03.1 - Aerodynamics – CFD Methods and Validation
EFFICIENT ROTOR AERODYNAMIC SHAPE DESIGN USING A MULTI-FIDELITY NEURAL NETWORK-BASED GLOBAL OPTIMIZATION METHOD
S. Li¹, W.P. Song¹, Z.Q. Shen¹; ¹ Northwestern Polytechnical University, China
This paper introduces a novel MFNN-based optimization method for helicopter rotor design, achieving a 150x speedup over traditional CFD while maintaining high accuracy (errors <1%). Its news value lies in demonstrating a breakthrough in balancing computational efficiency with precision, a challenge across many engineering fields like aerospace, automotive, and wind turbine design.
DOI: 10.71945/icas2026_0492