Adaptive chirplet transform for sensitive and accurate monitoring of pulsed gas metal arc welding process

Condition monitoring of welding processes have received considerable attention in recent years. The method proposed in this paper provides a novel and a better method for analysis of the weld joint strength, i.e., the adaptive chirplet transform. The presence of the nonlinearities in the various sensor outputs of the monitoring systems of the welding procedure demands a more precise signal processing method for a more accurate analysis of the weld joint strength. The adaptive chirplet method has been used here which produced much better results than the other statistical signal processing methods like the wavelet transform technique due to a better time-frequency resolution of the same. In nonlinear feature extraction, wavelet transform technique was first used to detect the weld joint strength using current as a sensor output during the welding. Then the similar procedure was followed using the adaptive chirplet analysis technique which not only showed better differencing capacity between various signals but also provided better time-frequency resolution for the experimental cases where the wavelet method could not predict the weld joint strength correctly. A thorough laboratory study shows that the diagnostic method proposed in this paper is much more accurate, has high sensitivity with respect to faults, and also has better diagnostic resolution.