Predictive maintenance in naval vessel propulsion systems for enhanced marine operations using a BiGMM-HMM framework with divergence-based clustering

This study introduces a BiGMM-HMM Integration Framework designed to improve predictive maintenance strategies for naval vessel propulsion systems, addressing the need for efficient and reliable operation in marine engineering applications. The framework effectively manages multimodal sensor data by leveraging a unique combination of Gaussian Mixture Models (GMMs) and Hidden Markov Models (HMMs) in a bidirectional architecture. It analyses the dynamic interactions between sensors and subsystems. Two preprocessing methods are evaluated: Method 1 focuses on subsystem interactions, employing divergence-based root cause analysis to identify key sensor variables by clustering of sensors and subsystems. In contrast, Method 2 processes the entire dataset directly. Experimental results demonstrate that Method 1 outperforms Method 2, achieving an accuracy of 91%, precision of 94%, recall of 91%, and an F1-score of 91%, compared to 87%, 90%, 87%, and 88% for Method 2, respectively. These findings underscore the role of feature selection, clustering, and dimensionality reduction in predictive analytics for marine systems. Utilizing two GMMs—one for label inference and another for transition and emission probabilities—the framework captures the multimodal characteristics of propulsion system data, resulting in improved health state prediction. Applied to a naval propulsion system dataset, this approach provides actionable insights into optimizing maintenance by understanding sensor interdependencies and subsystem interactions, offering substantial advancements in marine engineering operations through more effective maintenance strategies.