Machine Learning Regression Based Approach for Prediction of the Ultimate Tensile Strength of Tungsten Inert Gas Welded Joints

Abstract

Many leaders in technology education have shown that the main difference between the technical design process and the process of engineering construction analysis and efficiency [1–3]. The engineering analysis phase of the construction process is where the mathematical and scientific models principles are used to help the designer predict the design results. The engineering feasibility phase process is a systematic process that uses structural elements and conditions to allow the designer to get the best out of its solution. In the process of engineering design, both analysis and efficiency are applied before any production work has started. To achieve a desired weld quality, the weld structures such bead geometry and the mechanical properties were examined and related to the weld input parameters. Tuning welding parameters in welding processes is a big challenge because there are different welding process parameters and indices for welding performance. The traditional techniques of changing welding parameter parameters are time-consuming and arduous; moreover, the quality of the weld is dependent on the welders' expertise. To address the issue of parameter setting, a number of methods have been proposed to use the welding process [4–8]. Although there are review papers to summarize these methods, there is a lack of a systematic approach to analysing and summarizing these methods. The present research work deals with the application of machine learning based regression models such as Linear Regression for the prediction of the Ultimate Tensile Strength of the Tungsten Inert Gas Welded joints. The results showed that the Linear Regression based machine learning based model performed better.