I was interested to read of the use of computational fluid dynamics to design the Quicksilver boat that is to attempt to beat the water speed record (NCE 2 November). This seems to me an entirely appropriate way to design such a streamlined vehicle.
However I was disturbed by what I hope was a throwaway remark that such methods can also be used to determine wind loads on buildings.
Most current CFD methods are only able to predict average wind speeds and loads, and are not able to predict adequately surface loads in regions of highly separated flow. Since wind load calculations on buildings require extreme fluctuating wind loads, and usually have significant regions of separated flow, the use of conventional CFD methods is likely to produce erroneous answers that need not necessarily be conservative.
The development of CFD methods that can perform well in such circumstances is still in its infancy, and conventional physical modelling tests in wind tunnels are, at the present time, still the only way to obtain reliable wind load information of non-standard structures where codes are not applicable.
Professor CJ Baker (F), School of Civil Engineering, University of Birmingham