A Fault Diagnosis Method of Plug Doors Based on Sound Recognition

カテゴリ: 研究会(論文単位)

論文No: TER18008

グループ名: 【D】産業応用部門 交通・電気鉄道研究会

発行日: 2018/02/01

タイトル(英語): A Fault Diagnosis Method of Plug Doors Based on Sound Recognition

著者名: SUN Yongkui(Beijing Jiaotong University),CAO Yuan(Beijing Jiaotong University),ZHANG Yuzhuo(Beijing Jiaotong University)

著者名(英語): Yongkui SUN(Beijing Jiaotong University),Yuan CAO(Beijing Jiaotong University),Yuzhuo ZHANG(Beijing Jiaotong University)

キーワード: train plug doors|fault diagnosis|sound recognition|EMD|FCM|train plug doors|fault diagnosis|sound recognition|EMD|FCM

要約(日本語): train plug doors are the only channels for passengers getting on and off, of which the reliability has a direct impact on passengers’ safety and operational efficiency. Aimed at the defects of the poor real-time of current fault diagnosis method of plug doors, a fault diagnosis method based on sound recognition is proposed. To process the non-stationarity sound signals, the Empirical Mode Decomposition (EMD) method is applied on switching sound signals of train plug doors. Then, a series of Intrinsic Mode Functions (IMFs) are obtained. According to fast Fourier transform analysis, it is found that the spectral characteristics of the first frame IMF are different among different fault categories which can be selected as the fault identification characteristics. Dividing the spectral domain by interval of 500 points, electing the maximum value in the interval as the envelope value. Then, the envelopes of the spectral domain are obtained as the features for fault classification and discrimination. The clustering accuracy of 18 sets of envelope features using fuzzy C - means (FCM) clustering method reaches to 94.44%, indicating that the proposed method for train plug doors fault diagnosis is feasible.

要約(英語): train plug doors are the only channels for passengers getting on and off, of which the reliability has a direct impact on passengers’ safety and operational efficiency. Aimed at the defects of the poor real-time of current fault diagnosis method of plug doors, a fault diagnosis method based on sound recognition is proposed. To process the non-stationarity sound signals, the Empirical Mode Decomposition (EMD) method is applied on switching sound signals of train plug doors. Then, a series of Intrinsic Mode Functions (IMFs) are obtained. According to fast Fourier transform analysis, it is found that the spectral characteristics of the first frame IMF are different among different fault categories which can be selected as the fault identification characteristics. Dividing the spectral domain by interval of 500 points, electing the maximum value in the interval as the envelope value. Then, the envelopes of the spectral domain are obtained as the features for fault classification and discrimination. The clustering accuracy of 18 sets of envelope features using fuzzy C - means (FCM) clustering method reaches to 94.44%, indicating that the proposed method for train plug doors fault diagnosis is feasible.

原稿種別: 英語

PDFファイルサイズ: 1,021 Kバイト