Signals from the laser process must be as robust as possible when used for the online quality assurance of laser processing techniques. A single measuring method is often not enough, as the number of relevant characteristics provided by one signal is usually limited. The pseudo-error rate can, at the same time, be improved by a combined analysis of two signals. A synchronized observation of the laser process in different spectral ranges therefore represents a robust method of online quality assurance. This is demonstrated here using the example of laser brazing.

Method

The laser brazing process is recorded in the visible spectral range (VIS) using monochromatic external illumination. A real image of the process zone is obtained (Fig. 4). At the same time a second camera records the thermal radiation of the process as a broadband image in the near-infrared spectral range (NIR) (Fig. 5). The cameras are synchronized by an external clock.

Result

While the VIS image indicates the feed speed, wire position and geometrical dimensions of the seam, the NIR image provides information on the heat distribution of the two joining partners, the occurrence of pores (Fig. 5), wetting on one side and the position of the laser spot. Quality-relevant conclusions can be drawn by comparing NIR and VIS features. For example,the position of the laser spot (symmetry in the NIR image) relative to the position of the wire (VIS) has a decisive influence on the wetting behavior of the joining partners and thus on the occurrence of seam defects.

Furthermore, the algorithms for quality assurance gain a higher level of robustness through the combined analysis of both camera images.

Applications

The system makes it possible to conduct online quality assurance during laser brazing of two-part auto tailgates. Future applications include laser metal deposition and the SLM process. The heat distribution provides information about the geometry of the microstructure.