High performance simulation of ultrasound inspection of porosity in additive manufactured parts

Due to its versatility, additive manufacturing (AM) holds the potential to revolutionize the manufacturing industry. For aerospace applications however, varying part quality and the ability to manufacture complex AM components, creates new and substantial challenges for nondestructive evaluation (NDE) processes and methods. Inspecting for porosity is a crucial piece of the qualification of AM components and while X-ray computed tomography (CT) has proven to be effective at this task, CT has limits for large-scale application. Ultrasound testing (UT), on the other hand, is an attractive alternative to CT if it can be used to reliably characterize porosity in realistic scenarios. To advance this goal, a data-augmented approach to simulating UT inspections on fabricated titanium alloy (Ti-6Al-4V) fatigue specimens is developed by taking pore statistics that are derived from CT data to seed high-fidelity UT inspections under varying porosity conditions. UT attenuation metrics obtained from this series of simulations were compared to distributions of attenuation, measured from empirical UT scans. Simulated and empirical distributions were in reasonable agreement, serving as a strong indicator for the potential of high performance computing (HPC) NDE simulation for more complex AM components.