Two Reynolds-averaged Navier‐Stokes solvers, CFL3D and WIND, are applied to the subsonic turbulent jet flowfield originating from a six-lobed nozzle, with emphasis placed on turbulence modeling effects. The turbulence models investigated include linear one-equation and two-equation models and nonlinear two-equation explicit algebraic stress model (EASM) formulations. Two nozzle operating points are investigated, corresponding to exit Mach numbers of 0.30 and 0.94. Comparisons of calculated mean axial velocities and turbulence intensities are made with experimental data. All of the turbulence models were deficient in predicting the initial mixing rate exhibited by the experimental data. The one-equation model provided the best agreement with experimental data in the near field of the jet. The linear two-equation models and a modified EASM provided better agreement with data in the farfield. The Mach 0.30 k‐ω EASM calculation required a time-accurate calculation because of significant unsteadiness in the initial jet region, which is believed to be characteristic of the nozzle flow under consideration.

A multi-site team has completed an investigation of modeling high speed mixing layers using the computational methods currently being applied to predict HSCT nozzle flow fields. The objectives of this investigation were to: (1) calibrate the codes used by the various team members against benchmark experimental data, and (2) assess the accuracy of the Navier-Stokes codes in calculating turbulent flows having flow characteristics similar to those of HSCT engine nozzles. Two flow geometries were investigated using the five codes of NASTAR, PAB3D, GIF3D, NASTD, and NPARC. The first was the heated supersonic round jet of Seiner. For this configuration, with a jet exit Mach number similar to that of the primary flow from mixer chutes, three nozzle flow temperatures were investigated with the five codes. Using the same grid, boundary conditions, and k-e turbulence model (in the mixing region), very similar results were obtained for all codes, but the solutions did not agree well with the experimental velocity and temperature profiles. Further calculations using different turbulence models, compressibility corrections, and axisymmetric dissipation corrections improved the agreement with experimental data, but the corrections are not universally applicable. The second configuration was the two-dimensional supersonic mixing layer of Goebel & Dutton. For the flow case examined, with two supersonic streams, the five codes again produced very similar results using the same grid, boundary conditions, and turbulence model. The agreement with experimental data was better than for the Seiner round nozzle.

The leakage performance of a brush seal with gaseous working fluids at static and low rotor speed conditions was investigated. This report includes the leakage results for air, helium, and carbon dioxide at several bristle/rotor interferences. In addition, the effects of packing a lubricant into the bristles and also of reversing the pressure drop across the seal were investigated. Results were compared to that of an annular seal at similar operating conditions. In order to generalize the results, they were correlated using corresponding state theory. The brush seal tested had a bore diameter of 3. 792 cm ( 1. 4930 in. ), a fence height of 0. 0635 cm (0. 025 in. ), and 1800 bristles/cm-circumference (4500 bristles/in. -circumference). Various bristle/rotor radial interferences were achieved by using a tapered rotor. The brush seal reduced the leakage in comparison with the annular seal, up to 9.5 times. Reversing the pressure drop across the brush seal produced leakage rates approximately the same as that of the annular seal. Addition of a lubricant reduced the leakage by 2.5 times when compared to a nonlubricated brush seal. The air and carbon dioxide data were successfully correlated using the corresponding state theory. However, the helium data followed a different curve from the air and carbon dioxide data.