Aerodynamic sensitivity coefficients using the three-dimensional full potential equation

Abstract

The quasianalytical (QA) approach is applied to the three-dimensional full potential equation to compute wing aerodynamic sensitivity coefficients in the transonic regime. Symbolic manipulation is used and is crucial in reducing the effort associated with obtaining sensitivity equations, and the large sensitivity system is solved using sparse solver routines such as the iterative conjugate gradient method. The results obtained are almost identical to those obtained by the finite difference (FD) approach and indicate that obtaining the sensitivity derivatives using the QA approach is more efficient than computing the derivatives by the FD method, especially as the number of design variables increases. It is concluded that the QA method is an efficient and accurate approach for obtaining transonic aerodynamic sensitivity coefficients in three dimensions.

Preconditioned domain decomposition scheme for three-dimensional aerodynamic sensitivity analysis

“…When a three-dimensional sensitivity analysis is performed on a single grid, however, even the iterative methods may become impractical and sometimes inapplicable because of either a high computer memory requirement or a slow convergence rate. 4 In an attempt to alleviate such hindrances, the sensitivity analysis with domain decomposition (SADD) scheme was developed. 5 ' 6 This scheme divides the computational domain into smaller and nonoverlapping subdomains (multiblock grids) that are solved separately.…”

Multidisciplinary design optimization techniques - Implications and opportunities for fluid dynamics research

“…At a 1986 conference Sobieski 4 challenged the aerodynamics community to develop a general sensitivity analysis capability, and the first papers on exact aerodynamic sensitivity analysis appeared a few years afterwards. [5][6][7][8] Of course, there had been earlier related developments. Sensitivity analysis of a limited sort was implicit in the aerodynamic optimization methods utilizing adjoint equations 9,10 that originated in the late 1970s.…”