کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

Determining the Position of the Electric Arc Tool Tip in CO₂ Welding Using Arc Sound Signal and Artificial Intelligence

نویسندگان :

Amin Farahani¹ Majid Sanaeepour² Maryam Momeni³

1- دانشگاه اراک 2- دانشگاه اراک 3- دانشگاه اراک

کلمات کلیدی :

CO₂ welding،Artificial intelligence،Electric arc sound،Optimal distance،Electrode tip

چکیده :

Considering the direct effect of the distance between the welding electrode tip and the workpiece on weld quality, determining the electrode tip–to–workpiece distance during welding is of great importance. In this study, the sound of the electric arc was used to determine the electrode tip–to–workpiece distance using artificial intelligence. For this purpose, arc sound signals were recorded at three different ranges of distance between the electrode tip and the workpiece, and due to their non-stationary nature, time–frequency features were extracted. The results indicate that variations in the electrode–to–workpiece distance lead to the generation of distinct patterns in the electric arc sound signal. Accordingly, the electrode tip–to–workpiece distance was classified into three classes: near, normal, and far. Subsequently, several machine learning algorithms, including Support Vector Machine (SVM), Logistic Regression, and K-Nearest Neighbors (KNN), were employed to classify the arc sound signals. The results show that the SVM algorithm achieved the best performance with an accuracy of 95.9% in classifying the electric arc sound signals. Logistic Regression and KNN followed with accuracies of 94.7% and 93.5%, respectively. The proposed approach enables real-time monitoring of welding quality and, by providing immediate feedback to the operator, not only improves weld quality and reduces energy consumption but also minimizes the need for destructive testing.