A numerical analysis was conducted to study the effects of geometrically scaling scramjet inlet-combustor isolators. Three-dimensional fully viscous numerical simulation of the flow inside constant area rectangular ducts, with a downstream back pressure condition, was analyzed using the SolidWorks Flow Simulation software. The baseline, or 1X, isolator configuration has a 1” x 2.67” cross section and 20” length. This baseline configuration was scaled up based on the 1X configuration mass flow to 10X and 100X configurations, with ten and one hundred times the mass flow rate, respectively. The isolator aspect ratio of 2.67 was held constant for all configurations. To provide for code validation, the Flow Simulation program was first used to analyze a converging-diverging channel and a wind tunnel nozzle. The channel case was compared with analytical theory and showed good agreement. The nozzle case was compared with AFRL experimental data and showed good agreement with the entrance and exit conditions (Pi0= 40 psia, Ti0= 530ºR, Pe= 18.86 psia, Te= 456ºR, respectively). While the boundary layer thickness remained constant, the boundary layer thickness with respect to the isolator height decreased as the scale increased. For all the isolator simulations, a shock train was expected to form inside the duct. However, the flow simulation failed to generate this flow pattern, due to improper sizing of the isolator and combustor for a 3-D model or having a low pressure ratio of 2.38. Instead, a single normal shock wave was established at the same relative location within the length of each duct, approximately 80% of the duct length from the isolator entrance. The shape of the shock changed as the scale increased from a normal shock wave, to a bifurcated shock wave, and to a normal shock train, respectively for the 1X, 10X, and 100X models.

A computer model of a blow-down free-jet hypersonic propulsion test facility exists to validate facility control systems as well as predict problems with facility operation. One weakness in this computer model is the modeling of an air ejector diffuser system. Two examples of facilities that could use this ejector diffuser model are NASA Langley Research Center's 8-ft High Temp. Tunnel (HTT) and the Aero-Propulsion Test Unit (APTU) located at Arnold Engineering Development Center. Modeling an air ejector diffuser system for a hypersonic propulsion test facility includes modeling three coupled systems. These are the ejector system, the primary free-jet nozzle that entrains secondary airflow from the test cell, and the test article. Both of these facilities are capable of testing scramjets/ramjets at high Mach numbers. Compared with computer simulation data, experimental test cell pressure data do not agree due to the current modeling technique used. An improved computer model was derived that incorporates new techniques for modeling the ejector diffuser. This includes real gas effects at the ejector nozzles, flow constriction due to free-jet nozzle and ejector plumes, test article effects, and a correction factor of the normal shock pressure ratio in a supersonic diffuser. A method was developed to account for the drag and thrust terms of the test article by assuming a blockage factor and using a drag coefficient*Area term for both the test article and thrust stand derived from experimental data. An ideal ramjet model was also incorporated to account for the gross thrust of the test article on the system. The new ejector diffuser model developed improved the accuracy and fidelity of the facility model as compared with experimental test data while only negligibly affecting computational speed. Comparisons of the model data with experimental test data showed a close match for test cell pressure (within 1 percent for final test cell pressure). The model accurately simulated both the unstarted and started modes of ejector flow, in which test cell pressure increases with nozzle total pressure once in started mode.

Accurate measurement of turbine engine compressor inlet total temperature is paramount for controlling engine speed and pressure ratio. Various methods exist for measuring compressor inlet total temperature on turbojet engines with hydromechanical control. One method involves the use of an ejector-diffuser system (eductor) to pull air from the engine inlet in order to measure the incoming total temperature. Analysis of historical test data has revealed that the inlet temperature measurement can be biased at certain flight conditions causing engine mis-scheduling and off-nominal engine operation. This bias is characterized primarily by adverse heat transfer effects and secondly by poor flow quality in the eductor tubing. Alternate eductor system configurations have been proposed to mitigate temperature bias. A one-dimensional engineering model of the eductor system was developed to facilitate the analysis of baseline and alternate eductor configurations. The model is calibrated with results from Computational Fluid Dynamics and validated with ground test data. The validated model is used to quantify the performance of several eductor configurations throughout the range of expected operating conditions and to quantify the amount of compressor inlet temperature measurement bias mitigation each configuration provides.

Cross flow heat exchangers are a fairly common apparatus employed throughout many industrial processes. For these types of systems, correlations have been extensively developed. However, there have been no correlations done for very high enthalpy flows as produced by Arnold Engineering Development Center’s (AEDC) H2 facility. The H2 facility uses a direct current electric arc to heat air which is then expanded through a converging-diverging nozzle to impart a supersonic velocity to the air. This high enthalpy, high temperature air must be cooled downstream by the use of a cross flow heat exchanger. It is of interest to evaluate the actual performance of the air cooler to determine the effectiveness of possible facility upgrades. In order to characterize cooler effectiveness, a numerical model is built to calculate per-tube-row energy balances using real (temperature and pressure dependent) air and water properties and cross-flow Nusselt number calculations.

Various methods have been used, including airborne radars, LIDAR, observation of flying birds, towers, tethered balloons, and aircraft to gain both a qualitative and quantitative representation of how heat and moisture are transported to higher altitudes and grow the boundary or mixing layer by thermal updrafts. This paper builds upon that research using an instrumented glider to determine the structure and build a mathematical model of thermals in a desert environment. During these flights, it was discovered that the traditional view of a thermal as a singular rising plume of air did not sufficiently explain what was being observed, but rather another phenomenon was occurring. This paper puts forth the argument and a mathematical model to show that thermals actually take the form of a hexagonal convection cell at higher levels in the convective boundary layer when the thermal acts as if unrestrained by borders as in non-linear cases of free convection.

A numerical analysis was conducted to study the effects of geometrically scaling scramjet inlet-combustor isolators. Three-dimensional fully viscous numerical simulation of the flow inside constant area rectangular ducts, with a downstream back pressure condition, was analyzed using the SolidWorks Flow Simulation software. The baseline, or 1X, isolator configuration has a 1” x 2.67” cross section and 20” length. This baseline configuration was scaled up based on the 1X configuration mass flow to 10X and 100X configurations, with ten and one hundred times the mass flow rate, respectively. The isolator aspect ratio of 2.67 was held constant for all configurations. To provide for code validation, the Flow Simulation program was first used to analyze a converging-diverging channel and a wind tunnel nozzle. The channel case was compared with analytical theory and showed good agreement. The nozzle case was compared with AFRL experimental data and showed good agreement with the entrance and exit conditions (Pi0= 40 psia, Ti0= 530ºR, Pe= 18.86 psia, Te= 456ºR, respectively). While the boundary layer thickness remained constant, the boundary layer thickness with respect to the isolator height decreased as the scale increased. For all the isolator simulations, a shock train was expected to form inside the duct. However, the flow simulation failed to generate this flow pattern, due to improper sizing of the isolator and combustor for a 3-D model or having a low pressure ratio of 2.38. Instead, a single normal shock wave was established at the same relative location within the length of each duct, approximately 80% of the duct length from the isolator entrance. The shape of the shock changed as the scale increased from a normal shock wave, to a bifurcated shock wave, and to a normal shock train, respectively for the 1X, 10X, and 100X models.

A computer model of a blow-down free-jet hypersonic propulsion test facility exists to validate facility control systems as well as predict problems with facility operation. One weakness in this computer model is the modeling of an air ejector diffuser system. Two examples of facilities that could use this ejector diffuser model are NASA Langley Research Center's 8-ft High Temp. Tunnel (HTT) and the Aero-Propulsion Test Unit (APTU) located at Arnold Engineering Development Center. Modeling an air ejector diffuser system for a hypersonic propulsion test facility includes modeling three coupled systems. These are the ejector system, the primary free-jet nozzle that entrains secondary airflow from the test cell, and the test article. Both of these facilities are capable of testing scramjets/ramjets at high Mach numbers. Compared with computer simulation data, experimental test cell pressure data do not agree due to the current modeling technique used.An improved computer model was derived that incorporates new techniques for modeling the ejector diffuser. This includes real gas effects at the ejector nozzles, flow constriction due to free-jet nozzle and ejector plumes, test article effects, and a correction factor of the normal shock pressure ratio in a supersonic diffuser. A method was developed to account for the drag and thrust terms of the test article by assuming a blockage factor and using a drag coefficient*Area term for both the test article and thrust stand derived from experimental data. An ideal ramjet model was also incorporated to account for the gross thrust of the test article on the system.The new ejector diffuser model developed improved the accuracy and fidelity of the facility model as compared with experimental test data while only negligibly affecting computational speed. Comparisons of the model data with experimental test data showed a close match for test cell pressure (within 1 percent for final test cell pressure). The model accurately simulated both the unstarted and started modes of ejector flow, in which test cell pressure increases with nozzle total pressure once in started mode.

A turbulent channel flow was used to study the scaling of the combined effects of roughness and flow injection on the mean flow and turbulence statistics of turbulent plane Poiseuille flow. It was found that the additional momentum injected through the rough surface acted primarily to enhance the roughness effects and, with respect to the mean flow, blowing produced similar mean flow effects as increasing the roughness height. This was not found to hold for the turbulence statistics, as a departure from Townsend’s hypothesis was seen. Instead, the resulting outer-scaled streamwise Reynolds stress for cases with roughness and blowing deviated significantly from the roughness only condition well throughout the inner and outer layers. Investigation into this phenomena indicated that suppression of the large-scale motions due to blowing may have been contributing to this deviation.

The data-driven adaptive algorithms are explored as a means of increasing the accuracy of Reynolds-averaged turbulence models. This dissertation presents two new data-driven adaptive computational models for simulating turbulent flow, where partial-but-incomplete measurement data is available. These models automatically adjust (i.e., adapts) the closure coefficients of the Reynolds-averaged Navier-Stokes (RANS) k-ω turbulence equations to improve agreement between the simulated flow and a set of prescribed measurement data. The first approach is the data-driven adaptive RANS k-ω (D-DARK) model. It is validated with three canonical flow geometries: pipe flow, the backward-facing step, and flow around an airfoil. For all 3 test cases, the D-DARK model improves agreement with experimental data in comparison to the results from a non-adaptive RANS k-ω model that uses standard values of the closure coefficients. The second approach is the Retrospective Cost Adaptation (RCA) k-ω model. The key enabling technology is that of retrospective cost adaptation, which was developed for real-time adaptive control technology, but is used in this work for data-driven model adaptation. The algorithm conducts an optimization, which seeks to minimize the surrogate performance, and by extension the real flow-field error. The advantage of the RCA approach over the D-DARK approach is that it is capable of adapting to unsteady measurements. The RCA-RANS k-ω model is verified with a statistically steady test case (pipe flow) as well as two unsteady test cases: vortex shedding from a surface-mounted cube and flow around a square cylinder. The RCA-RANS k-ω model effectively adapts to both averaged steady and unsteady measurement data.

This thesis documents a series of three dimensional unsteady Reynolds Averaged Navier-Stokes CFD simulations used to investigate the influence of an upstream prolate spheroid body on tandem pitching hydrofoils. The model is validated by performing separate CFD simulations on the body and pitching hydrofoils and comparing results to existing experimental data. The simulations were run for a range of Strouhal numbers (0.2-0.5) and phase differences (0-π). Results were compared to identical simulations without an upstream body to determine how the body affects thrust generation and the unsteady flow field. The combined time-averaged thrust increases with Strouhal number, and is highest when the foils pitch out of phase with each other. At intermediate phase differences between φ = 0 and φ = π the leading foil produces significantly more thrust than the trailing foil, peaking at φ = π/2. For St = 0.5 this difference is 21.7%. Results indicate that adding an upstream prolate spheroid body does not significantly alter thrust results, though it does provide a small (nearly negligible) boost. Vorticity from the body is pulled downstream from the pitching foils, which interacts with the vortex generation when the vortex being generated is of the same sign as the body vorticity. This body vorticity does not affect the vorticity magnitude of the downstream vortex pairs.