A Probabilistic Model for Estimating Ship Performance in Ice Conditions

Loading...
Thumbnail Image

Access rights

openAccess
publishedVersion

URL

Journal Title

Journal ISSN

Volume Title

A4 Artikkeli konferenssijulkaisussa

Date

Major/Subject

Mcode

Degree programme

Language

en

Pages

11

Series

Proceedings of the 28th International Conference on Port and Ocean Engineering under Arctic Conditions, Proceedings of the International Conference on Port and Ocean Engineering under Arctic Conditions, POAC

Abstract

During the winter months, ice conditions can significantly impact ship performance, posing challenges for efficient navigation. Traditionally, deterministic models, such as semi-empirical approaches, have been used to estimate ship performance in ice, resulting in non-linear dependencies between ship's speed (v) and ice thickness (h), i.e. so-called h-v curve. However, these models often fall short in capturing the inherent variability and uncertainty associated with complex ice conditions. Given the complex and unpredictable nature of ice, there is a motivation to adopt probabilistic modeling techniques that can account for these uncertainties. As a preliminary exploration, this study models ship transit speed in ice using the h-v curve guided Gaussian Process Regression (HGP), providing a probabilistic relationship between ice conditions and ship speed. The proposed model is applied to independent navigation trips of merchant ships, and the outputs include the mean transit speed and the uncertainty of each estimation. Mean absolute error and the root mean squared error are used to evaluate the effectiveness of this approach. The output of the HGP is compared with that of the standard Gaussian Process model and the recorded ship speed, highlighting the need for combining the physics-based guidance with the data-driven insights for better supporting ship performance analysis in ice. The proposed approach could be further developed e.g., by refining the integration of physics-based guidance with data-driven approach, incorporating additional variables to represent ship maneuvers, and integrating observed ice data to mitigate inherent data uncertainties.

Description

Publisher Copyright: © 2025 Lulea University of Technology. All rights reserved.

Other note

Citation

Liu, C, Arif, A, Kujala, P, Suominen, M & Musharraf, M 2025, A Probabilistic Model for Estimating Ship Performance in Ice Conditions. in Proceedings of the 28th International Conference on Port and Ocean Engineering under Arctic Conditions. Proceedings of the International Conference on Port and Ocean Engineering under Arctic Conditions, POAC, Curran Associates Inc., International Conference on Port and Ocean Engineering under Arctic Conditions, St. John's, Newfoundland and Labrador, Canada, 13/07/2025.