Literature Review Analysis: Fault Diagnostics of Marine Engines
Abstract
Marine diesel engines remain central to ship propulsion and onboard power generation; however, fault monitoring, diagnosis, and effective maintenance continue to be challenging due to complex subsystem interactions, variable operating profiles, and the limited availability of real failure data. This review investigates recent research on fault detection, condition monitoring, health assessment, and predictive maintenance for marine diesel engines, marine gas turbines, and directly related combustion-engine subsystems. A Scopus-based search covering 2015–2026 was screened through title, abstract, and keyword analysis. The final database includes more than 200 papers, organized into seven methodological areas: model-based and simulation-assisted diagnostics, signal-processing methods, limited-data and transferability issues, data-driven diagnosis and predictive maintenance, reliability and maintenance optimization, lubrication and wear diagnosis, and combustion and emissions performance monitoring. The review shows a clear shift after 2020 towards machine learning, deep learning, digital twins, and hybrid diagnostic frameworks. Engine monitoring and turbocharger and exhaust systems are the most frequently addressed targets, while fuel injection, lubrication, wear, and cylinder combustion diagnostics remain less represented despite their relevance to early degradation. The results also highlight a persistent gap between algorithmic development and operational validation, since many studies rely on laboratory, test-bench, or engine datasets. The review also discusses how diagnostic results can inform maintenance decisions and highlights the need to validate future machine-learning algorithms using realistic simulation datasets that reflect actual shipboard operating conditions. Future progress requires stronger validation with operational data, reusable benchmark datasets and interpretable hybrid models that combine sensor data, physical engine knowledge and maintenance methodologies.