MACHINE LEARNING MODELS IN PULSE TIG WELDING OF TITANIUM ALLOY

Abstract

Titanium is one of the most important non-ferrous metals, which finds extensive application in the aerospace industry, because of its lightweight, excellent corrosion resistance, high strength level. Pulsed current tungsten–inert-gas welding is widely used in manufacturing important structures and products because it has quite a few advantages and practical benefits. With this process, it is easy and convenient to control heat input into parent materials and weld formation, to improve arc stability, to reduce thermal distortion and to achieve fine grains in the weld zone. Important process parameters from the previous investigation are identified and an experimental plan is derived to perform the experiments. Machine learning is a beneficial method to solve problems faced in welding processes. It can improve the welding process effectiveness as well as the process of quality monitoring. Machine learning models developed will optimize the parameters to give better mechanical properties.