Inference Of Transfer Functions And Prediction Of Vessel Responses Using Machine Learning

University essay from KTH/Marina system

Author: John Conall Harrington; [2020]

Keywords: ;

Abstract: Knowledge of vessel responses to waves is of the utmost importance for vessel operations.The responses affect which routes a vessel can take, what cargo it can carry, theconditions it’s crew will experience and much more. This can pose a problem for performanceoptimisation companies such as GreenSteam, partners in this project, for whomthe vessel transfer functions are generally not available.This project aims to use bayesian machine learning methods to infer transfer functionsand predict vessel responses. Publically available directional wave spectra are combinedwith highfrequencymotion measurements from a vessel to train a model to create thetransfer functions, which can be integrated to get the motions. If successful, this would bea relatively inexpensive method for computing transfer functions on any vessel for whichthe required measurements are available. Though not many vessels measure this datacurrently, the industry is moving towards more data collection, so that number is likely torise.The results identify a number of issues in the available data which must be overcome toproduce usable results from these methods. Though results are not optimal, they show apromising start and a route is proposed for future research in this area.

  AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)