Open AccessFault Diagnosis of Gearbox using Machine Learning and Deep Learning Techniques
Author(s) -
Jithin Jose,
O. Deepa,
M Saimurugan,
P. Krishnakumar
Publication year - 2019
Publication title -
international journal of engineering and advanced technology
Language(s) - English
Resource type - Journals
ISSN - 2249-8958
DOI - 10.35940/ijeat.a1486.109119
Subject(s) - fault (geology) , support vector machine , convolutional neural network , artificial neural network , computer science , component (thermodynamics) , artificial intelligence , feature extraction , decision tree , vibration , deep learning , machine learning , fault detection and isolation , condition monitoring , pattern recognition (psychology) , engineering , actuator , physics , quantum mechanics , seismology , electrical engineering , thermodynamics , geology
Gearbox is an important component used for automobiles, machine tools, industries etc. Failure of any component in gearbox will cause huge maintenance cost and production loss. Failure should be detected as early as possible in order to avoid sudden breakdown which even cause catastrophic failures. Vibration signals are used for machine condition monitoring for predictive maintenance and efficiently predicts fault in the gearbox. In this paper signals from vibration is used for diagnosis of gearbox fault. The experiment uses four different conditions of gearbox in four different load conditions. Then statistical feature extraction is done and obtained result is given to Decision Tree, Support Vector Machine (SVM), Convolutional Neural Network (CNN) and Deep Neural Network (DNN) for fault diagnosis. The efficiency of these four techniques is compared and shows that machine learning is better than deep learning in gearbox fault diagnosis.