In-process monitoring and characterization of arc welding

Inspection of weld quality is critical because it ensures the integrity of structures. There are many available methods for monitoring and diagnosis of the weld quality. However, most of them are off-line and thus make the quality monitoring and remedial measures difficulty and costly. In this research, a real-time quality monitoring and diagnosis method based on the input electrical impedance of the arc welding is proposed. It is obtained by taking the quotient of input voltage to current which both signals are measured simultaneously at the output terminal of welding machine. Two time record data, real part and imaginary part, or the resistance and reactance of impedance reflect the system property of arc welding which is represented by an equivalent circuit. This equivalent circuit consists of resistor, inductor and capacitor connected in series. Therefore, any abnormal change of arc welding will be reflected by the time variation of these components. Several major findings were obtained from this research. The mean and standard deviation of resistance and reactance of impedance are affected by the operating parameters such as welding voltage, current, welding speed, free wire length and leading angle of welding torch. Further investigation of impedance also leads to have a more accurate heat input per unit length by including the resistance of impedance. The implementation of Taguchi Method and Macro Testing shows that the proposed method can achieve 95% accuracy of detecting weld defects. Furthermore, the proposed method is also capable to diagnose the weld defect as result of abrupt change of arc length or wrong welding speed. The capability of proposed method is further extended to provide real-time and in-situ information of metal transfer. The time varying resistance and reactance curves reflect the dynamic change of metal transfer like the formation and detachment of droplet to the weld pool. This in-situ information helps to develop a new classification method of metal transfer which the metal transfer modes are quantified in percent weightage. Furthermore, the study of welding metallurgy indicates that the metal transfer mode shown in percent weightage is correlated with the macro structural change of weld but not the micro structural change of weld.DOCTOR OF PHILOSOPHY (MAE