基于SA-EMD-PNN的柱塞泵故障诊断方法研究
|
摘要
为了提高柱塞式液压泵的故障诊断效率和准确性,提出了SA-EMD-PNN柱塞泵故障诊断方法。提取各种状态下振动信号的特征参数,并对所提取特征参数进行敏感度分析(SA),找出敏感度较高的特征参数;对原始故障信号进行经验模态分解(EMD)结合,构造新的故障信号,再提取敏感度高的特征参数;将所提取特征参数以向量的形式输入概率神经网络(PNN)进行训练和测试。实验表明,SA-PNN方法能快速、有效的诊断出柱塞泵故障,减少诊断时间;而SA-EMD-PNN能在SA-PNN的基础上提高正确率。
In this paper, a plunger pump fault diagnosis method based on SA-EMD-PNN was proposed to improve the speed and accuracy of plunger pump fault diagnosis. First, feature parameters were extracted under a various states, and the sensitivity of the feature parameters was analyzed. Then, EMD was used to break down the original signal and reconstruct new signal. Feature parameters were extracted with higher sensitivity from the new fault signal. Finally, vectors with feature parameters that have the higher sensitivity were constituted. The vectors were used to train PNN. The trained PNN was used to diagnose the fault of a plunger pump. Experiment showed that SA-PNN can quickly and accurately diagnose the fault of the plunger pump. And compared to SA-PNN,the SA-EMD-PNN can has higher diagnosis accuracy.
In this paper, a plunger pump fault diagnosis method based on SA-EMD-PNN was proposed to improve the speed and accuracy of plunger pump fault diagnosis. First, feature parameters were extracted under a various states, and the sensitivity of the feature parameters was analyzed. Then, EMD was used to break down the original signal and reconstruct new signal. Feature parameters were extracted with higher sensitivity from the new fault signal. Finally, vectors with feature parameters that have the higher sensitivity were constituted. The vectors were used to train PNN. The trained PNN was used to diagnose the fault of a plunger pump. Experiment showed that SA-PNN can quickly and accurately diagnose the fault of the plunger pump. And compared to SA-PNN,the SA-EMD-PNN can has higher diagnosis accuracy.
引文
[1]许葆华,韩东,杜明.基于小波-平滑能量算子解调的液压泵故障分析[J].液压与气动,2014(8):32-35.XU Baohua,HAN Dong,DU Ming.Fault analysis based on wavelet-smoothed energy operator demodulation for hydraulic pump[J].Chinese Hydraulics&Pneumatics,2014(8):32-35.
[2]唐宏宾,吴运新,滑广军,等.基于EMD包络谱分析的液压泵故障诊断方法[J].振动与冲击,2012,31(9):44-48.TANG Hongbin,WU Yunxin,HUA Guangjun,et al.Fault diagnose of pump using EMD and envelope spectrum analysis[J].Journal of Vibration and Shock,2012,31(9):44-48.
[3]ZHAO Z,JIA M X,WANG F L,et al.Intermittent chaos and sliding window symbol sequence statistics-based early fault diagnosis for hydraulic pump on hydraulic tube tester[J].Mechanical Systems and Signal Processing,2009,23(5):1573-1585.
[4]马济乔,李洪儒,许葆华.基于多尺度局部最大样本熵的液压泵故障特征提取[J].液压与气动,2014(12):23-27.MA Jiqiao,LI Hongru,XU Baohua.The feature extraction of hydraulic diagnosis based onmultiscale local-maximum sample entropy[J].Chinese Hydraulics&Pneumatics,2014(12):23-27.
[5]敖银辉,汪宝生.基于小波包与概率神经网络的液压泵故障模式识别[J].机床与液压,2014,42(13):168-170.AO Yinghui,WANG Baosheng.Fault model recognition of hydraulic pump on wavelet packet and probability neural network[J].Machine Tool&Hydraulics,2014,42(13):168-170.
[6]全学海,丁宣浩,蒋英春.基于EMD和概率神经网络的说话人识别[J].桂林电子科技大学学报,2010,30(2):108-112.QUAN Xuehai,DING Xuanhao,JIANG Yingchun.Speaker recognition based on EMD and Probabilistic neural networks[J].Journal of Guilin University of Electronic Technology,2010,30(2):108-112.
[7]肖淑萍,孙亚红,杨建雄.基于EMD进化概率神经网络的纹理图像识别[J].计算机与数字工程,2014,9:1713-1716.XIAO Shuping,SUN Yahong,YANG Jianxiong.Texyure image recognition based on empirical mode decomposition and evolution probabilistic neural network[J].Computer&Digital Engineering,2014,9:1713-1716.
[8]李曜洲.经验模态分解的方法改进研究[D].湘潭:湖南科技大学,2016.
[9]BOUZID A,ELLOUZE N.Empirical mode decomposition of voiced speech signal[C]//First International Symposium on Control,Communications and Processing.Hammamet:IEEE,2004.
[10]BASSIUNYA A M,LI X L.Flute breakage detection during end milling using Hilbert Huang transform and smoothed nonlinear energy operator[J].International Journal of Machine Tools and Manufacture,2007,47(6):1011-1020.
[11]LAU B,RIEMENSCHNEIDERA S,XU Y.Gearbox fault diagnosis using empirical mode decomposition and Hilbert spectrum[J].Mechanical Systems and Signal Processing,2006,20:718-734.
[12]马孝江,王凤利,蔡悦,等.局域波时频分布在转子系统早期故障诊断中的应用研究[J].中国电机工程学报,2004,24(3):161-164.MA Xiaojiang,WANG Fengli,CAI Yue,et al.Application of local wave time-frequency distribution to early faults diagnosis of rotor system[J].Proceedings of the Csee,2004,24(3):161-164.
[13]ZHANG R R,MA S,SAFAK E,et al.Hilbert-Huang transform analysis of daynamic and earthquake motion recordings[J].Journal of Engineering Mechanise,2003,129(8):861-875.
[14]CHAPPELL M A,PAYNE S J.A method for the automated detection of venous gas bubbles in humansusing empirical mode decomposition[J].Annals of Biomedical Engineering,2005,33(10):1411-1421.
[15]曹端超.基于EMD-HMM的装备故障诊断与寿命预测研究[D].石家庄:军械工程学院,2012.
[16]陈波.基于粗糙集-概率神经网络结合的变压器故障诊断研究[D].南宁:广西大学,2008.
[17]王宏,徐长英,邓芳明.基于独立成分分析与概率神经网络的滚动轴承故障识别方法的研究[J].仪表技术与传感器,2016(12):161-164.WANG Hong,XU Changying,DENG Fangming.Research on fault identification for rolling bearing based on independent component analysis and probabilistic neural network[J].Instrument Technique and Sensor,2016(12):161-164.
[2]唐宏宾,吴运新,滑广军,等.基于EMD包络谱分析的液压泵故障诊断方法[J].振动与冲击,2012,31(9):44-48.TANG Hongbin,WU Yunxin,HUA Guangjun,et al.Fault diagnose of pump using EMD and envelope spectrum analysis[J].Journal of Vibration and Shock,2012,31(9):44-48.
[3]ZHAO Z,JIA M X,WANG F L,et al.Intermittent chaos and sliding window symbol sequence statistics-based early fault diagnosis for hydraulic pump on hydraulic tube tester[J].Mechanical Systems and Signal Processing,2009,23(5):1573-1585.
[4]马济乔,李洪儒,许葆华.基于多尺度局部最大样本熵的液压泵故障特征提取[J].液压与气动,2014(12):23-27.MA Jiqiao,LI Hongru,XU Baohua.The feature extraction of hydraulic diagnosis based onmultiscale local-maximum sample entropy[J].Chinese Hydraulics&Pneumatics,2014(12):23-27.
[5]敖银辉,汪宝生.基于小波包与概率神经网络的液压泵故障模式识别[J].机床与液压,2014,42(13):168-170.AO Yinghui,WANG Baosheng.Fault model recognition of hydraulic pump on wavelet packet and probability neural network[J].Machine Tool&Hydraulics,2014,42(13):168-170.
[6]全学海,丁宣浩,蒋英春.基于EMD和概率神经网络的说话人识别[J].桂林电子科技大学学报,2010,30(2):108-112.QUAN Xuehai,DING Xuanhao,JIANG Yingchun.Speaker recognition based on EMD and Probabilistic neural networks[J].Journal of Guilin University of Electronic Technology,2010,30(2):108-112.
[7]肖淑萍,孙亚红,杨建雄.基于EMD进化概率神经网络的纹理图像识别[J].计算机与数字工程,2014,9:1713-1716.XIAO Shuping,SUN Yahong,YANG Jianxiong.Texyure image recognition based on empirical mode decomposition and evolution probabilistic neural network[J].Computer&Digital Engineering,2014,9:1713-1716.
[8]李曜洲.经验模态分解的方法改进研究[D].湘潭:湖南科技大学,2016.
[9]BOUZID A,ELLOUZE N.Empirical mode decomposition of voiced speech signal[C]//First International Symposium on Control,Communications and Processing.Hammamet:IEEE,2004.
[10]BASSIUNYA A M,LI X L.Flute breakage detection during end milling using Hilbert Huang transform and smoothed nonlinear energy operator[J].International Journal of Machine Tools and Manufacture,2007,47(6):1011-1020.
[11]LAU B,RIEMENSCHNEIDERA S,XU Y.Gearbox fault diagnosis using empirical mode decomposition and Hilbert spectrum[J].Mechanical Systems and Signal Processing,2006,20:718-734.
[12]马孝江,王凤利,蔡悦,等.局域波时频分布在转子系统早期故障诊断中的应用研究[J].中国电机工程学报,2004,24(3):161-164.MA Xiaojiang,WANG Fengli,CAI Yue,et al.Application of local wave time-frequency distribution to early faults diagnosis of rotor system[J].Proceedings of the Csee,2004,24(3):161-164.
[13]ZHANG R R,MA S,SAFAK E,et al.Hilbert-Huang transform analysis of daynamic and earthquake motion recordings[J].Journal of Engineering Mechanise,2003,129(8):861-875.
[14]CHAPPELL M A,PAYNE S J.A method for the automated detection of venous gas bubbles in humansusing empirical mode decomposition[J].Annals of Biomedical Engineering,2005,33(10):1411-1421.
[15]曹端超.基于EMD-HMM的装备故障诊断与寿命预测研究[D].石家庄:军械工程学院,2012.
[16]陈波.基于粗糙集-概率神经网络结合的变压器故障诊断研究[D].南宁:广西大学,2008.
[17]王宏,徐长英,邓芳明.基于独立成分分析与概率神经网络的滚动轴承故障识别方法的研究[J].仪表技术与传感器,2016(12):161-164.WANG Hong,XU Changying,DENG Fangming.Research on fault identification for rolling bearing based on independent component analysis and probabilistic neural network[J].Instrument Technique and Sensor,2016(12):161-164.