Abstract: Based on high-accuracy computational fluid dynamics (CFD) methods and surrogate modeling optimization techniques, an optimization procedure for rotor aerodynamic shape design is established. Rotor aerodynamic performance is calculated by the CFD methods based upon Navier-Stokes/Euler equations and a Baldwin-Lomax (B-L) turbulence model is chosen to satisfy the flowfield characteristics, accuracy and efficiency requirements. Additionally, Roe-MUSCL scheme is employed to calculate the flux. In order to improve grid generating quality and facilitate the solution of the governing equations, the whole flowfield is divided into two zones, namely, a viscous zone around the rotor and an inviscous background zone. The blade grids are generated by a parameterized blade based on a 2D airfoil grid, and numerical examples demonstrate that the method can effectively improve grid generating quality and efficiency. By analyzing the flowfield of the reference rotor, design variables and constraints are selected to minimize the size of the optimization. To meet the needs of both optimization and mechanism investigation, an approach of optimization is established from permutated genetic algorithmic latin hypercube sampling (PermGA LHS) design and radial basis function (RBF). The Helishape 7A rotor case is first presented as validation of the numerical simulation method. Then, the optimization procedure is used to optimize rotor twist distribution. The results indicate that the aerodynamic performance of the optimized rotor in hover has been markedly improved.

Key words: rotor, aerodynamic optimization, aerodynamic performance, Navier-Stokes equations, overset grid, surrogate model