A method for the analysis of the MDTF data using neural networks

Ibrahim, Mohamed (2000) A method for the analysis of the MDTF data using neural networks. Masters thesis, Memorial University of Newfoundland.

[English] PDF (Migrated (PDF/A Conversion) from original format: (application/pdf)) - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (15MB) [English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. (Original Version)

Abstract

Numerical simulation techniques are widely used to investigate the behavior of submarines during the design stage. The accuracy of these techniques depend upon the accurate determination of the hydrodynamic coefficients for the model. -- The Marine Dynamic Test Facility (MDTF) is a new-six-degree-of-freedom forced motion testing rig. The rig has the ability to test underwater vehicles in a manner that makes it possible to determine the hydrodynamic coefficients in the equations of motion. Multi-variant linear regression is used to obtain the hydrodynamic coefficients from the experimental data. -- In this study a neural network technique to identify the hydrodynamic model from experimental data is investigated. The technique uses the model trajectory (motion history) to predict the hydrodynamic coefficients of the model. A single MDTF generated maneuver was used to train the network. The trained network was then tested using different maneuvers and the network predictions were compared to the actual MDTF measured forces and moments. -- Results obtained from the neural network technique indicate that the technique can be used to predict the hydrodynamic model of underwater vehicles. The use of this technique can dramatically cut the running costs to conduct experiments on new models.

Actions (login required)

View Item