Aerodynamic shape optimization using preconditioned conjugate gradient methodsIn an effort to further improve upon the latest advancements made in aerodynamic shape optimization procedures, a systematic study is performed to examine several current solution methodologies as applied to various aspects of the optimization procedure. It is demonstrated that preconditioned conjugate gradient-like methodologies dramatically decrease the computational efforts required for such procedures. The design problem investigated is the shape optimization of the upper and lower surfaces of an initially symmetric (NACA-012) airfoil in inviscid transonic flow and at zero degree angle-of-attack. The complete surface shape is represented using a Bezier-Bernstein polynomial. The present optimization method then automatically obtains supercritical airfoil shapes over a variety of freestream Mach numbers. Furthermore, the best optimization strategy examined resulted in a factor of 8 decrease in computational time as well as a factor of 4 decrease in memory over the most efficient strategies in current use.
Burgreen, Greg W. (NASA Langley Research Center Hampton, VA, United States)
Baysal, Oktay (Old Dominion Univ. Norfolk, VA, United States)
August 16, 2013
January 1, 1993
Publication: In: AIAA Computational Fluid Dynamics Conference, 11th, Orlando, FL, July 6-9, 1993, Technical Papers. Pt. 1 (A93-44994 18-34)