RESEARCH NEWS
Machine Learning for Atmospheric Re-Entry Optimization Drs. Riccardo Bevilacqua and Emanuela Gaglio Atmospheric re-entry presents complex aerodynamic and thermal challenges that demand rapid and robust design methods. Drs. Bevilacqua and Gaglio apply machine learning to develop optimal control profiles for targeted re-entry trajectories. These learning-based approaches identify control strategies that simplify downstream vehicle design, particularly in areas related to guidance, control and aerothermal load management. In collaboration with CFAL, the team is integrating improved aerodynamic models derived from CFD- informed machine learning methods, advancing both performance and computational efficiency in re-entry system design.
Experimental Aerodynamics and Data-Driven Flow Reconstruction Dr. Ebenezer Gnanamanickam The Experimental Aerodynamics Group, which operates the wind tunnel facility in MicaPlex, applies machine learning techniques based on Proper Orthogonal Decomposition (POD) to enhance aerodynamic measurement capabilities. Current research focuses on characterizing airflow behind naval vessels to better understand unsteady wind conditions that affect helicopter takeoff and landing operations. While modern experiments provide high- resolution flow data, practical constraints in instrumentation and facility configuration can limit measurable velocity components. Using POD-based machine learning methods, the team reconstructs difficult-to-measure flow quantities from available experimental data. Recent work published in Experiments in Fluids demonstrates how these techniques expand the functional capability of the wind tunnel, enabling more comprehensive ship-airwake analysis and broader aerodynamic investigations.
DEPARTMENT OF AEROSPACE ENGINEERING | 7
Made with FlippingBook - professional solution for displaying marketing and sales documents online