He Yu, Zai-ke Tian, Hong-ru Li, Baohua Xu, An Guoqing
{"title":"A Novel Deep Belief Network Model Constructed by Improved Conditional RBMs and its Application in RUL Prediction for Hydraulic Pumps","authors":"He Yu, Zai-ke Tian, Hong-ru Li, Baohua Xu, An Guoqing","doi":"10.20855/ijav.2020.25.31669","DOIUrl":null,"url":null,"abstract":"Residual Useful Life (RUL) prediction is a key step of Condition-Based Maintenance (CBM). Deep learning-based techniques have shown wonderful prospects on RUL prediction, although their performances depends on heavy structures and parameter tuning strategies of these deep-learning models. In this paper, we propose a novel Deep Belief Network (DBN) model constructed by improved conditional Restrict Boltzmann Machines (RBMs) and apply it in RUL prediction for hydraulic pumps. DBN is a deep probabilistic digraph neural network that consists of multiple layers of RBMs. Since RBM is an undirected graph model and there is no communication among the nodes of the same layer, the deep feature extraction capability of the original DBN model can hardly ensure the accuracy of modeling continuous data. To address this issue, the DBN model is improved by replacing RBM with the Improved Conditional RBM (ICRBM) that adds timing linkage factors and constraint variables among the nodes of the same layers on the basis of RBM. The proposed model is applied to RUL prediction of hydraulic pumps, and the results show that the prediction model proposed in this paper has higher prediction accuracy compared with traditional DBNs, BP networks, support vector machines and modified DBNs such as DEBN and GC-DBN.","PeriodicalId":49185,"journal":{"name":"International Journal of Acoustics and Vibration","volume":"25 1","pages":"373-382"},"PeriodicalIF":0.8000,"publicationDate":"2020-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"5","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Acoustics and Vibration","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.20855/ijav.2020.25.31669","RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"ACOUSTICS","Score":null,"Total":0}
引用次数: 5
Abstract
Residual Useful Life (RUL) prediction is a key step of Condition-Based Maintenance (CBM). Deep learning-based techniques have shown wonderful prospects on RUL prediction, although their performances depends on heavy structures and parameter tuning strategies of these deep-learning models. In this paper, we propose a novel Deep Belief Network (DBN) model constructed by improved conditional Restrict Boltzmann Machines (RBMs) and apply it in RUL prediction for hydraulic pumps. DBN is a deep probabilistic digraph neural network that consists of multiple layers of RBMs. Since RBM is an undirected graph model and there is no communication among the nodes of the same layer, the deep feature extraction capability of the original DBN model can hardly ensure the accuracy of modeling continuous data. To address this issue, the DBN model is improved by replacing RBM with the Improved Conditional RBM (ICRBM) that adds timing linkage factors and constraint variables among the nodes of the same layers on the basis of RBM. The proposed model is applied to RUL prediction of hydraulic pumps, and the results show that the prediction model proposed in this paper has higher prediction accuracy compared with traditional DBNs, BP networks, support vector machines and modified DBNs such as DEBN and GC-DBN.
期刊介绍:
The International Journal of Acoustics and Vibration (IJAV) is the refereed open-access journal of the International Institute of Acoustics and Vibration (IIAV). The IIAV is a non-profit international scientific society founded in 1995. The primary objective of the Institute is to advance the science of acoustics and vibration by creating an international organization that is responsive to the needs of scientists and engineers concerned with acoustics and vibration problems all around the world.
Manuscripts of articles, technical notes and letters-to-the-editor should be submitted to the Editor-in-Chief via the on-line submission system. Authors wishing to submit an article need to log in on the IJAV website first. Users logged into the website are able to submit new articles, track the status of their articles already submitted, upload revised articles, responses and/or rebuttals to reviewers, figures, biographies, photographs, copyright transfer agreements, and send comments to the editor. Each time the status of an article submitted changes, the author will also be notified automatically by email.
IIAV members (in good standing for at least six months) can publish in IJAV free of charge and their papers will be displayed on-line immediately after they have been edited and laid-out.
Non-IIAV members will be required to pay a mandatory Article Processing Charge (APC) of $200 USD if the manuscript is accepted for publication after review. The APC fee allows IIAV to make your research freely available to all readers using the Open Access model.
In addition, Non-IIAV members who pay an extra voluntary publication fee (EVPF) of $500 USD will be granted expedited publication in the IJAV Journal and their papers can be displayed on the Internet after acceptance. If the $200 USD (APC) publication fee is not honored, papers will not be published. Authors who do not pay the voluntary fixed fee of $500 USD will have their papers published but there may be a considerable delay.
The English text of the papers must be of high quality. If the text submitted is of low quality the manuscript will be more than likely rejected. For authors whose first language is not English, we recommend having their manuscripts reviewed and edited prior to submission by a native English speaker with scientific expertise. There are many commercial editing services which can provide this service at a cost to the authors.