Xuan Hu , Peihao Zheng , Hao Wu , Yongming Han , Zhiqiang Geng
{"title":"Variational masking progressive learning method for multi-rate industrial processes soft sensor","authors":"Xuan Hu , Peihao Zheng , Hao Wu , Yongming Han , Zhiqiang Geng","doi":"10.1016/j.jprocont.2025.103488","DOIUrl":null,"url":null,"abstract":"<div><div>Deep learning has been widely used in industrial processes to predict critical quality indicators. However, existing methods assume that industrial process data are uniformly sampled, which is far from real industrial scenarios. To solve the problem of multi-rate sampling in industrial processes, a variational masking progressive learning (VMPL) method is proposed for multi-rate industrial processes soft sensor. In the VMPL, a multi-rate decomposition strategy (MDS) is first developed to construct generalized multi-rate data and corresponding masking matrix. Then, based on the MDS, a variational masking network (VMN) is designed to represent the uncertain distribution information of industrial process data. Meanwhile, the progressive learning (PL) algorithm is derived to assist the VMN in transferring process features from high-rate to low-rate data. Therefore, the VMPL can progressively mine features in different rates data without changing the structure of the VMN to improve soft-sensing accuracy. Finally, compared with the state-of-the-art multi-rate soft sensor model on the three key quality variable datasets of the catalytic cracking process, the VMPL achieves more accurate soft sensing results.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103488"},"PeriodicalIF":3.3000,"publicationDate":"2025-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Process Control","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0959152425001167","RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"AUTOMATION & CONTROL SYSTEMS","Score":null,"Total":0}
引用次数: 0
Abstract
Deep learning has been widely used in industrial processes to predict critical quality indicators. However, existing methods assume that industrial process data are uniformly sampled, which is far from real industrial scenarios. To solve the problem of multi-rate sampling in industrial processes, a variational masking progressive learning (VMPL) method is proposed for multi-rate industrial processes soft sensor. In the VMPL, a multi-rate decomposition strategy (MDS) is first developed to construct generalized multi-rate data and corresponding masking matrix. Then, based on the MDS, a variational masking network (VMN) is designed to represent the uncertain distribution information of industrial process data. Meanwhile, the progressive learning (PL) algorithm is derived to assist the VMN in transferring process features from high-rate to low-rate data. Therefore, the VMPL can progressively mine features in different rates data without changing the structure of the VMN to improve soft-sensing accuracy. Finally, compared with the state-of-the-art multi-rate soft sensor model on the three key quality variable datasets of the catalytic cracking process, the VMPL achieves more accurate soft sensing results.
期刊介绍:
This international journal covers the application of control theory, operations research, computer science and engineering principles to the solution of process control problems. In addition to the traditional chemical processing and manufacturing applications, the scope of process control problems involves a wide range of applications that includes energy processes, nano-technology, systems biology, bio-medical engineering, pharmaceutical processing technology, energy storage and conversion, smart grid, and data analytics among others.
Papers on the theory in these areas will also be accepted provided the theoretical contribution is aimed at the application and the development of process control techniques.
Topics covered include:
• Control applications• Process monitoring• Plant-wide control• Process control systems• Control techniques and algorithms• Process modelling and simulation• Design methods
Advanced design methods exclude well established and widely studied traditional design techniques such as PID tuning and its many variants. Applications in fields such as control of automotive engines, machinery and robotics are not deemed suitable unless a clear motivation for the relevance to process control is provided.