Special Session-10

Special Session Organizers: 

Prof. Zong Meng, Yanshan University, China

Assoc. Prof. Yuejian Chen, University of Manitoba, Canada

Assoc. Prof. Xingkai Yang, Hunan University, China

Assoc. Prof. Rui Yuan, Wuhan University of Science and Technology, China

Introduction and Topics:

The rapid deployment and continuous operation of modern industrial systems, such as aircraft engines, railway transportation equipment, and large scale manufacturing machinery, enhance production efficiency, and system reliability. However, increasing system complexity under harsh and highly variable operating conditions poses significant challenges to the reliable operation and safe maintenance of these systems. Traditional prognostics and health management (PHM) methods, including purely physics based models and data driven approaches, exhibit inherent limitations when dealing with multimodal data, non stationary operating conditions, and continuously evolving degradation patterns. Data-model driven remaining useful life (RUL) prediction and health management methods, which integrate data analysis, dynamical models, physical insights, and domain knowledge, are emerging as an effective approach to enhancing their interpretability, robustness, and adaptability. This hybrid paradigm balances the interpretability of physics based models with the adaptability of data driven approaches, demonstrating significant advantages in degradation modeling and life prediction of complex engineering systems. With the continuous evolution of deep learning technologies and the emergence of new paradigms such as digital twins and edge intelligence, data-model driven RUL prediction and health management will play an increasingly critical role in enhancing the operational reliability of complex engineering systems.

The topics include, but are not limited to: 

• Data-Model Driven Hybrid Modeling for RUL Prediction and Health Management
• Data-Model Driven Physics Informed Deep Learning for RUL Prediction
• Data-Driven Physics-Inspired Fault Identification and Degradation Modeling
• Multimodal Sensor Data Fusion for Degradation Modeling in Complex Systems
• Multivariate Signal Processing Methods for Fault Diagnosis and Prognostics
• Physics Informed Deep Learning for Interpretable and Robust PHM Systems
• Domain Knowledge Guided Deep Learning for PHM Systems
• Industrial Case Studies on Data-Model Driven PHM Applications