X-ray and neural network based in-situ identification of the melt pool during the additivemanufacturing of a stainless steel part

Laser Metal Deposition with Powder (LMDp) is an additive manufacturing techniqueused for repairing metal components or producing parts with intricate geometries.However, a comprehensive understanding of the melt pool dynamics, whichsignificantly influences the final properties of LMDp-fabricated parts, remains limited.Non-destructive testing is highly valuable for conducting in-situ controls duringmanufacturing. X-ray imaging offers the ability to penetrate metallic parts and detectdefects such as porosity. In the context of additive manufacturing, X-rays can beemployed to visualize the shape of the melt pool during the fabrication process. Thecontrast between the liquid and solid phases, due to their density differences, shouldbe observable in the radioscopy images.The experimental setup required to perform such a test on an industrial additivemanufacturing installation consists of a movable X-ray source that producespolychromatic beams, a detector, and extensive lead shielding to ensure X-ray safety.In-situ observations of the melt pool were conducted during the deposition of tensuccessive layers of stainless steel 316L (SS316L). The polychromatic nature of the X-ray beam, however, rendered traditional image analysis methods ineffective fordetecting contrast variations. To address this challenge, neural networks trained onsimulated data (thermal and X-ray) were employed, providing a solution to identify themelt pool in low-contrast radioscopic images. The architecture inspired by VGG16demonstrated promising results, confirming the potential for in-situ non-destructivetesting using X-ray imaging in industrial additive manufacturing processes.