Search NASA⌕ Search

Engineering topics

Delphine Duquette

Publications and source records attributed to Delphine Duquette.

7 records
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

ASTM International Conference on Additive Manufacturing – 2022

The In-Situ Project: Correlating In-situ monitoring data and Non-Destructive Evaluation (NDE) methods to characterize defect populations in Laser Powder Bed Fusion (L-PBF) material NDE is critical to enable the safe use of additive manufactured (AM) parts in aerospace applications. However, traditional NDE methods are not fully adaptable to complex AM geometries to appropriately screen for critical defects, and L-PBF processes inherently generate defects. In-situ monitoring techniques can give information about the quality of the build process and indicate process variations that could potentially create defects. Proven, causal correlations between in-situ monitoring indications and resulting defects are needed to inform about the final quality of the part. Thus, the development of a new approach is necessary. NASA started a two-phase study that would investigate techniques for characterizing defect populations in L-PBF material, first in material with induced defects, then in a nominal L-PBF process, by correlating in-situ, NDE, and serial sectioning data. Objectives also include the evaluation of seeded defect methodologies for realistic defect creation, creating a baseline of defect populations in a Qualified Material Process (QMP), characterizing the effect of heat treatment, and correlating defect populations with tensile and fatigue properties. By comparing computed tomography images, in-situ monitoring images, and metallography images, the team has been able to draw preliminary conclusions. They include a better understanding of the limitation of the NDE tools, learning about the “healing” of the samples observed during layer building and a comparison study on different laser power variation samples. Those preliminary conclusions will help dictate the follow-up work that will be presented at the conference.

additive manufacturing↗
NASA NTRSDate not supplied

Convolutional Autoencoder for Defect Detection in Additive Manufacturing

The core idea behind using machine learning (ML) for defect detection is that it can be used to detect flaws as they are being formed in an AM part. As the part is being made, a near-infrared (NIR) sensor records each layer and creates an image of the entire build layer. These images, usually thousands, can be compiled into a ‘3D’ array of the entire part. ML tools, such as a convolutional autoencoder (CAE) can go through these images and highlight potential anomalous regions of your part.

In-Situ Monitoring↗
NASA NTRSDate not supplied

Three-Dimensional Reconstruction of Defects and Structures in Additively Manufactured Parts with Automated Serial Sectioning

Metal additive manufacturing (AM) processes have been demonstrated to be effective at reducing costs and lead times associated with complex components for space flight applications. Laser powderbed fusion (L-PBF) is a commonly used AM technology due to the ability to produce complex parts with fine feature resolution in a wide variety of alloys and applications. L-PBF, like many other manufacturing processes, can produce minor flaws in parts when in nominal operation as well as process-escape defects when process abnormalities occur. The effects of the flaws and methods of detecting the flaws are a subject of interest to understand the difficulties in detecting these flaws with current technology and how much risk the flaws or defects pose to potential flight parts. Using a RoboMet.3D automated serial sectioning system, seeded defects as well as minor process flaws can be imaged and reconstructed in three dimensions to compare to non-destructive evaluation (NDE) techniques, such as x-ray computed tomography (CT), neutron CT, and in-situ monitoring. The RoboMet automates the metallography process by automatically grinding, polishing, and imaging samples in a single system and providing the control data for NDE comparisons to know the real size of defects built into coupons. These comparisons provide an understanding behind the technological limitations of the NDE techniques for different alloys. The same serial sectioning methods have also been utilized to characterize the surfaces of parts to reconstruct the surfaces and take measurements of internal features not easily examined with non-destructive methods. Using the RoboMet, fine lattice structures built with L-PBF have been characterized to determine the actual thicknesses of struts and density of the lattice structures. These structures have been used as finer build supports for the L-PBF process, designs for fine catalysts, and other design considerations for small components. The RoboMet data helps to inform the modeling and design efforts around these fine components.

additive manufacturing↗
NASA NTRSDate not supplied

Nondestructive Evaluation and In-Situ Monitoring Efforts for Additive Manufacturing

Nondestructive evaluation (NDE) is required to determine the quality of additive manufactured (AM) parts due to the inherent variability of AM processes. In-situ monitoring technologies endeavor to characterize the process and part quality during production. Qualifying an in-situ monitoring technology requires a proven, causal correlation between indications in the monitoring data and flaws in the finished part.

James Mavo↗