Development of prediction system model for mechanical property in friction stir welding using Support Vector Machine (SVM) / Arman Syah ... [et al.]

• Main Authors: Syah, Arman (Author), Astuti, Winda (Author), Saedon, Juri (Author)
• Format: Book
• Published: Faculty of Mechanical Engineering Universiti Teknologi MARA (UiTM), 2018.
• Subjects: Article; PeerReviewed
• Online Access: Link Metadata

Summary

The cost and efficiency of experiments and tests, are still a major issue in manufacturing. In this work, a machine learning technique i.e. support vector machine (SVM) is applied to develop a system model for prediction the mechanical property in friction stir welding (FSW) of a 6mm thick AA6061 welded joint. Experimental works are performed to measure the welded structure properties, which focus on the tensile strength based on the governing parameters. In the development of the prediction system model the data obtained from the governing parameters and the tensile strength measurement are classified into two different classes, high and low tensile strength, as input for the SVM classifier through training and testing the data from both classes for pattern classification and model development. The result of the testing and evaluation stage of the proposed system model shown the achievement of the prediction accuracy of 100% for each training and testing system on both classes. The study proved the proposed system model is fully accurate. The methodology given in this paper delivers a useful tool to predict the coming tensile strength of friction stir welded structure based on the governing parameters without conducting experiment and test.

Item Description

https://ir.uitm.edu.my/id/eprint/40606/1/40606.pdf