Statistical models for the speed prediction of a container ship. (1st November 2016)

• Record Type: Journal Article
• Authors: Mao, Wengang; Rychlik, Igor; Wallin, Jonas; Storhaug, Gaute
• Is Part Of: Ocean engineering. Volume 126(2016)
• Page Start: 152
• Page End: 162
• Publication Date: 2016-11-01
• Subjects: Performance measurement -- Ship speed prediction -- Engine RPM -- Regression -- Autoregressive model -- Mixed effects model; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2016.08.033
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: 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. Highlights: Autoregressive process and Mixed Effect statistical models are introduced to derive a ship's speed/RPM characteristics. The speed/RPM characteristics are described in terms of encountered sea environments. One year's performance data of a 2800TEU containership is used to demonstrate the applications of above statistical models. Sea environments are extrapolated from climate reanalysis data, with wind/waves from ECMWF and current from NOAA. The prediction errors are less than 1% for all analysed voyages.