智慧型診斷與希爾伯特黃變換於變轉速旋轉機械故障特徵分析與瞬時頻率之研究;Investigation of Intelligent Diagnostics and Hilbert-Huang Transform on Characteristic Instantaneous Frequencies of Faulted Rotary System under Variable Rotating Speed

NCU Institutional Repository > 研究中心 > 數據分析方法研究中心 > 研究計畫 > Item 987654321/63192

jsp.display-item.identifier=請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/63192

题名:	智慧型診斷與希爾伯特黃變換於變轉速旋轉機械故障特徵分析與瞬時頻率之研究;Investigation of Intelligent Diagnostics and Hilbert-Huang Transform on Characteristic Instantaneous Frequencies of Faulted Rotary System under Variable Rotating Speed
作者:	吳天堯
贡献者:	國立中央大學數據分析方法研究中心
关键词:	機械工程
日期:	2012-12-01
上传时间:	2014-03-17 14:21:18 (UTC+8)
出版者:	行政院國家科學委員會
摘要:	研究期間：10108~10207;The objective of this research proposal is to extend the research of gear defects identification of the last NSC research project to other malfunction diagnosis in rotating machinery under variable rotation speed. Based on the preliminary results of the last NSC research project, the Hilbert-Huang Transform method combining the dimensionless instantaneous frequency normalization will be applied to diagnose the rotary malfunctions, such as bearing defects, shaft misalignment as well as component looseness. Through the proposed approach, the study will provide the detailed phenomenal explanation as well as the meaningful physical insight to the faulty vibration behaviors. Moreover, the collected fault features will be sifted for classification to achieve the intelligent fault detection and diagnosis of rotating machinery. Since the fault features of rotating machinery are governed by the shaft rotation speed, the vibration signals are highly non-stationary while the rotation speed is variable. Therefore, the characteristic frequencies of machine faults are not fixed. This makes the vibration analysis and fault diagnosis more difficult. To resolve this problem, the novel approach combining the dimensionless instantaneous frequency normalization and Hilbert-Huang Transform time-frequency analysis is proposed to examine the vibration signals of faulted machine in case of variable rotation speed. The non-stationary vibration signals become stationary through the dimensionless instantaneous frequency normalization. The factor of variable rotation speed will be also removed. The meaningful physical insight will be thus provided by the analysis results. A rotary test rig will be built in this research project. Different faults, such as defects in bearings, shaft misalignment and component looseness, are first artificially made in the experiment. The faults and malfunctions of rotating machinery will be simulated through experiment. With the faulted components in rotating machinery, the vibration signals are measured and analyzed in case of variable rotation speed. The fault features will be observed in Hilbert dimensionless frequency-time-energy distribution, marginal dimensionless frequency spectrum and dimensionless envelope spectrum. The degree of nonlinearity (DN) of the vibration signals will be calculated to quantify the levels of deterioration. The fault features and DN values will be collected in the fault diagnosis knowledge base of rotating machinery. The sifting process (principal component analysis, adaptive feature selection) is to select the significant fault features for classification. The classifiers (support vector machine, decision tree and artificial neural network) are utilized to identify the different fault types. Through the proposed approach, the intelligent fault detection and diagnosis of rotating machinery can be achieved, and hence it can provide the maintenance instructions for industrial applications.
關聯:	財團法人國家實驗研究院科技政策研究與資訊中心
显示于类别:	[數據分析方法研究中心 ] 研究計畫

文件中的档案:

档案	描述	大小	格式	浏览次数
index.html		0Kb	HTML	835	检视/开启

在NCUIR中所有的数据项都受到原著作权保护.

社群 sharing

数据加载中.....