What do we need to substitute experiments with simulations in turbulence?

摘要

The cost and future prospects of direct numerical simulations are discussed in the context of their use in the development of turbulence theory. It is concluded that, even today, simulations are comparable to laboratory experiments in many aspects, including Reynolds number, and that they offer advantages with respect to experiments whenever both are available. The postprocessing of the data is also discussed, and it is noted that it often implies the analysis of individual flow fields whose storage requirement greatly exceed the central memory of the postprocessing computers. It is argued that, in spite of those problems, simulation of turbulent flows at 'asymptotically large' Reynolds numbers will probably become available within the next decade.