Eakkachai WarinsirirukFuad MahfudiantoMahidol University2020-01-272020-01-272019-01-22TIMES-iCON 2018 - 3rd Technology Innovation Management and Engineering Science International Conference. (2019)2-s2.0-85062593478https://repository.li.mahidol.ac.th/handle/123456789/50454© 2018 IEEE. The objective of this study is to design the feature extraction (time-frequency domain analysis) for detecting the mock-up weld defects (normal, porosity, irregular weld bead, and incomplete filling groove) in real-time at GMAW. Welch algorithm was implemented and improved for welding quality inspection by dividing into six frequency bands with a satisfactory sample size (512 samples/second). Frequency band 2 (frequency ranges 26-52 Hz) was selected for designing the internet of things (IoT) welding embedded system by proving the most sensitive to detect the mock-up weld defects.Mahidol UniversityBusiness, Management and AccountingComputer ScienceEngineeringConference PaperSCOPUS10.1109/TIMES-iCON.2018.8621641