您好, 访客   登录/注册

面向广义数学形态颗粒特征的灰色马尔科夫剩余寿命预测方法

来源:用户上传      作者:

  摘要: 在滚动轴承状态监测与故障预测领域中,针对滚动轴承退化特征提取这一关键问题,提出了一种基于广义数学形态颗粒的特征提取新方法,该方法以数学形态颗粒分析为理论基础,在形态运算中引入腐蚀和膨胀算子,以计算出的广义数学形态颗粒值作为特征指标,定量地反映滚动轴承的性能退化程度。分别通过仿真信号和实例信号对该方法进行了有效性验证。在此基础上,为准确拟合滚动轴承性能退化过程的整体趋势与随机波动规律,将灰色马尔科夫模型应用到滚动轴承剩余寿命预测中,从而建立一种基于广义数学形态颗粒与灰色马尔科夫模型的剩余寿命预测方法。依托杭州轴承试验研究中心进行了滚动轴承疲劳寿命强化试验,以采集得到的轴承内圈全寿命试验数据验证了方法的有效性。关键词: 故障诊断; 滚动轴承; 广义数学形态颗粒; 灰色马尔科夫模型; 剩余寿命预测
  中图分类号:TH163+.3; TP306文献标志码: A文章编号: 10044523(2015)02031608
  DOI:10.16385/j.cnki.issn.10044523.2015.02.019
  引言
  滚动轴轴承一直都是旋转机械状态监测与故障预测领域的热门研究对象[1]。随着维修理论和相关技术的发展,基于状态的维修越来越得到人们关注。故障预测技术则是实现基于状态维修的核心,它能够估计损坏的部件或者子系统的剩余使用寿命,对于状态检修、预测与健康管理(CBM/PHM)系统是至关重要的[2]。退化特征提取是实现剩余寿命预测的基础,它着眼于分析特征参数能否反映设备从完好到完全失效的这一连续的性能退化过程,直接关系到预测的可信性[3]。文献[3]提取小波相关特征尺度熵作为性能退化的描述信息;文献[4,5]提取多尺度形态分解谱熵描述滚动轴承性能退化程度;文献[6]将小波包分解的节点能量构成特征向量,定量评估样本的退化程度;文献[7]则将通过循环平稳分析得到的组合切片累积能量作为预测特征值,取得了较好的效果。
  滚动轴承故障振动信号是一种典型的非平稳、非线性信号,从信号变化的本质分析,滚动轴承的振动信号在全寿命周期中的变化过程,就是其内部随机成分所占比例不断变化的过程[8]。利用数学形态颗粒分析等多尺度信号分析方法对信号进行分析描述,更易于从不同层次“剖析”信号的本质。
  本文对滚动轴承的退化特征提取与剩余寿命预测方法展开研究,详细阐述数学形态学颗粒分析的基本原理,提出基于广义数学形态颗粒的退化特征提取方法,针对滚动轴承退化过程的特点,研究基于广义数学形态颗粒与灰色马尔科夫模型的剩余寿命预测方法,并应用实测数据进行有效性验证分析。
  1数学形态颗粒分析
  1.1数学形态颗粒分析数学形态颗粒分析是一种有效处理图像粒度和形状特征的方法,是利用不同尺寸和形状的结构元素处理图像,了解其内部特征,目前广泛应用于图像形状和纹理特征描述、图像分割、图像复原和图像降噪等领域[9,10]。
  4结论
  本文提出了一种基于广义数学形态颗粒的退化特征提取方法,该方法所提取的广义数学形态颗粒能够有效地表征轴承性能退化程度,其有效性通过仿真和实例信号数据进行了验证。以其为基础,结合灰色马尔科夫预测模型,可以准确地预测滚动轴承的剩余寿命。应用轴承全寿命退化试验数据验证了该方法的有效性。
  参考文献:
  [1]Patil M S,Mathew J,Rajendra Kumar P K. Bearing signature analysis as a medium for fault detection: a review[J].Journal of Tribology,2008,130( 1) : 01400117.
  [2]George Wachtsevanos, Frank L.Lewis, Michael Roemer,et al. Intelligent Fault Diagnosis and Prognosis for Engineering Systems[M].袁海文,译. 北京:国防工业出版社,2013.
  [3]曾庆虎,邱静,刘冠军. 基于小波相关特征尺度熵的HSMM设备退化状态识别与故障预测方法[J]. 仪器仪表学报,2008,29(12):2 559―2 564.
  Zeng Qinghu, Qiu Jing, Liu Guanjun. Equipment degradation state recognition and fault progonsis method based on wavelet correlation feature scale entropy and HSMM[J]. Chinese Journal of Scientific Instrument, 2008,29(12):2 559―2 564.
  [4]王冰,李洪儒,许葆华. 基于多尺度形态分解谱熵的电机轴承预测特征提取及退化状态评估[J]. 振动与冲击,2013,32(22):124―128.
  WANG Bing,LI Hongru,XU Baohua.Motor bearing forecast feature extracting and degradation status identification based on multiscale morphological decomposition spectral entropy[J].Jourual of Vibration and Shock,2013,32(22),124―128.
  [5]王冰,李洪儒,许葆华. 基于多尺度形态学分解谱熵的电机轴承性能退化特征提取[J]. 轴承,2013,(8):43―47.   Wang Bing, Li Hongru,XU Baohua.Extraction of performance degradation feature for motor bearings base on multiscale morphological decomposition spectrum entropy[J].Bearing,2013,(8),43―47.
  [6]潘玉娜,陈进. 小波包支持向量数据描述在轴承性能退化评估中的应用研究[J]. 振动与冲击,2009,28(4):164―167.
  Pan Yuna, CHEN Jin. Wavelet package support vector date description applied in bearing performance degradation assessment[J]. Jourual of Vibration and Shock,2009,28(4):164―167.
  [7]潘玉娜,陈进. 结合循环平稳和支持向量数据描述的轴承性能退化评估研究[J]. 机械科学与技术,2009,28(4):442―445.
  Pan Yuna, Chen Jin. Assessment of bearing performance degradation by cyclostationarity analysis and support vector data description[J].Mechanical Science and Technology for Aerospace Engineering,2009,28(4):442―445.
  [8]潘玉娜. 滚动轴承的性能退化特征提取及评估方法研究[D]. 上海: 上海交通大学,2011.
  Pan Yuna.Study on feature extraction and assessment method of rolling element bearing performance degradation[D].Shanghai:Shanghai Jiao Tong University,2001.
  [9]Valero S, Chanussot J, Benediktsson J A, et al. Advanced directional mathematical morphology for the detection of the road network in very high resolution remote sensing images[J]. Pattern Recognition Letters, 2010,31(10):1 120―1 127.
  [10]Aptoula E, Lefuvre S. Morphological description of color images for content―based image retrieval[J]. IEEE Transactions on Image Processing, 2009,18(11):2 505―2 517.
  [11]徐东,徐永成,陈循,等. 基于EMD的灰色模型的疲劳剩余寿命预测方法研究[J]. 振动工程学报,2011,24(1):104―110.
  XU Dong, XU Yongcheng. CHEN Xun, et al. Residual fatigue life prediction based on grey model and EMD[J]. Journal of Vibration Engineering,2011,24(1):104―110.
  [12]郭阳明,姜红梅,翟正军. 基于灰色理论的自适应多参数预测模型[J]. 航空学报,2009,30(5):925―931.
  Guo Yanming,Jiang Hongmei,Zhai Zhengjun. Adptive mulitiparamenter predication model based on grey theory[J]. Acta Aeronautica et Astronautical Sinica, 2009,30(5):925―931.
  [13]李嵩松. 基于隐马尔可夫模型和计算智能的股票价格时间序列预测[D]. 哈尔滨:哈尔滨工业大学,2011.
  Li Songsong. Stock price time series prediction based on Hidden Markov Model and computational intelligence[D]. Harbin:Harbin Institute of Technology,2011.
  [14]Ramasso E, Gouriveau R. Prognostics in switching systems: evidential Markovian classification of realtime neurofuzzy predictions damage propagation modeling for aircraft engine runtofailure simulation[A]. IEEE Frontiers of Prognostics and Health Management Conference[C]. Besancon, France, 2010:1―10.
  [15]徐国平. 基于支持向量机的动调陀螺仪寿命预测方法研究[D]. 上海: 上海交通大学,2008.   Xu Guoping. Research on life forecasting methods of a DTG based on support vector machine[D].Shanghai: Shanghai Jiao Tong Unversity, 2008.
  Abstract: In the domain of rolling bearing condition monitor and fault prognosis, to solve the key problem of rolling bearing degenerate feature extraction, a new approach based on generalized mathematical morphological particle is proposed in the paper, the new approach, which is founded on the mathematical morphological particle analysis, introduces corrosion and dilation operators in morphological calculation and takes the calculated generalized mathematical morphological particle as feature indicator, therefore, the performance degenerate degree could be reflected in quantity. The effectiveness of this approach is test and verified with simulation and actual signal. On this basis, in order to describe the whole tendency and random fluctuating feature for rolling bearings, grey Markov model is applied in the remaining service life prediction for rolling bearing, A method of remaining service life prediction based on generalized mathematical morphological particle and grey Markov model is proposed thereby. Rolling bearing fatigued life testing was proceeded with Hangzhou Bearing Test & Research Center, the approach is proved effective with the collecting bearing inner race whole life data in fatigued life testing.
  Key words: fault diagnosis; rolling bearing; generalized mathematical morphological particle; grey Markov model; remaining service life prediction
转载注明来源:https://www.xzbu.com/8/view-11942595.htm