Repository logo
 

Prediction of Friction Stir Weld Quality Using Self Organising Maps

Files

Whole thesis(4.77 MB)

Date

2010

Authors

Supervisor

Littlefair, Guy
Chen, Zhan

Item type

Thesis

Degree name

Master of Philosophy

Journal Title

Journal ISSN

Volume Title

Publisher

Auckland University of Technology

Abstract

Interest in using artificial neural networks for predicting & forecasting has led to a tremendous surge in research activities in the past decade. Self organising maps, also commonly known as unsupervised neural networks, are known to generate their topology during learning. This leads to a network structure which converts complex, nonlinear relationships between multi-dimensional data into simple geometric relationships on a low-dimensional display. In this study, a Self Organising Map (SOM) is employed to predict the quality of welds using the Friction Stir Welding (FSW) process. FSW is a relatively novel welding technology, which has caught the interest of many industrial sectors due to its many advantages and clear industrial potential. Despite the successful deployment of FSW in industry, research relating to friction stir weld quality is developing rather gradually. This is mainly due to the non-deterministic nature of the environment in which the system must function. The study is aimed to demonstrate and apply the most important property of the SOMs to FSW - orderliness of input-output mappings. Experimental data was collected by performing a series of FSW trials on Aluminium alloy AA2024 and A253 against a selected range of parameters. The SOM system was trained using the prepared training set. The generated model was tested against an unseen set of data captured during the FSW trials. The network results were in good agreement with the previously unseen data. It has been demonstrated that the SOM algorithm can be used as reliable tool for predicting weld quality.

Description

Keywords

Friction Stir Welding, Artificial neural networks

Source

DOI

Publisher's version

Rights statement

Permanent link

https://hdl.handle.net/10292/1229

Collections

Masters Theses