Cutting-edge Technologies for Intelligent Operation and Maintenance of Mechanical System Health

    Deadline for Manuscript Submissions: December 30, 2025

     

     

    Special Issue Editors

     

    Dr. Guoqiang Li Website  E-Mail: lgq@jmu.edu.cn
    School of Marine Engineering, Jimei University, China
    Interests: Intelligent fault diagnosis driven by physical information networks; Contrastive learning; Deep reinforcement learning

     

    Dr. Weijie Bao Website  E-Mail: baowenjie@cumt.edu.cn
    School of Mechanical and Electrical Engineering, China University of Mining and Technology, China
    Interests: Fault diagnosis for mechanical systems; nonstationary signal processing and time-frequency analysis; structural health monitoring

     

     

    Special Issue Information

    This special issue focuses on cutting-edge technologies for intelligent operation and maintenance of mechanical system health. It includes a transformative framework integrating digital twin degradation simulation, physically constrained neural operators, and multimodal sensor fusion for self-aware predictive maintenance of mechanical systems across the dynamic range of operation.

    The escalating intricacy of contemporary mechanical systems, combined with stringent requirements for operational resilience, sustainability, and lifecycle optimization, has catalyzed transformative progress in intelligent condition monitoring and predictive maintenance. This special collection seeks cutting-edge contributions that bridge theoretical innovation and industrial implementation in asset health management. We encourage submissions demonstrating cross-disciplinary approaches integrating physics-informed models, data-driven techniques, and intelligent frameworks to advance fault detection, diagnostics, prognostics, and maintenance optimization. Particular emphasis lies in developing explainable AI solutions for multi-component systems operating under dynamic, uncertain environments to optimize maintenance strategies and extend asset longevity.

    Topics of interest include, but are not limited to:

    • Signal Processing Techniques: Novel approaches for multimodal sensor streams (vibration, acoustic, imaging) leveraging physics-guided representation learning, spectral-graph neural architectures, and multivariate time-series decomposition to enable discriminative feature mining.
    • Zero-Shot and Few-Shot Fault Diagnosis: Meta-learning frameworks and cross-domain adaptation strategies addressing scarce annotated samples through knowledge distillation, synthetic fault generation, and self-supervised representation alignment, optimizing generalization across equipment variants.
    • Physics-Informed Machine Learning: Neural architectures encoding domain-specific governing equations (e.g., Paris' law for crack propagation, Arrhenius-based thermal aging) via physics-constrained loss formulations and symbolic regressors, enabling uncertainty-aware prognostics with sparse training samples through knowledge distillation.
    • Advanced Remaining Life Prediction: Multimodal sensor fusion frameworks integrating physics-constrained neural operators, hybrid neuro-physical architectures, and self-supervised temporal transformers to achieve uncertainty-aware RUL projections under non-stationary operational regimes.
    • Signal Processing-Informed Machine Learning: Co-designed architectures integrating adaptive wavelet transforms, time-frequency atoms, and attention-based neural operators to achieve noise-robust feature distillation from nonstationary vibration/acoustic signals, optimized via multimodal sensor fusion.
    • Digital Twin for PHM: Multiscale asset modeling enabling cross-domain synchronization between virtual replicas and physical entities, facilitating real-time degradation mirroring, fault scenario emulation, and maintenance action validation through embedded physics-dynamics constraints.
    • Multi-Modal Data Fusion: Integration of multi-sensor or heterogeneous data sources to improve accuracy and robustness.
    • PHM for Variable Operational Conditions: Adaptive signal processing techniques or PHM models addressing non-stationary environments, load variations, and transient states.
    • IoT-Driven System Monitoring: Edge computing, wireless sensor networks, and cloud-based architectures for real-time health monitoring.
    • Interpretable Machine Learning: Transparent and explainable AI/ML techniques for trustable PHM decision support.
    • Reliability Assessment: Probabilistic and statistical methods for system reliability evaluation and risk management.
    • Uncertainty Quantification: Techniques to model and calibrate uncertainties in PHM decisions.

     

    Keywords

    prognostics and health management

    fault diagnosis

    fault detection

    signal processing

    remaining useful life

    digital twin

    interpretable machine learning

    reliability assessment

    uncertainty quantification