旋转机械振动信号处理算法研究与程序设计

发布时间：2018-03-03 00:35

本文选题：旋转机械　切入点：振动信号　出处：《西安工业大学》2016年硕士论文　论文类型：学位论文

【摘要】：旋转机械是电力、石油、化工等部门关键设备,长期处于高速运行状态,易出现机械故障引发安全事故。对其振动信号进行滤波、谱估计及特征提取等处理,在此基础上建立故障诊断系统是保障可靠运行的有效手段。论文对旋转机械在线监测系统中振动信号处理算法进行研究,并完成处理软件设计。论文首先对振动信号处理需求进行分析,确定程序的设计方案,包括通信程序、去噪程序、谱估计程序及界面程序。其次,针对滤波要求,提出离散余弦变换域循环维纳滤波算法,对算法原理及实现过程进行了介绍,并对算法性能进行了定性与定量分析。在去噪功能实现的基础上,采用最小方差谱估计算法进行振动信号谱估计,在谱估计的分辨力性能与方差性能之间取得了较好的平衡。最后,在Microsoft Visual Studio 2010环境中基于C#与MATLAB混合编程设计信号分析与处理程序,建立与在线监测设备之间的数据通信,实现对振动信号的实时与离线分析、处理,具有良好的人机交互功能。所设计的程序经功能测试,通信功能、数据存储功能、参数传递及波形显示功能均可正常运行。论文研究与测试结果表明,采用离散余弦变换域循环维纳滤波算法对振动信号进行滤波能保证线性相位,且有良好滤波效果；采用最小方差谱估计对信号进行谱分析既有较好分辨力性能又能保证良好的方差性能；采用C#与MATLAB混合编程开发振动信号分析与处理程序能够实现较好的人机交互功能,提高振动信号处理的灵活性。
[Abstract]:Rotating machinery is the electric power, petroleum, chemical industry and other departments of key equipment, in high speed running state for a long time, prone to mechanical failure caused by accidents. To filter the vibration signal, spectrum estimation and feature extraction processing, established on the basis of fault diagnosis system is the effective means to ensure the reliable operation. This paper studies on the vibration signal processing the algorithm of on-line monitoring system of rotating machinery, and complete the software design. The paper first carries on the analysis to the vibration signal processing requirements, determine the design procedures, including communication procedures, denoising procedure, spectrum estimation program and interface program. Secondly, according to the filtering requirements, put forward the discrete cosine transform domain cyclic Wiener filtering algorithm and Realization of the the principle of the algorithm process were introduced, and the performance of the algorithm is analyzed qualitatively and quantitatively. Based on denoising function, using minimum variance spectrum Estimation algorithm for vibration signal spectrum estimation, achieved a good balance between the performance of spectral estimation resolution and variance performance. Finally, C# and MATLAB mixed programming signal analysis and processing program based on Microsoft in Visual Studio 2010 environment, establishment and data communication between online monitoring equipment, real-time and off-line analysis of the vibration signal processing, and has good human-computer interaction. The program is designed by function test, communication function, data storage, waveform display and parameter transfer functions can be normal operation. The research and test results show that using the discrete cosine transform domain cyclic Wiener filter algorithm can guarantee the linear phase of the vibration signal. And have a good filtering effect; the minimum variance spectrum estimation spectrum analysis has good resolution performance and can guarantee the good variance of signal It can use C# and MATLAB mixed programming to develop vibration signal analysis and processing program, which can achieve better man-machine interaction function and improve the flexibility of vibration signal processing.

【学位授予单位】：西安工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH17

【相似文献】