DATLAS : a new approach for monitoring the laser welding process

Laser welding is nowadays a widely used, established welding technology. Critical components often require highest weld quality without any defects. The reliability can be sensitive to changes of the process conditions, thus costly post inspection is often applied. Automated process monitoring could eliminate post-inspection. However, due to the complexity of the welding process commercial systems neither provide a comprehensive documentation as reference nor a systematic and safe method for determining a monitoring criterion. Therefore new correlation criteria have to be empirically identified for each new application. The recently launched research project DATLAS (Data Interactive Process Monitoring for Laser Welding) aims at understanding the context between dynamic process changes causing laser welding defects and the dynamic sensor signal obtained for commercial inprocess monitoring by photodiodes or cameras. Seven Swedish companies selected different cases, specified by the metal, thickness, joint type, laser and a typical welding defect to be detected and studied. The context between the welding process dynamics and the sensor signal is investigated by high speed filming of motions of the melt pool surface as well as of the plasma plume followed by numerical simulation of their thermal radiation that eventually illuminates the sensor and leads to a signal versus time. The research hypothesis assumes dynamic correlations either by transients of the geometry or the temperature field of the welding process. From the cases a more general theory shall be developed in order to explain the context between process and signal dynamics, thus supporting monitoring systems and their industrial use.