摘要伴随着现代化工业的发展,旋转机械在电力、化工等行业中的地位显得日益重要。机械设备故障诊断技术对于保证设备安全、可靠、高效的运行具有重要的理论意义和应用价值。旋转机械的故障诊断是由故障信号提取、故障特征信息提取以及模式识别三部分构成。故障特征提取作为最关键的问题之一,对故障诊断准确性和故障早期预报可靠性的影响举足轻重。基于经验模态分解(Empirical Mode Decomposition, EMD)和基于神经网络的故障诊断方法是近年来发展起来的两种方法。本文将经验模态分解的时频分析方法与神经网络技术相结合,构成基于EMD和神经网络的旋转机械故障诊断系统。22770
关键词 旋转机械 故障诊断 特征提取 经验模态分解
毕业设计说明书(毕业论文)外文摘要
Title Implementation and verification of rotating
machinery fault feature extraction method
Abstract
With the development of modern industry, the status of rotating machinery in the power, chemical and other industries is becoming increasingly important. Machinery equipment fault diagnosis technology for ensuring safe, reliable and efficient operation has important theoretical significance and application value. Fault diagnosis of rotating machinery consists of three parts, i.e. fault signal acquisition, fault feature information extraction, and pattern recognition. Fault feature extraction as one of the most critical issues affects the accuracy of fault diagnosis and fault early warning reliability. Methods based on empirical mode decomposition (Empirical Mode Decomposition, EMD) and neural network have been developed in recent years. In this thesis, the empirical mode decomposition time-frequency analysis and neural network technology are combined to form a rotating machinery fault diagnosis system.
Keywords Rotating Machinery Fault diagnosis Feature Extraction Empirical Mode Decomposition
目 次
1 引言(或绪论)1
1.1旋转机械故障诊断的研究意义 1
1.2旋转机械故障诊断的国内外研究现状1
1.3基于EMD时频分析方法3
1.4本文的主要内容和结构安排4
2 基于EMD时频分析基本理论和算法6
2.1基本概念 6
2.2EMD方法的基本原理 6
3.人工神经网络9
4基于EMD和BP网络的旋转机械故障诊断方法13
4.1振动信号层13
4.2EMD分解层16
4.3特征向量提取层16
4.4神经网络训练层16
4.5故障模式分类层16
5算法18
5.1EMD算法18
5.2 BP神经网络算法20
6实验内容23
6.1数据处理23
6.1.1数据分析23
6.1.2特征值提取24
6.1.3数据规格化25
6.2实验步骤27
6.3实验结果28
结论 30
致谢 31
参考文献32