Aerospace Simulation Shifts Toward AI Models

⊲ Aerospace America ⊳ At AIAA SciTech Forum 2026 in Orlando, panellists from NVIDIA, Neural Concept and Leonardo Helicopters discussed how machine learning surrogates now speed up aerospace design and manufacturing. Leonardo demonstrated a geometric deep learning model predicting ground noise 540 times faster than standard CFD. The panelists emphasised that success requires customised, problem-specific models rather than generic solutions and that AI functions as an accelerator complementing physics rather than replacing it. Operational applications face regulatory obstacles with EU certification guidelines anticipated in 2027.
⊲ Calculations such as cockpit fogging fixes that once waited on overnight simulations can now be tested and rejected within the same design session. ⊳
⊲ Image – ⊳
www.simulation.news