Aeroelastic stability of a twin-box deck: Comparison of different procedures to assess the effect of geometric details

Focus of this paper is the comparison of three different approaches to assess the aerodynamic stability of twin-box decks: experimental free-motion tests, numerical simulations using quasi-steady theory, and numerical simulations using flutter derivatives. To this end, these approaches are used to compare three deck configurations: a nominal twin-box deck configuration, a variation that includes a different walkway porosity, and a variation that consists in a different shape of the twin-box internal corner. Even though these configurations are geometrically very similar, their aeroelastic behaviour is quite different in terms of critical flutter speed, eigenvalue/eigenvector trend and static deformation as a function of wind speed. It is shown that free-motion tests provide the reference aeroelastic solution for each configuration, but do not provide any insight on the reason of the differences; simulations based on aerodynamic flutter derivative coefficients provide accurate results and insights, but only if unsteady coefficients are measured over all the mean angles of attack of interest; quasi-steady simulations, based on static coefficients slopes, can define quite precisely the aeroelastic behaviour in the nearby of flutter, provided that torsional aeroelastic damping is measured experimentally. Each method has therefore advantages and disadvantages that are highlighted and discussed throughout the paper.