Are Tornado Vortex Generators fit for purpose?

In recent years a number of Tornado Vortex Generators (TVGs) have been constructed and tested, with a view to providing facilities that can be used to determine wind loads on a variety of structures in tornado conditions. The scaling of TVGs has however proved to be contentious and different authors have taken different approaches. In this paper we address this issue and firstly present a formal dimensional analysis of the flow within full scale tornadoes and TGVs, which identifies a number of important dimensionless groups. We then consider a range of full-scale tornado data and, as far as possible, derive values of these dimensionless groups for each tornado. This analysis is then used to define the ranges of the dimensionless parameter for three tornado types (all of the two cell form) that can be used as simulation targets, rather unimaginatively naming them small, medium and large tornadoes. We then consider the performance of four medium to large TVGs in achieving these simulations. The analysis shows that the larger TVGs can achieve a range of geometrical similarities for the small and medium simulation targets, but none are able to achieve kinematic similarity, in that the ratio of circumferential to radial velocities are significantly lower than at full scale. Dynamic similarity (based on the Reynolds number) is of course not possible for physical models, but the analysis shows that in almost all cases the Reynolds number and model scales fall well below what would be considered acceptable in atmospheric boundary layer wind tunnels. We thus regretfully conclude that current TVGs are not wholly fit for purpose at the moment and are in need of significant modification if they are to be used to give reliable loading data. Alternatively it may be that the wind engineering community should consider different types of simulation to obtain the required information.