Keynote Speech 01——Professor Dr.-Ing. S. X. Ding

Application of control theory-informed machine learning to process monitoring, detection and control of dynamic systems

Professor Steven X. Ding

 senior professor of the University of Duisburg-Essen, Germany

Abstract：Even in an era where data-driven and machine learning (ML) methods dominate research in technical fault diagnosis and process monitoring, model-based monitoring, detection and fault-tolerant control techniques remain capable tools for addressing engineering challenges in dynamic, particularly feedback control systems. The difficulties associated with modeling nonlinear processes, managing uncertainties, and performing performance monitoring and optimization are often cited as motivations for exploring and promoting data-driven and ML-based methods to tackle these issues. Nonetheless, ML techniques frequently fall short when applied to dynamic control systems, largely due to their specific configurations and the inherent causal relationships among process inputs, internal states, and outputs.

Physics-informed machine learning (PIML) is an emerging and innovative technique that combines domain-specific physical knowledge with the strengths of machine learning. It serves as a bridge between data-driven ML methods and fundamental physical laws. Recent studies have reported promising applications of PIML in control system modeling, analysis, and design, as well as in fault diagnosis. These works have motivated and inspired our endeavour to investigate control theory-informed machine learning (CTIML), which aims to enable reliable monitoring, fault detection, and control of dynamic systems under uncertainty. In this presentation, we outline core ideas, conceptional schemes as well as basic algorithms of CTIML.  The major focuses are on its application to handling of nonlinear dynamic processes, detection, performance monitoring and control in uncertain dynamic systems.

Biography：Steven X. Ding is currently a senior professor of the University of Duisburg-Essen, Germany. He received Ph.D. degree in electrical engineering from the Gerhard-Mercator University of Duisburg, Germany, in 1992. From 1992 to 1994, he was a R&D engineer at Rheinmetall GmbH. From 1995 to 2001, he was a professor of control engineering at the University of Applied Science Lausitz in Senftenberg, Germany, and served as vice president of this university during 1998 – 2000.  Between 2001 – March 2025, he was a chair professor of control engineering and the head of the Institute for Automatic Control and Complex Systems (AKS) at the University of Duisburg-Essen, Germany. His research interests include model-based, data-driven and machine learning aided fault diagnosis and control as well as their applications in industry with focus on chemical processes, renewable energy systems, smart buildings, automotive systems as well as secure CPSs.

Keynote Speech 02——Professor Shen Yin

Prediction of Remaining Useful Life: Insights from Traditional and Machine Learning Methods

Professor Shen Yin

the DNV Endowed Professor in the Department of Mechanical and Industrial Engineering at the Norwegian University of Science and Technology, Norway

Abstract：This presentation delves into approaches for calculating the Remaining Useful Life (RUL) of industrial machinery and systems, a crucial aspect in optimizing maintenance protocols and ensuring the reliability of equipment. The dialogue starts with conventional, model-based techniques, such as mechanistic modeling, which are typically employed for determining the life expectancy of batteries through a thorough understanding of the physical properties and failure dynamics of components. When these conventional methods are inadequate, attention turns to machine learning (ML) and other data-driven strategies that facilitate RUL predictions even without comprehensive domain-specific knowledge.

The discourse addresses three significant challenges within this domain. The initial challenge is the paucity of data, examining tactics to secure trustworthy predictions of RUL even under constraints of limited data. The subsequent challenge pertains to the robustness and sensitivity of models, concentrating on ameliorating performance and diminishing computational requirements through the adjustment of data inputs and configurations of models. The final challenge involves the amalgamation of diverse data types, with a discussion on techniques to integrate temporal and spatial data to derive critical insights that enhance prediction accuracy.

Biography：Shen Yin is the DNV Endowed Professor in the Department of Mechanical and Industrial Engineering at the Norwegian University of Science and Technology, Norway. He earned his Dr.-Ing. and MSc degrees from the University of Duisburg-Essen, Germany. His research focuses on fault diagnosis, prognosis, and fault-tolerant control within systems and control theory, as well as the prediction of remaining life concerning the reliability, safety and maintenance of technical processes.

He has been elected as an IEEE Fellow and a member of the Norwegian Academy of Technological Sciences (NTVA).

Keynote Speech 03——Professor Yaguo Lei

Research on Intelligent Operation and Maintenance for High-End Equipment and Large Models

Professor Yaguo Lei

Full Professor of the School of Mechanical Engineering at Xi'an Jiaotong University

Abstract：High-end equipment plays an important role in the fields such as aerospace, energy and power, and transportation. Faults are the potential threats to their safe and reliable operation. Intelligent operation and maintenance is a vital means to ensure the safe operation of equipment and high-quality production. The speaker will first introduce the methodologies and technologies established by his research team in the field of intelligent equipment operation and maintenance. Then, the application scenarios and typical cases of the developed intelligent diagnosis and operation and maintenance systems will be shared. Finally, the latest research work on large models for intelligent operation and maintenance will be reported.

Biography：Prof. Yaguo Lei is a Full Professor of the School of Mechanical Engineering at Xi'an Jiaotong University. He had held the research position as an Alexander von Humboldt Fellow at the University of Duisburg-Essen, Germany, and as a Postdoctoral Research Fellow at the University of Alberta, Canada. He is a Fellow of ASME, IET, and ISEAM, as well as a Senior Member of IEEE, CAA, ORSC, and CMES. He serves as a Senior Editor for Mechanical Systems and Signal Processing and an Associate Editor for IEEE Transactions on Industrial Electronics. Additionally, he is an editorial board member of over ten leading journals.

His research interests focus on big data-driven intelligent maintenance, intelligent fault diagnostics and prognostics, reliability evaluation and remaining useful life prediction. He has published four monographs and more than 100 peer-reviewed papers. His work has been cited over 32,000 times, with an H-index of 75 according to Google Scholar. His most-cited paper has received over 2,200 citations. His proposed methodologies and techniques have been widely applied in intelligent condition monitoring and diagnostic systems for renewable energy systems and other industrial domains, such as wind turbines, new-energy vehicles, and high-speed trains, etc.

Prof. Lei has received the Xplorer Prize from the New Cornerstone Science Foundation. He has been recognized as a Global Highly Cited Researcher by Clarivate Analytics and a Chinese Most Cited Researcher by Elsevier. He is also listed in the Stanford/Elsevier Global Top 2% Scientists and holds the distinction of being ranked 1st in the field of Acoustics.

Keynote Speech 04——Professor Chunhui Zhao

Large Model-Empowered Zero-Shot Fault Diagnosis for Industrial Processes: Knowledge Transfer and Semantic Reasoning from Known to Unknown Faults

Professor Chunhui Zhao

Qiushi Distinguished Professor, Recipient of the National Outstanding Youth Fund

Abstract：Fault diagnosis systems are crucial for ensuring the safe and reliable operation of industrial processes. While data-driven fault diagnosis modeling typically relies on collected historical fault data, practical industrial scenarios often face the widespread challenge of process faults lacking both samples and labels. To address this, we investigate a highly challenging fault diagnosis task: performing diagnosis without historical fault samples available for model training. Aiming at the bottleneck of traditional zero-shot diagnosis methods that depend on manual annotation of fault semantic attributes, this study introduces industrial large model technology as a breakthrough innovation, establishing an automated semantic annotation and representation framework based on existing professional documentation. Specifically, we propose a large model-driven attribute annotation method that automates fault attribute labeling, reducing semantic annotation time from days to minutes; a test-time adaptive semantic adjustment strategy is developed to optimize semantic annotation updates for unseen fault categories; by integrating cross-modal matching with knowledge distillation, we establish a bidirectional alignment method between temporal fault representations and semantic descriptions, enabling knowledge transfer and semantic reasoning from known to unknown faults. Validated in real-world applications including thermal power generation and chemical processes, this research demonstrates for the first time that large models can effectively replace manual fault semantic modeling. The proposed framework provides a new paradigm for developing autonomous cognitive industrial health management systems.

Biography：Chunhui Zhao, Qiushi Distinguished Professor, Recipient of the National Outstanding Youth Fund. She received Ph.D. degree from Northeastern University, China, in 2009. From 2009 to 2012, she was a Postdoctoral Fellow with the Hong Kong University of Science and Technology and the University of California, Santa Barbara, Los Angeles, CA, USA. From 2012 to 2014, she was a distinguished researcher with Zhejiang University and since Dec. 2014, she has been a Professor with the College of Control Science and Engineering, Zhejiang University, Hangzhou, China.

Her research interests include statistical machine learning and data mining for industrial application. She has authored or coauthored more than 250 papers in peer-reviewed international journals. She has published 3 monographs and two big data textbooks. She authorized more than 70 invention patents. She is principal investigator of a Distinguished Young Scholar Program supported by the Natural Science Foundation of China. She has hosted more than 20 scientific research projects, including the NSFC funds, National key R&D project, provincial projects and corporate cooperation projects. She has been awarded multiple research accolades, including the First Prize in Natural Science of Zhejiang Province (ranked first), the Natural Science Award from the Ministry of Education, the inaugural Youth Science and Technology Award of Zhejiang Province, and the First Prize in Natural Science from the Chinese Association of Automation. She has been honored with more than ten academic awards, including the China Young Female Scientist Award, the Model Woman Pacesetter of Zhejiang Province, the Fellow of the Chinese Association of Automation, and the inaugural CAA Young Female Scientist Award. She has served AE of three International Journals, including Journal of Process Control, Control Engineering Practice and Neurocomputing, and three domestic journals, including Control and Decision, etc.