aiaa-2002-4409 supersonic aerodynamic characteristics of

For permission to copy or republish, contact the American Institute of Aeronautics and Astronautics370 L'Enfant Promenade, S.W., Washington, D.C. 20024

AIAA-2002-4409

Supersonic Aerodynamic Characteristics of Proposed Mars '07Smart Lander Configurations

Kelly J. Murphy, Thomas J. Horvath,Gary E. EricksonNASA Langley Research CenterHampton, VA 23681

Joseph M. GreenMississippi State UniversityMississippi State, MS 39762

AIAA Atmospheric Flight Mechanics Conference and Exhibit 5-8 August 2002/ Monterey, California

1

Supersonic Aerodynamic Characteristics of ProposedMars ’07 Smart Lander ConfigurationsKelly J. Murphy*, Thomas J. Horvath*†, Gary E. Erickson‡

NASA Langley Research Center, Hampton, VA 23681

Joseph M. Green§

Mississippi State University, Mississippi State, MS 39762

Supersonic aerodynamic data were obtained for proposed Mars ’07 Smart Landerconfigurations in NASA Langley Research Center’s Unitary Plan Wind Tunnel. The primaryobjective of this test program was to assess the supersonic aerodynamic characteristics of thebaseline Smart Lander configuration with and without fixed shelf/tab control surfaces. Datawere obtained over a Mach number range of 2.3 to 4.5, at a free stream Reynolds Number of 1x 106 based on body diameter. All configurations were run at angles of attack from -5 to 2 0degrees and angles of sideslip of –5 to 5 degrees. These results were complemented withcomputational fluid dynamic (CFD) predictions to enhance the understanding o fexperimentally observed aerodynamic trends. Inviscid and viscous full model CFD solut ionscompared well with experimental results for the baseline and 3 shelf/tab configurations. Overthe range tested, Mach number effects were shown to be small on vehicle aerodynamiccharacteristics. Based on the results from 3 different shelf/tab configurations, a fixed controlsurface appears to be a feasible concept for meeting aerodynamic performance metricsnecessary to satisfy mission requirements.

Nomenclature

CA axial-force coefficientCD drag-force coefficientCL lift-force coefficientCLa lift-force curve slopeCl rolling-moment coefficientClb rolling-moment beta derivative, dCl/dbCm pitching-moment coefficientCma pitching-moment alpha derivative, dCm/daCn yawing-moment coefficientCnb yawing-moment beta derivative, dCn/dbCN normal-force coefficientCNa normal-force alpha derivative, dCN/daCY side-force coefficientLref longitudinal reference lengthM Mach numberp• static pressure of free stream, psiapt tunnel stagnation pressure, psiaq• free stream dynamic pressure, psiaRe• free stream unit Reynolds number/ftSref reference areaSAmes “Ames” tab reference areaSCanted “Canted” tab reference areaSShelf “Shelf” reference areaT• static temperature of free stream, °RTt tunnel stagnation temperature, °Ra angle of attack, degb angle of sideslip, deg

*Aerospace Technologist, Aerothermodynamics Branch,Aerodynamics, Aerothermodynamics and Acoustic Competency‡ Aerospace Technologist, Research Facilities Branch,Aerodynamics, Aerothermodynamics and Acoustic Competency§ Aerospace Engineering Cooperative Education Student,Mississippi State University.† Member AIAA

Introduction

As part of NASA’s on-going MarsExploration Program, several planned missions includesample return requirements. These future missions willrequire a highly precise entry through the Martianatmosphere and a “smart” landing with the ability todetect and avoid hazards on the planet’s surface. A keycomponent to enabling these highly accuratetrajectories and precise landings is reduction ofuncertainties in the aerodynamic characteristics of thevehicle’s aeroshell. The Mars ’07 Smart Landerprogram was established to land a vehicle on theMartian surface with the dual goals of furthering basicscientific research on Mars as well demonstrating a hostof “smart” technologies needed for a Mars SampleReturn mission. Smart Lander technology goalsinclude demonstration of a precision landing within a±5-kilometer footprint and a 100-meter siteredesignation capability for hazard avoidance, enablingsurface mobility of a rover to successfully navigatefrom the landing site to a pre-determined targetlocation.1

Both low-L/D and mid-L/D aeroshell shapeswere initially considered among the numerous ‘07Lander trade studies, 2 but the focus was eventuallyturned to a low-L/D shape with Pathfinder/Vikingheritage. The Mars ’07 Lander baseline aeroshell is a4.05-meter diameter spherically blunted 70-degree conewith a biconic backshell (Fig. 1). The large cone half-angle is required to produce the necessary drag todecelerate the Lander’s entry mass (approximately 2300kg) at sufficiently high altitudes to permit parachutedeployments in both the supersonic and subsonic flightregimes. To provide sufficient control authority andthe necessary cross range capability, the configuration

2

must achieve sufficient trimmed L/D, which isgenerally obtained via geometric asymmetry and/oroffset of the vehicle center of gravity (c.g.).Minimizing radial c.g. offsets can reduce ballastrequirements, provide internal packaging benefits, andthus yield corresponding optimization in weight andperformance. One proposed solution to attaininghigher L/D values with less mass penalty is a fixedcontrol surface, a shelf or tab, attached in the corner orshoulder region and of a suitable geometry to satisfyaerodynamic requirements.1,2

Initial design goals for the ’07 Smart Landeraeroshell with a fixed shelf/tab control surface include atrimmed L/D of 0.25 with no radial center of gravityoffset. Naturally it is desired to minimize shelf/tabmass to minimize corresponding ballast mass andtherefore increase payload capability. An additionaldesign constraint is that a fixed shelf/tab configurationmust fit within the maximum payload diameter(approximately 4.5m) of the Delta IV launch vehicle’sfairing. Current trajectories show supersonic parachutedeployment at a nominal Mach number of 2.2. Thus,obtaining supersonic aerodynamic wind tunnel databecame a priority to reduce uncertainties in initialdesign/trade studies for the ’07 Lander. Thisexperimental data, in conjunction with ballistic rangedata, supersonic and hypersonic CFD predictions, andmechanical design/packaging requirements, will be usedto optimize a proposed control surface design for the’07 Smart Lander.

Objectives

To provide vehicle designers with necessarysupersonic aerodynamic data, a sub-scale model of aproposed Smart Lander Configuration was designed,fabricated, and tested in Leg II of the Unitary Plan WindTunnels at the NASA Langley Research Center. Thetest program described in this paper was designed tosatisfy two primary objectives:

(1) Provide aerodynamic data in thesupersonic speed regime to establish thestability and performance characteristics o fthe proposed baseline Mars ’07 Smart Landerand to validate CFD predictions; (2) Providedata to assess aerodynamic characteristics o fthree shelf/tab configurations.

To accomplish the first objective, a 0.0376-scalealuminum force and moment Smart Lander model wasfabricated for testing in the UPWT. An extensivematrix of 6-component force and moment data wastaken on the model using a strain-gage force andmoment balance. Over 80 runs were made in theUPWT over a range of supersonic Mach numbers tocharacterize baseline vehicle stability and control andperformance characteristics at relevant angles of attackand sideslip. A number of these runs were part of anextensive data quality program to ensure the smallestpossible data uncertainties. To address this secondobjective, three different shelf/tab configurations were

fabricated at LaRC for testing in the UPWT. Over 60runs were conducted to assess the control effectivenessand performance characteristics of these controlsurfaces.

Experimental Program

Model Description All experimental aerodynamic data presented in

this report were obtained with a 0.0376-scale metallicforce-and-moment model designed and fabricated in-house at NASA Langley. The model diameter was 6inches, and the nominal reference areas and lengths usedto calculate aerodynamic coefficient data for the full-scale vehicle and the 0.00376-scale model are presentedin Table 1.

All model components were fabricated fromaluminum and included 3 removable shelf/tabconfigurations. Schematic drawings of these controlsurfaces are presented in Fig. 2. The “Ames Recreated”(which will be referred to as “Ames”) surface is asomewhat square-shaped tab, located at maximumvehicle diameter, with a full-scale area of 9.372 ft2.This tab shape was one generated in initial parametricstudies. It was retained for these tests for comparisonsto Ames ballistic range data as well as comparisons tothe blended canted tab and blended shelf configurations,which are refined versions of earlier designs. Theblended canted tab (referred to as “canted”) has a full-scale area 5.659 ft2. It is located at the maximumdiameter of the vehicle and is angled at 80 degrees tothe axis of symmetry of the 70-degree sphere coneheatshield. The blended shelf (referred to as “shelf”) isa continuation of the windward aeroshell angled at thesame 70-degree slant to the axis of symmetry. Theshelf full-scale area is 7.628 ft2. Control surface areascorresponding to both full-scale and a 0.0376 model-scale are summarized in Table 2.

A non-metric aluminum balance sleeve wasfabricated to shield the portion of the force balance thatprotruded from the model from the tunnel flow tominimize balance heating and loading due to flowimpingement. A photograph showing design of thissleeve is shown in Fig 3.

The outer mold lines of the aerodynamic modelwere extensively checked against Lander geometry filesby Langley’s Surface Verification Laboratory. Thex,y,z-location of numerous surface contour points weremeasured using both global and discrete pointtechniques to characterize surface coordinate fidelity: Allof the measured contour points were within ±.005 in.of configuration outer-mold-line definition. Thesesurface measurements were also used to preciselydetermine balance bore alignment and balance electricalcenter (i.e. c.g) location for accurate calculation ofmodel attitude and moment transfer distances.

Facility Description Unitary Plan Wind Tunnel-Leg II: The

UPWT is a closed-circuit, continuous-running, pressure

3

tunnel with two test sections that are nominally 4 ft by4 ft in cross section and seven ft long. The stagnationpressure, pt, can be varied up to a maximum ofapproximately 50 psia in Test Section I andapproximately 100 psia in Test Section II. The nozzlethroat-to-test-section area ratio is varied by a lowerasymmetric sliding nozzle block that providescontinuous variation of the Mach number. The Machnumber range is nominally 1.5 to 2.86 in Test SectionI and 2.3 to 4.63 in Test Section II. Tunnel stagnationtemperatures, Tt, are typically 125 °F and 150 °F foreach test section, respectively. Reynolds numbers from1.0 to 5.0 million per foot are possible. The basicmodel support mechanism is a horizontal wall-mountedstrut that is capable of forward and aft travel of over 3 ftin the streamwise direction. A main sting supportattached to the strut can transverse laterally ±20 in andcan provide a yaw capability of ±12°. Forward of themain sting support is the angle-of-attack mechanismthat provides pitch motion from –15° to +30°. A rollmechanism can be installed ahead of the pitchmechanism to provide continuous roll motion over a310° range. The history and test capabilities of theUPWT are discussed in Ref. 3.

Instrumentation and Data UncertaintyThree aerodynamic forces and three

aerodynamic moments were measured using the 6-component strain gage balance designated as theLangley 2008. Wind-off balance readings weremonitored before and after each run, and balancecomponents were monitored during the tunnel run fordrift caused by thermal gradients across the balancegages. Due to model and balance geometries, asignificant portion of the force balance extended beyondthe model base (Fig. 3). Thus, despite the presence ofthe balance sleeve to shield the model from the flow,relatively large temperature gradients were recordedacross the length of the balance. For this reason runswere reduced throughout the test program using "hot-zeros," i.e., wind-off balance zero readings taken beforeand after each run (rather than only at the beginning ofeach shift) to capture slight zero shifts due totemperature gradients. This practice was shown tosignificantly reduce temperature effects on aerodynamicdata and produce repeatabilities within quoted balanceaccuracies (further explained below). The balanceuncertainties for representative flow conditions areshown in Table 3. Static pressure measurements weremade at the base of the model within the sting shroudusing an offboard electronically scanned pressure (ESP)module.

To verify that data were repeating withinquoted balance accuracies, multiple repeat runs wereobtained throughout the test program for the baselineconfiguration at Mach 2.3. Mach 2.3 produced themost severe (highest pressure and temperature) testconditions, and thus it was presumed that these datawould be a worst case estimate on repeatability.“Residual” plots of six-component body-axis coefficientdata were calculated by subtracting the coefficient value

in a given run, interpolated to a nominal angle ofattack, from average values computed at those samenominal angles. The variation was compared to quotedaccuracies for all six components based on the ±95%confidence level from balance calibration reports. Allresidual data were shown to be bounded well by thesequoted balance accuracies through the angle of attackrange.

It should be noted that the aforementionedbalance accuracies represent only the uncertaintiesassociated with the balance itself. This would be ameasure of the overall uncertainty on the wind tunnelmeasurements only in the absence of all othervariations. An extensive test program involving largenumbers of repeat runs on all configurations at all testconditions over multiple test entries would be requiredto obtain rigorous uncertainty values. In the screeningand development phase of a configuration, this level oftesting is simply not feasible. Thus a factor of 3 timesthe quoted balance uncertainties has been recommendedas a conservative and physically reasonable estimate ofoverall uncertainties for the data generated in these teststo account for all other sources of random variation.4

Test ParametersThe test matrix for all of the supersonic data

obtained in this test entry is shown in Table 4. Themodel angle of attack ranged from –5 to 20 degrees, andthe model angle of sideslip ranged from –5 to 5 degrees.Data were obtained at Mach 2.3, 2.7, 3.5, and 4.5 forthe baseline, the Ames, the canted, and the shelfconfigurations.

Computational Methods

CFD calculations were performed for selectedconfigurations over a range of angle of attack and Machnumber to complement the experimental database andto provide data at flight conditions. An inviscid Eulercode, FELISA, and a finite-volume, Navier-Stokessolver, USM-3D, were used to obtain aerodynamic data.

The FELISA5 (Finite Element LangleyImperial College Swansea Ames) software packagecombines a series of codes that generate unstructuredtetrahedral grids over complex three-dimensionalgeometries and solve the steady three-dimensional Eulerequations on these grids. Unstructured mesh methodsare ideally suited for rapid analysis, as was required forearly parametric assessment of control surfaceeffectiveness, because they have the capability togreatly reduce the time associated with grid generationon complex configurations as compared to traditionalblock-structured methods. The mesh generator withinthe FELISA package carries out the discretization of thethree-dimensional computational domain intotetrahedra. The FELISA unstructured mesh flow solveruses an edge-based finite-volume formulation. Fluxesare computed using a flux vector splitting scheme thatis capable of representing constant enthalpy solutions.

4

USM3D6 is a three-dimensional, cell-centered,finite volume Euler and Navier-Stokes flow solver.Computations are done on unstructured meshes usingthe tetrahedral grid generator VGRIDns. Inviscid fluxquantities are computed across each cell face usingRoe’s flux-difference splitting. Spatial discretization isaccomplished using an analytical formulation forcomputing solution gradients within tetrahedral cells.The solution is advanced to a steady state condition byan implicit backward Euler time-stepping scheme.

Computational results were obtained for allfour configurations at Mach numbers of 2.3, 2.7, 3.5,and 4.5. Inviscid calculations were performed on theforebody alone as well as the full wind tunnel model.Viscous solutions were obtained for the wind tunnelmodel with a turbulent boundary layer. A subset ofthis data will be shown and compared with theexperimental data in subsequent sections. In Ref. 7Prabhu presents a thorough discussion of inviscidmethods, models, and data. Viscous methods andresults are presented in Ref. 8.

Results and Discussion

PrefaceIn a fast-paced study to satisfy a request for

supersonic experimental aerodynamic data on proposedSmart Lander configurations, over 140 runs wereobtained in Langley’s UPWT (Fig. 4). The mostrelevant data will be presented in the sections thatfollow. The standard aerodynamic coordinate system isused for all measurements and analysis (Fig. 5). Allmoment data is reduced about the vehicle nose (x,y,z-location (0,0,0) shown in Fig. 5). Longitudinal dataare of primary interest and are presented in coefficientform for both body axes, CN, CA, and Cm, and stabilityaxes CL, CD, and L/D.

Mach Number EffectsAll configurations were run at sideslip to obtain

lateral-directional data, and these data showed littlevariation with Mach number. Data showed allconfigurations (baseline and asymmetric) to beessentially neutrally stable in roll (Clb≈0) and stable inyaw (Cnb≈0.002).

Baseline Configuration: Longitudinalaerodynamic coefficients are shown versus angle ofattack in Figs. 6(a)-6(f). Data is presented for fourMach numbers from 2.3 to 4.5 at a free streamReynolds number based on model maximum diameterof 1x106. Axial force is presented in coefficient formin Fig. 6(a). All CA data presented in this report areuncorrected for sting/base interference effects, which areexpected to be small based on the low pressures (Cpvalues were on the order of -1/M∞

2, as seen in earlierstudies on the Viking configuration in UPWT)measured within the sting shroud. Data at all Machnumbers show a maximum axial force at zero degreesangle of attack, as expected, with nearly symmetrictrends (within balance accuracy limitations) between ±5

degrees angle of attack. Axial force decreases rapidlywith increasing (or decreasing, as one could infer fromsymmetry considerations) angle of attack. Two distinctgroupings of axial force trends are noticed. Data atMach 2.3 and 2.7 lie in an upper band with amaximum CA of approximately 1.54, while data atMach 3.5 and 4.5 lie in a lower band with a maximumof just over 1.51. Inviscid calculations on the forebodyonly7 show curves of similar shape but with nearlyequally spaced increments in CA with a correspondingchange in Mach number. One can look to viscouseffects, aftbody effects, or some combination of both toexplain the small but non-linear Mach effects seen forbaseline axial force measurements. Due to test timelimitations, no diagnostics were performed toinvestigate the state of the boundary layer or aftbodyflow separation/reattachment patterns on theexperimental model. Both would be recommended forfurther study to better understand the aforementionedaxial force trends. As expected for a blunt body at lowto moderate angles of attack, the coefficient of drag, CD,shown in Fig. 6(e), mirrors the behavior of CA for allangles of attack and Mach numbers tested.

Fig. 6(b) shows normal force as a function ofangle of attack for all supersonic Mach numbers tested.Magnitudes of CN are very small compared with thoseof CA. Data for all Mach numbers show near zeronormal force at a=0°, as expected for a symmetricconfiguration. However, the slope of the normal forcecurve at a=0° is negative for M=2.3 and M=2.5. AtM=3.5 and M=4.5. CNa is positive in and around a=0°and only slighter smaller in magnitude than at largerangles of attack. For the two lowest Mach numbers, anegative CNa at a=0° leads to a negative normal forcefor small (less than approximately 2°) angles of attack.While somewhat counter-intuitive, this trend wasobserved for UPWT data on the Viking aeroshell9 atlow supersonic Mach numbers. Inviscid forebody CFDresults show no evidence of these trends for any Machnumber while viscous solutions8 (to be presented anddiscussed in the section that follows), shows slightlynegative values of CNa for 0°<a<5°. This againhighlights likely effects of viscosity and aftbodygeometry on configuration aerodynamics. Whileinteresting from a fluid dynamics perspective, themagnitudes of CN are so small as not to show anysignificant effects on lift, drag, and pitching momentcharacteristics as seen in Figs. 6(c)-6(f).

Pitching moment for the baseline configuration,shown in Fig. 6(c), is stable and nearly linear for allangles of attack and Mach numbers. Cm at a=0° isnominally zero and Cma is approximately -0.0024 forall data. Any variations in Cm and Cma with Machnumber are over an order of magnitude smaller than thepitching moment increments of any control surfacetested (to be shown in later sections of this paper).Figs 6d and 6f show lift coefficient, CL, and lift-to-dragratio, L/D, also to be nearly linear with angle of attackand independent of Mach number variation for most ofthe data set. Slight Mach number effects are seen in CLfor a>10°, a direct result of the previous Mach number

5

effects for CA. Lift is zero at a=0° for allconfigurations. For a<10°, CLa is approximately–0.024 and for a>10° CLa is approximately –0.016.Fig. 6(f) shows that the baseline achieves its targetmagnitude of L/D=0.25 at a≈16°.

Asymmetric Configurations: The detaileddiscussion in the preceding section on the effects ofangle of attack and Mach number on the baselineconfiguration is equally applicable to all threeasymmetric configurations. The coefficient trends aresimilar for all configurations but differ in magnitudedue to the various fixed control surfaces. The increasesin axial force due to increased planform area have aprominent effect on CL, CD, but a lesser effect on L/D.The asymmetry in each configuration yields a non-zerotrim angle, as desired. Comparison of the magnitude ofthe aforementioned changes in lift, drag, and pitchingmoment for the asymmetric configurations will bepresented in a subsequent section.

Comparison with Computational ResultsFigures 7(a)-7(f) show comparison of the

previously discussed experimental results for thebaseline configuration with both inviscid and viscouscomputational data at Mach 2.3. Inviscid calculationswere performed initially for the forebody only toquickly generate solutions on multiple configurationsfor trade study requirements. To evaluate thecontributions of the aftbody, additional inviscidcomputations were performed on the complete windtunnel configuration (full model with a support sting).Both sets of inviscid axial force data are shown in Fig.7(a). Although trends are similar, inviscid forebody-alone solutions fall well below measured values ofaxial force across the entire angle of attack range.Corresponding inviscid solutions on the complete windtunnel model show very good agreement withexperimentally obtained CA data for all angles of attack.The fully viscous solutions from USM-3D also showexcellent axial force agreement with experimental data,especially at the highest angles of attack.

Figure 7(b) shows CN versus alpha for thebaseline configuration with accompanying inviscid andviscous calculations. Neither forebody nor full-modelinviscid solutions show evidence of the non-lineartrends for normal force observed around a=0°. Asmentioned previously, the viscous solutions show achange in sign for CNa at low angles of attack, and thusa more detailed analysis of viscous computations mayprovide some insight into experimentally observedtrends. CNa exhibits a noticeable increase for a>15°,which is tracked well by both inviscid and viscous full-model solutions, while the inviscid forebody alone datashows no change in slope. This increase in normalforce may be due to increased pressures on the aftbodyat high angles of attack. Again, values of CN are quitesmall for all angles of attack and have very smallcontributions towards lift, drag, and pitching momentcharacteristics.

Figure 7(c) shows very little difference inpitching moment characteristics between forebody and

full-model inviscid solutions. All computationalpredictions for Cm agree well with experimental dataover the entire angle of attack range. Drag predictions(Fig. 7(e)) are similar to those for axial force, while liftcoefficient and lift-to-drag ratio data are predicted wellby the computational data for the full body. Based onthe agreement between inviscid and viscous solutionsfor the wind tunnel configuration and their agreementwith experimental data, it appears that viscous effect donot make a first order contribution to the aerodynamiccharacteristics of this configuration.

The level and nature of agreement betweencomputational and experimental data is similar for otherMach numbers and configurations. The reader is againreferred to References 7 and 8 for further comparison ofcomputational and experimental results.

Configuration EffectsOne of the stated objectives of this test

program was to provide aerodynamic data for threeasymmetric tab/shelf configurations to allow vehicledesigners to assess their feasibility for meeting missionrequirements. The aerodynamic performance of theseconfigurations is only one of many areas where tradestudies will be performed to optimize vehicle missionperformance. Thus, it is outside the scope of this workto make final determinations on the “goodness” orfeasibility of one configuration versus another.

Figures 8(a)-8(f) present comparisons oflongitudinal data for all four configurations tested: thebaseline, the Ames, the canted tab, and the shelf. Theaxial force characteristics (Fig. 8(a)) remain essentiallythe same in trend, but with a positive increment thatscales proportionally with projected control surface area(see Table 2). The normal force characteristic (Fig.8(b)) for the three asymmetric configurations can alsobe viewed as incrementally different from the baselinevalues. The Ames configuration has the largest area ofthe three control surfaces and thus shows the largestnegative increment in normal force. Normal forceincrements for all configurations are quite small inmagnitude and vary only slightly with angle of attack.

Pitching moment characteristics for all fourconfigurations are shown in Fig. 8(c). As statedpreviously, the goal for these fixed control surfaces isto allow the configuration to trim at a non-zero angleof attack in order to attain sufficient L/D values tosatisfy mission requirements. All surfaces do show ashift in trim angle of attack, which again showscorrelation to control surface area. The canted tabprovides an 11° shift in trim angle, the shelf shows a13° shift in trim angle, and the Ames tab shows a 16°shift in trim angle. All asymmetric configurationsshow a nearly constant positive increment in pitchingmoment, maintaining similar stability levels for allconfigurations. It is important to note that thepitching moment data presented herein are reduced abouta non-realistic center of gravity location, namely thevehicle nose. When transferred to a realistic (further aft)position, the trim angle increments would become evenlarger and the asymmetric configurations would show a

6

lesser level of stability for all angles of attack. Forexample, Prabhu7 presents both measured data andinviscid CFD pitching moment results reduced about areference point at full-scale axial station –0.8659m(with no radial offset). This aft center of gravitylocation was derived from initial performance andpackaging estimates. The Ames surface yields a trimpoint at a≈17°, the Canted tab yields a trim point ata≈13°, and the Shelf yields a trim point at a≈14° forthis aft c.g. location. The corresponding L/D valuesfor these trim points would meet or exceed statedperformance metrics.

Corresponding to normal and axial force, CLand CD (Figs. 8(d) and 8(e)) show well-behavedincrements over the angle of attack range. Fig. 8(f)shows that L/D varies only slightly among the onesymmetric and three asymmetric configurations and thatthe trim angle to obtain L/D values in the 0.22 to 0.25range are nominally 13 to 16 degrees. With a realisticaft c.g. placement, these data have shown this to be anattainable trim angle range for all control surfaces.

For completeness, Figs. 9, 10, and 11 showcomparison of the four configurations at Mach 2.7,3.5, and 4.5. No experimental data were obtained forthe Ames configuration at M=4.5. As Mach numbereffects were shown to be small, the discussion of datatrends at M=2.3 are generally applicable to the higherMach number data. As Mach number increases CA andCD curves tend to be generally smoother, while CN, Cm,and L/D tend to be more linear.

Concluding Remarks

A 0.0376-scale model of a proposed Mars ’07Smart Lander configuration was tested in Leg II of theUnitary Plan Wind Tunnel at the NASA LangleyResearch Center. The objectives of the testing were toestablish the stability and control characteristics of theproposed baseline vehicle and to assess the aerodynamicfeasibility of three fixed control surfaces. Over 140runs were obtained on four configurations, the baseline,the Ames surface, the Canted tab, and the Shelf, over aMach range of 2.3 to 4.5. Mach effects were shown tobe small on vehicle aerodynamic characteristics.Inviscid and viscous CFD predictions for the windtunnel configuration geometry at wind tunnelconditions agreed well with measured data. Based onresults from 3 different shelf/tab configurations, a fixedcontrol surface appears to be a feasible concept formeeting aerodynamic performance metrics necessary tosatisfy mission requirements.

Acknowledgements

The authors gratefully acknowledge thefollowing persons/groups for their contributions to thiswork: Ramadas K. Prahbu, Paresh Parikh, Mary KaeLockwood, Glen J. Bobskill, Richard Wheless, John

Micol, Scott Goodliff, and The staff at the UPWTtunnel.

References

1. Lockwood, M.K., et al.: “Entry Configurations andPerformance Comparisons for the Mars SmartLander “AIAA 2002-4407, August 2002.

2. Lockwood, M.K., Powell R.W., Graves, C.A., andCarman, G.L.: “Entry System DesignConsiderations for Mars Landers, “ AmericanAstronautical Society Paper No. 01-023, 24th AASGuidance and Control Conference, January 31-February 4 2001, Breckenridge, CO.

3. Erickson, G.E.: “Overview of SelectedMeasurement Techniques for Aerodynamic Testingin the NASA Langley Unitary Plan Wind Tunnel,”AIAA 2000-2396, June 2000.

4. Hemsch, M.J.: private communication, June 2002.5. Peiro, J, Peraire, J., and Morgan K.: “FELISA

System Reference Manual and Users Guide,”NASA CP-3291, May 1995.

6. Frink, N.T., and Pirzadeh, S.Z.: “TetrahedralFinite-Volume Solutions to the Navier-StokesEquations on Complex Configurations,” NASATM-1998-208961, December 1998.

7. Prabhu, R.K.: “An Inviscid Computational Studyof Three ’07 Mars Lander Aeroshell ConfigurationsOver a Mach Number Range of 2.3 to 4.5,” NASACR-2001-211266.

8. Bobskill, G.J, Parikh, P.C., Prabhu, R.K., Tyler,E.D.: “Aerodynamic Database Development forMars Smart Lander Vehicle Configurations,” AIAA2002-4411, August 2002.

9. Blake, W.W.: "Experimental AerodynamicCharacteristics of the Viking Entry Vehicle overthe Mach Range 1.5 –10.0,” NASA TR-3720106,April 1971.

7

Table 1. Reference Dimensions

Dimension Full-Scale .0376-ScaleSref 138.9 ft2 28.274 in2

Lref (vehicle diameter) 13.3 ft 6 inXc.g.ref (at nose) 0 ft 0 inYc.g.ref (at nose) 0 ft 0 inZc.g.ref (at nose) 0 ft 0 in

Table 2. Control Surface Areas

Dimension Full-Scale .0376-ScaleSAmes 9.372 ft2 1.908 in2

SCanted 5.659 ft2 1.152 in2

SShelf 7.628 ft2 1.553 in2

Table 3. Balance Uncertainties for LaRC 2008

NF(lbs) AF(lbs) PM(in-lbs) RM(in-lbs) YM(in-lbs) SF(lbs)60 180 150 30 120 600.1 0.32 0.11 0.27 0.15 0.05

Maximum LoadAccuracy (%full-scale)

Accuracy (Load) 0.06 0.576 0.165 0.081 0.18 0.03Mach q∞ (psi) CN ±2s

AccuracyCA ±2sAccuracy

CM ±2sAccuracy

CRM ±2sAccuracy

CYM ±2sAccuracy

CY ±2sAccuracy

2.3 2.972 ±0.00071 ±0.00686 ±0.00033 ±0.00016 ±0.00036 ±0.000362.7 2.704 ±0.00078 ±0.00753 ±0.00036 ±0.00018 ±0.00039 ±0.000393.5 2.110 ±0.00101 ±0.00965 ±0.00046 ±0.00023 ±0.00050 ±0.000504.5 1.587 ±0.00134 ±0.01284 ±0.00061 ±0.00030 ±0.00067 ±0.00067

Table 4. UPWT Test Matrix for Mars ’07 Smart Lander Test 1735

Configuration Re/ftX106

Grit a b Mach 2.3 Mach 2.7 Mach 3.5 Mach 4.5

Baseline 2.0 None A11 0 Run 59 Run 64 Run 69 Run 76Baseline 2.0 None A22 2 Run 60 Run65 Run 70 Run 77Baseline 2.0 None A2 4 Run 61 Run66 Run 71 Run 78Baseline 2.0 None A33 B14 Run 149 Run 151 - -Ames 2.0 None A1 0 Run 81 Run 84 Run 88 -Ames 2.0 None A2 2 Run 82 Run 85 Run 89 -Ames 2.0 None A2 4 Run 83 Run 86 Run 90 -Ames 2.0 None A3 B1 - - - -Shelf 2.0 None A1 0 Run 93 Run 99 Run 103 Run 114Shelf 2.0 None A2 2 Run 94 Run 100 Run 105 Run 116Shelf 2.0 None A2 4 Run 95 Run 101 Run 106 Run 117Shelf 2.0 None A3 B1 Run 96 Run 102 Run 107 Run 115Canted 2.0 None A1 0 Run 119 Run 123 Run 128 Run 134Canted 2.0 None A2 2 Run 120 Run 124 Run 129 Run 135Canted 2.0 None A2 4 Run121 Run 125 Run 130 Run 136Canted 2.0 None A3 B1 Run 122 Run 126 Run 131 Run 1381A1: Alpha –5,-4,-3,-2,-1,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,202A2: Alpha –5,-2,0,2,4,6,8,10,11,12,13,14,16,18,203A3: Trim Angle of Attack for each configuration4B1: Beta –5,-4,-3,-2,-1,0,1,2,3,4,5

8

Figure 1. Schematic of Proposed Mars ’07 LanderConfiguration

Figure 2. Schematic of Proposed ControlSurfaces for Smart Lander

Figure 3. Photograph of Wind Tunnel Modelwith Balance Shroud

Figure 4(a). Installation of Mars Smart LanderBaseline Model in the UPWT at LaRC.

Figure 4(b). Schlieren Photograph of Mars SmartLander Model in the UPWT at LaRC

Figure 5. Aerodynamic Coordinate System.

9

1.35

1.4

1.45

1.5

1.55

-5 0 5 10 15 20

Mach = 2.3Mach = 2.7Mach = 3.5Mach = 4.5

CA

a (deg)

± Balance Accuracy for M=2.3

Figure 6(a). Effect of Mach Number on Measured AxialForce Coefficient for Baseline Configuration

-0.02

0

0.02

0.04

0.06

0.08

-5 0 5 10 15 20


CN

a (deg)


Figure 6 (b) . Effect of Mach Number on MeasuredNormal Force Coefficient for Baseline Configuration

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

-5 0 5 10 15 20


Cm

a (deg)


Figure 6 (c ) . Effect of Mach Number on MeasuredPitching Moment Coefficient for Baseline Configuration

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

-5 0 5 10 15 20


CL

a (deg)Figure 6(d) . Effect of Mach Number on Measured LiftForce Coefficient for Baseline Configuration

1.3

1.35

1.4

1.45

1.5

1.55

-5 0 5 10 15 20


CD

a (deg)Figure 6(e). Effect of Mach Number on Measured DragForce Coefficient for Baseline Configuration

-0.4

-0.3

-0.2

-0.1

0

0.1

-5 0 5 10 15 20


L/D

a (deg)Figure 6(f). Effect of Mach Number on Measured Lift-to-Drag Ratio for Baseline Configuration

10

1.25

1.3

1.35

1.4

1.45

1.5

1.55

1.6

-5 0 5 10 15 20

Experimental Data -UPWTCFD - FELISA Forebody OnlyCFD - FELISA Full ModelCFD - USM3D Full Model

CA

a (deg)


Figure 7(a). Comparison of Measured and PredictedAxial Force Coefficient for Baseline Configuration atMach 2.3

-0.02

0

0.02

0.04

0.06

0.08

-5 0 5 10 15 20


CN

a (deg)


Figure 7(b). Comparison of Measured and PredictedNormal Force Coefficient for Baseline Configuration atMach 2.3

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

-5 0 5 10 15 20


Cm

a (deg)


Figure 7(c). Comparison of Measured and PredictedPitching Moment Coefficient for Baseline Configurationat Mach 2.3

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

-5 0 5 10 15 20


CL

a (deg)Figure 7(d). Comparison of Measured and PredictedLift Force Coefficient for Baseline Configuration atMach 2.3

1.2

1.25

1.3

1.35

1.4

1.45

1.5

1.55

1.6

-5 0 5 10 15 20


CD

a (deg)Figure 7(e). Comparison of Measured and PredictedDrag Force Coefficient for Baseline Configuration atMach 2.3

-0.4

-0.3

-0.2

-0.1

0

0.1

-5 0 5 10 15 20


L/D

a (deg)Figure 7(f). Comparison of Measured and PredictedLift-to-Drag Ratio for Baseline Configuration at Mach2.3

11

1.45

1.5

1.55

1.6

1.65

-5 0 5 10 15 20

BaselineAmesCantedShelf

CA

a (deg)


Figure 8(a). Effect of Fixed Control Surface onMeasured Axial Force Coefficient at Mach 2.3

-0.04

-0.02

0

0.02

0.04

0.06

0.08

-5 0 5 10 15 20


CN

a (deg)


Figure 8(b). Effect of Fixed Control Surface onMeasured Normal Force Coefficient at Mach 2.3

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

-5 0 5 10 15 20


Cm

a (deg)


Figure 8(c). Effect of Fixed Control Surface onMeasured Pitching Moment Coefficient at Mach 2.3

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

-5 0 5 10 15 20


CL

a (deg)Figure 8(d). Effect of Fixed Control Surface onMeasured Lift Force Coefficient at Mach 2.3

1.35

1.4

1.45

1.5

1.55

1.6

1.65

-5 0 5 10 15 20


CD

a (deg)Figure 8(e). Effect of Fixed Control Surface onMeasured Drag Force Coefficient at Mach 2.3

-0.4

-0.3

-0.2

-0.1

0

0.1

-5 0 5 10 15 20

Baseline AmesCantedShelf

L/D

a (deg)Figure 8(f). Effect of Fixed Control Surface onMeasured Lift-to-Drag Ratio at Mach 2.3

12

1.4

1.45

1.5

1.55

1.6

1.65

-5 0 5 10 15 20


CA

a (deg)Figure 9(a) . Effect of Fixed Control Surface onMeasured Axial Force Coefficient at Mach 2.7

-0.04

-0.02

0

0.02

0.04

0.06

0.08

-5 0 5 10 15 20


CN

a (deg)Figure 9 (b) . Effect of Fixed Control Surface onMeasured Normal Force Coefficient at Mach 2.7

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

-5 0 5 10 15 20


Cm

a (deg)Figure 9 (c ) . Effect of Fixed Control Surface onMeasured Pitching Moment Coefficient at Mach 2.7

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

-5 0 5 10 15 20


CL

a (deg)Figure 9(d) . Effect of Fixed Control Surface onMeasured Lift Force Coefficient at Mach 2.7

1.35

1.4

1.45

1.5

1.55

1.6

1.65

-5 0 5 10 15 20


CD

a (deg)Figure 9 (e ) . Effect of Fixed Control Surface onMeasured Drag Force Coefficient at Mach 2.7

-0.4

-0.3

-0.2

-0.1

0

0.1

-5 0 5 10 15 20


L/D

a (deg)Figure 9 ( f ) . Effect of Fixed Control Surface onMeasured Lift-to-Drag Ratio at Mach 2.7

13

1.4

1.45

1.5

1.55

1.6

1.65

-5 0 5 10 15 20


CA

a (deg)Figure 10(a ) . Effect of Fixed Control Surface onMeasured Axial Force Coefficient at Mach 3.5

-0.04

-0.02

0

0.02

0.04

0.06

0.08

-5 0 5 10 15 20


CN

a (deg)Figure 1 0 ( b ) . Effect of Fixed Control Surface onMeasured Normal Force Coefficient at Mach 3.5

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

-5 0 5 10 15 20


Cm

a (deg)Figure 1 0 ( c ) . Effect of Fixed Control Surface onMeasured Pitching Moment Coefficient at Mach 3.5

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

-5 0 5 10 15 20


CL


1.3

1.35

1.4

1.45

1.5

1.55

1.6

1.65

-5 0 5 10 15 20


CD

a (deg)Figure 1 0 ( e ) . Effect of Fixed Control Surface onMeasured Drag Force Coefficient at Mach 3.5

-0.4

-0.3

-0.2

-0.1

0

0.1

-5 0 5 10 15 20


L/D

a (deg)Figure 1 0 ( f ) . Effect of Fixed Control Surface onMeasured Lift-to-Drag Ratio at Mach 3.5

14

1.35

1.4

1.45

1.5

1.55

1.6

-5 0 5 10 15 20

BaselineCantedShelf

CA

a (deg)Figure 11(a ) . Effect of Fixed Control Surface onMeasured Axial Force Coefficient at Mach 4.5

-0.04

-0.02

0

0.02

0.04

0.06

0.08

-5 0 5 10 15 20

BaselineCantedShelf

CN

a (deg)Figure 1 1 ( b ) . Effect of Fixed Control Surface onMeasured Normal Force Coefficient at Mach 4.5

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

-5 0 5 10 15 20

BaselineCantedShelf

Cm

a (deg)Figure 1 1 ( c ) . Effect of Fixed Control Surface onMeasured Pitching Moment Coefficient at Mach 4.5

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

-5 0 5 10 15 20

BaselineCantedShelf

CL


1.3

1.35

1.4

1.45

1.5

1.55

1.6

-5 0 5 10 15 20

BaselineCantedShelf

CD

a (deg)Figure 1 1 ( e ) . Effect of Fixed Control Surface onMeasured Drag Force Coefficient at Mach 4.5

-0.4

-0.3

-0.2

-0.1

0

0.1

-5 0 5 10 15 20

BaselineCantedShelf

L/D

a (deg)Figure 1 1 ( f ) . Effect of Fixed Control Surface onMeasured Lift-to-Drag Ratio at Mach 4.5

aiaa-2002-4409 supersonic aerodynamic characteristics of

Documents