Uncertainty Analysis of Transonic Aerodynamics for Wing-Mounted Aircraft

Abstract

Since random uncertainty may cause severe aerodynamic performance fluctuations for the wing-mounted aircraft, the Gaussian process regression (GPR) surrogate model method based on was proposed. The strategy of adding sample points by active-learning method can effectively reduce model uncertainty and improve the accuracy of uncertainty prediction. Focusing on incoming flow with uncertainty, the polynomial chaos expansion (PCE) method based on Smolyak sparse grid and the GPR surrogate model method based on the strategy of adding sample points by active-learning method were used to analyze the uncertainty of the wing-body-nacelle-pylon geometry combined with the Sobol sensitivity analysis method. Results show that the uncertainty of angle of attack and Mach number will cause dramatic fluctuation in lift coefficient and drag coefficient of the wing-mounted aircraft under transonic condition. The fluctuation of lift coefficient is affected by angle of attack and Mach number, and the fluctuation of drag coefficient is mainly determined by Mach number.

Keywords

• uncertainty analysis
• gaussian process regression(gpr)
• polynomial chaos expansion (pce)
• sensitivity analysis
• aerodynamic performance