Computation of a confined compressible mixing layer

Fundamental mechanisms of supersonic mixing layers are being investigated in both computational and experimental efforts at Penn State. This paper describes the direct simulation of a particular set of experimental conditions. A low pressure, moderate Reynolds number supersonic shear layer with a high speed stream Mach number of 3.0 and a low speed stream Mach number of 1.2 is examined. The supersonic shear layer is simulated using dense grids to provide a fine-grain resolution of the mixing layer. Grid densities are chosen to resolve the fundamental instability mode of the shear layer. Massively parallel computer technology (the CM-200 and CM-5) provide the computing power needed. The MacCormack 2-4 scheme developed by Gottlieb and Turkel is used to achieve the numerical accuracy needed to resolve the shear layer spatial and temporal features. The MacCormack 2-4 method is second-order accurate in time and fourth-order accurate in space. The numerical method is supplemented by the use of artificial viscosity. A variation of the Jameson scheme is used. The dynamic behavior of the shear layer is computed using the non-linear Euler equations.