New explainable artificial intelligence model predicts internal defects and mechanical performance in metal three dimensional printing

KIMS and Max Planck Institute develop an explainable AI model to correlate internal defect morphology with the mechanical performance of metal 3D-printed parts.

By: AXL Media

Published: Mar 3, 2026, 4:52 AM EST

Source: The information in this article was sourced from National Research Council of Science & Technology

Addressing industrial limitations in metal additive manufacturing

Metal additive manufacturing has emerged as a critical technology for producing complex, high-value components in the aerospace and defense sectors. However, its widespread industrial adoption has been hindered by microscopic internal defects that can lead to unexpected component failure. Traditional quality control methods often rely on simple porosity indicators, which fail to account for how the specific shape, size, and location of defects impact mechanical integrity. To solve this, researchers have developed a model that identifies potential failures during the initial process design stage.

Development of an explainable AI framework

A research team led by Dr. Jeong Min Park of the Korea Institute of Materials Science (KIMS), alongside colleagues from the Max Planck Institute, created an "Explainable AI" model. Unlike conventional black-box AI systems, this framework provides transparent insights into the decision-making process. It systematically analyzes the relationships between laser powder bed fusion (LPBF) process conditions and resulting defect morphology. This allows engineers to understand not just if a part will fail, but exactly why certain process variables lead to performance degradation.

Morphological analysis of internal pores

The core innovation of the developed model is its ability to categorize defects based on their morphological characteristics. By analyzing microstructural images, the AI automatically evaluates pore size, non-circularity, and spatial distribution. These factors are then directly correlated with mechanical properties, enabling a quantitative explanation of how specific defect patterns influence the durability and strength of the printed metal. This level of detail is essential for ensuring the reliability of parts used in high-stress environments.