05 - Propulsion
DEEP LEARNING BASED FAST PREDICTION OF AERODYNAMIC PARAMETERS FOR DUCTED PROPELLERS
L. Liu¹, L. Zeng¹, T. Wang¹, Z. Gao¹, X. Shao¹; ¹Zhejiang University, China
A deep learning-based aerodynamic prediction surrogate model for ducted propellers is developed, which combines different physical properties to improve prediction precision, and has a certain generalization ability and interpretability, solves the drawbacks of traditional surrogate models, and greatly shortens the global optimization design cycle of the aircraft.