Open Access

In-situ captured visual images of the laser powder bed fusion process (PBF-LB/M) provide valuable insights into process dynamics. Automatic analysis of after-recoating images using machine learning algorithms enables the detection of process deviations to reduce scrap production. However, current industrial monitoring systems for PBF-LB/M are limited by low image resolution. While higher resolutions enable the system’s ability to capture smaller features, they increase storage and computational demand. Edge devices offer a solution by enabling near-real-time, on-premises image analysis within the machine and company network. In this study, high-resolution after-recoating images, captured with a spatial resolution of 17 µm/pixel and an image size of 9344 x 7000 pixels, were processed on an Nvidia Jetson Orin NX16 edge device. The images were downscaled, and anomaly detection algorithms were used to identify regions of interest for segmentation and classification at full resolution. To address computational constraints, state-of-the-art anomaly detection algorithms were evaluated and an appropriate downscaling factor for the on-edge implementation was determined. The EfficientAD algorithm achieved promising results, detecting anomalies within an inference time of less than 10 seconds. The presented framework enables anomaly detection with a maximum delay of one layer. This lays the foundation for the future development of near-real-time intervention in the PBF-LB/M process.

Laser powder bed fusion of metals (PBF-LB/M) offers great potential for the production of new and spare parts for stationary gas turbines made of nickel superalloys such as Inconel 939 (IN939). In order to enable integration into existing assemblies and overcome design limitations, the additive manufacturing process chain must be expanded by suitable joining techniques. This study compares the electron beam welding of cast IN939 sheets and sheets produced additively using PBF-LB/M. The investigation focuses on the achievable seam quality with regard to geometric irregularities and internal defects in the form of liquation cracks on the microscale in the heat-affected zone. The evaluation of the welded samples shows no differences in the formation of the seam shape between the additively manufactured material and the cast material. For both materials, the highest quality category for beam-welded seams according to DIN EN ISO 13,919–1 was achieved at high welding speeds of 20 mm/s. Regardless of the manufacturing method, both materials show an increase in crack formation with increasing welding speed. However, due to its microstructure, the PBF-LB/M material exhibits significantly fewer microcracks overall. Final crack propagation tests on welded PBF-LB/M samples that were treated using HIP also show stable crack growth without sudden failure, which opens up potential for practical application.

High-throughput density functional theory (DFT) calculations have become a vital element of computational materials science, enabling materials screening, property database generation, and training of “universal” machine learning models. While several software frameworks have emerged to support these computational efforts, new developments such as machine learned force fields have increased demands for more flexible and programmable workflow solutions. This manuscript introduces atomate2, a comprehensive evolution of our original atomate framework, designed to address existing limitations in computational materials research infrastructure. Key features include the support for multiple electronic structure packages and interoperability between them, along with generalizable workflows that can be written in an abstract form irrespective of the DFT package or machine learning force field used within them. Our hope is that atomate2's improved usability and extensibility can reduce technical barriers for high-throughput research workflows and facilitate the rapid adoption of emerging methods in computational material science.

Parameter studies are a common step in selecting process parameters for powder bed fusion of metals with laser beam (PBF-LB/M). Density cubes manufactured with varied process parameters exhibit distinguishable surface structures visible to the human eye. Industrial visual in-situ monitoring systems for PBF-LB/M currently have limited resolution and are incapable of reliably capturing differences in the surface structures. For this work, a 65 MPixel high resolution monochrome camera is integrated in an industrial PBF-LB/M machine together with a high intensity led bar. Post-exposure images are taken to analyze differences in light reflection on the specimen’s surface. The grey level co-occurrence matrix is used to quantify the in-situ measured visual surface structure of nickel-based super alloy IN939 density cubes. The properties of the grey level co-occurrence matrix correlate to the energy input and resulting porosity of specimens. Low energy samples with lack of fusion flaws show an increased contrast in the grey level co-occurrence matrix compared to specimens with an optimal energy input. The potential of high-resolution images as reference data in in-situ process monitoring in PBF-LB/M is discussed.

The layerwise geometry build-up of additive manufacturing (AM) enables the possibility of in-situ process monitoring. The objective is the detection of irregularities during the build cycle, ensuring component quality and process stability. Focus of this work is the visual in-situ monitoring of the process of powder bed fusion with laser beam of metals (PBF-LB/M). Current state of the art visual monitoring systems for PBF-LB/M are limited by low resolution, allowing the detection of gross flaws. In this work a 65 Mpixel high-resolution monochrome camera is integrated into a commercial PBF-LB/M machine enabling a spatial resolution of approx. 17.2 µm/Pixel. The observed inhomogeneities are clustered into directly detectable irregularities, and indirectly detectable irregularities that can be inferred from the surface. In parallel, two different illumination techniques are realized in the process chamber and compared. The impact of the distinct illumination technique, direct light and dark field, on the identification of irregularities is evaluated.