Abstract
In laser powder bed fusion (LPBF) additive manufacturing, the melt pool (MP) characteristics are key indicators for process and part defects. For example, the laser scan location on the build plate can slightly change the MP size and shape due to the f-θ optics of the printer. The laser scan parameters can cause variations in MP signatures that may indicate lack-of-fusion and keyhole regimes. However, the effects of these process parameters on MP monitoring (MPM) signatures and part properties are not yet fully understood, especially during a multilayer big-part printing. In this study, our objective is to comprehensively evaluate the dynamical changes of MP signatures (location, intensity, size, and shape) under realistic printing scenarios-printing multilayer objects at different build plate locations with various print process settings. To accomplish this, we developed a coaxial high-speed camera-based MPM system for a commercial LPBF printer (EOS M290), to capture MP images continuously throughout a multilayer part. From our experimental data and results, we find that the MP image position on the camera sensor is not stationary as reported in the literature and is partly subjected to scan location. Its correlations to process deviation or part defect need to be determined. Also, the MP image profile can significantly reflect the changes in print process conditions. The developed system and analysis method can be used to establish a comprehensive profile of MP image signatures for online process diagnosis and part properties prediction, thus ensuring quality assurance and control in LPBF.