基于QPSO-SVR和声发射信号的机械密封寿命预测
|
摘要
针对现有机械密封监测技术难以有效预测剩余使用寿命的问题,提出基于声发射特征融合的退化指标和QPSO-SVR寿命预测模型的机械密封剩余寿命预测方法。首先通过实验采集多组机械密封的全寿命数据,进行小波阈值降噪处理,从原始声发射信号中提取出能表征机械密封运行状态的特征,利用KPCA分析优化得到的声发射特征,然后通过马氏距离对得到的特征进行融合进而得到能够表征机械密封退化的指标,利用QPSO优化SVR模型参数,建立寿命预测模型。实验结果显示:基于退化指标和QPSO-SVR模型的寿命预测方法有着较好的泛化能力和较高的精度,具有良好的工业前景。
Mechanical seal life is an important index for the operation of large rotating equipment in process industries such as petrochemical industry.Aimed at the deficiencies of existing study of mechanical seal in prediction of using life,and the unknown of regression model,a new prediction method of mechanical seal was proposed based on extracting features from AE signal and support vector regression.Several groups of full life data in mechanical seal were acquired in experiment and were processed by wavelet denoising,the features in time domain and frequency domain representing the mechanical seal operation state were extracted,the mechanical seal degradation index was acquired by treating the high-dimensional features with KPCA and Mahalanobis distance,and a prediction model was got by inputting the index into QPSO to estimate SVR parameters.The experiment result of application shows that life predication method based on AE signal features and QPSO-SVR model has strong generalization and high prediction precision.
Mechanical seal life is an important index for the operation of large rotating equipment in process industries such as petrochemical industry.Aimed at the deficiencies of existing study of mechanical seal in prediction of using life,and the unknown of regression model,a new prediction method of mechanical seal was proposed based on extracting features from AE signal and support vector regression.Several groups of full life data in mechanical seal were acquired in experiment and were processed by wavelet denoising,the features in time domain and frequency domain representing the mechanical seal operation state were extracted,the mechanical seal degradation index was acquired by treating the high-dimensional features with KPCA and Mahalanobis distance,and a prediction model was got by inputting the index into QPSO to estimate SVR parameters.The experiment result of application shows that life predication method based on AE signal features and QPSO-SVR model has strong generalization and high prediction precision.
引文
[1]张尔卿,傅攀,陈侃,等.机械密封端面接触状态的声发射监测方法[J].润滑与密封,2014,39(8):8-13.ZHANG E Q,FU P,CHEN K,et al.Mechanical fluid seal face contact state monitoring based on acoustic emission[J].Lubrication Engineering,2014,39(8):8-13.
[2]SINGLETON R K,STRANGAS E G,SELIN A.Extended Kalmanfiltering for remaining-useful-life estimation of bearings[J].IEEE Transactions on Industrial Electronics,2015,62(3):1781-1790.
[3]YIN S,DING S X,XIE X.et al.A review on basic data-driven approaches for industrial process monitoring[J].IEEE Transactions on Industrial Electronics,2014,61(11):6148-6428.
[4]王录才,王芳.泡沫金属制备、性能及应用[M].北京:国防工业出版社,2015.
[5]CHOUDHURY A,TANDON N.Application of acoustic emission technique for the detection of defects in rolling element bearings[J].Tribology International,2000,33(1):39-45.
[6]孙见君,涂桥安.接触式机械密封基本性能研究进展[J].流体机械,2015(2):41-47.SUN J J,TU Q A.Research progress in basic properties of contacting mechanical seal[J].Fluid Machinery,2015(2):41-47.
[7]郭福平,李衡,沈书乾.基于马尔可夫模型的滚动轴承声发射故障诊断研究[J].机床与液压,2018,46(17):177-179.GUO F P,LI H,SHEN S Q.Research on fault diagnosis of the rolling bearing acoustic emission Based on the Markov model[J].Machine Tool&Hydraulics,2018,46(17):177-179.
[8]焦宏超.基于SVR的旋转机械耦合故障诊断方法研究[D].北京:华北电力大学,2016.
[9]蒋恩超,傅攀,张思聪.小波包和GA-SVM在轴承故障诊断中的应用[J].计算机测量与控制,2017,25(10):7-10.JIANG E C,FU P,ZHANG S C.Application of wavelet packet and GA-SVM in bearing fault diagnosis[J].Computer Measurement&control,2017,25(10):7-10.
[10]ZHANG X Y,LIANG Y T,ZHONG J.A novel bearing fault diagnosis model integrated permutation entropy,ensemble empirical mode decomposition and optimized SVM[J].Measurement,2015,69:164-179.
[11]刘珍翔.基于改进GA-SVR的机械关键部件寿命预测及维修策略研究[D].重庆:重庆大学,2016.
[12]李军.机械密封失效机理分析与效能延长策略[J].工业技术创新,2017,4(3):199-201.LI J.Mechanism analysis of mechanical sealing failure and its efficiency extension strategy[J].Industrial Technology Innovation,2017,4(3):199-201.
[13]姚力.几种典型声发射信号的特征参数分布分析[J].无损探伤,2004,28(2):19-22.
[14]SUN J,XU W B.Aglobal search strategy of quantum behaved particle swarm optimization[C]//Proceedings of IEEE Conference on Cybernetics and Intelligent Systems.Singapore:IEEE,2004:111-116.
[2]SINGLETON R K,STRANGAS E G,SELIN A.Extended Kalmanfiltering for remaining-useful-life estimation of bearings[J].IEEE Transactions on Industrial Electronics,2015,62(3):1781-1790.
[3]YIN S,DING S X,XIE X.et al.A review on basic data-driven approaches for industrial process monitoring[J].IEEE Transactions on Industrial Electronics,2014,61(11):6148-6428.
[4]王录才,王芳.泡沫金属制备、性能及应用[M].北京:国防工业出版社,2015.
[5]CHOUDHURY A,TANDON N.Application of acoustic emission technique for the detection of defects in rolling element bearings[J].Tribology International,2000,33(1):39-45.
[6]孙见君,涂桥安.接触式机械密封基本性能研究进展[J].流体机械,2015(2):41-47.SUN J J,TU Q A.Research progress in basic properties of contacting mechanical seal[J].Fluid Machinery,2015(2):41-47.
[7]郭福平,李衡,沈书乾.基于马尔可夫模型的滚动轴承声发射故障诊断研究[J].机床与液压,2018,46(17):177-179.GUO F P,LI H,SHEN S Q.Research on fault diagnosis of the rolling bearing acoustic emission Based on the Markov model[J].Machine Tool&Hydraulics,2018,46(17):177-179.
[8]焦宏超.基于SVR的旋转机械耦合故障诊断方法研究[D].北京:华北电力大学,2016.
[9]蒋恩超,傅攀,张思聪.小波包和GA-SVM在轴承故障诊断中的应用[J].计算机测量与控制,2017,25(10):7-10.JIANG E C,FU P,ZHANG S C.Application of wavelet packet and GA-SVM in bearing fault diagnosis[J].Computer Measurement&control,2017,25(10):7-10.
[10]ZHANG X Y,LIANG Y T,ZHONG J.A novel bearing fault diagnosis model integrated permutation entropy,ensemble empirical mode decomposition and optimized SVM[J].Measurement,2015,69:164-179.
[11]刘珍翔.基于改进GA-SVR的机械关键部件寿命预测及维修策略研究[D].重庆:重庆大学,2016.
[12]李军.机械密封失效机理分析与效能延长策略[J].工业技术创新,2017,4(3):199-201.LI J.Mechanism analysis of mechanical sealing failure and its efficiency extension strategy[J].Industrial Technology Innovation,2017,4(3):199-201.
[13]姚力.几种典型声发射信号的特征参数分布分析[J].无损探伤,2004,28(2):19-22.
[14]SUN J,XU W B.Aglobal search strategy of quantum behaved particle swarm optimization[C]//Proceedings of IEEE Conference on Cybernetics and Intelligent Systems.Singapore:IEEE,2004:111-116.