一种基于支持向量机的宽带声谱故障监测方法

doi:10.16180/j.cnki.issn1007-7820.2025.06.006

摘要/Abstract

摘要：

针对复杂背景噪声环境下的机电设备状态监测问题,文中提出了一种基于支持向量机(Supported Vector Machine, SVM)的宽带声谱故障监测方法。与传统基于加速度传感器和振动分析的窄带处理故障监测方法不同,所提方法采用宽带声学传感器进行声谱信号采集。利用功率谱熵和多尺度熵运算进行背景噪声抑制和声纹提取,基于支持向量机进行宽带声纹识别和分类处理,将支持向量机的模式识别能力与功率谱熵和多尺度熵的背景噪声抑制及声纹提取能力有机结合。所提方法的故障信息宽带拾取能力强,干扰噪声抑制能力强,设备状态信息利用率高,声纹跟踪匹配准确,能够准确监测和识别复杂背景噪声环境下的机电设备运行故障。仿真分析和实验结果均证明了所提方法的有效性。

关键词: 声纹识别, 声谱分析, 宽带匹配, 状态监测, 支持向量机, 机电设备, 故障监测, 振动分析

Abstract:

In view of the problem of condition monitoring of electromechanical equipment under complex background noise, a wideband sound spectrum fault monitoring method based on SVM (Supported Vector Machine) is proposed in this study. Different from the traditional narrowband fault monitoring method based on acceleration sensor and vibration analysis, the proposed method uses a broadband acoustic sensor to collect the sound spectrum signal. Power spectral entropy and multiscale entropy computations are employed for background noise suppression and voiceprint extraction. Based on SVM, the pattern recognition capability of SVM is combined with the background noise suppression capability of power spectrum entropy and multi-scale entropy. The proposed method has strong broadband fault information picking ability, strong interference noise suppression ability, high equipment status information utilization rate, accurate voiceprint tracking and matching, and can accurately monitor and identify the operation faults of electromechanical equipment under complex background noise environment, which is verified by simulation and experimental results.

Key words: voiceprint recognition, spectrum analysis, broadband matching, condition monitoring, SVM, electromechanical equipment, fault monitoring, vibration analysis

中图分类号:

TP181

竹小锋, 白兴宇. 一种基于支持向量机的宽带声谱故障监测方法[J]. 电子科技, 2025, 38(6): 39-44.

ZHU Xiaofeng, BAI Xingyu. Broadband Acoustic Spectrum Fault Monitoring Technology Based on Support Vector Machine[J]. Electronic Science and Technology, 2025, 38(6): 39-44.

图/表 9

图1

图2

图3

图4

表1

图5

图6

图7

图8

参考文献 19

[1]	任贵粉, 刘增力, 孙伟栋. 机械故障诊断技术研究综述[J]. 农业装备与车辆工程, 2021, 59(2):69-73,78.
	Ren Guifen, Liu Zengli, Sun Weidong. Review of research on mechanical fault detection technology[J]. Agricultural Equipment & Vehicle Engineering, 2021, 59(2):69-73,78.
[2]	吴景红. 煤矿机械故障诊断研究现状及发展趋势[J]. 煤炭工程, 2023, 55(6):187-192.
	Wu Jinghong. Research status and development trend of mining machinery fault diagnosis[J]. Coal Engineering, 2023, 55(6):187-192.
[3]	陈人教. 旋转机械故障诊断方法及其应用研究[J]. 中国新技术新产品, 2023(12):60-63.
	Chen Renjiao. Fault diagnosis method and application of rotating machinery[J]. New Technologies and New Products of China, 2023(12):60-63.
[4]	王俊波, 武利会, 罗容波, 等. 基于小波变换—振动起始时刻分析的高压断路器故障诊断方法[J]. 高压电器, 2020, 56(6):211-217.
	Wang Junbo, Wu Lihui, Luo Rongbo, et al. Fault diagnosis method of high voltage circuit breaker based on wavelet transform-vibration start time analysis[J]. High Voltage Apparatus, 2020, 56(6):211-217.
[5]	虎雷. 复杂机械振动系统噪声源诊断技术研究[D]. 哈尔滨: 哈尔滨工程大学,2009:1-4.
	Hu Lei. A research of diagnostic technique based on noise source in complex system of mechanical vibration[D]. Harbin: Harbin Engineering University,2009:1-4.
[6]	潘华锋, 叶勇, 王军, 等. 在线振动监测、诊断系统在行星减速机机组中的应用[J]. 聚酯工业, 2020, 33(2):51-55.
	Pan Huafeng, Ye Yong, Wang Jun, et al. Application of on-line vibration monitoring and diagnosis system in planetary reducer unit[J]. Polyester Industry, 2020, 33(2):51-55.
[7]	梁斌, 刘泽, 牛延博. 基于多尺度模糊熵和支持向量机的采煤机截割模式识别[J]. 煤炭工程, 2021, 53(5):131-137. doi: 10.11799/ce202105025
	Liang Bin, Liu Ze, Niu Yanbo. Shearer cutting pattern recognition based on multi-scale fuzzy entropy and support vector machine[J]. Coal Engineering, 2021, 53(5):131-137.
[8]	蔡坤烨, 蔡景, 周迪, 等. 基于SVM方法的APU故障预测方法[J]. 南京航空航天大学学报, 2019, 51(4):466-473.
	Cai Kunye, Cai Jing, Zhou Di, et al. APU fault prediction based on SVM method[J]. Journal of Nanjing University of Aeronautics & Astronautics, 2019, 51(4):466-473.
[9]	Hasan T, Hansen J H L. A study on universal background model training in speaker verification[J]. IEEE Transactions on Audio,Speech and Language Processing, 2011, 19(7):1780-1789.
[10]	Harne R L, Wang T Y, Leng Y G, et al. Application of stochastic resonance signal recovery[J]. Chinese Journal of Mechanical Engineering, 2009, 21(3):425-434.
[11]	Kim B, Kong S H, Kim S. Low computational enhancement of STFT-based parameter estimation[J]. IEEE Journal of Fourier Analysis and Applications, 2015, 6(8):155-198.
[12]	Scanlon P, Kavanagh F, Boland F M. Residual life prediction of rotating machines using acoustic noise signals[J]. IEEE Transactions on Instrumentation & Measurement, 2013, 62(1):79-102.
[13]	李可, 姚忠远, 王枭, 等. 基于声纹的GIS断路器机械故障诊断[J]. 振动、测试与诊断, 2023, 43(5):866-871.
	Li Ke, Yao Zhongyuan, Wang Xiao, et al. Mechanical fault diagnosis of GIS circuit breaker based on voiceprint[J]. Journal of Vibration,Measurement & Diagnosis, 2023, 43(5):866-871.
[14]	汪徐阳, 易映萍, 李田丰. 一种用于变压器故障诊断的卷积神经网络优化方法[J]. 电子科技, 2023, 36(12):79-86.
	Wang Xuyang, Yi Yingping, Li Tianfeng. A convolutional neural network optimization method for fault diagnosis of power transformer[J]. Electronic Science and Technology, 2023, 36(12):79-86.
[15]	谢爱平, 张雨生, 刘莹, 等. 基于瞬时特征参数和功率谱熵的联合调制识别[J]. 电子科技, 2022, 35(11):104-110.
	Xie Aiping, Zhang Yusheng, Liu Ying, et al. Joint modulation recognition based on instantaneous feature and power spectrum entropy[J]. Electronic Science and Technology, 2022, 35(11):104-110.
[16]	柴兴亮. 基于改进CEEMDAN和多尺度模糊熵的气阀故障诊断[J]. 组合机床与自动化加工技术, 2020, 10(33):140-145.
	Chai Xingliang. Gas valve fault diagnosis based on improved CEEMDAN and multi-scale fuzzy entropy[J]. Modular Machine Tool & Automatic Manufacturing Technique, 2020, 10(33):140-145.
[17]	郑近德, 陈焱, 童靳于, 等. 精细广义复合多元多尺度反向散布熵及其在滚动轴承故障诊断中的应用[J]. 中国机械工程, 2023, 34(11):1315-1325.
	Zheng Jinde, Chen Yan, Tong Jinyu, et al. RGCMvMRDE and its applications in rolling bearing fault diagnosis[J]. China Mechanical Engineering, 2023, 34(11):1315-1325. doi: 10.3969/j.issn.1004-132X.2023.11.007
[18]	范玉刚, 张由振. 基于多尺度排列熵和正则化RVFL的高压隔膜泵单向阀故障诊断[J]. 云南大学学报(自然科学版), 2023, 45(1):38-47.
	Fan Yugang, Zhang Youzhen. Fault diagnosis of check valve of high-pressure diaphragm pump based on multi-scale permutation entropy and regularized RVFL[J]. Journal of Yunnan University(Natural Science Edition), 2023, 45(1):38-47.
[19]	于磊, 陈森, 张瑞. 深度支持向量机在齿轮故障诊断中的应用[J]. 机械传动, 2019, 43(8):150-156. doi: 10.16578/j.issn.1004.2539.2019.08.028
	Yu Lei, Chen Sen, Zhang Rui. Application of deep support vector machine in gear fault diagnosis[J]. Journal of Mechanical Transmission, 2019, 43(8):150-156. doi: 10.16578/j.issn.1004.2539.2019.08.028

信号状态	标签	多尺度熵均值	功率谱熵值
正常信号S(t)	A	0.523 1	0.891 3
正常信号S(t)	A	0.531 8	0.904 2
正常信号S(t)	A	0.491 2	0.920 2
正常信号S(t)	A	0.501 1	0.884 0
正常信号S(t)	A	0.510 2	0.914 5
故障信号S₁(t)	B	0.822 3	1.210 3
故障信号S₁(t)	B	0.819 8	1.205 4
故障信号S₁(t)	B	0.782 3	1.209 7
故障信号S₁(t)	B	0.793 4	1.183 4
故障信号S₁(t)	B	0.784 2	1.194 2
故障信号S₂(t)	C	1.210 2	1.402 3
故障信号S₂(t)	C	1.182 4	1.523 1
故障信号S₂(t)	C	1.191 4	1.518 2
故障信号S₂(t)	C	1.217 8	1.492 4
故障信号S₂(t)	C	1.198 4	1.490 2