风机轴承故障诊断中的振动信号特征提取方法研究

随着风电产业的快速发展以及对风机系统高可靠性、易维护性等的各方面要求，风机状态监测与故障诊断技术引起了学术界和工业界的广泛关注。轴承作为风机机械传动系统和发电机系统的核心部件，其运行状态的实时监测和准确分析，对整个风机的故障诊断和运行维护均具有重要的意义。本文针对风机轴承故障诊断中的振动信号特征提取问题展开研究，运用局部均值分解(Local Mean Decomposition, LMD)瞬态信号分解技术、信息熵和非线性动力学参数分析，分别从瞬态特征描述和非线性特征分析两个角度，对风机轴承振动信号特征提取方法进行理论研究和实验验证，为轴承状态监测和故障诊断提供了有效的理论方法。论文主要工作如下：
(1)  在分析风机轴承运行特点、故障机理及其振动故障特征的基础上，针对复杂工况下风机轴承振动信号非平稳、非线性特征难以提取及量化问题，研究基于 LMD的瞬态信号分解技术和基于信息熵的信号特征定量描述方法，实现风机轴承振动信号特征的有效提取和准确描述。
(2)  研究基于 LMD 和信息熵的 Wigner-Ville 谱熵的振动信号瞬态能量特征提取方法，用于定量刻画轴承不同状态下振动信号的时频能量分布的规律，并设计基于LS-SVM 的智能故障诊断模型，实现轴承状态和故障类型的自动分类与识别。仿真分析和风机轴承诊断实验验证了该方法和模型较好的特征提取与故障诊断效果。
(3)  从非线性动力学角度出发，提出基于 LMD 的多尺度排序熵的轴承振动信号非线性特征提取方法，有效刻画轴承振动信号的非线性复杂度特征，实现了轴承内不同故障程度的有效识别。仿真分析和风机轴承故障诊断实验验证了该方法的有效性。 
关键词：振动信号；特征提取；风机轴承；局部均值分解；排序熵；故障诊断

石家庄风机厂石家庄风机石家庄风机销售石家庄风机维修

Abstract
With  the  rapid  development  of  wind  power  industry,  the  reliability  and maintainability of the wind turbine system are very urgent. Condition monitoring and fault diagnosis technology of wind turbine system cause the extensive concern of the academia and  industry.  Bearings  are  as  the  core  components  of  wind  turbine  mechanical transmission  system  and  generator  system,  there  has  a  realistic  significance  to  make  a condition monitoring and fault diagnosis to them. In this paper, aimed at feature extraction methods  of  wind  turbine  bearing  diagnosis, applied  Local  Mean  Decomposition  (LMD), Shannon  entropy  and  nonlinear  dynamic  parameters,  in  view  of  the  transient characteristics description and nonlinear feature analysis research, the proposed methods are verified by simulation experiment and experimental platform. The proposed methods provide a solution for wind turbine bearing condition monitoring and fault diagnosis. The specific research ways are as follows: Firstly, the wind turbine bearing operation characteristics and the failure mechanism are  discussed,  besides,  aiming  at  nonstationary  and  nonlinear  characteristics  of  bearings, the transient signal decomposition technique based on LMD are studied; the quantitative description method based on information entropy are analyzed. Both of that is in order toeffective extraction and accurate description of wind turbine bearing vibratory signals.Secondly,  a  transient  characteristic  extraction  method  based  on  LMD  and Wigner-Ville  spectral  entropy  is  proposed,  in  order  to  quantitatively  describe  thetime-frequency energy distribution of bearing vibratory signals under different condition. After  that,  a  intelligent  fault  diagnosis  model  based  on  LS-SVM  is  used  for  automaticclassification  and  recognition  of  bearing  faults.  Simulation  experiment  and  experimental platform verified the proposed method and diagnosis model. Finally, in view of nonlinear dynamics, a nonlinear feature extraction method named a multi-scale permutation entropy based on LMD is proposed. The proposed method can effectively represent nonlinear complexity characteristics of bearing vibratory signals and identify  different  fault  degree  of  bearing.  Simulation  experiment  and  experimental  III platform verified the proposed method.
Keywords:  vibratory  signals;  feature  extraction;  wind  turbine  bearings;  local  mean decomposition (LMD); permutation entropy; fault diagnosis