A Novel Stability Quantification for Disk Laser Welding by Using Frequency Correlation Coefficient Between Multiple-Optics Signals

This paper presents a multioptics sensing system consisting of two photodiode sensors and two visual sensors for multisignals extraction. Five welding features, including visible light (V), laser reflection (R), keyhole size (K), metallic vapor on top of the material (PT) and at the bottom (PB), were obtained by way of signal processing and image processing. A new correlation coefficient of low-frequency component has been designed to quantify the stability of laser welding process. Under the same welding conditions, the frequency correlation between different features was analyzed. Also, with the same features, the frequency correlation between different conditions was then researched. A three-phase welding experiment was devised to present the practical application of frequency-domain correlation to welding monitoring. Experimental results showed that welding status was well reflected by the correlation of frequency-domain component that is within the range of 11-1000 Hz. The welding process remained stable when frequency-domain correlation was low. However, when welding process turned unstable the correlation would increase. The current research has devised an approach of analyzing frequency-domain correlation that could effectively quantify the stability of laser welding process.