Vision-Based Welding Control and Quality Assurance – Year two

Manual welding is a highly demanding task which requires extensive expertise for certain applications such as pipe welding. To facilitate the welding process, increase productivity, and decrease welders required level of skill, Novarc Technologies has designed and manufactured a collaborative Spool Welding Robot (SWR) equipped with a laser assistant weld path tracking. Our proposed research takes into account the fact that currently, welders rely on their eyes and the limited view of the weld pool through the helmet to control the welding process. This is a very arduous task which requires training and experience. Therefore, machine vision and AI can effectively take over the task of visual inspection and detection. Our goals in this industrial research are to use weld image for weld path tracking and torch distance control. We also aim to extract information about weld quality using computer vision and machine learning techniques.

Faculty Supervisor:

Farrokh Sassani

Student:

Neda Eskandari

Partner:

Novarc Technologies Inc

Discipline:

Engineering - mechanical

Sector:

Advanced manufacturing

University:

Program: