Abstract
At present, deep neural network (DNN) technology is often used in intelligent diagnosis research. However, the huge amount of calculation of DNN makes it difficult to apply in industrial practice. In this paper, an advanced multiscale dense connection deep network MSDC‐NET is designed. A well‐designed multiscale parallel branch module is used in the network. This module can greatly improve the acceptance domain of MSDC‐NET, so as to learn useful information from input samples more effectively. Based on the inspiration of Densely Connected Convolutional Networks, MSDC‐NET designed a similar dense connection technology, so that the model will not have the problem of gradient vanishing because of the deep network. The experimental data of MSDC‐NET on MFPT, SEU, and Pu datasets show that our method has higher performance than other latest technologies. At the same time, we carried out knowledge distillation based on the high‐precision classification level of MSDC‐NET, which makes the diagnosis ability and robustness of the lightweight CNN model improve significantly.
Highlights
The mechanical equipment in modern industry often needs to work in the complex environment of high temperature, fatigue, and heavy load for a long time, which may cause incalculable production accidents and economic property losses
Zhang et al used support vector machine (SVM) to identify the working state of bearing [1]; Banerjee and Das used SVM fusion multisensor signal to detect motor fault [2]; Bugharbee and Trendafilova used the nearest neighbor classifier to identify faults [3]; Keskes et al studied the method of feature extraction using stationary wavelet packet transform, and support vector machine (SVM) is applied to rotor fault classification. [4]; Mustafa et al proposed that the spectrum analysis method be used in feature extraction and SVM be used to detect mechanical faults
We distill the knowledge of MSDC-NET according to the algorithm in Figure 6, so that the classification Acc of the Convolutional Neural Net (CNN) model is significantly improved without any change in the internal structure
Summary
The mechanical equipment in modern industry often needs to work in the complex environment of high temperature, fatigue, and heavy load for a long time, which may cause incalculable production accidents and economic property losses. As a mature unsupervised system, intelligent diagnosis can solve this problem for related enterprises. The traditional intelligent diagnosis method mainly uses a machine learning algorithm to classify the feature data from the sensor signal. Zhang et al used support vector machine (SVM) to identify the working state of bearing [1]; Banerjee and Das used SVM fusion multisensor signal to detect motor fault [2]; Bugharbee and Trendafilova used the nearest neighbor classifier to identify faults [3]; Keskes et al studied the method of feature extraction using stationary wavelet packet transform, and support vector machine (SVM) is applied to rotor fault classification. The signal obtained by sensor needs complex feature extraction, which is difficult to achieve instantaneity, and depends on professional knowledge and relevant experience
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
