Statistical models for the speed prediction of a container ship
Journal article, 2016

Accurate prediction of ship speed for given engine power and encountering sea environments is one of the key factors for ship route planning to ensure expected time of arrivals (ETA). Traditional methods need first to compute a ship's total resistance based on theoretical calculations, which are often associated with large uncertainties. In this paper, two statistical approaches are investigated to establish models for a ship's speed prediction. The measurement data of a containership during one year's sailing are used for the demonstration and validation of the presented statistical methods. The pros and cons of the methods are compared in terms of capability, robustness, and accuracy of the prediction. By means of the measured engine Revolutions Per Minute (RPM) and extracted sea environments along the ship's sailing routes, the statistical methods are shown to be able to give reliable speed predictions. Further investigation is needed to test the capability of the statistical methods for the speed prediction using engine power instead of RPM.

Engine RPM

Performance measurement

Ship speed prediction

Autoregressive model

Regression

Mixed effects model

Author

Wengang Mao

Chalmers, Shipping and Marine Technology, Marine Technology

Igor Rychlik

University of Gothenburg

Chalmers, Mathematical Sciences, Mathematical Statistics

Jonas Wallin

University of Gothenburg

Chalmers, Mathematical Sciences, Mathematical Statistics

Gaute Storhaug

Det Norske Veritas (DNV)

Ocean Engineering

0029-8018 (ISSN)

Vol. 126 152-162

Areas of Advance

Energy

Subject Categories

Probability Theory and Statistics

DOI

10.1016/j.oceaneng.2016.08.033

More information

Latest update

11/7/2022