Analysis and Modeling of GTAW Weld Pool Geometry - Intelligent Computing for Sustainable Energy and Environment - page 284

Analysis and Modeling of GTAW Weld Pool Geometry

XueWu Wang

Key Laboratory of Advanced Control and Optimization for Chemical Processes,
Ministry of Education, East China University of Science and Technology,
200237 Shanghai, P.R. China
[e-mail]

Abstract. A three dimensional vision sensing system was used to mimic the
human vision system to observe the three-dimensional weld pool surface in pipe
Gas Tungsten Arc Welding (GTAW) process. Novel characteristic parameters
containing information about the penetration state specified by its back-side
weld pool width and height were proposed based on the reconstructed three
dimensional weld pool surfaces. In order to obtain the penetration status in real
time conveniently, a neural network model was established to estimate the
penetration based on the proposed characteristic parameters. It was found that
the top-side characteristic parameters proposed can reflect the back-side weld
pool parameters accurately and the neural network is capable of predicting the
penetration status in real time by observing the three-dimensional weld pool
surface which is beneficial for penetration control of GTAW process.

Keywords: GTAW, Three-dimensional, Characterization, Model, PCA-BP.

1 Introduction

In manual welding process, skilled welders can ensure the welding quality through
compensation for deviation in the process. This can be achieved by observing the
weld pool surface, which contains sufficient information about the weld quality. In
contrast, more rapid and accurate welding can be achieved by automatic welding
machines, which are more convenient to adjust several welding parameters
simultaneously. Besides, welding manipulators and robots can also substitute for
workers to finish the task in some severe environments. However, machines typically
lack the visual feedback ability welders possess. Hence, a capable sensing system is
the first step to realize intelligent penetration control. Some previous works were done
to meet this increasing demand. These methods can be divided into indirect methods,
two dimensional vision methods, and three dimensional vision methods.
Indirect methods include line scanner [1], ultrasonic [2], pool oscillation [3],
infrared [4]-[5], X-ray [6], and non-transferred arc methods [7], etc. Information
acquired by above indirect methods can only reflect one characteristic of weld pool,
and it is often not sufficient. Besides, high temperature, contamination, extensive arc
light, and noise in welding process will make most above sensing methods invalid in
some conditions. Vision sensing method is more similar to the welder's visual system,