Author(s)

Giuseppe G. Lobriglio, Antonio J. Álvarez, Felix Nieto, Santiago Hernández & José Á. Jurado

Abstract

Today, the assessment of the safety of long-span bridges relies on wind tunnel testing, although CFD methods are steadily penetrating in research and industrial practice. The evaluation of force coefficients and flutter derivatives presents multiple uncertainties, related with inflow boundary conditions, mechanical and mathematical models or parameter choices. In this work, we focus on one single uncertainty parameter that is the wind angle of incidence, which has been studied for instance in building aerodynamics. The assumed input probability density function adopted for the angle of incidence has been uniform in the range of angles considered. Uncertainty quantification tools, such as the stochastic collocation method, are used to propagate the uncertainty in the wind angle of attack for the force coefficients and flutter derivatives of a twin-box bridge deck. To this end, 5 2D URANS static simulations have been completed to quantify the uncertainty in the force coefficients, and 70 2D URANS forced oscillation simulations have been required to obtain the stochastic mean and standard deviation of the flutter derivatives, applying nested Clenshaw–Curtis quadrature points at level 3. It has been found that for the force coefficients, the stochastic standard deviation has been up to 0.032 for the lift coefficient. Furthermore, for the aeroelastic response, the flutter derivatives H1*, A1*, H2* and A2* show important stochastic standard deviations relative to the stochastic mean value for reduced velocities above 10.

Keywords

aerodynamic derivatives, flutter, force coefficients, stochastic collocation, twin-box deck, uncertainty quantification.