Ensembled mechanical fault recognition system based on deep learning algorithm

Abstract

Primary detection and removal of mechanical fault is vital for the recovery of mechanical and electrical equipment. The conventional mechanical fault recognition modules are not able obtain highly sensitive feature attributes for mechanical fault classification in the absence of prior knowledge. The fault diagnosis via data-driven methods have become a point of expansion with recent development in smart manufacturing and fault recognition techniques using the concept of deep learning. In this work, a combination of feature selection with Artificial Intelligence (AI) algorithm is presented for the mechanical fault recognition to deal with smart machine tools. This article proposes a CNN based fault recognition and classification framework that uses the combination of feature extraction, feature vector decomposition using Empirical Mode Decomposition (EMD) and deep neural network (DNN) for recognising the different fault states of the rotating machinery. The experimental outcomes obtained by the combination of EMD, feature selection module and Convolutional Neural Network (CNN) provides the detailed fault information by selecting the sensitive features from large number of faulty feature attributes. The proposed fault recognition and classification method performs better in terms of all the parameters yielding 99.01 % accuracy with respective cross-entropy loss of 0.325 and time complexity of 18 mins and 31 seconds. The comparative analysis is also done with other mainstream models and other state of the art methods, which reveals that the maximum improvement of 12.29 % is attained in terms of accuracy for the proposed fault recognition method. The presented method is robust in terms of reduction of network size, improvement of mechanical fault recognition, providing classification accuracy along with high fault diagnostic solution.