Ensuring process repeatability and consistent part certification remains one of the key barriers to scaling metal additive manufacturing (AM) for regulated industries such as aerospace and medical technology. Current inspection and qualification methods—X-ray computed tomography (CT), ultrasonic scanning, and destructive testing—can increase production costs by up to 50 % and extend lead times by months. In this joint research effort between Gravity Pull Systems and Amiquam, we investigate how in-situ monitoring and AI-assisted analytics can fundamentally change the economics and efficiency of AM part certification. The project combines eddy current sensing for compliant real-time detection of surface and near surface discontinuities and inhomogeneities with intelligent toolpath adaptation using the PAAM optimization framework. This integration enables real-time, automated detection of surface and near surface defects during production, eliminating the need for slow and costly post-process X-ray computation tomography. Non-destructive, continuous assessment of process quality during laser powder bed fusion (LPBF) effectively bridges the gap between process monitoring and part inspection according to established standards. Our approach aims not only to detect defects as they arise but also to proactively adjust process parameters and scan strategies to mitigate them. The joint solution is being validated on representative geometries and materials used in aerospace-grade LPBF processes. Early results demonstrate a clear potential to reduce post-process inspection requirements while maintaining or improving part reliability. Embedding traceable monitoring data into the digital manufacturing record further supports the concept of a digital certification workflow, aligning with emerging standards for data-centric qualification - potentially eliminating the dependency on post-process CT and destructive testing. This presentation will detail the sensing methodology, AI analytics pipeline, integration with toolpath logic, and implications for industrial qualification frameworks. The overarching aim is to contribute to a more efficient, data-driven, certifiable AM ecosystem—open to collaboration with interested research and industrial partners.
Learning Objectives:
Understand the advantages of a hybrid eddy-current + AI toolpath system for real-time defect detection and adaptive correction in LPBF.
Evaluate the potential to replace or minimize post-process CT and destructive testing through data-driven quality assurance.
Outline a pathway toward digital certification, improving traceability, cost efficiency, and qualification readiness for critical AM parts.
