Combining HMM for state recognition and Hertzian Contact Theory for 'blind-spot' bearing wear estimation.
Read Time: 9 min read | Published: 2025-03-15
Conventional vibration monitoring often fails for offshore cranes due to the variable duty cycles and the unique nature of slewing bearings\u2014large, low-speed, heavy-load components that behave differently from standard bearings (Journal of Vibroengineering, 2016). Standard sensors often have 'blind spots' regarding the internal wear of these critical components.
We implement a double-layer Hidden Markov Model (HMM) to distinguish operational intentions (lifting, slewing, idling) from noise, mirroring methodologies used for container crane driver identification. For health assessment, we use a multi-physical signal model (torque, temperature, vibration) coupled with Hertzian contact stress distribution analysis to predict Residual Useful Life (RUL).
This approach fills a critical gap where no prior literature existed for condition-based life prediction of slewing bearings. The HMM layer correctly segments operational states, allowing the physical model to accurately estimate contact forces and defect progression using Hertzian theory, providing a validated method for large-bearing prognostics.
This technology is mature and ready for immediate fleet-wide rollout. It serves as a low-risk, high-reward entry point for physics-informed predictive maintenance.
Multi-Physics PINNs for layer-by-layer stress and thermal fade prediction on multi-layer drums....
Generative Design for Heavy Lift CranesLeveraging topology optimization and evolutionary algorithms to discover structural efficiencies that defy conventional engineering intuition....
Data Scarcity in Maritime Predictive MaintenanceCritical offshore assets do not fail often enough to feed classical machine learning, so most pure data driven predictive maintenance initiatives stal...