A framework to compare wind loads on low-rise buildings in tornadoes and atmospheric boundary layers

Abstract

This paper presents a framework to compare the building aerodynamics measured in a tornado-like vortex simulator and a typical boundary layer wind tunnel. The method relies on the hypothesis that the aerodynamic and static pressures can be separated and on simultaneous wind and pressure measurements. Test data from a tornado-vortex simulator and a boundary layer wind tunnel are presented to illustrate the method. The simulated tornado had a swirl ratio such that the core had multiple vortices. A building model was placed to the side of the translation path, at a position 0.8 core radius diameters from the tornado center. Surface pressures were measured on both a circular ground plate and the building model surfaces, synchronized with the velocities measured at four cobra probes placed near each corner of the building. A large set of nominally identical runs were used to obtain ensemble statistics of both pressures and velocities. The static pressure induced by the tornado was removed from the overall building pressures using a conditional average of the ground plane tornado pressure field. A quasi-steady (QS) vector model was developed in boundary layer results for the same building and then used to compare the aerodynamic results in the tornado, based on the velocities measured by the probes near the building in the tornado wind field. The QS model is found to predict the order of pressure variations but misses the local pressure distributions. The estimations are greatly improved when the upstream wind azimuth angles are shifted in the QS model to the direction of tornado rotation. The degree of the required azimuth shift for better predictions are shown to be governed by the curvature of the tornado wind field in the horizontal plane near the building. Furthermore, this curvature of the tornado wind field is observed to reduce the size of separation bubbles on the leeward walls, increasing the magnitude of suctions on these walls. The rotating multi-cell vortices confined in the tornado core also produce high pressure fluctuations that are not well captured by the QS model.