| An improved algorithm is presented for performing lift-constrained aerodynamic optimization. The angle of attack is set in the flow solution stage, ensuring the desired lift coefficient is achieved. Changes to the flow solver are carried through to the adjoint calculation of the gradient. For this project, the drag coefficient is the objective function, but the algorithm could easily be extended to other lift-constrained optimization problems. Work on normalizing objective functions for multi-point optimization ensures that the weight given to each design point is independent of the initial objective function value, and that the thickness constraints have the same meaning for all design points.; Several design cases are examined. One case is of particular interest, a multipoint optimization with eighteen design points, which is compared to optimized airfoils for each individual design point, to examine the benefits of morphing airfoils for a realistic design case. |
