The classical inspection methods used for detecting and finding disturbances in welding process are based on direct measurement of its parameters as arc voltage, welding current, wire feed speed, etc. Using these inspection methods implies sensors insertion around the welding process and its presence could alter the metallic transference behavior and consequently an uneven quality as well as it can increase the production cost. For reducing these implications is necessary using a non intrusive inspection method. In this paper we will show nonintrusive methods to the weld quality inspection. These methods are based on sensor fusion, the extraction of global information coming from the interrelation data given by each sensor that, for example, sensing the spectroscopy radiation emission, the acoustic sensing of the electrical arc, the infrared emissions indicating the heat content of the weld. Finally, the fusion data will be applied to a statistical control for detecting and finding welding disturbances. The results will show that sensor fusion could be used as a tool to measure indirectly the weld quality in the GMAW process.
Quality Control in Welding; Spectroscopy; Acoustic sensing; Infrared emission; Sensor fusion
