Data analysis to assess part quality in DED-LB/M based on in-situ process monitoring

Abstract

In the field of Laser beam Direct Energy Deposition (DED-LB/M) for metal additive manufacturing, the implementation of qualification strategies parts from monitoring data and reduced order models is presently at low level of maturity. In this work, a methodology and a suite of novel data analysis tools targeting the joint analysis of multimodal data: process parameters, coaxial thermal imaging and part quality by Computer Tomography scans is presented. To demonstrate the proposed approach, a set of stainless-steel coupons were built with varying process parameters (power, process speed) and path planning strategies. Exploratory data analysis and feature engineering was performed on the dataset: process indicators, thermal and geometrical monitoring features are correlated to spatially resolved defects (mainly cracks) as well as the overall part quality obtained from the inspection phase paving the way for further implementation of in-situ process monitoring as a reliable tool for process optimization and qualification.