Enabling Real-Time Flaw Detection in Laser Powder Bed Fusion Using In-Situ Infrared Monitoring
Abstract
Laser powder bed fusion (L-PBF) additive manufacturing (AM) is increasingly being utilized by the Navy and manufacturing industry to manufacture parts. A key issue is that porous defects and flaws can be generated due to local part geometry variation, pre-existing voids in the feedstock, spattering, recoater blade interference during powder spreading process. To address this issue more efficiently and economically, this project aims to enable real-time flaw detection in L-PBF using in-situ infrared (IR) monitoring. To achieve this objective, the following technical approaches are proposed: 1) Developing additional IR image processing algorithms to identify key thermal signatures for defect detection. 2) Developing ML model to correlate various thermal signatures from the IR data with defect characteristics on the test artifact. 3) Developing hardware accelerators for IR image processing and ML inference on AMD/Xilinx Versal Adaptive Compute Acceleration Platforms (ACAP). 4) Developing an efficient, scalable and sustainable cyberinfrastructure for real-time end-to-end intelligent L-PBF build quality monitoring. This project addresses one of the top priorities for the quality assurance (QA) of L-PBF processed Naval components and reduced use of post-build non-destructive evaluation (NDE). Current QA method for L-PBF processed Naval components requires extensive ex-situ NDE to detect porous defects and flaws, which is expensive and timeconsuming. The proposed real- time in-situ defect detection method will significantly reduce the number of exsitu NDE tests needed to inspect L-PBF components, thereby reducing inspection time and costs by 1-2 orders of magnitude. In longer term, the proposed real-time technology will potentially enable “detect and repair” and save numerous defected components from becoming waste, resulting in substantial savings in material and manufacturing costs. All in all, the proposed technology is expected to significantly enhance the Navy’s ability to produce legacy and replacement parts in small quantity and fabricate high-valued complex parts reliably at much lower costs.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 12, 2025
- Source ID
- N001742310011
Entities
People
- Albert To
Organizations
- United States Navy
- University of Pittsburgh