Robust reliability-based aerodynamic shape optimization

Abstract

This paper presents a methodology for robust and reliable shape optimization of aerodynamic bodies under uncertainties. The mean value and standard deviation of the drag coefficient objective function with respect to flow related uncertain parameters are computed using sparse grid quadrature rules. On the other hand the lift coefficient inequality constraint is dealt with a probabilistic manner based on the reliability index approach by not allowing the probability of unacceptable performance to be larger than a predefined value. The sensitivity derivatives of the two statistical moments and the probability of unacceptable performance with respect to the shape controlling design parameters are computed using the adjoint approach and an augmented Lagrangian method is used to obtain the robust and reliable optimal shape. The methodology is applied to pure aerodynamic shape optimization, comparing the robust solutions with the single-point optimum. © 2015 Taylor & Francis Group, London.