AHTS Shaft Alignment Database

Published: 3 March 2023| Version 1 | DOI: 10.17632/cn6dk6f77y.1
Diego Landim


The files here uploaded were created from the random variation of the bearings offstets of an AHTS ship and its influence coefficients matrix. They include 3 different databases (namely: the first one with 800 samples; another with 1600 samples; and the las rith 1678 sampes), and 3 different conditions (namely: ship at drydock, cold ship, and hot ship). This database were used feeding Machine Learning models in order to try to classificate if a determined set of offsets would align or not the studied shaft line system.


Steps to reproduce

The influence coefficent matrix was calculates in a finite element software, after representig the AHTS shaft line in this tool. Than, creating randomly set of bearings' offsets and applying the influence equation, it was possibel to calculate the reactions in the bearings. By comparing the forces and moments achieved with the FE models, it was determined a classification of the samples. After that, making the pre-processement some features could be discarted and others were used in Machine Learning models trying to solve a multiclass classification problem.


Universidade Federal do Rio de Janeiro


Engineering, Machine Learning, Naval Architecture