L-PBF High-Throughput Data Pipeline Approach for Multi-Modal Integration

Kristen J. Hernandez, Case Western Reserve University
Thomas G. Ciardi, Case Western Reserve University
Rachel Yamamoto, Case Western Reserve University
Mingjian Lu, Case Western Reserve University
Arafath Nihar, Case Western Reserve University
Jayvic Cristian Jimenez, Case Western Reserve University
Pawan K. Tripathi, Case Western Reserve University
John J. Lewandowski, Case Western Reserve University
Roger H. French, Case Western Reserve University
Laura S. Bruckman, Case Western Reserve University

Abstract

Metal-based additive manufacturing requires active monitoring solutions for assessing part quality. Multiple sensors and data streams, however, generate large heterogeneous data sets that are impractical for manual assessment and characterization. In this work, an automated pipeline is developed that enables feature extraction from high-speed camera video and multi-modal data analysis. The framework removes the need for manual assessment through the utilization of deep learning techniques and training models in a weakly supervised paradigm. We demonstrate this pipeline’s capability over 700,000 high-speed camera frames. The pipeline successfully extracts melt pool and spatter geometries and links them to corresponding pyrometry, radiography, and processparameter information. 715 individual prints are examined to reveal melt pool areas that exceeds 0.07 mm2 and pyrometry signal over a threshold (375 pyrometry units) were more likely to have defects. These automated processes enable massive throughput of characterization techniques.