基于EMD-PNN网络的刚性罐道故障诊断方法
|
摘要
针对刚性罐道故障种类识别困难,提出了一种基于经验模态分解-概率神经网络(EMD-PNN)的刚性罐道故障诊断方法。首先,搭建立井提升实验平台,使用北京东方振动和噪声技术研究所的INV3062T0设备采集罐道的振动信号,然后对含噪的振动信号进行EMD降噪;其次,提取降噪后振动信号的能量参数、偏度参数、峰度参数、波形参数、峰值参数、峭度参数、脉冲参数、裕度参数构成特征向量,作为PNN网络输入层的训练样本和测试样本;最后,利用训练样本建立PNN网络模型,选取测试样本检测概率神经网络的模式识别效果。实验证明,本方法对台阶凸起故障、接头错位故障和正常状态3种模式的识别率达到100%,为立井提升等非线性非平稳复杂系统的故障诊断提供一种通用可行的解决方案。
Aiming at the difficulty in identifying fault types of rigid cage guide, a fault diagnosis method for rigid cage guide based on EMD-PNN network is proposed. Firstly, the vertical shaft lifting experiment platform is set up, the INV3062 T0 equipment of China Orient Institute of Noise & Vibration is used to collect the vibration signals of the rigid cage guide, and then EMD noise reduction on the noisy vibration signals is performed. Secondly, the energy parameter, skewness parameter, kurtosis parameter, waveform parameter, peak value parameter, kurtosis parameter, pulse parameter and margin parameter of the vibration signal after noise reduction are extracted to form the feature vector, which will be used as the training sample and test sample of PNN network input layer. Finally, the training samples are used to establish the PNN network model, and the test samples are selected to detect the patterns recognition effect of the probabilistic neural network. The experiment proves that the recognition rate of the three patterns of step protrusion fault, joint misalignment fault and normal state can reach 100%, which provides a universal and feasible solution for fault diagnosis of nonlinear non-stationary complex systems such as vertical shaft lifting.
Aiming at the difficulty in identifying fault types of rigid cage guide, a fault diagnosis method for rigid cage guide based on EMD-PNN network is proposed. Firstly, the vertical shaft lifting experiment platform is set up, the INV3062 T0 equipment of China Orient Institute of Noise & Vibration is used to collect the vibration signals of the rigid cage guide, and then EMD noise reduction on the noisy vibration signals is performed. Secondly, the energy parameter, skewness parameter, kurtosis parameter, waveform parameter, peak value parameter, kurtosis parameter, pulse parameter and margin parameter of the vibration signal after noise reduction are extracted to form the feature vector, which will be used as the training sample and test sample of PNN network input layer. Finally, the training samples are used to establish the PNN network model, and the test samples are selected to detect the patterns recognition effect of the probabilistic neural network. The experiment proves that the recognition rate of the three patterns of step protrusion fault, joint misalignment fault and normal state can reach 100%, which provides a universal and feasible solution for fault diagnosis of nonlinear non-stationary complex systems such as vertical shaft lifting.
引文
[1] 李爱民,包从望.基于EMD-SVM的刚性罐道故障诊断研究[J].煤炭技术,2016,35(4):264-266.LI AI M,BAO C W.Research on fault diagnosis of rigid tank road based on EMD-SVM[J].Coal Technology,2016,35(4):264-266.
[2] CHEN J Y,ZHOU D,LV C.An integrated method based on CEEMD-SampEn and the correlation analysis algorithm for the fault diagnosis of a gearbox under different working conditions[J].Mechanical Systems and Signal Processing,2018(113):102-111.
[3] 年夫顺.关于故障预测与健康管理技术的几点认识[J].仪器仪表学报,2018,39(8):1-14.NIAN F SH.Some understandings about fault prediction and health management techniques [J].Chinese Journal of Scientific Instrument,2018,39 (8):1-14.
[4] 韩建群,石旭东.基于EMD航空发电机定子绕组故障检测方法[J].国外电子测量技术,2016,35(10):34-37,49.HAN J Q,SHI X D.Detection method of stator winding fault based on EMD aeronautical generator[J].Foreign Electronic Measurement Technology,2016,35(10):34-37,49.
[5] 宋伟,尹涛.基于概率神经网络(PNN)的矿井提升机故障诊断研究[J].电子测量技术,2016,39(11):187-189,194.SONG W,YIN T.Research on fault diagnosis of mine hoist based on probabilistic neural network (PNN)[J].Electronic Measurement Technology,2016,39(11):187-189,194.
[6] 贾瑞生,赵同彬,孙红梅,等.基于经验模态分解及独立成分分析的微震信号降噪方法[J].地球物理学报,2015,58(3):1013-1023.JIA R SH,ZHAO T B,SUN H M,et al.Micro-seismic signal denoising method based on empirical mode decomposition and independent component analysis[J].Chinese Journal of Geophysics,2015,58(3):1013-1023.
[7] 蒋永华,焦卫东,李荣强,等.应用自适应带宽信号的BS-EMD混叠消除[J].振动与冲击,2018,37(16):83-90.JIANG Y H,JIAO W D,LI R Q,et al.BS-EMD aliasing cancellation using adaptive bandwidth signals[J].Journal of Vibration and Shock,2018,37(16):83-90.
[8] 赵志宏,杨绍普,申永军.一种改进的EMD降噪方法[J].振动与冲击,2009,28(12):35-37,62,200-201.ZHAO ZH H,YANG SH P,SHEN Y J.An improved EMD noise reduction method[J].Journal of Vibration and Shock,2009,28(12):35-37,62,200-201.
[9] 陈法法,刘帅,肖文荣,等.混沌粒子群优化RVM的滚动轴承早期故障诊断[J].电子测量与仪器学报,2018,32(8):9-16.CHEN F F,LIU SH,XIAO W R,et al.Early fault diagnosis of rolling bearings based on chaotic particle swarm optimization RVM[J].Journal of Electronic Measurement and Instrumentation,2018,32(8):9-16.
[10] WANG D,ZHAO Y,YI C,et al.Sparsity guided empirical wavelet transform for fault diagnosis of rolling element bearings[J].Mechanical Systems and Signal Processing,2018(101):292-308.
[11] SUN X Y,KANG F N,WANG M M,et al.Improved probabilistic neural network PNN and its application to defect recognition in rock bolts[J].International Journal of Machine Learning & Cybernetics,2016,7(5):909-919.
[12] CHEN X Y,ZHOU J Z,XIAO H.Fault diagnosis based on comprehensive geometric characteristic and probability neural network[J].Applied Mathematics and Computation,2014,230(3):542-554.
[13] 张坤硕,齐华,任安虎,等.超声信号的改进EMD阈值方法降噪研究[J].国外电子测量技术,2017,36(11):15-20.ZHANG K SH,QI H,REN AN H,et al.Research on denoising of improved EMD threshold method for ultrasonic signals[J].Foreign Electronic Measurement Technology,2017,36(11):15-20.
[14] 张玲玲,肖静.基于MATLAB的机械故障诊断技术案例教程[M].北京:高等教育出版社,2016.ZHANG L L,XIAO J.A Case Study of Mechanical Fault Diagnosis based on MATLAB [M].Beijing:Higher Education Press,2016.
[15] 程皓,张志杰,张浩,等.改进EMD算法在爆炸冲击波后处理中的应用[J].电子测量技术,2018,41(23):78-81.CHENG H,ZHANG ZH J,ZHANG H,et al.Application of improved EMD algorithm in post-explosion shock wave processing[J].Electronic Measurement Technology,2018,41(23):78-81.
[16] LIU J X,WEN X B,YUAN L M.A robust approach of watermarking in contourlet domain based on probabilistic neural network[J].Multimedia Tools and Application,2017,76(22):24009-24026.
[17] 范虹.非平稳信号特征提取方法及其应用[M].北京:科学出版社,2013.FAN H.Non-Stationary Signal Feature Extraction Method and Its Application[M].Beijing:Science Press,2013.
[18] 马天兵,王孝东,杜菲.基于小波包和BP神经网络的刚性罐道故障诊断[J].工矿自动化,2018,44(8):76-80.MA T B,WANG X D,DU F.Fault diagnosis of rigid tank road based on wavelet packet and BP neural network[J].Industrial and Mining Automation,2018,44(8):76-80.
[19] 蒋玉强.立井刚性罐道系统的非线性耦合特性及状态评估研究[D].北京:中国矿业大学,2011:1-151.JIANG Y Q.Vertical shaft rigid tank system nonlinear coupling characteristics and state assessment [D].Beijing:China University of Mining and Technology,2011:1-151.
[20] 赵光权,葛强强,刘小勇.基于DBN的故障特征提取及诊断方法研究[J].仪器仪表学报,2016,37(9):1946-1953.ZHAO G Q,GE Q Q,LIU X Y.Research on fault feature extraction and diagnosis based on DBN[J].Chinese Journal of Scientific Instrument,2016,37(9):1946-1953.
[21] 陈明.MATLAB神经网络原理与实例精解[M].北京:清华大学出版社,2013.CHEN M.Principles and Examples of MATLAB Neural Network [M].Beijing:Tsinghua University Press,2013.
[2] CHEN J Y,ZHOU D,LV C.An integrated method based on CEEMD-SampEn and the correlation analysis algorithm for the fault diagnosis of a gearbox under different working conditions[J].Mechanical Systems and Signal Processing,2018(113):102-111.
[3] 年夫顺.关于故障预测与健康管理技术的几点认识[J].仪器仪表学报,2018,39(8):1-14.NIAN F SH.Some understandings about fault prediction and health management techniques [J].Chinese Journal of Scientific Instrument,2018,39 (8):1-14.
[4] 韩建群,石旭东.基于EMD航空发电机定子绕组故障检测方法[J].国外电子测量技术,2016,35(10):34-37,49.HAN J Q,SHI X D.Detection method of stator winding fault based on EMD aeronautical generator[J].Foreign Electronic Measurement Technology,2016,35(10):34-37,49.
[5] 宋伟,尹涛.基于概率神经网络(PNN)的矿井提升机故障诊断研究[J].电子测量技术,2016,39(11):187-189,194.SONG W,YIN T.Research on fault diagnosis of mine hoist based on probabilistic neural network (PNN)[J].Electronic Measurement Technology,2016,39(11):187-189,194.
[6] 贾瑞生,赵同彬,孙红梅,等.基于经验模态分解及独立成分分析的微震信号降噪方法[J].地球物理学报,2015,58(3):1013-1023.JIA R SH,ZHAO T B,SUN H M,et al.Micro-seismic signal denoising method based on empirical mode decomposition and independent component analysis[J].Chinese Journal of Geophysics,2015,58(3):1013-1023.
[7] 蒋永华,焦卫东,李荣强,等.应用自适应带宽信号的BS-EMD混叠消除[J].振动与冲击,2018,37(16):83-90.JIANG Y H,JIAO W D,LI R Q,et al.BS-EMD aliasing cancellation using adaptive bandwidth signals[J].Journal of Vibration and Shock,2018,37(16):83-90.
[8] 赵志宏,杨绍普,申永军.一种改进的EMD降噪方法[J].振动与冲击,2009,28(12):35-37,62,200-201.ZHAO ZH H,YANG SH P,SHEN Y J.An improved EMD noise reduction method[J].Journal of Vibration and Shock,2009,28(12):35-37,62,200-201.
[9] 陈法法,刘帅,肖文荣,等.混沌粒子群优化RVM的滚动轴承早期故障诊断[J].电子测量与仪器学报,2018,32(8):9-16.CHEN F F,LIU SH,XIAO W R,et al.Early fault diagnosis of rolling bearings based on chaotic particle swarm optimization RVM[J].Journal of Electronic Measurement and Instrumentation,2018,32(8):9-16.
[10] WANG D,ZHAO Y,YI C,et al.Sparsity guided empirical wavelet transform for fault diagnosis of rolling element bearings[J].Mechanical Systems and Signal Processing,2018(101):292-308.
[11] SUN X Y,KANG F N,WANG M M,et al.Improved probabilistic neural network PNN and its application to defect recognition in rock bolts[J].International Journal of Machine Learning & Cybernetics,2016,7(5):909-919.
[12] CHEN X Y,ZHOU J Z,XIAO H.Fault diagnosis based on comprehensive geometric characteristic and probability neural network[J].Applied Mathematics and Computation,2014,230(3):542-554.
[13] 张坤硕,齐华,任安虎,等.超声信号的改进EMD阈值方法降噪研究[J].国外电子测量技术,2017,36(11):15-20.ZHANG K SH,QI H,REN AN H,et al.Research on denoising of improved EMD threshold method for ultrasonic signals[J].Foreign Electronic Measurement Technology,2017,36(11):15-20.
[14] 张玲玲,肖静.基于MATLAB的机械故障诊断技术案例教程[M].北京:高等教育出版社,2016.ZHANG L L,XIAO J.A Case Study of Mechanical Fault Diagnosis based on MATLAB [M].Beijing:Higher Education Press,2016.
[15] 程皓,张志杰,张浩,等.改进EMD算法在爆炸冲击波后处理中的应用[J].电子测量技术,2018,41(23):78-81.CHENG H,ZHANG ZH J,ZHANG H,et al.Application of improved EMD algorithm in post-explosion shock wave processing[J].Electronic Measurement Technology,2018,41(23):78-81.
[16] LIU J X,WEN X B,YUAN L M.A robust approach of watermarking in contourlet domain based on probabilistic neural network[J].Multimedia Tools and Application,2017,76(22):24009-24026.
[17] 范虹.非平稳信号特征提取方法及其应用[M].北京:科学出版社,2013.FAN H.Non-Stationary Signal Feature Extraction Method and Its Application[M].Beijing:Science Press,2013.
[18] 马天兵,王孝东,杜菲.基于小波包和BP神经网络的刚性罐道故障诊断[J].工矿自动化,2018,44(8):76-80.MA T B,WANG X D,DU F.Fault diagnosis of rigid tank road based on wavelet packet and BP neural network[J].Industrial and Mining Automation,2018,44(8):76-80.
[19] 蒋玉强.立井刚性罐道系统的非线性耦合特性及状态评估研究[D].北京:中国矿业大学,2011:1-151.JIANG Y Q.Vertical shaft rigid tank system nonlinear coupling characteristics and state assessment [D].Beijing:China University of Mining and Technology,2011:1-151.
[20] 赵光权,葛强强,刘小勇.基于DBN的故障特征提取及诊断方法研究[J].仪器仪表学报,2016,37(9):1946-1953.ZHAO G Q,GE Q Q,LIU X Y.Research on fault feature extraction and diagnosis based on DBN[J].Chinese Journal of Scientific Instrument,2016,37(9):1946-1953.
[21] 陈明.MATLAB神经网络原理与实例精解[M].北京:清华大学出版社,2013.CHEN M.Principles and Examples of MATLAB Neural Network [M].Beijing:Tsinghua University Press,2013.