Laser welding is a very important factor in modern industrial production. Especially in automotive industry, it can gain advantages to fulfil the current demands in lightweight design and meet high productivity together. But laser welding is also a highly sophisticated process. To reach good welding results, a lot of different influences have to be considered. In this regard, process monitoring plays a major role. Almost every modern in-process monitoring approach is based on electromagnetic techniques. All investigations of acoustic emissions were concentrating on lower frequency ranges (< 100 kHz air-borne; < 500 kHz structure-born). In order to set up a welding process with new parameters, metallography is applied offline to measure all important quality parameters of the laser welds.
In this paper, a new in-process and a new offline monitoring strategy are presented. The in-process monitoring technique is based on high frequency structure- and air-borne acoustic emissions, which are measured during the welding process. The investigated frequency range lies between 400 kHz – 1200 kHz for structure-borne acoustic emissions and between 700 kHz – 1000 kHz for air-borne acoustic emissions. The correlation to variables of the laser welding process (e.g. laser power, welding speed, material, coating) as well as to important quality parameters of the welds (e.g. penetration depth) is shown. Also the influence of environmental noise, which was the major problem in further investigations, is discussed. In addition a new offline monitoring technique is presented. High frequency immersion ultrasonic testing was validated in order to measure important quality parameters for overlapping welding seams (e.g. welding width, width of connection, lack of connection, welding surface, pores, false friends). In comparison to metallography, this method delivers testing results for the whole volume of the welding seam. Due to the high testing speeds, it helps to find out suitable welding parameters, if new welding processes should be validated.