Stability Control of Weld Formation in Variable Polarity Plasma Arc Welding

点击翻译按钮获取中文摘要

This project will investigate a novel approach in regulating the heat and arc force applied to the workpiece as a basis for closed-loop control of the keyhole stability in variable polarity plasma arc welding. A machine vision system will be used to capture clear keyhole images from the front side of a workpiece and correlate these images with the dynamics of weld formation. A closed-loop control system based on a neural network model that relates the keyhole size with control variables such as wire feeding speed and welding current was established and tested. Variable polarity plasma arc welding is an ideal arc welding process for aluminum alloys. It features gas pore-free, full-penetration weld in a single pass of thick sections ranging from 3 mm to 25 mm, and relatively low cost. Currently, the stability of weld formation is still dependent on open-loop control. The critical features of this process can be lost due to the existence of unexpected interference. This is unacceptable for the fabrication of key products where welding accuracy must be achieved in the first trail. The project is expected to advance the plasma arc welding technology and result in significant impact on the production of high-quality key aluminum alloy products in the defense, avioanics, automobile, and shipbuilding industries.