Abstract
Typically, fault detection using deep learning is performed based on the features extracted from only one well-trained deep model. However, our results show that large-scale data is complicated and originates from different schemas, which will cause great pressure on deep neural networks, furthermore, the quality of the extracted features will be affected, and the training complexity and time will also be increased. Conversely, deep models would feel comfortable to extract features from raw data that contain less complex relationships and the quality of extracted features are higher and more representative. Hence, variables from large-scale industrial processes in this study are reasonably divided into various schemas with simple relationships by mutual information. Then, the corresponding deep belief network (DBN) models are established under a lighter pressure state to sufficiently extract the abstract and high-order information from data in each schema. Experimental analysis shows that the training efficiency, the accuracy of extracted features and the monitoring performance based on the proposed model system are all better than using only one DBN. What’s more, a comparison with those of representative and state-of-the-art methods on numerical and Tennessee Eastman processes also demonstrates the high performance of the proposal called M-DBN.
Highlights
Industrial production processes have become complicated with the increase in the quality requirements of products, and many sensors and data gathering equipment have been introduced
On the other hand, Clustering the original variables and extracting the local feature information and monitoring the blocks have elicited much attention [18]–[23]. These proposed methods are based on the local information in the data and using the blocking technique and multivariable statistical process monitoring (MSPM) method for feature extraction. which results in that the extracted features are shallow and not optimal. Motivated by this situation and the problems of time-consuming and accuracy in deep model training, this study partition the process variables into multiple schemas based on the mutual information (MI) between variables to simplify the complexity of the relationship between variables
In this study, a novel process monitoring method called M-deep belief network (DBN) was proposed based on MI, DBN, kernel density estimation (KDE), and support vector data description (SVDD)
Summary
Industrial production processes have become complicated with the increase in the quality requirements of products, and many sensors and data gathering equipment have been introduced. Which results in that the extracted features are shallow and not optimal Motivated by this situation and the problems of time-consuming and accuracy in deep model training, this study partition the process variables into multiple schemas based on the mutual information (MI) between variables to simplify the complexity of the relationship between variables. The primary advantages of the proposed novel method that utilizes multiple DBNs (M-DBN) to extract the features of each schema of the measurement variables can be summarized as follows: 1) The MI technique is a measurement index based on information entropy and considers various relationships among the variables to partition all variables into various schemas, making the partition result highly objective. A classifier, such as soft max, is added to the end of DBN to train the labeled data in a supervised manner
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.
