变工况大型旋转机械多重故障耦合机理及早期微弱故障特征提取与诊断研究

Under variable working conditions, the multiple faults coupling vibration signal of large rotating machinery becomes more complex in the time domain and frequency domain. Diagnosis is difficult. Especially in the stage of early fault, fault characteristic signal is weak, and is affected by the noise, make the diagnosis more difficult. This project takes gear and rolling bearing of large rotating machinery as the object of study and focuses on the problems of multiple fault coupling and early weak fault feature extraction and fault diagnosis under variable condition. By virtue of combining theoretical investigation with numeric calculation and experimnets, multiple fault coupling mechanism and fault characteristics are investigated. The feature extraction method of early weak fault signal under strong noise background, especially the characteristics of multi frequency weak signal extraction technology, is studied. Combining the empirical mode decomposition method and the analytical modal decomposition method, the high precision signal processing method is studied. The new denoising method based on multi stable stochastic resonance is studied. The new method to extract features of weak signal based on stochastic resonance and AMD-EMD is proposed. The fault diagnosis method based on time-frequency analysis and nonlinear dynamic characteristics is proposed. Finally, the experimental research and application examples and promotion are carried out. the project provides new theory and technology to solve the problem of accuracy and reliability of monitoring and diagnosis in multiple fault large rotating machinery.

在变工况条件下，大型旋转机械多重耦合故障振动信号在时域和频域中变得更加复杂，诊断难度加大。尤其是在早期故障阶段，故障特征信号微弱，同时受到噪声的影响，导致诊断变得更加困难。本项目以大型旋转机械齿轮-滚动轴承为研究对象，针对变工况条件下多重故障耦合及早期故障特征提取与诊断问题，通过理论分析、数值计算与实验相结合，揭示多重故障耦合机理及故障特征规律；研究强噪声背景下的机械故障早期微弱信号特征提取方法，尤其是多频微弱信号特征提取技术。探索将解析模态分解和经验模态分解相结合的高精度信号处理方法；研究多稳态随机共振去噪新方法，提出基于随机共振和AMD-EMD的微弱信号特征提取新方法；研究基于时频分析和非线性动力特征的故障诊断方法。最后进行实验研究和工程应用。项目研究成果将为有效解决大型旋转机械在多重故障耦合作用下的监测与诊断的准确性和可靠性问题提供新理论和新技术。

大型旋转机械多重故障振动信号在时域和频域中特征复杂，故障诊断难度加大。尤其是在早期故障阶段，故障特征信号微弱，同时受到噪声的影响，导致诊断变得更加困难。本项目以大型旋转机械齿轮-滚动轴承为研究对象，针对变工况条件下多重故障及早期故障特征提取与诊断问题，研究强噪声背景下的机械故障早期微弱信号特征提取方法，尤其是多频微弱信号特征提取技术。研究基于随机共振微弱信号检测方法和基于经验模态分解、解析模态分解等高精度信号处理方法及故障诊断方法。项目取得的主要研究成果如下：建立了轴承故障、削齿故障、齿根裂纹故障及削齿和齿根裂纹耦合故障动力学模型并进行数值仿真，研究了故障的动力学机理。通过变载荷激励下动力学特性对比，分析了不同故障类型在变载荷作用的下的动力学特性。提出变尺度多稳随机共振、参数补偿随机共振和级联多稳态随机共振的微弱信号检测方法；在此基础上，提出一种基于小波变换和参数补偿带通多稳随机共振的多频微弱信号检测方法；并研究了白噪声、乘性白噪声和加性色噪声、Levy噪声等多种噪声驱动下的多稳系统的随机共振模型。研究了基于解析模态分解、经验模态分解和变分模态分解时频分析方法，提出了多种用于噪声背景下的机械故障早期微弱故障特征提取及诊断的有效方法。本研究为噪声背景下的机械故障早期微弱故障特征提取及诊断提出多种方法。项目研究成果为大型旋转机械多重故障诊断提供了新的理论和技术。本项目通过四年的研究工作，已完成项目计划内容。共发表科技论文35篇，其中SCI收录16篇，EI收录15篇；获批发明专利3项。培养硕士研究生9名。