Dean R. Eklund

Dual-Mode Scramjet Combustor: Numerical Sensitivity and Evaluation of Experiments

Abstract : Experiments were performed at the Air Force Research Laboratorys Propulsion Directorate (AFRL/RZ) in Research Cell 22 (RC22). Twelve cases from the experiment were computationally analyzed and each case varied in either engine operating condition and/or combustor configuration. Initial computations were performed on all twelve cases to establish a baseline computational approach. Computations were performed on one of the cases to test sensitivity to turbulent Schmidt number, reaction rate, and grid resolution. Improvements to the baseline analysis using the results from the sensitivity analyses were extended to two additional cases. It was shown that adjustments in Schmidt number, reaction rate, or grid refinement improved the agreement with experimental data for two cases relative to the baseline results, but worsened agreement for the third case. It was left undetermined that grid refinement was a better approach to improving the baseline analysis as compared to calibrations in Schmidt number and/or reaction rate. Improvement to the grid using local refinement in regions with chemical reactions produced better results for one case and was computationally less expensive than globally refining the grid. Negligible differences were shown between results that were obtained using wall functions with Y+ value as high as 38 or results obtained using wall integration with Y+ values around one. Negligible differences were shown between periodic results that were obtained by averaging results using either a constant CFL or a constant time step. CFL-averaging a result using the constant CFL approach was 5.4 times less computationally expensive than using the constant time step approach. Computations showed that 2.53-lbm/sec. of air leaked into the exhauster housing at the exit of the combustor in RC22s test apparatus.

Dual-Mode Scramjet Combustor: Numerical Analysis of Two Flowpaths

A PRELIMINARY numerical characterization of a U.S. Air Force Research Laboratory (AFRL) test facility was performed. Research cell 22 (RC22) is a supersonic wind-tunnel facility at AFRL. RC22 is presently simulating dual-mode combustion for circular (axisymmetric) combustors. Previous efforts in RC22 have studied combustion in rectangular flowpaths. The current effort in RC22 is to examine the benefits of axisymmetric flowpaths as compared with rectangular. For the same cross-sectional area in which circular flowpaths eliminate the challenges involved with corner flow effects, they have increased structural efficiency (heat load distribution) and have reducedweight. Circular combustors also pose challenges that involve effective fuel penetration and flame propagation. For more information regarding the rectangular experiments performed in RC22, refer to [1].

Dual Mode Scramjet Combustor: Analysis of Two Configurations

** Two engine configurations were analyzed numerically and experimentally. The first configuration had a constant area backstep and the second configuration had a tapered (divergent) combustor with no backstep. Both combustors burned gaseous ethylene fuel. The analysis used a simulated low enthalpy flight condition corresponding to Mach 3.0 flight. Numerical results showed good agreement with experimental data in terms of performance, isolator heat loss and pressure distribution. For the same fuel equivalence ratio and fuel split the tapered configuration outperformed the backstep configuration by approximately 1% in terms of stream thrust, but the backstep configuration had more isolator margin (before unstart) which allowed for higher total equivalence ratios. The backstep configuration achieved stoichiometric fueling. Model calibration via constant turbulent Schmidt number was performed for the first case in order to better match experimental results. Two different k-epsilon turbulence models were analyzed and yielded differences in pressure distribution. A conceptual design analysis involving a reduction in the number of injectors showed better fuel penetration but less circumferential coverage. Future design concepts could include changing the axial placement of the secondary injector, which could have a positive impact on operability for both configurations. Numerically, future analysis could include a variable turbulent Schmidt number which would be more representative of the flow physics and could alleviate the need for calibration in modeling reacting flows.

TDLAS-based Measurements of Temperature, Pressure, and Velocity in the Isolator of an Axisymmetric Scramjet

Using tunable diode laser absorption spectroscopy (TDLAS), water-based measurements of path-integrated static temperature, static pressure, and local velocity are made in the isolator of a direct-connect, ethylene-fueled, axisymmetric scramjet engine. Three optical lines of sight are directed across the midplane of the isolator just downstream of the facility nozzle. For the data set presented here, the scramjet engine was operated to simulate relatively low flight Mach number conditions of M ~ 3. The flight conditions were achieved using a dry, heated, but non-vitiated inflow which provided a challenging environment for the water-based direct-absorption TDLAS implementation. The TDLAS-derived temperature and velocity agree well with facility operating conditions while the derived pressure measurements highlight the need for improved pressure-broadening data. With a pressure-correction factor, the TDLAS-derived values for mass flux agree well with the measured facility mass flow. The measurements are discussed in the context of a 1-D analytical model and 3-D CFD calculations. Notably, the dynamic behavior of the velocity measurements under conditions of low isolator margin are consistent with the presence of a local shock structure predicted by the CFD computations.

Instrumentation and Performance Analysis Plans for the HIFiRE Flight 2 Experiment

Supersonic combustion performance of a bi -component gaseous hydrocarbon fuel mixture is one of the primary aspects under investigation in the HIFiRE Flight 2 experiment. In -flight instrumentation and post -test analyses will be two key elements used to determine the combustion performance. Pre -flight computational fluid dynamics (CFD) analyses provide valuable information that can be used to optimize the placement of a constrained set of wall pressure instrumentation in the experiment. The simulations also allow pre -flight assessments of performance sensitivities leading to estimates of overall uncertainty in the determination of combustion e fficiency. Based on the pre -flight CFD results, 128 wall pressure sensors have been located throughout the isolator/combustor flowpath to minimize the error in determining the wall pressure force at Mach 8 flight conditions. Also, sensitivity analyses show that mass capture and combustor exit stream thrust are the two primary contributors to uncertainty in combustion efficiency.

A Computational Assessment of Independent Stage Control of a Cascade Injector

‡§ ‡ A computational assessment of independent stage control of a cascade injector was performed. This investigation used computational fluid dynamics to gain understanding of the mechanics governing the penetration characteristics of the cascade injector. Comparison to experimental data, the effects of the first injection stage, and a turbulence model study are presented. The computational solutions predicted penetration height with good accuracy compared to the experimental data, based on planar laser-induced fluorescence of the NO molecule, but were not as successful predicting injectant plume width. Penetration in the near-field was found to correspond with the position of the shock wave generated by the injectant. The Menter turbulence model produced higher values of eddy viscosity in the vicinity of the plume and more spreading, which agreed better with the experimental data. The computational data was also used to generate a synthetic laserinduced fluorescence signal, which was used to improve the penetration boundary prediction.

The Advantages and Costs of Higher-Fidelity Turbulence Modeling

This paper compares the results and the computational efficiency of a research code with that of a commercial code on the same problem. The research code (REACT-MB) is tested using its unsteady hybrid large-eddy simulation/Reynolds-averaged Navier-Stokes (LES/RANS) method as well as a more common steady-state Menter RANS method. CFD++ of Metacomp is tested on the same problem using its realizable k-ε turbulence model. Normal sonic ethylene injection through a circular injector into a Mach 2 cross-flow was simulated by each code. Time-averaged statistics of the hybrid LES/RANS computations and converged solutions from the RANS computations are compared with experimental contours of time-averaged mixture fraction. Scalability of the codes is also compared.

Uncertainty Quantification for a Scramjet Inlet Flow

A numerical framework is presented for CFD-based uncertainty quantification (UQ) involving propagation of combined aleatory and epistemic uncertainties to relevant output quantities, with intended application to scramjet inlet analysis. Latin hypercube sampling routines are used in combination with surrogate response surface methodology and nested Monte Carlo sampling to generate probability boxes and other UQ results based on CFD output data. For a demonstration case involving subscale Mach 6 semi-freejet testing of an inward turning inlet configuration, a total of 21 CFD simulations are performed, and probability boxes are computed for three output quantities through consideration of four different input uncertainties. Global sensitivity analysis calculations are also performed to assess relative contributions of each input uncertainty. For the current demonstration case, we find Mach number uncertainty from thermal expansion of the facility nozzle to be a dominant contributor to uncertainty in the inlet mass flow rate, while mounting angle uncertainties account for large contributions to uncertainty in wall pressure at the isolator throat.

Improved Simulation of Inflow Distortion for Direct-Connect Scramjet Combustor

A technique to develop flight-like inflow conditions for direct-connect testing has been extended in the current work. This technique utilizes an expansion/contraction section upstream of the engine isolator in a direct-connect test facility to create shock waves consistent with those generated in flight. The current work improves upon the earlier work by better matching the inlet shock structures, the average conditions at the engine throat, and boundary layer properties. RANS CFD simulations have been performed for isolators downstream of a flight inlet, a direct-connect facility nozzle without distortion, and a directconnect facility nozzle with distortion. The RANS CFD simulations predict similar shock positions in all cases, however a change in the mass flux distribution at the exit plane is observed.

Dual Mode Scramjet Combustor: Design and Operational Analysis

†‡ Numerical analysis was performed on a Dual-Mode Scramjet isolator-combustor with emphasis on the combustor. The investigation was performed to analyze overall performance and observe isolator-combustor interaction. In this investigation four studies were analyzed. It was shown by removing a flame holding cavity at the second stage of a two stage combustor, yielded negligible performance loss with a decrease in combustor complexity. A comparison study outlined marginal differences in the analysis between models performed in 2D as opposed to 3D. Another study revealed minimal performance loss by reducing the flight Mach number from Mach 3.0 to Mach 2.5 without unstarting the isolator. The final study revealed a decrease in stream thrust by introducing a less reactive fuel such as methane into an ethylene fuel mixture.