利用多传感器信号指标的风扇状态监测方法研究

吴川辉; 程华利; 陈英武; 李明超; 刘力源

doi:10.13433/j.cnki.1003-8728.20220033

利用多传感器信号指标的风扇状态监测方法研究

doi: 10.13433/j.cnki.1003-8728.20220033

珠海市精实测控技术有限公司，广东珠海　519000

详细信息

作者简介:
吴川辉（1987−），中级工程师，硕士，研究方向为机电系统故障诊断，eric.wu@prmeasure.com

通讯作者:
程华利，中级工程师，硕士，huali.cheng@prmeasure.com

中图分类号: TH165 ⁺ .3
计量
- 文章访问数: 92
- HTML全文浏览量: 49
- PDF下载量: 15
- 被引次数: 0
出版历程
- 收稿日期: 2021-05-27
- 刊出日期: 2023-06-25

Study on Fan Condition Monitoring Method Using Multi-sensor Signal Index

P & R Measurement Technology Co., Ltd., Zhuhai 519000, Guangdong, China

摘要

摘要: 本文提出一种基于多传感器信号指标的风扇状态监测方法。首先使用多种常用的传感器对风扇进行拾取信号，采集不同状态下的多种传感器信号，并且计算出信号的指标；然后采用主成分分析法（PCA）对多传感器信号指标进行降维处理，提取可以表征风扇运行状态的主要成分指标；最后使用对非线性特征指标进行学习具有优势的循环神经网络（RNN）进行预测分析，实现对风扇状态的有效监测和故障识别，并通过实验验证了本文方法的有效性。
- 风扇 /
- 主成分分析 /
- 循环神经网络 /
- 状态监测
Abstract: This paper presents a fan condition monitoring method based on multi-sensor signal indexes. Firstly, a variety of commonly used sensors are used to pick up the signals of the fan, collect the signals of various sensors in different states, and calculate the indicators of the signals. Secondly, the principal component analysis is used to reduce the dimension of multi-sensor signal indexes, simplify the state detection model and extract the main component indexes which can represent the running state of the fan. Finally, the RNN, which has the advantage of learning the nonlinear characteristic indexes, is used to predict and analyze the fan state, so as to realize the effective monitoring of the fan state. Experimental results show the effectiveness of the present method.
- fan /
- principal component analysis /
- recurrent neural network /
- condition monitoring

HTML全文

图 1 RNN基本结构图

下载: 全尺寸图片幻灯片

图 2 方法流程图

下载: 全尺寸图片幻灯片

图 3 采集流程图

下载: 全尺寸图片幻灯片

图 4 准确率和损失曲线图

下载: 全尺寸图片幻灯片

表 1 相关系数矩阵

指标	C₂	C₁	T	C₃	C₄	C₅
C₂	1	0.95	0.92	0.87	0.85	0.93
C₁	0.95	1	0.87	0.83	0.81	0.88
T	0.92	0.87	1	0.80	0.78	0.85
C₃	0.87	0.83	0.80	1	0.74	0.81
C₄	0.85	0.81	0.78	0.74	1	0.79
C₅	0.93	0.88	0.85	0.81	0.79	1

下载: 导出CSV

表 2 各指标贡献率

特征指标	成分	贡献率/%	累计贡献率/%
C₂	1	42.61	42.6
C₁	2	16.39	59.0
C₅	3	15.20	74.2
T	4	9.30	83.5
C₃	5	8.90	92.4
C₄	6	7.60	100.0

下载: 导出CSV

参考文献(11)

[1]	JIN X H, AZARIAN M H, LAU C, et al. Physics-of-failure analysis of cooling fans[C]//2011 Prognostics and System Health Management Conference. Shenzhen: IEEE, 2011: 1-8
[2]	JIN X H, MA E W M, CHOW T W S, et al. An investigation into fan reliability[C]//Proceedings of the IEEE 2012 Prognostics and System Health Management Conference. Beijing: IEEE, 2012
[3]	MIAO Q, AZARIAN M, PECHT M. Cooling fan bearing fault identification using vibration measurement[C]//2011 IEEE Conference on Prognostics and Health Management. Denver: IEEE, 2011
[4]	OH H, AZARIAN M H, PECHT M. Estimation of fan bearing degradation using acoustic emission analysis and mahalanobis distance[C]//Virginia Beach, 2011
[5]	程礼, 杨武奎, 梁涛, 等. 基于声模态的压气机/风扇气路故障诊断[J]. 机械工程学报, 2019, 55(13): 38-44. doi: 10.3901/JME.2019.13.038 CHENG L, YANG W K, LIANG T, et al. Failure diagnosis on gas path of fan/compressor based on acoustic mode[J]. Journal Of Mechanical Engineering, 2019, 55(13): 38-44. (in Chinese) doi: 10.3901/JME.2019.13.038
[6]	姚景瑜, 孟海洋, 杨京, 等. 基于BP人工神经网络的轴流风扇气动噪声预测[J]. 南京大学学报(自然科学), 2020, 56(6): 900-908. YAO J Y, MENG H Y, YANG J, et al. Aerodynamic noise prediction of tube-axial fan based on BP artificial neural network[J]. Journal of Nanjing University (Natural Science), 2020, 56(6): 900-908. (in Chinese)
[7]	杨林源. 工程机械水温高故障分析与解决方法[J]. 内燃机与配件, 2021(8): 110-112. doi: 10.3969/j.issn.1674-957X.2021.08.047 YANG L Y. Analysis and solution of high water temperature fault in construction machinery[J]. Internal Combustion Engine & Parts, 2021(8): 110-112. (in Chinese) doi: 10.3969/j.issn.1674-957X.2021.08.047
[8]	赵仕通, 肖黎, 屈文忠. 框架结构非线性损伤的主成分分析识别方法研究[J]. 机械科学与技术, 2018, 37(1): 8-12. ZHAO S T, XIAO L, QU W Z. Study on nonlinear damage identification of frame structure with a principal component analysis method[J]. Mechanical Science and Technology for Aerospace Engineering, 2018, 37(1): 8-12. (in Chinese)
[9]	冯人海, 赵政, 谢生, 等. 基于主成分分析和凸优化的低压配电网拓扑识别方法[J]. 天津大学学报(自然科学与工程技术版), 2021, 54(7): 746-753. FENG R H, ZHAO Z, XIE S, et al. Topology identification for low voltage network based on principal component analysis and convex optimization[J]. Journal of Tianjin University (Science and Technology), 2021, 54(7): 746-753. (in Chinese)
[10]	张成帆, 江泽鹏, 曹伟, 等. 一维多尺度卷积神经网络及其在滚动轴承故障诊断中的应用[J]. 机械科学与技术, 2022, 41(1): 120-126. doi: 10.13433/j.cnki.1003-8728.20200309 ZHANG C F, JIANG Z P, CAO W, et al. One-dimensional multi-scale convolution neural network and its application in rolling bearing fault diagnosis[J]. Mechanical Science and Technology for Aerospace Engineering, 2022, 41(1): 120-126. (in Chinese) doi: 10.13433/j.cnki.1003-8728.20200309
[11]	冯志鹏, 褚福磊. 行星齿轮箱故障诊断的扭转振动信号分析方法[J]. 中国电机工程学报, 2013, 33(14): 101-106. FENG Z P, CHU F L. Torsional vibration signal analysis method for fault diagnosis of planetary gearboxes[J]. Proceedings of the CSEE, 2013, 33(14): 101-106. (in Chinese)