Correlation analysis between the beam propagation and the vapor capillary geometry by machine learning

Low Coherence Interferometry allows for a direct inline measurement of the capillary depth during laser material processing. To enable a robust interpretation of the measurement results, a large number of influencing factors have to be considered. The geometry of the vapor capillary significantly affects the reflection behavior of the measuring radiation. Particularly, with capillary dimensions in the order of the magnitude of the measuring beam, scattering and absorption effects become significant. This paper describes the utilization of a ray tracing algorithm to analyze the beam propagation within the capillary. A keyhole generator and Machine Learning methods were applied for systematic investigations of the propagation of the measuring radiation and thus enabled an improved interpretation of capillary depth measurements.