Proactive Fault Prediction in Marine Diesel Engines Using Edge Computing Technology
Contributors
Prosanjeet Sarkar
Keywords
Proceeding
Track
General Track
License
Copyright (c) 2026 Sustainable Global Societies Initiative

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Abstract
Modern marine machinery is inherently complex, comprising interconnected systems and subsystems, including the ship management system and the safety and protective systems for the vessel and crew. These subsystems are interconnected and collectively contribute to the ship’s operational purpose; their maintenance may be critical to ensuring smooth system operation. Marine diesel engines used to produce the electricity on board, are the primary source of electricity and supply power to all essential and non-essential equipment, if any unplanned breakdown occurs in the machinery it disturb the ship operation and impact of huge financial, environmental and life loss. Increasingly rigorous international regulations regarding ship and crew safety, as well as environmental protection, encourage ship owners to implement the proactive maintenance system to ensure interrupt operation without delay. This paper presents a framework that utilizes real-time auxiliary-engine sensor data ollected from a cargo ship over a one-year monitoring period. Edge-computing hardware was installed aboard the ship, and five machine-learning models—Logistic Regression, Extra Trees, Random Forest, LightGBM, and XGBoost—were developed; the XGBoost model appeared to outperform the other models. The model achieved a reported accuracy of 98.89% following hyperparameter tuning, suggesting strong predictive e performance. This research may offer a scalable, interpretable, and data-driven framework for predictive maintenance in complex marine environment, contributing to the advancement of ship machinery and sustainable asset management.