Progress on the Calculation of Large-Scale Separation at High Reynolds Numbers

The computation of high Reynolds number laminar separation has been one of the central issues in fluid mechanics over the past two decades. Laminar separation has eluded a description based on Prandtl’s boundary layer theory primarily due to the presence of the Goldstein [10J singularity at separation (see also Stewartson [28J). However, this state of affairs changed abruptly with the development of the tripledeck theory by Stewartson & Williams [27] and Neiland [14] for supersonic flows. Drawing upon Lighthill’s -[12] earlier work with viscous sublayers, Stewartson & Williams [27] were able to show that a boundary layer in a supersonic flow could “spontaneously” separate by setting up a local interaction between a viscous sublayer lying within the depths of the boundary layer and the local inviscid flow just outside the boundary layer. The work of Neiland [14] and Stewartson & Williams [27] introduced a very new concept to viscous flow theory, that of a freeinteraction. Prior to this time it had generally been thought that separation occurs due to a gradual reaction of the boundary layer to an externally imposed adverse pressure gradient.