In modern industry, welding plays an important role for jointing materials and arc welding has been the dominant process for many years. In general, in order to produce a high-quality weld, the system should have two main functionalities. First, it should have the capacity of directing the welding torch based on the detection of the joint location and geometry, namely “seam tracking”. Second, the system should be able to adjust the welding parameters, namely “adaptive control of the welding condition". Both functionalities heavily rely on the sensor systems. The analysis of welding technological processes, adaptive control and intelligent systems proved the opportunity of utilization artificial neural networks in the procedure of weld seam tracking and weld head movement control. Welding process engineering and adaptive correction with artificial neural networks are considered in the paper both with the computer vision system based on video monitoring and laser profilоmetry. The paper is focused on the development of a proper sensor system to facilitate the arc welding automation and neural network-based control to provide precise seam tracking and high-quality joints.
Keywords: Adaptive control system, automatic welding, advanced welding equipment, adaptive welding technology , welding, pipeline.