Toggle Main Menu Toggle Search

Open Access padlockePrints

Vertical Motion Control of Twin-Hull Vessels Using Neural Optimal Control

Lookup NU author(s): Dr Farhad Kenevissi, Professor Mehmet Atlar, Professor Ehsan Mesbahi


Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


This paper presents a novel Neural Optimal Controller (NOC) to improve the heave and pitch motion responses of a twin-hull vessel operating in regular head seas. A practical time domain model is used to simulate the motion responses of the vessel in the presence of a pair of actively controlled forward and aft fins. Initially an On-line Switching procedure is introduced to govern a number of Linear Quadratic Regulator (LQR) optimal controllers, designed for different operating conditions of the vessel, to improve the system robustness. Although the on-line switching offered better robustness and performance characteristics, in between switching operating points, it still remained sub-optimal. Therefore, an Artificial Neural Network (ANN) controller has been proposed as an alternative. The ANN has been initially trained to emulate the same level of control at a number of design operating points and implemented on benchmark test vessel, SWATH6A, as a NOC. The advantage of this novel application is that the on-line switching procedure is no longer required and more importantly, the ANN has been capable of non-linear generalisation to give a near optimal solution away from the trained operating conditions.

Publication metadata

Author(s): Kenevissi F, Atlar M, Mesbahi E

Publication type: Conference Proceedings (inc. Abstract)

Publication status: Published

Conference Name: FAST 2001: the 6th International Conference on Fast Sea Transportation

Year of Conference: 2001

Number of Volumes: 3

Pages: 271-281

Publisher: Royal Institution of Naval Architects

Notes: "This conference supported by IZAR"

Library holdings: Search Newcastle University Library for this item

ISBN: 0903055708