Physics meets machine learning in shaft power prediction

Accurate shaft power prediction is central to voyage optimisation, energy efficiency assessment and IMO’s target of a 30% emissions reduction from the global fleet by 2030. Environmental effects, hull fouling, loading variations and the transient behaviour of large vessels such as bulk carriers all introduce complexities that neither hydrodynamic nor machine learning models handle well alone.

A paper presented at HullPIC 2026, Comparing AI and Ship Hydrodynamics for Ship Performance Estimation, addresses this directly, using five years of full-scale operational data from a 200m deep-sea bulk carrier, logged at 15-minute intervals from 2020 to 2025. The team evaluated five approaches: