Hypersonic vehicles experience extreme aerodynamic forces and intense heating. Knowing those forces in real time is important for national security because it supports safer, more reliable flight control, helps prevent structural failure, and improves the ability to assess vehicle health during missions.
A team of researchers from Sandia and The University of Texas at Austin, with principal investigator Professor Noel Clemens of UT’s Department of Aerospace Engineering and Engineering Mechanics, improved a sensing method that estimates aerodynamic pressure loads using only a small number of strain sensors embedded in the vehicle’s structure. This allows the vehicle to “feel” how it’s bending and stretching, and that information can be used to infer the air pressures causing those loads.
While aerodynamic pressure loads can be estimated in advance using heavy computation, hypersonic heating complicates the picture. Temperature changes affect material behavior and structural response in a nonlinear way, which means precomputed system models no longer work.
To overcome this, the teams from Sandia and UT Austin created a neural-operator technique called the Neural Matrix Operator (NEMO), in work led by UT graduate student Julie Pham. NEMO is designed to respect the underlying physics equations while learning how the system changes with temperature. This allows the method to keep a fast, closed-form style inverse solution for the aerodynamic pressure loads that can be computed quickly onboard, even when temperature effects matter. This new method was demonstrated in numerical experiments on the Initial Concept 3.X hypersonic vehicle.
Results show NEMO can accurately represent the temperature-dependent behavior and produce accurate aerodynamic load estimates. However, additional work is needed to further reduce errors in the thermal-strain correction.
Sandia researchers linked to work
- Patrick Joseph Blonigan
- Omar Ghattas
Sponsored by
The Air Force Office of Scientific Research; The DOE’s Office of Advanced Scientific Computing Research; The National Science Foundation Graduate Research Fellowship
Associated Publications
Pham, J., Blonigan, P.J., O’Leary Roseberry, T., Ghattas, Omar, and Willcox, K. E. “Neural Operator-Enabled Aerodynamic Load Estimation for Hypersonics,”AIAA 2026-1210. AIAA SCITECH 2026 Forum. January 2026. https://doi.org/10.2514/6.2026-1210
September 21, 2026