Equipment and Machine Learning in Welding Monitoring

Abstract

Laser welding has been widely applied to various industries, effective real time monitoring can help to improve the welding efficiency and production quality. This paper makes a short review on the signal detection equipment and machine learning algorithms. It starts with a detailed introduction to some basic monitoring sensors and methods, some special monitoring methods like ICI, MOI and multiple sensor fusion technology are also talked over. The commonly used machine learning algorithms like BPNN and SVM used in welding data mining, weld defects classification and weld seam features forecasting are summarized in the end section. This fundamental work aims to provide a guideline for the selection of monitoring equipment, chose suitable machine learning algorithms for effectively classifying weld defects and realizing welding process intelligent real time monitoring.