Scientific Knowledge–Guided Machine Learning for Vessel Power Prediction: A Comparative Study
Orfeas Bourchas, George Papalambrou · 2026
Read Scientific Knowledge–Guided Machine Learning for Vessel Power Prediction: A Comparative Study by Orfeas Bourchas, George Papalambrou (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-18403.txt.
Description
Orfeas Bourchas and one co-author predict ship engine power by mixing physics with machine learning. Their 2026 paper notes accurate main engine power prediction drives performance tuning, fuel savings, and emissions compliance. Pure data-driven models capture nonlinearities but often break the propeller law linking power and speed, extrapolating poorly outside training data.
The hybrid framework starts from calm-water power curves of power proportional to speed to a power, then trains a nonlinear regressor on the residual, the gap left by weather and operations. Constraining learning to residuals simplifies training, improves generalization, and keeps physics consistent. An XGBoost model, a plain neural net, and a physics-informed neural net, each with and without the baseline, are compared. On in-service data the hybrid wins in sparse-data regions and matches the baseline where data are dense, offering a practical tool for routing, trim, and energy planning.
Catalog record
| CREATOR | Orfeas Bourchas, George Papalambrou |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-18403