基于噪声分析的机械故障诊断方法研究.docVIP

基于噪声分析的机械故障诊断方法研究 摘要 基于噪声分析的机械故障诊断方法可以非接触地获得机械信号，适用于众多不便于使用振动传感器的场合，如某些高温、高腐蚀环境，是一种常用而有效地故障诊断方法。但在实际应用中，由于不相干噪声和环境噪声的影响，我们需要的待测信号往往被淹没在这些混合噪声中，信号的信噪比较低。 盲源分离作为数字信号处理领域的新兴技术，能利用观测信号恢复或提取独立的各个机械信号，在通讯、雷达信号处理、图像处理等众多领域具有重要的实用价值及发展前景，已经成为神经网络学界和信号处理学界的热点研究课题之一。 本文分析总结了盲源分离技术的相关研究现状，对盲源分离的原理、算法、相关应用作了探讨和研究。并就汽轮机噪声问题运用了盲源分离技术进行机械故障诊断，试验表明，该方法能将我们需要的故障信号从混合信号中分离出来，成功实现汽轮机部件的故障诊断。 关键词：声信号，机械故障诊断，独立分量分析 Investigation of Mechanical Fault Diagnosis Based on Noise Analysis Abstract You can obtain a non-contact method of mechanical fault diagnosis based on noise analysis of mechanical signals , not suitable for many occasions to facilitate the use of vibration sensors , such as certain high temperature , highly corrosive environment , is a common and effective fault diagnosis method . However, in practice , the effects of noise and extraneous ambient noise , the signal under test often need to be submerged in the mixed noise , lower signal to noise ratio .??Blind source separation as an emerging field of digital signal processing technology to take advantage of the observed signal recovery or extraction of various mechanical signals independently in many communications, radar signal processing , image processing has important practical value and development prospects , has become a neural network one of the hot research topic in academic circles and signal processing .??In this paper summarizes the research status of blind source separation techniques , the principles of blind source separation algorithms, related applications and research were discussed . Turbine noise problems and to use the blind source separation techniques for mechanical fault diagnosis, tests showed that the method we need fault signal can be separated from the mixed signal , fault diagnosis of steam turbine components successfully . Key Words：Mechanical Fault Diagnosis，Independent Component Analysis 目录 1绪论 1.1 选题背景 1.2 国内外研究发展现状 1.2.1机械故障诊断技术发展现状 1.2.2 声学故障诊断发展现状 1.2.3 盲源分离技术发展现状 1.3课题研究内容及意义 2 噪