Estimating Average Power of Welding Process With Emitted Noises Based on Adaptive Neuro Fuzzy Inference System

Abstract

In this study, the average power consumption of an electrode welding machine during the welding process was estimated using the features of the sound emitted during welding. First, the instantaneous values of electrode current and voltage and the sound emitted during the welding process were recorded simultaneously. The minimum, maximum, average, root mean square (RMS), and energy values of the sound data were found and feature extraction was performed, and the instantaneous power and average power values were calculated using the instantaneous current and voltage values. Three Adaptive Neuro-Fuzzy Inference Systems (ANFIS) using the sound features as inputs and average power values as outputs were created, and their results were compared. The average power values consumed during the welding process have been successfully estimated at a rate of 87-95%.

Keywords

• Welters
• average power
• emitted noise
• neuro-fuzzy inference
• data acquisition