A tertiary review of condition-based maintenance literature: insights and future directions

Condition-Based Maintenance (CBM) optimizes asset management by integrating multisensor data, advanced analytics, and decision-making. This tertiary review analyzes CBM’s evolving architecture across sensing, analytics, and decision layers. It highlights mature vibration sensing, challenges in applying deep learning models industrially, and complexities in edge-cloud integration and digital twin deployment. Multimodal sensor fusion improves fault diagnosis but faces practical issues like synchronization and cost. Despite academic progress, gaps remain between research and industrial use due to data variability and lack of standardization. Future work should focus on physics-informed learning, federated learning, autonomous sensing, and scalable digital twins. This review identifies key challenges and paths for broader CBM adoption.