Browse by author
Lookup NU author(s): Dr Martyn Lightfoot, Professor George Bruce
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
A current trend for ship construction is to minimise the thickness of the ships structure, in order to minimise the vessels weight and maximise its payload. This reduction in plate thickness has lead to an increased amount of panel distortion during welding, resulting in significant additional costs during the subsequent assembly of the vessel. Prediction of the weld induced distortion has been investigated using Artificial Neural Networks ( ANN ) and has demonstrated on a series of Im x 1m trial samples that the approach is capable of predicting the resulting distortion. In order to collect accurate data for the ANN a measurement process called close ranged photogrammetry has been employed. The method is versatile, scalable, relatively inexpensive to setup and provides accuracy appropriate to, and often above, the needs of the shipbuilder. The series of trials included both simple butt welds and stiffened plates to cover the typical structures encountered in the initial stages of ship fabrication on the panel lines. The structure is measured after every stage of the fabrication process to record when the distortion occurs. It is intended that once the shipbuilders know where the location and amount of distortion is occurring they can alter production parameters or consider different production techniques so as to try and reduce the distortion and minimise the cost of correction at later stages. This paper concludes that photogrammetry is an inexpensive and straightforward manner to record surface deformation and that the ANN approach has been successfully demonstrated on a series of trials with further work currently being undertaken to expand the process for larger 2 metre square samples and then full scale production.
Author(s): Lightfoot MP, Bruce GJ
Publication type: Conference Proceedings (inc. Abstract)
Publication status: Published
Conference Name: RINA: International Conference on Computer Applications in Shipbuilding
Year of Conference: 2007
Pages: 269-273
Library holdings: Search Newcastle University Library for this item
ISBN: 9781905040384
