+VGG16	Sallen-key低通滤波器	100.00	-	-
IHHO+BP	Sallen-key低通滤波器	-	-	95.77
IHHO+BP	三运放CSTV低通滤波电路	-	-	94.85
PSO+PF+SVM	Sallen-key低通滤波器	95.60	-	-
PSO+PF+SVM	四运放二阶高通滤波器	94.40	-	-
ELM-AE	四运放二阶高通滤波器	100.00	-	-
SE-MSCNN	Leapfrog电路(五运放)	-	-	83.58
GMKL-SVM+QPSO	Leapfrog电路(五运放)	-	97.30	-
FCM+LIE	二级四运放双二阶低通滤波器	93.70	-	-
DBN+SVM	二级四运放双二阶低通滤波器	98.13	-	-
单测点+CEEMDAN+伪彩色+AlexNet	二级四运放双二阶低通滤波器	97.67	94.75	79.00
双测点+CEEMDAN+伪彩色+AlexNet	二级四运放双二阶低通滤波器	100.00	99.50	99.00