Feature extraction for gas metal arc welding based on EMD and time-frequency entropy

摘要

Hilbert-Huang transform (HHT) and time-frequency entropy were used to estimate the stability of short-circuiting gas metal arc welding (GMAW). First, the current signals were divided by empirical mode decomposition (EMD) into several intrinsic mode functions (IMFs). Then the IMFs were converted by Hilbert transform to Hilbert-Huang spectrum which describes the instantaneous time-frequency distribution of welding current signals. Since the uniformity of energy amplitude distribution with time-frequency reflects the stability, we introduced time-frequency entropy to quantify the energy distribution with time-frequency range in the HHT spectrum. We found HHT can effectively depict the amplitude with time and frequency distribution of welding current signals, and the welding was more stable when the time-frequency entropy was larger. Thus, this is a new way to assess and quantify the stability of short-circuiting GMAW.