1. Even though the problems associated with interacting flowshave been studied extensively. less attention has been devoted to the interaction in the merged layer and more rarefied flow regimes. Even so,a substantialamount ofexperimental effort has been made to clarify some of the issues associated with merged-layerflows. Thiswas particularlytrue during the 1960s where several experimentalstudies(Ref.2-9.for example)were made for flowover sharp,flat plates. One of the earlier numerical simulations For such flowsisdiscussed in Ref. 10. Also, Metcalf and Berry 5 studied experimentally the corner flow problem for supersonic, low-density flows. Recently. Chunl has focused on this problem by studying the merged-layerinteractions that occur for hypersonic nitrogen flow over two-dimensional models consisting of a flat plate followed by a compression ramp. The ramp comer islocated 71.4 mm from the plate leading edge. andramp anglesOF 15'. 25". and 35"were considered. Thetest conditions were produced in
2. a two-dimensional version of the direct simulation Monte Carlo (DSMC)method of Bird.12-14 Details concerning flowfield structure and surface quantitiesare provided. Results ol the calculations provide additional information that canbe used to interpret the experimental findings. Also,the experimental data can be used to assess the abilityof the DSMCmethod to calculate viscous interacting flows and in particular, the onset and extent ofthe separated flow region induced by the ramp. However.the comparisons canbe viewed as qualitativeat best. since the experimental measurements show the flow to be three-dimensional while the calculations assume the flow to be two-dimensional to present calculations detail the sensitivity of the flow structure and flow separation as a function of flow rarefaction. Furthermore. the sensitivity of the simulation to variation in numerical parameters is investigated. two hypersonic low-density facilities for Mach
3. The gasglow dischargetechnique isused to determine the location of shock waves and to provide qualitative information concerning the boundaly-layer thickness. The test model serves asa ground electrode,whilethe other electrode w consists of twometallic plates placed above and beneath the model in shallow side cavities of the test section. rhe gasglow discharge,suppliedby a short-wave generator, emits light with an intensity that isroughly proportional to the gas density, provided the field between the electrodes isuniform. Oil flowpatterns were used to detect the flow separation ahead of the ramp. This was achieved by first coating the models with approximately a 1-mn-thick layer of fine titanium dioxidepowder dispersed in silicon oil. Once the flow conditions are established, then the coated model isinjected rapidly (0.5s)intothe test sectionand exposed to the test environment. A video tape recorder isused to record the events from the time'of model injection until the end of the oil layer movement,which requires approximately 0.5to 1.0minutes. Finally. photographs are taken from above while looking perpendicular to the horizontal flat plate. Additional details concerning the experiments and data reduction procedures are given in Ref. 11. m t a t i o n a lMethodandProcedurs
4. The direct simulation Monte Carlo (DSMC)methodcode used inthe current study is the general two-dimensional/auisymmetrkcode of Bird.12-14 Themethod and requirements for application have been presented in previous publications (Ref. 14for example)and will not be repeated here.