Owe Eld Research Articles

Experimental Analysis of Aerodynamic Interactions Occurring on Hypersonic Spacecraft

Experimentalinvestigationshavebeenachievedin theONERAresearch facilitiesof theChalais ‐Meudon Center in order to validate design tools used for the dee nition of a scramjet for a hypersonic spacecraft. Aerodynamic interactions occurring on such a spacecraft have been thoroughly studied, notably 1 ) the shock/shock interactions ahead of the air inlets, 2 ) the shock-wave/boundary-layer interaction and its control inside the air inlet, and 3)the e ow cone uence between an external e ow and an aerospike nozzle jet in an underexpansion condition. These studies contribute to the knowledge of the e ow physics of aerodynamic interactions that occur on hypersonic spacecraft, compensate for the lack of published detailed measurements of such e owe eld patterns, and provide well-documented test cases to validate computer codes.

Unstructured Method for Flows past Bodies in General Three-Dimensional Relative Motion

A three-dimensional methodology was developed for unsteady e ows past bodies that are in relative motion and where the trajectory of the motion was determined from the instantaneous aerodynamic e eld. This method coupled the equations of e uid e ow and those of rigid-body dynamics and then captured the unsteady aerodynamic interference between the stationary and moving boundaries. The time-dependent, compressible Euler equations weresolvedondynamic,unstructuredmeshesbyanexplicit,e nitevolume,andupwindmethod.Thegridadaptation was performed within a window placed around a moving body. The Euler equations of dynamics were then solved by a Runge‐Kutta integration scheme. The e ow solver and the adaptation scheme were validated by simulating the transonic, unsteady e ow around a wing undergoing a forced, periodic pitching motion and then comparing the results with the experimental data. Finally, the overall methodology was demonstrated by simulating the unsteady e owe eld and the trajectory of a store dropped from a wing. The methodology, with its computational cost notwithstanding, has proven to be accurate, automated, easy for dynamic gridding, and relatively efe cient for the required work hours. Nomenclature at

Effect of Reynolds Number on Separation Bubbles on Compressor Blades in Cascade

A detailed experimental investigation of second-generation, controlled-diffusion, compressor stator blades at an off-design inlet-e ow angle was performed in a low-speed-cascade wind tunnel primarily using laser Doppler velocimetry (LDV). The object was to characterize the e owe eld in the Reynolds number range of 2 :1‐6:4 £ 10 5 and to obtain LDV measurements of the suction surface boundary-layer separation that occurred near midchord. Surfacee owvisualizationshowedthatatthelowReynoldsnumberthemidchordseparationbubblestartedlaminar and reattachedturbulentwithin 20%chordonthesuctionsideoftheblade.Theextentofthebubblecomparedvery well with the measured blade surface pressure distribution, which showed a classical plateau and then diffusion in the turbulent region. LDV measurements of the e ow reversal in the bubble were performed. At the intermediate Reynolds number, the boundary layer was transitional upstream of the separation bubble that had decreased signie cantly in size (down to 10% chord ). At thehighestReynoldsnumber, the e owwas turbulentfrom close to the leading edge, and three-dimensional e ow reversal as a result of endwall effects appeared at approximately 80% chord. These data, particularly the low Reynolds number data, are an excellent test case for either large-eddy simulation or direct numerical simulation of cascade e owe elds.

Oxygen Atoms' Effect on Vibrational Relaxation of Nitrogen in Blunt-Body Flows

Numerical simulations are presented of the steady-state aire ow over a hemisphere cylinder of 1-m radius having hypersonic Mach numbers, where vibrational relaxation is the dominant mechanism and the dissociation of oxygen is small. A Mach 6.5 e ow was analyzed at freestream pressure of 50 Pa with a nonequilibrium freestream translational temperature of 300 K and vibrational temperature of 4000 K; a Mach 1.5 e ow was also studied to delineate effects of vibration ‐translation (V‐T)energy losses due to N 2‐O collisions. The effects on the vibrational population distribution, temperature, and pressure in the e owe eld were studied for various media: pure nitrogen and air mixtures of 0.0001, 0.1, and 1% oxygen atoms. Code validation was performed with previously reported computational results and experimental data for equilibrium e ow in freestream, but nonequilibrium in the shock layer. An upwind difference numerical scheme was used to solve the inviscid Euler equations coupled to a vibrational kineticsmodel ofN 2,assumedasananharmonicoscillatorof40quantum levels. Theshock-standoffdistance comparison with experimental data for a Mach 7.7 and 8.6 aire ow past a blunt body showed good agreement. For the Mach 1.5 e ow at nonequilibrium freestream conditions, the high efe ciency of the V ‐T rates of N2‐O collisions introduces additional heating in the shock layer for 0.1% and higher atomic oxygen, thus increasing the shockstandoff distance; for the Mach 6.5 e ow, a 0.1% atomic oxygen in air decreases the translational temperature in air compared to that of pure nitrogen in the stagnation region.

Development of Ultraviolet Raman Diagnostics for Rocket Engine Injector Analysis

The development of a laser-based combustion diagnostics system and its application to the analysis of a highpressure rocket engine injector e owe eld are described. One dimensionally spatially resolved (linewise), multispecies, single-pulseimages are obtained with an imaging spectrograph/digital camera detection setup that images spontaneousvibrational Raman light scattering caused by a pulsed ultraviolet laser, used to enhanceRaman signal strength by the inverse fourth power of laser wavelength scaling forRaman scattering cross sections. These images provide temporally and spatially resolved qualitative information about the presence of gaseous major species (water vapor and molecular hydrogen, nitrogen, and oxygen ) during a hot-e re, high-pressure (6.04-MPa) test of a single-element liquid oxygen/gaseous hydrogen swirled injector operating at a high oxidizer/fuel mixture ratio (100)inatestarticlepressurizedbyanitrogencoe ow.Ramanimages,takenevery40ms,aregenerallyuncorrelated with respect to each other during the constant pressure, main-stageportion of the test, but occasionally temporally adjacent imagepairsarecorrelated, suggesting large-scalemotion in thee ow, possibly from transverse movements of the reactant jet within the diluent coe ow. Simultaneous measurements of major species show unburnt ignitor gas (hydrogen) during the startup transient and excess diluent (nitrogen) during the shutdown transient, along with signie cant variations in the oxygen/nitrogen ratio during the test. Stimulated Raman and ozone production problems are investigated.

Computational Investigation of a Method to Compress Air Fluidically in Supersonic Inlets

The inlet is an important component in high-speed airbreathing engines. Whereas a variable-geometry inlet maximizes performance, it adds weight and complexity to the engine. Fixed-geometry inlets are limited to low contraction ratios to allow the initial normal shock to beswallowed and the engine to start. A novel e xed-geometry inlet cone guration that uses a high-speed jet to act as the contact surface and compress the air is computationally investigated. Adjusting the jet pressure alters the contact surface and allows shock-on-lip conditions to be maintained over a range of e ight Mach numbers without a movable e owpath. This cone guration allows a low-blockage e owpath forlowe ightspeeds, resultinginlowspillage, whilemaintaining highcompression inscramjetmode. Flow solutions are obtained using a Navier ‐Stokes solver. Flowe eld details are presented and discussed, including the shock patterns and mixing between the jet and air. Shock-on-lip is demonstrated at varying e ight Mach numbers, and associated effects on the e owe eld are presented. Global analysis indicates that minimizing jet mass e ow is a key to applying this concept to practical engines.

Accelerating Computational Fluid Dynamics Based Aeroelastic Predictions Using System Identification

System identie cation is evaluated as an efe cient and accurate technique for modeling unsteady aerodynamic forces for use in time-domain aeroelastic analysis. In the system identie cation methodology, the constant coefe cients of a linear system model are e t to the computed response time histories from a three-dimensional, unsteady computational e uid dynamics (CFD) solver. The resulting model of the unsteady CFD solution is independent of both dynamic pressure and structural parameters. Hence, this methodology has the advantage that only one CFD e owe eld computation for each Mach number must be completed to determine the aeroelastic instability boundary. Results show that system identie cation can accurately model the unsteady aerodynamic forces for complex aerospace structures of practical interest. The methodology results in a substantial savings in computational time when predicting aeroelastic instabilities.

Chemical-Kinetic Parameters of Hyperbolic Earth Entry

Chemical-kinetic parameters governing the e ow in the shock layer over a heat shield of a blunt body entering Earth’ s atmosphere from a hyperbolic orbit are derived. By the use of the assumption that the heat shield is made of carbon phenolic and by allowing for an arbitrary rateof pyrolysis-gasinjection, chemical reactions occurring in the shock layer are postulated, and the collision integrals governing the transport properties, the rate coefe cients of the reactions, and the parameters needed for the bifurcation model and for the e nite-rate kinetic wall boundary conditions are determined using the best available techniques. Sample e owe eld calculations are performed using this set of parameters to show that the heating and surface removal rates are substantially smaller than calculated using theexisting setofsuch parameters and traditionalassumptionsof gas ‐surfaceequilibrium and quasi-steadystate ablation.

Simulation of Wake Vortex Detection with Airborne Doppler Lidar

A simulation approach is used to demonstrate that an airborne, forward-looking lidar can detect trailing wake vortices generated by a leading aircraft. Computational e uid dynamics techniques are used to generated e owe elds containing wakevortices.Thesee owe eldsareappliedtoaDopplerlidarsimulation model.Thesimulated geometry hastheforward-lookinglidarmountedonafollowingaircraft,whereasthewakevorticesaregeneratedbyaleading aircraft. In such a cone guration, it is not possible to directly detect the strong rotational velocity components of a wake vortex with a Doppler system; detection relies on the presence of induced axial velocity signatures. We demonstrate that axial e ows are induced as the vortices evolve, and that these axial e ows can be detected with a Doppler lidar system.

Application of Bethe-Zel'dovich-Thompson Fluids in Organic Rankine Cycle Engines

Thenovel concept ofimproving organicRankinecycleengines (ORCEs)by exploiting the unusual e uid dynamic behaviorofaspecialclassofworkinge uidsknownasBethe ‐Zel’dovich‐Thompson(BZT)e uidsisexplored.ORCEs are currently manufactured and used for numerous remote terrestrial applications requiring reliable, unattended power.OneofthemajorlossmechanismsinORCEsisshockwavesgenerated intheturbinestage.Operatingunder theproperthermodynamicconditions, a BZT working e uid can potentially weaken oreliminate shock waves. This would reduce losses due to both the wave drag from shocks as well as losses from boundary-layer separation due to shock ree ections on the turbine blades. Two-dimensional e owe elds through realistic symmetric impulse turbinecascades arecompared using a e ux-limited, e nite differencenumerical schemeto assess differences in e ow efe ciencies. Resultsshowsignie cantimprovementsin turbineefe cienciesforBZTworkinge uidsoverconventional ORCE working e uids.

Numerical Simulation of Transient Jet-Interaction Phenomenology in a Supersonic Freestream

The objective is to evaluate the transient effects of a reaction control jet on the aerodynamic performance of a genericinterceptormissileoperating at supersonice ightconditions.Three-dimensionalcomputationsofthehighly turbulent e owe eld produced by a pulsed, supersonic, lateral-jet control thruster interacting with the supersonic freestream and missile boundary layer of a generic interceptor missile are evaluated at different altitudes and thruster conditions. A generic missile interceptor cone guration consisting of a long, slender body containing e xed dorsal and tail e ns is simulated. Parametric computational e uid dynamic solutions are obtained at altitude conditions corresponding to 19.7 and 35.1 km for 1 ) steady-state conditions with the lateral control jet turned off, 2) steady-state conditions with the lateral control jet turned on, 3 ) transient jet startup conditions, and 4 ) transient jet shutdown conditions. A thermally and calorically perfect gas with a specie c heat ratio equal to 1.4 was assumed for both the Mach number 5 freestream and Mach number 3 lateral jet. Vehicle forces and moments are assessed from each solution by integrating the surface pressures and viscous shear stresses computed on the missile surfaces. These results are used to determine the ine uence of the jet-interaction effects on the transient aerodynamicperformanceofthemissile.Theanalysispredictsstrongtransientine uencesfortheintegratednormal force and pitching moment.

Automated Design of a Three-Dimensional Subsonic Diffuser

A novel methodology is developed to integrate state-of-the-art computational e uid dynamics analysis, NURBS, and optimization theory to reduce total pressure distortion and sustain total pressure recovery within a curved three-dimensional subsonic S-duct diffuser by automated redesign of the diffuser shape. Two independent design variables are used. The change of the surface shape is assumed to be Gaussian. GASP with the modie ed Baldwin ‐ Lomax turbulence model (Baldwin, B. S., and Lomax, H., “ Thin Layer Approximation and Algebraic Model for Separated Turbulent Flows,” AIAA Paper78-257, 1978 )is employed for the e owe eld prediction and proved to give good agreement with the experimental surface pressure for the baseline S-duct diffuser geometry. The automated design optimization is performed with a gradient-based method to minimize the total pressure distortion based on the two design variables. The best cone guration obtained reduced distortion by typically 70% while keeping the total recovery essentially the same. The results indicate that the mechanism responsible for improved diffuser performance is the suppression of detrimental secondary e ows by changing the surface shape to redirect the e ow.

Hover Performance Prediction Using Full-Potential Method and Comparison with Experiments

A computational and experimental effort to investigate the e ow around rotor blades in hover is presented. The experimental analysis of the e owe eld in the immediate vicinity of the blade is supported by laser velocimetry (LV) performed on the three-dimensional velocity e eld around the bladesection and in its nearwake. A specie c LV data processingmethod based onthemomentumequationapplied around a bladesectionhasbeen developed to identify the contribution of the induced-drag component to the total sectional drag coefe cient. From the computational approach, some improvements to a vortex embedding full-potential method are checked by direct comparisons withtestdata. Comparisonsareperformed foroverallrotorairloads (thrustand torquecoefe cients ),localsectional coefe cients (liftanddragcontributions ),ande owe eldcharacteristics,includingthetipvortexpathandthespanwise circulation distribution.

Unsteady Viscous Flow Causing Rotor-Stator Interaction in Turbines, Part 1: Data, Code Pressure

A coupled experimental and computational study of the effects of the nozzle wake-rotor interaction is carried out to understand the cause and effects of the unsteady e ow in turbine rotors. The paper is organized in two parts. Part 1 deals with the experimental and numerical program and interpretation of the blade pressure e eld. The experiment was carried out in a low-speed turbine at 22.6% of the nozzle chord spacing between the rotor and the nozzle. The laser Doppler velocimetry (LDV) data, acquired earlier, is integrated with the dynamic pressure measurements and numerical simulation to provide an integrated interpretation. A systematic study has been carried out to evaluatethe effect of the turbulence model and artie cialdissipation on the accuracy of thenumerical prediction. The steady e owe eld, unsteady pressure at design and off-design conditions on both blade surfaces, and the hub are presented, interpreted, and compared with the predictions from the Navier ‐Stokes code. The predictions are in a good agreement with the data. The second part deals with the integrated interpretation of the unsteady velocity and pressure e eld, as well as the e ow physics associated with the nozzle wake transport and decay.

Unsteady Viscous Flow Causing Rotor-Stator Interaction in Turbines, Part 2: Simulation, Integrated Flowfield, and Interpretation

Numerical simulation has been performed to investigate the unsteady e ow physics associated with the nozzle ‐ rotorwakeinteractioninturbines.Acomparison between thepredictedand theexperimentalunsteadypressureon the blade surface presented by Lakshminarayana et al. (Lakshminarayana, B., Chernobrovkin, A., and Ristic, D., “ Unsteady Viscous Flow Causing Rotor-Stator Interaction in Turbines, Part 1: Data, Code Pressure,” Journal of Propulsion and Power , Vol. 16, No. 5, 2000, pp. 744 ‐750) shows good agreement. The predicted e owe eld also correlates well with the blade-to-blade laser Doppler velocimetry data. Results of the numerical prediction in conjunction with the experimental data have been used to gain a better understanding of the unsteady e ow physics. An assessment of the viscous and the inviscid contribution to the nozzle wake decay and the unsteady loss distribution in the rotor passage reveals the dominant effect of the viscous decay upstream of the leading edge. Inside the passage, the inviscid effects have a signie cant ine uence. The numerical solver has been able to predict mostof the features associated with theunsteady transition on a turbineblade. The predicted e ow at theoff-design conditionhasbeeninterpreted tounderstandthenatureoftheunsteadye owe eldata high negativeincidenceangle.

Oscillatory Temperature Measurements in a Pulsed Microgravity Diffusion Flame

Results from an experimental study of a pulsed, laminar, propane gas-jet diffusion e ame in microgravity are presented. The primary objective of the study was to investigate the interaction of the imposed vortex with the jet and e ame e owe eld under momentum-dominated conditions. A novel technique, which used a sinusoidally driven iris placed concentrically around the e ame base, was used to provide axisymmetric, controlled, and repeatable disturbances to interact with the e ame. Measurements of the oscillatory temperature were used to elucidate the evolution of the imposed vortex as it convects downstream. Regionsof the e ame where energy transferbetween the oscillations and the mean thermal e eld occur and the direction of the energy transfer were identie ed. Differences with normal gravity behavior were noted. Finally, correlations between the temperature and e ame radiation oscillations were provided.

Comparison of Induced Velocity Models for Helicopter Flight Mechanics

Modeling of rotor-induced velocity receives continued attention in the literature as the rotorcraft community addresses limitations in thee delity of simulations of helicopterstability, control,andhandling qualities. A comparison is presented of results obtained using a rigid-blade rotor-fuselage model cone gured with two induced velocity models: a conventional, e rst-order, e nite state, dynamic ine ow model and a wake model that solves a vorticitytransport equation on a computational mesh enclosing the rotorcraft. Differences between the two models are quantie ed by comparing predictions of trimmed rotor blade e ap, lag and feather angles, airframe pitch and roll attitudes, cross-coupling derivatives, response to control inputs, and airframe vibration. Results are presented in the context of measurements taken on a Puma aircraft in steady e ight from hover to high speed. More accurate predictions of the cross-coupling derivatives, response to control, and airframe vibration obtained using the vorticity transport model suggest that incorporation of real e owe eld effects is important to extending the bandwidth of applicability of helicopter simulation models. Unexpectedly small differences in some of the trim predictions obtainedusing thetwo wakemodelssuggest thatanoverall improvement in simulatione delity may notbeachieved without equivalent attention to the rotor dynamic model.

Convergent-Flow-Derived Waveriders

Results are presented of a study on the aerodynamics of a new type of waverider derived from supersonic axisymmetric e ows inside constricting ducts, specie cally conical trumpet ducts. In such a duct, an initial shock wave arises from its leading edge, and the compression e ow downstream of this shock has streamlines converging toward the e ow axis. This e ow is chosen as a basic e ow for the waverider design. The simplest convergent-e owderived waveriders are constructed with a lifting surface with a transverse-concave arc-shaped contour. They are compared with known types of waveriders constructed based on uniform e ows behind plane oblique shock waves or divergent e ows behind axisymmetric conical shock waves. The characteristics of convergent waveriders as lifting cone gurations are analyzed with the lift and drag coefe cients, the lift-to-drag ratio, and the integral heat e uxes through the waverider surfaces determined. The possibilities of using these new waveriders as forebodies for hypersonic vehicles powered by airbreathing engines are also estimated. The e owe eld characteristics near the lower lifting surface of the waverider as a precompression surface arranged upstream of the inlet are considered in this connection.

Numerical Simulation of Transonic Flow over Wing-Mounted Twin-Engine Transport Aircraft

A numerical method has been developed for computing the e owe eld around advanced transport aircraft with wing-mounted nacelles. The method is based on a multiblock point-matched grid-generation approach combined with zonal solving strategy for complex e owe eld. In this study the e owe eld is divided into a number of nonoverlapped blocks by a cutout technique. H-type grids are generated independently in each block using an elliptic grid-generation method, in which the control of the grid quality is accomplished by the forcing-function technique of Hilgenstock. The e owe eld is simulated by solving the Euler equations. An explicit three-stage Runge ‐Kutta algorithm based on the Jameson’ s e nite volume scheme for the Euler equations has been developed that is applied to the multiregion H-type grids. The present method has been applied to isolated powered engine nacelles and complex transport aircrafts consisting of low-wing/fuselage with wing-mounted pylon/nacelles. On the wing surfaces the viscous effects are simulated by the employment of the viscous/inviscid interaction (VII) technique of two-dimensional strip boundary layer. In this study the boundary-layer program uses an integral method to calculate turbulent boundary layers. With the concept of an equivalent inviscid e ow, the model of blowing velocity is employed in the VII technique. The effect of the boundary layeron the outer inviscid e ow is represented through a transpiration boundaryconditionderived from theboundary-layerparameters.Themain benee tofthistreatment isthatthegridisgeneratedonlyonceinoverallcomputingprocedure.Computationalresultsandcomparisonswith experimental data are presented. The good agreement indicates that the present method is effective in predicting the e ows about powered engine nacelles and/or complex transport aircrafts.

Eigenmode Analysis and Reduced-Order Modeling of Unsteady Transonic Potential Flow Around Airfoils

An eigenmode analysis and reduced-order models of the unsteady transonic aerodynamic e ow around isolated airfoilsarepresented.Theunsteadye owismodeledusingthetime-linearizedfrequency-domainunsteadytransonic full potential equation. The full potential was discretized in space using a e nite element method. The resulting equations are linear in the unknown velocity potential and quadratic in the reduced frequency of excitation. The dominant eigenfrequencies and corresponding mode shapes of the discretized potential model are computed, and theeffect ofdifferent parametersthat determinethesteady and unsteady e owe eld (e.g.,thefar-e eld Mach number, theangleofattack,and theairfoilshape )areinvestigated.Anormalmodeanalysisanda staticcorrectiontechnique are then used to construct a low degree-of-freedom, reduced-order model of the unsteady e owe eld. Depending on the range of frequencies of interest, a relatively small number of eigenmodes are required. An alternative reducedorder modeling technique based on Arnoldi ‐Ritz vectors is also presented. For the case where the structural excitations are known a priori, the latter method is more efe cient. Using the aerodynamic reduced-order models, we construct aeroelastic reduced-order models and compute e utter boundaries for different airfoils at several different Mach numbers.