I cannot let Professor Baker's letter on the use of computational fluid dynamics for wind loading (NCE 23 November) pass without comment.
It is certainly true that making accurate CFD predictions in any area of application is a non-trivial process and much remains to be learned. Nonetheless CFD does have a useful role to play in the design of effective structures, complementing rather than replacing physical simulation methods.
It is instructive to look at the aerospace industry, where CFD is now essential to the design of modern aircraft, even though it still cannot be trusted to give accurate absolute drag predictions. What CFD has done, however, is to change and limit the use of expensive physical testing by allowing early comparison of possible designs to eliminate the least promising and avoid costly mistakes.
It is in this way that highly efficient wings are now designed that would have been far too expensive or even impossible to design using wind tunnel testing alone. Such a philosophy is already starting to be seen in the design of longspan suspension bridges.
It is also important to recall that wind tunnel testing is itself a simulation method and as such is prone to problems especially in relation to scaling (particularly problematic for smooth shapes) and consistent definition of the atmospheric boundary layer.
Nicholas Waterson, senior engineer, simulation group, Mott MacDonald, NPW@mm-croy.mottmac.com