【作者】 邢钧；

【导师】 高立新；

【作者基本信息】 北京工业大学 ， 计算机应用技术， 2002， 硕士

【摘要】 小波分析是近年来迅速发展起来的一门理论，在图象处理、通信和地球物理研究方面取得了成功的应用。小波包变换是小波变换的推广，共同构成了小波分析。本论文围绕小波变换和小波包变换在齿轮箱故障诊断中的应用进行了理论研究和实验验证，将小波分析这一时频分析方法应用到机械设备故障诊断领域。主要内容如下： 首先，研究了小波分析在故障诊断中的意义，从诊断理论和故障信号特点两方面说明对故障信号进行小波分析是诊断过程的内在要求；其次，研究了小波分析的理论基础，重点说明其工程解释，并通过分析具体的性能指标得到适合故障信号分析的小波母函数；第三、研究了小波分析在非平稳信号分析方面的实际应用，说明了小波分析对此类信号分析有着傅立叶分析不可比拟的优势。成功的通过小波分析提取出武钢大型厂高速线材轧机齿轮箱输出接手螺栓拉断故障特征，捕捉到故障信息，为今后判断、预防同类事故及事后确定故障部位提供了有效的分析手段。第四、研究了小波分析在信号消噪方面的实际应用，分别对仿真信号、实际信号进行了小波消噪及小波包消噪，取得了较好的结果。最后研究了小波分析在频带能量特征提取中的实际应用，通过实例分析，说明其可以较好的诊断故障。 综上所述：将小波分析应用在齿轮箱等机械的故障诊断，可以提取其它方法无法得到的一些重要特征。为发现设备隐患，进行故障诊断提供了一种可行及有效的方法。更多还原

【Abstract】 Wavelet Analysis is a theory developed in the past decade. It has been successfully used in image processing, telecommunication and Geophysics. Wavelet Packet Transform is a generation of Wavelet Transform, and both of them are the main block of Wavelet Analysis. In this thesis, some research and application work with wavelet and wavelet packet in the field of fault diagnosis of GearBox has been conducted. And wavelet analysis, a new time and frequency analysis method, is used in the field of Mechanical Fault Diagnosis. This work comprises of the following parts :Firstly, some research on diagnosis theory and fault signal is carried out and concluded that the diagnosis process needs wavelet analysis; Secondly, theory basis of wavelet analysis is researched. Emphasis to its explanation of engineering is given. According to analyze concrete performance, the appropriate wavelet function for diagnosis is obtained; Thirdly, research on the application of wavelet transform in nonsteady signal is conducted. It is made out that wavelet analysis in these signal has a great advantage over fourier analysis. From wavelet analysis, the fault feature of high speed wire rod processer in the Heavy Section Mill of WISCO is successfully extracted and from this effective method the same kind accident can be avoided and the fault situation can be confirmed for the future. Fourthly, research on the application of wavelet transform in signal noise reduction is conducted. The approach of noise reduction based on wavelet denoise and wavelet packet denoise is proposed. The results of experiments with simulation signals and real signals show that the approach can be applied to noise reduction effectively. Lastly, research on the application of wavelet transform in feature extraction is conducted and demonstrated its merit on fault diagnosis according to the experiments of real signals.From above all, some important feature that can not be obtained by other method like fourier analysis can be got by wavelet analysis in mechanical fault diagnosis. A feasible and effective method is provided for detecting hidden trouble of equipment and analyzing fault.更多还原