The rolling bearing is one of the most important parts of any rotating machinery. A bearing operating condition determines the safety during working condition. So it is very important to do condition monitoring in real time and diagnosis of rolling bearing for preventing failure. The main challenges in traditional method of fault diagnosis are finding good fault features. To address this problem, author have used data augmentation technique for bearing fault diagnosis. This proposed method converts 1-D time domain vibration signals into 2-D time frequency image by using short-term fourier transformation (STFT). However, a major problem in deep convolution neural network (CNN) is that in CNN large number of samples data is required to obtain well trained model. Proposed method introduces data augmentation for generating 2-D time frequency image for preparing large number of data set, and Deep Neural Network is used for rolling bearing fault classification. Major purpose of this study is to identify and successfully classify the different severity level faults of rolling bearing. Author apply this approach on rolling bearing dataset of Case Western Reserve University to verify the effectiveness of the proposed method. Results show that proposed methodology give higher accuracy and is fast to train the model. Also this method achieve better accuracy with different fault conditions as compare to other traditional method of fault classification.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Identification of Fault Severity of Rolling Element Bearing Using Image Augmentation and Mobile Net V_2 Convolutional Neural Network


    Weitere Titelangaben:

    Lect.Notes Mechanical Engineering


    Beteiligte:
    Narasimham, G. S. V. L. (Herausgeber:in) / Babu, A. Veeresh (Herausgeber:in) / Reddy, S. Sreenatha (Herausgeber:in) / Dhanasekaran, Rajagopal (Herausgeber:in) / Akhenia, P. (Autor:in) / Jamani, H. (Autor:in) / Vakharia, V. (Autor:in)


    Erscheinungsdatum :

    2022-03-02


    Format / Umfang :

    11 pages





    Medientyp :

    Aufsatz/Kapitel (Buch)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch