Data-driven approach to evaluate the impact of hull roughness on main engine load of river-sea ships

Hull roughness increases the ship's resistance during navigation, significantly increasing fuel consumption and emissions. The study proposed the hull fouling factors (HFFs) to evaluate the time-varying impact of hull roughness on main engine load (MEL) for the three bulk carriers. Initially, ships' Automatic Identification System data, operational data, and weather and sea state data were fused to develop predictive models for MEL. Specifically, the Random Forest (RF) model outperformed others in predicting MEL. Then, by adopting a controlled variable experiment, the RF model was utilized to evaluate the impact of hull roughness on MEL as the days since clean (DSC) increased. Hull roughness analysis results showed that the HFFs of ship operated primarily on sea routes showed an increasing trend with the increasing DSCs . In contrast, the HFFs of ships primarily operated on the river-sea routes did not change significantly. The approach revealed significant temporal variations in hull roughness for ships operating in different routes. Besides, the proposed HFFs could capture the impact of sea temperature, salinity , current velocity and ship speed on hull roughness. Overall, the finding provides new insights into the time-varying hull roughness impacts on sea-to-river ships, and supports hull cleaning strategies, energy management and emission estimates. • Hull roughness differs between sea and Yangtze River routes. • Hull roughness equations help to determine the optimal interval for cleaning. • Fouling factors can capture the impact of environmental and operational conditions.