powered by:
MagicWare, s.r.o.

Robust Sensing and Control of Weld Pool Surface

Authors:Zhang YuMing, University of Kentucky, United States
Lu Wei, University of Kentucky, United States
Zhang Chuan, University of Kentucky, United States
Walcott Bruce, University of Kentucky, United States
Topic:5.1 Manufacturing Plant Control
Session:Advanced Manufacturing Applications
Keywords: Predictive control; Adaptive control; Uncertainty; Manufacturing processes; Sensor system.

Abstract

Gas tungsten arc welding is the primary process for precision joining of metals. To develop a simple yet durable sensing and control system for this process, a novel sensor which requires no additional attachment is proposed to sense the depth of the weld pool surface using a non-transferred arc. To assure its effectiveness and accuracy, a flat surface is periodically established to provide a real-time reference so that the weld pool surface depth can be accurately measured. Because of possible large process variations, an interval model control algorithm with adaptively updated intervals is adopted. Closed-loop experiments showed that the developed control system is capable of achieving quality welds and is robust with respect to different variations.