The rapid expansion of international maritime transportation has led to rising greenhouse gas emissions, exacerbating climate change and environmental sustainability concerns. According to the International Maritime Organization, carbon dioxide (CO₂) emissions from vessels are projected to increase by over 17% by 2050. Traditional emission estimation methods are prone to inaccuracies due to uncertainties in emission factors, and inconsistencies in fuel consumption data. This study proposes deep learning-based CO₂ emission prediction models leveraging engine operation data. Unlike previous approaches that primarily relied on fuel consumption, this model incorporates multiple parameters capturing the relationship between combustion characteristics and emissions to enhance predictive accuracy. We developed and evaluated individual models—convolutional neural network (CNN), long short-term memory (LSTM), and temporal convolutional network (TCN)—as well as hybrid model (TCN–LSTM). The hybrid model achieved the highest predictive performance, with a coefficient of determination of 0.9726, outperforming other models across multiple quantitative metrics. These findings demonstrate the potential of deep learning for vessel emission assessment, providing a scientific basis for carbon management strategies and policy development in the international shipping industry. This study thus holds major academic and industrial value, advancing the field of deep learning-based emission prediction and extending its applicability to diverse operational scenarios.

Greenhouse gas emissions from ships are estimated to be approximately 1002 million tons per year; this is the largest carbon dioxide (CO₂) emission source among nonroad transportation. Previous studies have generally estimated CO₂ emissions using fuel- or power-based emission factors based on fuel consumption or engine power. In this study, CO₂ emissions from vessels were measured using a portable emission measurement system. Emission characteristics were analyzed according to the vessel’s operation conditions and compared with the results of other studies. Generally, the higher the rpm value, the more CO₂ is emitted, and the emissions at the maximum rpm differ depending on the type and size of the engine. In order to minimize the emissions by ships, those from high seas should be reduced rather than nearby ports. In addition, a method of establishing optimal operating conditions in consideration of economic and environmental perspectives was proposed. Fuel-based emission factors elicited in this study were constant regardless of engine rpm. The fuel-based emission factors of each engine were found to be similar at 3144.22 and 3150.58 kg-CO₂/tonne-fuel. Therefore, distinguishing CO₂ emission factors according to engine type is not necessary, and additional research is required to understand the emission factors of each fuel type.

Ships carrying large amounts of cargo and passengers are larger and slower than other modes of transportation. They are mostly foreign flagged and operate at sea far from coasts for 20 years or more, incurring more operating costs than construction costs. Therefore, an efficient maintenance system is necessary for stable, economical ship operation. Researchers are attempting to equip ships with predictive maintenance technology, which is used proactively in other modes of transportation to predict the maintenance time of machines through data monitoring and analysis. However, due to the nature of ship operation, data collection is difficult, and most studies focus on fault detection, hindering the application of predictive maintenance to ships. In this study, we developed a maintenance time prediction algorithm using the revision generator engine condition criterion (RGCCV) value and the cylinder exhaust gas temperature, as developed in a previous study for marine generator engines. And through comparison and verification using machine learning, the average mean absolute error (MAE) across all cylinders was 2.916 for the RGCCV-based method and 8.138 for the temperature-based method, demonstrating a 64% improvement. These findings establish a practical foundation for implementing predictive maintenance in ship engines by enabling more reliable and condition-based maintenance.