Bypass Transition Modelling: A New Method Which Accounts for Free-Stream Turbulence Intensity and Length Scale

Abstract

The publication of the present authors’ boundary layer transition data in 1992 (now widely known as the ERCOFTAC test case T3) has led to a spate of new experimental and modelling efforts aimed at improving our understanding of this problem. This paper describes a new method of determining boundary layer transition with zero mean pressure gradient. The approach examines the development of a laminar boundary layer to the start of transition, accounting for the influences of free-stream turbulence and test surface geometry. It is presented as a “proof of concept”, requiring a significant amount of work before it can be considered as a practically applicable model for transition prediction. The method is based upon one first put forward by G.I. Taylor in the 1930’s, and accounts for the action of local, instantaneous pressure gradients on the developing laminar boundary layer. These pressure gradients are related to the intensity and length scale of turbulence in the free-stream using Taylor’s simple isotropic model. The findings demonstrate the need to account for the separate influences of free-stream turbulence intensity and length scale when considering the transition process. Although the length scale has less of an effect than the intensity, its influence is, nevertheless, significant and must not be overlooked. This fact goes a long way towards explaining the large scatter to be found in simple correlations which involve only the turbulence intensity. Intriguingly, it is demonstrated that it is the free-stream turbulence at the leading edge of the test surface which is important, not that found locally outside the boundary layer. The additional influence of leading edge geometry is also shown to play a major role in fixing the point at which transition begins. It is suggested that the leading edge geometry will distort the incident turbulent eddies, modifying the effective “free-stream” turbulence properties. Consequently, it is shown that the scale of the eddies relative to the leading edge thickness is a further important parameter, and helps bring together a large number of test cases.