Abstract The present work presents a new method for classifying welding faults in real time during gas metal arc welding based on signature images. The through the arc technique classifies faults when they occur during welding. The method is generally applicable and utilises angle measures in the space of the welding signature images to recognise particular faults, accurately identifying them by comparison with the characteristics of known fault signatures in a database which can be extended as new faults appear. Overall, 10 categories representing common industrial faults are considered for pulse welding of a fillet joint. It is shown that the classification technique generalises fault categories in a reasonable way when unknown faults are encountered, and is capable of discriminating between similar faults when necessary. The method can reduce down time due to faults in production welding, as well as providing information which is useful for reducing rates of fault occurrence.
[1]
C. S. Wu,et al.
Intelligent monitoring and recognition of the short-circuiting gas—metal arc welding process
,
2004
.
[2]
Chuansong Wu,et al.
Real-time sensing and monitoring in robotic gas metal arc welding
,
2006
.
[3]
S. W. Simpson.
Signature images for arc welding fault detection
,
2007
.
[4]
Tena I. Katsaounis,et al.
Analyzing Multivariate Data
,
2004,
Technometrics.
[5]
P. Green,et al.
Analyzing multivariate data
,
1978
.
[6]
D. Rehfeldt,et al.
Gas metal arc welding process monitoring and quality evaluation using neural networks
,
2000
.
[7]
G. G. Stokes.
"J."
,
1890,
The New Yale Book of Quotations.
[8]
Aaas News,et al.
Book Reviews
,
1893,
Buffalo Medical and Surgical Journal.
[9]
K. Luksa,et al.
Influence of weld imperfection on short circuit GMA welding arc stability
,
2006
.