An experimental investigation of the behavior of a turbulent boundary layer subjected to adverse and favorable pressure gradients was conducted at Mach number 4 for a free-stream Reynolds number of 0.500,000 per inch. Two severe pressure gradients were imposed on the boundary layer by interchangeable contoured centerbodies inside a large hollow cylinder for cold-wall and adiabatic-wall temperature conditions. Imposition of either of the adverse pressure gradients significantly decreased the natural growth rate of the boundary-layer displacement thickness, whereas the favorable pressure gradient had opposite effects; momentum thickness was relatively unaffected by pressure gradient. A pressure gradient increase of about 30 percent caused relatively small changes in the skin friction, heat-transfer rate, and the characteristic boundary-layer parameters. Wall cooling effects (T sub w/t sub 0 approximately 0.3) on the boundary-layer thickness parameters were nearly insignificant, in comparison with the adiabatic-wall results. Heat-transfer distributions were similar to the local skin friction results based on free-stream conditions. (Author).

The problem of predicting the behavior of the incompressible turbulent boundary layer in an adverse pressure gradient is re-examined. An outline of the problem is given along with a brief summary of the work that has already been done, including both experimental investigation are presented for a separating turbulent boundary layer with various pressure distributions. An approximate theory is developed in which the momentum integral equation is satisfied for each half of the boundary layer. The velocity profiles used in the analysis consist of the well known wall and wake regions, resulting in a two-parameter family with the Reynolds number as one parameter. It is assumed, with some experimental justification, that the eddy viscosity can be reasonably approximated from zero pressure gradient experimets. The numerical calculations, using the Runge-Kutta procedure, show good agreement with the experiments. The reliability that can be expected of such approximate methods is discussed. (Author).