AI-based Surrogate Model for the Prediction of Ship Fuel Consumption Reflecting Hydrometeorological Conditions

Accurate prediction of ship fuel consumption is essential for optimizing ship performance and minimizing environmental impact. This study presents the development and validation of an artificial intelligence (AI)-based surrogate model specifically designed to predict Ship Fuel Consumption (SFC) in the case of a bulk carrier. The surrogate model employs a cutting-edge approach by combining deep learning techniques, specifically incorporating attention mechanisms into Bidirectional Long Short-Term Memory (Bi-LSTM) networks. This advanced model leverages a rich and diverse dataset comprising crucial operational parameters, including ship navigation, ship operational conditions, engine operational status, and Metocean data, to achieve highly accurate predictions of SFC. The dataset used for training and validation is sourced directly from realistic bulk carrier operations, ensuring the relevance and practical applicability of the model. Extensive generalization tests were conducted to evaluate the performance of the developed surrogate model. The results indicate that the AI-based surrogate model achieves long-term high accuracy in predicting ship fuel consumption under varying operational conditions. The developed surrogate model may serve as a valuable tool for bulk carrier operators, offering insights into fuel efficiency improvements and enhancing the overall sustainability of ship operations.