Unsteady Blade Shape Optimization for Rotorcraft

Design Optimization for rotorcraft blades is challenging given that inherently unsteady phenomena and varying freestream conditions affect the aerodynamic performance of these blades. Consequently, steady design optimization methods may be of limited use in carrying out design studies for rotorcraft blades. On the other hand, the process of unsteady design optimization is handicapped by the cost of computing the unsteady aerodynamic objective function. The authors have developed a novel unsteady optimization approach that combines Computational Fluid Dynamics (CFD), a modified Proper Orthogonal Decomposition Procedure (POD) and an Artificial Neural Network (ANN) to evaluate the unsteady objective function with the accuracy of a time-spectral method but at a fraction of the cost. Subsequently, the unsteady optimization procedure has been used to carry out design optimization studies of pitching airfoils with dynamic stall. Traditionally these studies have been performed under constant freestream condition much like the manner in which experimental studies are undertaken to gauge performance of airfoil sections. In this paper, an unsteady optimization study of airfoil sections has been carried out for alleviating dynamic stall with varying freestream Mach number thereby incorporating the realistic effect of a rotating blade in forward flight. Knowledge extraction or the identification of patterns/rules from common properties of airfoil designs and their performance has been carried out with the help of Self-Organizing Maps (SOMs). Consequently, design rules can be identified that help in generating rotorcraft airfoil shapes that show superior airfoil performance including the delayed onset of dynamic stall.