Aerodynamic shape optimisation based on gradient methods requi-res the calculation of the sensitivities of aerodynamic coefficients. These sensitivities can be evaluated using finite differences, but at a cost which grows linearly with the number of design variable. Alternatively, these sensitivities may be calculated directly with the adjoint method, which offers a better accuracy for constant computational expenses, roughly equal to those needed for the CFD calculation. Recent developments made by DSNA and DAAP departments within the ?DTP Optimisation? project, supported by the DPAC and the SPAe, enable the use of the adjoint technique through the e/sA and ffd softwares. Three applications of the adjoint technique for shape optimisation, performed in the DAAP department are presented in this article.
展开▼
