Quantitative Assessment of Additive Manufacturing Processes with Artificial Intelligence Integrated Ultrasound

Purpose and goal

The project aims to develop interpretable and reliable AI-based ultrasonic methods for quantitative non-destructive characterization (QNDE) of additively manufactured (AM) metal components. The main goal is to meet the stringent quality control requirements in critical sectors such as aerospace and nuclear power, where conventional NDE techniques fall short due to the complex and anisotropic microstructures of AM materials.

Expected effects and result

• Interpretable and reliable AI models that link ultrasound measurements to material properties and manufacturing process parameters have been implemented and tested in production-like environments at the industrial partners´ facilities. • A method for quantifying the effects of heat treatment on AM components for the nuclear industry has been validated for AM processes in insdustry. • The feasibility of using laser-ultrasound on as-built surfaces has been investigated.

Planned approach and implementation

The project is set up around two industrial use cases, one from the nuclear energy industry (Ringhals) and one from the aerospace industry (GKN). Both industrial partners have an interest in the methodologies developed for the other use case, but the respective efforts will be focused on either of the two.