Physics-Informed Mixture Models and Surrogate Models for Precision Additive Manufacturing

TL;DR

This paper addresses defect detection in laser-based additive manufacturing by integrating physics knowledge into probabilistic clustering via physics-informed Gaussian Mixture Models. The authors embed physics through surrogate features, notably a normalized energy proxy defined as $P/(C_p V)$, to bias learning toward physically meaningful parameter variations. They validate the approach on two AM processes—Laser Powder Bed Fusion and Directed Energy Deposition—and on public alloy datasets, using real data from SS316-1 and Inconel 625 alongside additional metals. The results indicate that physics-guided mixture models can reveal underlying physical behavior, with material-specific defect patterns and several directions for improving generalization and extending to multi-modal observations.

Abstract

In this study, we leverage a mixture model learning approach to identify defects in laser-based Additive Manufacturing (AM) processes. By incorporating physics based principles, we also ensure that the model is sensitive to meaningful physical parameter variations. The empirical evaluation was conducted by analyzing real-world data from two AM processes: Directed Energy Deposition and Laser Powder Bed Fusion. In addition, we also studied the performance of the developed framework over public datasets with different alloy type and experimental parameter information. The results show the potential of physics-guided mixture models to examine the underlying physical behavior of an AM system.

Figures (1)

• Figure 1: Example of the data distribution and classification results.