Surrogate-based robust design optimization of airfoil using inexpensive Monte Carlo method

This paper presents an efficient robust design optimization under the uncertainty quantification (UQ) of aleatory flight conditions using inexpensive Monte Carlo (IMC) method. Robust design optimization (RDO) technique is used to reduce the statistical moments of drag coefficient of an airfoil subject to some aerodynamic and geometric constraints. IMC approach with kriging surrogate model has been utilized to evaluate the robustness of a candidate design for prediction of statistical parameters and probability density functions of output response quantities. CFD calculations of RAE2822 airfoil have been evaluated using in-house code, PMNS2D. Combination of CFD input flight parameters such as free stream Mach number and angle of attack are varied and the propagation of their corresponding aleatory uncertainties on output properties of interest is studied. The reduction in objective function (Cd) is about 37.48% and 39.19% for combination of Mach and alpha uncertainties under Gaussian and LHS distributions respectively. Probability density function (PDF) of an output uncertainty is also compared with real PDF and found in good agreement with each other.