Special Session #03 Embodied Intelligent Artificial Joints: Sensing, Measurement, and Control
Special Session #06 Interpretable diagnosis and prediction technology for rotating machinery faults
Special Session #07 PHM for Transportation Infrastructure and Equipment Assets
Special Session #08 Advanced monitoring and intelligent maintenance of power equipment
Special Session #11 Advanced Sensing, Monitoring and Diagnosis of Renewable Energy batteries
Physics-informed and Data-driven Health Assessment for Complex Electromechanical Systems
Session Organizers:
Zhibin Zhao, Xi’an Jiaotong University
Email: zhaozhibin@xjtu.edu.cn
Tianfu Li, Kunming University of Science and Technology
Email: tianfu.li@kust.edu.cn
Fujin Wang, Kunming University of Science and Technology
Email: wangfujin@kust.edu.cn
Download: Special Session #1.pdf
With the rapid development of intelligent sensing and AI analytics, health assessment of complex electromechanical systems (e.g., high-end machinery, robotics, batteries and energy systems) is increasingly required to work under time-varying conditions, scarce fault labels, strong noise. Purely data-driven approaches typically rely on large-scale, high-quality training data and heavy computational resources, yet they may still lack interpretability and physical consistency when deployed in real-world systems. Therefore, it is urgent to develop physics-informed and hybrid methods that fuse physical knowledge (mechanisms, constraints, conservation laws) with data-driven learning to achieve reliable, interpretable and physics-consistent diagnosis/prognostics for real engineering applications. This special session aims to bring together researchers and practitioners working on sensing, measurement, and AI-enabled analytics for mechanical equipment, batteries, and energy systems, emphasizing physics-data fusion for robust state estimation, fault diagnosis, and prognostics.
Suitable topics for this special session include but are not limited to:
• Physics-informed machine learning for diagnostics/prognostics
• Hybrid digital twins: mechanism models + data-driven residual/compensation learning
• Degradation mechanism modeling and parameter identification with multi-source measurements
• Sensor fusion and multi-modal sensing for complex equipment
• Remaining useful life prediction with physics constraints and interpretable models
• Battery state estimation, fault diagnosis, thermal runaway early warning using physics-data coupling
• Health assessment for energy systems: wind turbines, power electronics, grid-connected converters
• Data quality, missing data, drift detection, and robust measurement-informed learning
• Standardization, benchmarking datasets, and evaluation metrics for physics-informed PHM
Multivariate signal analysis and information fusion for fault diagnosis and prognostics
Session Organizers:
Prof. Zong Meng, Yanshan University
Email: mzysu@ysu.edu.cn
Assistant Prof. Yuejian Chen, University of Manitoba
Email: yuejian.chen@umanitoba.ca
Assistant Prof. Xingkai Yang, Hunan University
Email: xkyang1992@163.com
Associate Prof. Rui Yuan, Wuhan University of Science and Technology
Email: yuanrui@wust.edu.cn
Download: Special Session #2.pdf
The advancement of multivariate signal analysis and information fusion technologies in modern industrial systems has significantly strengthened the perception and dynamic characterization of equipment operating conditions. Multivariate signal analysis facilitates the extraction of critical information on system state evolution, structural coupling characteristics, and performance degradation processes. Information fusion enables the effective integration of complementary information from diverse sensing sources and data modalities, thereby enhancing the completeness of health-state representation, the reliability of diagnostic outcomes, and the stability of prognostic analysis. In this context, multivariate signal analysis and AI-enabled information fusion serve not only as key technological pathways for improving the performance of fault diagnosis and prognostics, but also as essential enablers for the transformation of equipment health management from experience-based and corrective maintenance toward intelligent perception, proactive prediction, and lifecycle-oriented management. This session highlights recent advances in fault diagnosis, prognostics and health management, and structural health monitoring for industrial systems.
The topics of interest include, but are not limited to:
• Industrial Case Studies on PHM Applications with Multimodal Sensor Data
Embodied Intelligent Artificial Joints: Sensing, Measurement, and Control
Session Organizers:
Prof. Huiming Cheng, Beijing University of Technology
Email: chenghuiming295@126.com
Prof. Bo Yu, Beijing University of Technology
Email: yubo@bjut.edu.cn
Prof. Naipeng Li, Xi’an Jiaotong University
Email: naipengli@mail.xjtu.edu.cn
Associate Prof. Laihao Yang, Xi’an Jiaotong University
Email: yanglaihao@xjtu.edu.cn
Associate Prof. Jun Dai, Beijing Institute of Technology
Email: daijun@bit.edu.cn
Download: Special Session #3.pdf
The advancement of embodied intelligence has placed increasingly stringent demands on the sensing, measurement, and control of artificial joint systems. As fundamental units for motion generation and environmental interaction, artificial joints are required to achieve high-precision perception, adaptive actuation, and robust control under complex conditions. Advanced sensing and measurement technologies enable the acquisition of multi-source information, including force, motion, and interaction characteristics, providing critical insights into system dynamics and coupling behaviors. Intelligent control strategies enable real-time regulation of joint dynamics, enhancing motion accuracy, energy efficiency, and operational reliability. The integration of multi-physics sensing and control further enhances system responsiveness and adaptability in uncertain environments. In this context, the deep integration of sensing, measurement, and control in artificial joints not only promotes a more comprehensive understanding of electromechanical coupling mechanisms, but also drives the evolution of joint systems toward higher levels of autonomy, adaptability, and intelligence. This session highlights recent advances in multi-source sensing, precision measurement, intelligent control and system integration for embodied intelligent artificial joints.
The topics of interest include, but are not limited to:
• Humanoid Robot Joint and System Integration
Domain Adaptation Theory and Methods: Recent Advances in Industrial Equipment Fault Diagnosis and Prognostics
Session Organizers:
Chenyang Ma, Associate Professor, School of Computer Science, Xi'an University of Posts and Telecommnications
Email: machenyang@xupt.edu.cn
Shengkang Yang, Lecturer, School of Artificial Intelligence, School of Automation, Xi'an University of Posts and Telecommnications
Email: skyang@xupt.edu.cn
Bo Zhao, Post-Doctor, Department of Data Science, City University of Hong Kong
Email: bzhaopap@163.com
Zhilin Dong, Lecturer, College of Engineering, Zhejiang Normal University
Email: d18133679022@zjnu.edu.cn
Guowei Zhang, Post-Doctor, Tsinghua University
Email: zhanggw13@163.com
Download: Special Session #4.pdf
In the actual operation of modern industrial equipment, the complex and variable operating environments and task demands often lead to significant distribution shifts and label scarcity in equipment service performance data. This issue directly restricts the generalization capability of traditional data-driven operation and maintenance models, making it difficult for them to adapt to diverse scenarios in engineering practice and severely reducing their engineering application value. Domain adaptation methods, as one of the effective techniques to address the above issues, can achieve effective alignment of data distributions between source and target domains by deeply mining the shared deep invariant features across domain data, thereby significantly mitigating model performance degradation caused by factors such as operating condition transfer, equipment individual differences, and environmental disturbances. Nowadays, with the rapid development of artificial intelligence technologies, domain adaptation strategies integrating advanced techniques such as adversarial learning, causal inference, and meta-learning have emerged continuously, further enhancing the adaptive diagnosis capability and decision-making reliability of models in unknown and complex scenarios. In this context, domain adaptation methods have not only become a core support for breaking through the technical bottlenecks of cross-domain intelligent operation and maintenance but also serve as a key means to promote the transformation and upgrading of industrial intelligent operation and maintenance paradigms—driving intelligent maintenance from a mode confined to single operating conditions toward cross-scenario generalization applications, and from traditional supervised learning frameworks toward weakly supervised and self-supervised collaborative evolution. Based on this, this session focuses on the latest research achievements and advances in domain adaptation methods for fault diagnosis, prognostics, and health management in industrial systems, providing a reference for related research and engineering applications in relevant fields.
The topics of interest include, but are not limited to:
• Source-free domain adaptation methods for fault diagnosis and prognostics
• Domain generalization methods and their applications in fault diagnosis
• Causal representation domain adaptation for fault diagnosis
• Domain adaptation on Few-Shot / Zero-Shot fault diagnosis scenario based on meta-learning
• Incremental domain adaptation and domain generalization methods for incremental fault type and domain scenario
• Partial, Open-set, and Universal Domain Adaptation
• Multi-source Domain Adaptation
• Federated Domain Adaptation & Generalization
• Generative and Adversarial Transfer Learning
Nonstationary Signal Analysis and Intelligent Diagnosis of Electromechanical Systems under Complex Operating Conditions
Session Organizers:
Prof. Hao Zhang, Hebei University of Technology.
Email: zhanghao@hebut.edu.cn
Associate Prof. Jinzhen Kong, Hebei University of Technology.
Email: jinzhenkong@hebut.edu.cn
Associate Prof. Bingchang Hou, Chongqing University.
Email: bingchanghou@cqu.edu.cn
Assistant Prof. Bingyan Chen, Southwest Jiaotong University.
Email: bingyanchen@swjtu.edu.cn
Dr. Yang Guan, Hebei University of Technology.
Email: 2025907@hebut.edu.cn
Download: Special Session #5.pdf
With the rapid development of advanced equipment, robotics, wind turbine systems, rail transportation, and intelligent manufacturing systems, electromechanical transmission systems often operate under complex conditions such as variable speed, variable load, impact disturbances, and strong background noise. The condition monitoring signals of these systems typically exhibit nonstationary, nonlinear, non-Gaussian, and multi-source coupling characteristics. In particular, at the incipient fault stage, fault-induced impulses are weak, while characteristic components are easily modulated, submerged, or distorted by speed fluctuations, load variations, and background noise. These factors make fault feature representation and robust cross-condition diagnosis challenging.
This special session focuses on nonstationary signal analysis, weak fault feature extraction, and intelligent diagnosis methods for electromechanical transmission systems under complex operating conditions. Particular attention is given to condition monitoring and fault identification of key components such as gears, bearings, motors, reducers, and transmission chains. The session aims to provide a forum for discussions on time-frequency analysis, adaptive signal decomposition, resonance demodulation, blind deconvolution, time-varying feature representation, cross-condition transfer diagnosis, and physics-data fusion, with the goal of advancing health monitoring, fault diagnosis, and intelligent maintenance technologies for complex electromechanical transmission systems.
The topics of interest include, but are not limited to:
• Nonstationary Signal Analysis for Electromechanical Transmission Systems
• Weak Fault Feature Extraction and Enhancement
• Time-Varying Feature Representation under Complex Operating Conditions
• Cross-Condition Fault Diagnosis and Domain Adaptation
• Physics-Informed Fault Diagnosis and Condition Assessment of Electromechanical Transmission Systems
• Robust Diagnosis, Uncertainty Evaluation, and Interpretable Intelligent Maintenance
Interpretable diagnosis and prediction technology for rotating machinery faults
Session Organizers:
Xiaoan Yan, Nanjing Forestry University
Email: yanxiaoan@njfu.edu.cn
Yasong Li, Nanjing Forestry University
Email: liyasong@njfu.edu.cn
Yi Wang, Chongqing University
Email: wyyc@cqu.edu.cn
Chuancang Ding, Soochow University
Email: ccding@suda.edu.cn
Changfeng Yan, Lanzhou University of Technology
Email: changf_yan@163.com
Download: Special Session #6.pdf
Rotating machinery is a core equipment in fields such as energy, manufacturing, and rail transportation, and its intelligent operation and maintenance urgently need to move from black-box prediction to trusted decision-making. This topic focuses on the forefront of interpretable diagnosis and prediction technology for rotating machinery faults. In response to the black-box problem of deep learning, it systematically explores the theory and methods of explainable artificial intelligence (XAI) in fault feature extraction, pattern recognition, and remaining life prediction. The content covers interpretative models based on attention, gradient attribution, causal reasoning, and symbolic rule learning; Mechanism consistency modeling of the fusion of physical priors and deep learning; Explainable transfer learning under weak signs, compound faults, and variable operating conditions; And the visualization of diagnostic results, uncertainty assessment and verification system. The theme aims to promote the transition of fault diagnosis from high-accuracy to high-interpretability and high-reliability, reveal the correlation between data characteristics and fault mechanisms, provide reliable, verifiable, and scalable theoretical methods and technical paths for intelligent health management of major equipment, and serve the intelligent development of high-end equipment.
The topics of interest include, but are not limited to:
• Multi-source heterogeneous information fusion for fault diagnosis and prediction
• Fault diagnosis and prediction empowered by large models
• Reliable fault diagnosis and prediction based on signal model
• Signal processing driven machine learning for fault diagnosis and prediction
• Physical priors guided machine learning for fault diagnosis and prediction
• Machine learning with physical feature embedding for fault diagnosis and prediction
• Mechanism and data fusion driven for fault diagnosis and prediction
• Physics inspired machine learning for fault diagnosis and prediction
• Digital twin technology for fault diagnosis and prediction
• Engineering applications and case sharing in the field of fault diagnosis and prediction
PHM for Transportation Infrastructure and Equipment Assets
Session Organizers:
Hongrui Wang, Professor, School of Electrical Engineering, Southwest Jiaotong University
Email: hongruiwang1990@163.com
Dandan Peng, Professor, School of Mechanical Engineering, Northwestern Polytechnical University
Email: dandan.peng@nwpu.edu.cn
Te Han, Associate Professor, School of Management, Beijing Institute of Technology
Email: hante@bit.edu.cn
Xiaoxi Hu, Post-Doctor, State Key Laboratory of Intelligent Green Vehicle and Mobility, Tsinghua University
Email: xiaoxihurail@gmail.com
Download: Special Session #7.pdf
With the rapid development of intelligent sensing, digital twins, artificial intelligence, and asset management technologies, the operation and maintenance of transportation infrastructure and equipment assets is shifting toward condition-based, predictive, and intelligent decision-making. Rail, aviation, road/automotive, and maritime transportation systems involve numerous safety-critical and high-value assets operating under complex environments, time-varying loads, aging mechanisms, and strict safety requirements. Traditional maintenance strategies may cause delayed fault detection, high costs, and insufficient asset utilization, while purely data-driven methods often suffer from scarce failure samples, heterogeneous data, domain shifts, missing data, and limited interpretability. This special session aims to bring together researchers and practitioners working on intelligent sensing, health monitoring, fault diagnosis, prognostics, digital twins, maintenance decision-making, and lifecycle asset management for transportation infrastructure and equipment assets.
Suitable topics for this special session include but are not limited to:
• Intelligent PHM for transportation infrastructure and equipment assets
• Condition monitoring, fault diagnosis, and prognostics for diverse transportation systems
• Predictive maintenance and lifecycle asset management for safety-critical transportation systems
• Multi-source sensing, inspection, and measurement technologies for transportation asset health assessment
• Physics-informed, knowledge-guided, and data-driven methods for transportation prognostics and health management
• Digital twins for transportation infrastructure, vehicles, aircraft, ships, ports, and related equipment
• Degradation modeling, reliability analysis, and remaining useful life prediction of transportation assets
• Risk-based inspection, maintenance scheduling, and decision optimization under uncertainty
• Data quality, missing data, sensor faults, domain shift, and robust learning in transportation monitoring
• Health assessment and maintenance of railway tracks, turnouts, bridges, tunnels, rolling stock, and signaling equipment
• Aircraft structural health monitoring, engine health management, and airport infrastructure maintenance
• Intelligent vehicle health monitoring, road infrastructure assessment, and connected vehicle maintenance
• Ship, port, offshore equipment, and maritime infrastructure health monitoring and maintenance
• Asset-level and system-level maintenance decision-making for integrated transportation systems
• Standardization, benchmarking datasets, evaluation metrics, and practical deployment of intelligent transportation O&M systems
Advanced monitoring and intelligent maintenance of power equipment
Session Organizers:
Yu-Ling He, Professor, Department of Mechanical Engineering, North China Electric Power University.
Email: heyuling1@ncepu.edu.cn
Nai-Chao Chen, Professor, College of Energy and Mechanical Engineering, Shanghai University of Electric Power
Email: yeiji_chen@126.com
Zhi-Ying Zhu, Professor, School of Electric Power Engineering, School of Shenguorong, Nanjing Institute of Technology
Email: zyzhu@njit.edu.cn
Download: Special Session #8.pdf
Power equipment is the key component of electric power system. The condition monitoring and maintenance of the power equipment is pretty significant for the steady running of the power system. This session focuces on the advanced condition monitoring and intelligent maintenance technologies for power equipment, including the latest signal analysis and processing, the advanced structure design and improvement, the intelligent monitoring system development, etc, for wind turbines, generators, motors, transformers, breakers, etc.
The scope includes (but not limited to) the following ones.
• Advanced signal analysis and processing for power equipment.
• Strength analysis and structural improvement of key components in power equipment
• Intelligent fault diagnosis and maintenance for power equipment.
• Latest design and predictive manufacturing technology for power equipment.
• Vibration test and control theories and technologies.
• Sensor technologies.
Advanced Vibration Sensing and Intelligent Diagnosis Methods for Equipment Health Monitoring
Session Organizers:
Haidong Shao, Associate Professor, School of Mechanical and Vehicle Engineering, Hunan University
Email: hdshao@hnu.edu.cn
Zhiliang Liu, Professor, Glasgow College, University of Electronic Science and Technology of China
Email: Zhiliang_Liu@uestc.edu.cn
Jinrui Wang, Professor, College of Mechanical and Electronic Engineering, Shandong University of Science and Technology
Email: wangjinrui@sdust.edu.cn
Download: Special Session #9.pdf
Driven by the rapid development of artificial intelligence and intelligent manufacturing, modern industrial equipment presents the characteristics of high automation, high efficiency and high integration, which puts forward higher requirements for operation reliability, stability, service life and availability. In this context, vibration-based equipment health monitoring has become one of the most important research directions in industrial intelligent maintenance. Advanced vibration sensing and intelligent diagnosis provide indispensable data support for accurate fault feature extraction, health state assessment and service life prediction. However, complex working conditions, strong environmental interference and weak fault characteristics still bring severe challenges to vibration signal acquisition and intelligent analysis. This Special Session focuses on advanced vibration sensing and intelligent diagnosis methods for equipment health monitoring, aiming to provide a professional academic exchange platform for researchers and engineers to share the latest theoretical, methodological and engineering application achievements.
The topics of interest include, but are not limited to:
Intelligent Diagnosis and Prediction for Data-Scarce Scenarios: Generative AI and Few-Shot/Zero-Shot Learning
Session Organizers:
Prof. Yongbin Liu, Anhui University
Email: ybliu@ahu.edu.cn
Prof. Juan Xu, Anhui University
Email: xujuan@ahu.edu.cn
Associate Prof. Xu Ding, Hefei University of Technology
Email: dingxu@hfut.edu.cn
Download: Special Session #10.pdf
In real-world industrial operations, the severe scarcity of fault data remains the most fundamental bottleneck limiting the generalization and deployment of traditional data-driven fault diagnosis and prognosis models. While deep learning has achieved remarkable success in laboratory settings with balanced datasets, these models consistently fail when confronted with imbalanced fault categories, unseen fault types, or cross-domain operational shifts, severely undermining their engineering applicability. Recent advances in generative artificial intelligence (e.g., GANs, diffusion models, and large language models) alongside few-shot and zero-shot learning paradigms (including metric-based meta-learning, model-agnostic meta-learning, and semantic embedding methods) offer transformative pathways to overcome data scarcity by synthesizing high-fidelity fault samples or enabling diagnosis with minimal or no labeled fault examples. As these techniques rapidly evolve with the integration of physics-guided generation, cross-modal alignment, and open-set recognition, they have become a core enabler for next-generation intelligent maintenance systems—moving from data-hungry laboratory models toward data-efficient, adaptable, and industrially deployable diagnosis solutions. This session focuses on the latest research achievements and advances in generative AI, few-shot learning, and zero-shot learning for intelligent fault diagnosis and prediction in industrial systems, providing a dedicated platform for methodological innovation and engineering practice in data-scarce scenarios.
The topics of interest include, but are not limited to:
• Generative AI for fault data synthesis and augmentation
• Few-shot learning and meta-learning for fault diagnosis and prognosis
• Zero-shot learning, generalized zero-shot learning, and open-set recognition
• Physics-guided generative AI and digital twin-informed augmentation
• Multi-modal data fusion with generative and few-shot methods
• Explainable and interpretable models for data-scarce diagnosis and prognosis:
• Transfer learning, domain generalization, and source-free adaptation for fault diagnosis and prognosis
Advanced Sensing, Monitoring and Diagnosis of Renewable Energy batteries
Session Organizers:
Prof. Lei Mao, HeFei University of Technology
E-mail: leimao82@ustc.edu.cn
Assistant Prof. Zhiyong Hu, Anhui University
E-mail: hzyllwen@ahu.edu.cn
Dr. Hang Wang, Anhui University
E-mail: hangwang@ahu.edu.cn
Download: Special Session #11.pdf
With the deepening of clean and low-carbon transformation of energy, large-scale development and utilization of renewable energy require the safe operation of renewable energy batteries, which prompts advanced sensing technology to detect and monitor the security of renewable energy batteries. Based on advanced sensing technology, the operation condition of batteries can be in-situ detected and monitored. Once abnormal states and faults are occurred, state identification and diagnosis will be further implemented. More importantly, advanced sensing technology provides multi-dimensional data reflecting the detailed status of battery, from which data processing methods (including artificial intelligence algorithms) can be applied to explore more internal information of batteries for investigating state evolution mechanism and fault early warning.
The topics of interest include, but are not limited to:
20th August 2026 -
Manuscript Submission
30th September 2026 -
Acceptance Notification
15th October 2026 -
Camera Ready Submission
15th October 2026 -
Early Bird Registration
Website:
https://icsmd2026.aconf.org/
Email:icsmd_adm@163.com