On-Orbit Propellant Resupply Options for Mars Exploration

8/7/2019 On-Orbit Propellant Resupply Options for Mars Exploration

1/15

1

IAC-06-D1.1.01

On-Orbit Propellant Resupply Options for Mars ExplorationArchitectures

Christopher TannerGeorgia Institute of Technology, USA

[email protected]

James Young, Robert Thompson, Dr. Alan WilhiteGeorgia Institute of Technology, USA

[email protected], [email protected], [email protected]

ABSTRACTA report detailing recommendations for a transportation architecture and a roadmap for U.S.exploration of the Moon and Mars was released by the NASA Exploration Systems ArchitectureStudy (ESAS) in November 2005. In addition to defining launch vehicles and various aspects of alunar exploration architecture, the report also elaborated on the extent of commercial involvement infuture NASA activities, such as cargo transportation to the International Space Station. Anotherpotential area of commercial involvement under investigation is the delivery of cryogenic propellantsto low-Earth orbit (LEO) to refuel NASA assets as well as commercial assets on orbit. The ability toresupply propellant to various architecture elements on-orbit opens a host of new possibilities withrespect to a Mars transportation architecture first and foremost being the ability to conduct aMartian exploration campaign without the development of expensive propulsion systems such as

nuclear thermal propulsion. In-space propellant transfer in the form of an orbiting propellant depotwould affect the sizing and configuration of some currently proposed vehicles such as the EarthDeparture Stage (EDS) and the Mars Transit Vehicle (MTV). In addition, it would influence theoverall affordability and sustainability of a long-term Mars exploration campaign. To assess theseconsequences, these vehicles and their various stages are modeled to approximate the ESASperformance figures using a combination of analogous systems and physics-based simulation. Wellestablished modeling tools -- such as POST for trajectory optimization, APAS for aerodynamics,NAFCOM for cost modeling, and Monte Carlo analysis for technology advancement uncertainty --are used to perform these analyses. To gain a more complete view of the effects of an on-orbitpropellant refueling capability, a reference Mars mission is developed and compared to an equivalentmission without refueling capability. Finally, the possibility of propellant resupply in Mars orbit is alsodiscussed along with its implications on the sustainability of a long-term Mars exploration

architecture.

NOMENCLATUREC3 Excess hyperbolic energy squared (km2/s2)DDT&E Design, Development, Testing, and EvaluationDRM Design Reference MissionEDL Entry, Descent, and LandingEDS Earth Departure Vehicle


2/15

2

EOR Earth-Orbit RendezvousERV Earth Return VehicleESAS Exploration Study Architecture StudyFY Fiscal YearIMLEO Initial Mass in Low Earth OrbitISS International Space Station

JSC Johnson Space CenterKSC Kennedy Space CenterLEO Low Earth OrbitLH2 Liquid HydrogenLOX Liquid OxygenMOI Mars Orbit InsertionMPLM Multi-Purpose Logistics ModuleMT Metric TonsNAFCOM NASA / Air Force Cost ModelNASA National Aeronautics and Space AdministrationNTR Nuclear Thermal RocketO/F Oxidizer-to-Fuel Ratio

POST Program to Optimize Simulated TrajectoriesPRM Propellant Resupply ModuleSTS Shuttle Transportation SystemSVLCM Spacecraft / Vehicle Level Cost ModelTCM Trajectory Correction ManeuverTFU Theoretical First UnitTMI Trans-Mars InjectionVAB Vehicle Assembly BuildingVSE Vision for Space Exploration

INTRODUCTIONThe ambitious goal of the return of humans

to the Moon and the eventual transportationof humans to Mars was put forth in thePresidents Vision for Space Exploration(VSE)[1]. To develop an idea as to how thiscould be accomplished, the ExplorationSystems Architecture Study (ESAS) wasperformed, resulting in first-cuts at therecently labeled Orion crew vehicle andAres I and V launch vehicles. The launchvehicles have undergone some changessince the ESAS Final Report[2], but theoverall mission of using the vehicles to

place humans and their associated cargosafely on the Moon has not changed. Sincethe development of two new launch vehicleswill likely be an arduous and expensivetask, one must look at how the proposedvehicles will fit into a human-Marscampaign.

FRAMING THE PROBLEMArchitecture Elements

To establish a point of reference, the DesignReference Mission (DRM) v3.0[3]developed by NASA JSC is used as abaseline Martian architecture to manifestthe various launches and trans-Marsinjection (TMI) stages. This architecture issummarized in Table 1.

MassPayload

lbm kgEarth Return Vehicle (ERV) 163,301 74,072Cargo Lander 145,600 66,043Crew Lander 134,054 60,806

NTR TMI Stage (max) 168,874 76,600Table 1: DRM payloads and nuclear thermal rocket

(NTR) TMI stage[3].

Given the large payloads outlined by theDRM, the primary focus of this study will beon using the proposed heavy lift launchvehicle, the Ares V, and its Earth Departure


3/15

3

Stage (EDS). The assumed characteristicsof the Ares V are detailed in Table 2.

Boosters 2x 5-seg RSRMBooster Propellant 1,400 klbm eachCore Engines 5x RS-68

Isp (vac) 410 sThrust (vac) 745 klbf eachArea Raito 21.5Weight 14.6 klbm each

Core Propellant 3,000 klbmCore Diameter 33 ftEDS Engine 1x J-2X

Isp (vac) 448 sThrust (vac) 274 klbfArea Ratio 40.0Weight 3.0 klbm

EDS Propellant 508 klbmEDS Diameter 27.5 ftFairing Weight 13 klbmDelivery Orbit 30 x 160 nmi

Table 2: Ares V configuration and performancespecifications.

The DRM baselines a nuclear thermalrocket (NTR) TMI stage that is launchedseparately from each payload. However,forthis study, the EDS is used as the TMIstage in lieu of the DRM NTR stage. Asdesigned, the Ares V uses the EDS as bothan upper stage, to insert the payload into aparking orbit in LEO, and as an Earthdeparture stage, to inject the payload intothe proper transfer orbit. However, apartially used EDS does not have thecapability to inject all of the DRM payloadsinto a Mars transfer orbit. One way ofremedying this problem is to refuel the EDSin LEO prior to the injection burn using apre-positioned asset, such as a propellantdepot similar to Figure 1. There are severalunique advantages to the development andimplementation of a propellant depot:

1. Commercial launch services could beutilized to re-supply the depot

2. The depot could be used to re-fuelother manned or unmannedspacecraft

3. Enabling depot technologies, such aslow-boiloff and autonomousrendezvous & docking capabilities,could be used to enhance otherplanetary or Earth science spacecraft.

Figure 1: OASIS cryogenic propellant depot[5].

To fully understand the usefulness of arefueled EDS, the Earth departure C3 fromthe ESAS reference 30 x 160 nmi orbit isplotted versus the injected payload in Figure2. This plot assumes that the payload massindicated along the abscissa is launchedfrom KSC into a 28.5o inclination and that

the launched payload is the same as theinjected payload (i.e. no Earth-orbitrendezvous to add additional payload to thedeparture stage). The minimum departureC3 indicated is the minimum Earthdeparture C3 that occurs within the 2030-2040 timespan.

0

50

100

150

200

250

0 20 40 60 80 100 120 140

Payload Mass (MT)

C3(km

2/s

2)

No Refuel

Full Refuel

D

RME

RV

D

RMCrew

D

RMCargo

Min Departure C3

Figure 2: Available Earth departure C3 versuspayload mass for a partially expended EDS and acompletely fuelled EDS.

Clearly, a fully fuelled EDS has much higher

capability but that also means a propellantdepot would have to sustain over 230 MT ofpropellant (a complete EDS refuel) on-orbit,which could be prohibitive depending on thepropellant boiloff rate and capability of thecommercial re-supply missions. Thus, thisstudy focused on the minimum amount ofre-fuel needed to accomplish a givenmission i.e. the DRM mission with the


4/15

4

goal of providing the minimum (andsubsequently the most affordable) depotarchitecture needed to send humans toMars.

Defining the Trade Studies

Three separate architectures areinvestigated in this study given the abovearchitectural elements. These trades arecompared against one another on aperformance and cost basis.

1. Aerocaptured payloads the DRMbaselines aerocapture for all threepayloads into an elliptical parking orbitat Mars. The only change in this tradeis the replacement of the NTR stagewith an EDS and a propellant depot inLEO.

2. Propulsively captured payloads aerocapture could require largedevelopment resources. If the EDScan perform a propulsive capture atMars after being refuelled in LEO, thenaerocapture may not necessarily berequired.

3. NTR stage it is not entirely clear ifthe use of the EDS and a propellantdepot is less expensive than usingNTR, as the DRM originally suggests.Thus, the NTR is independently

analyzed and compared against thefirst two architectures.

Interplanetary Trajectory OptimizationSlight modification of the DRM is necessaryto put it in the scope of the VSE. The DRMassumes the first Mars departure of itsEarth Return Vehicle (ERV) and CargoLander would occur on the 2011conjunction-class opportunity, with the crewsubsequently launched on the 2014opportunity. However, given the timeline of

the Space Shuttle retirement and lunarexploration program[2], a more realistictimeframe for the first Mars mission wouldrange from 2030 to 2040. Optimizedtrajectories from the Earth to Mars withinthis timeframe were analyzed usingJAQARs Swing-by Calculator[4] resulting inthree different trajectories per opportunity:

1. Minimum Earth Departure C3Earth

DepartureDate

DepartureC3

(km2/s

2)

MarsArrivalDate

ArrivalC3

(km2/s

2)

TransitTime

(days)02/24/31 8.237 01/11/32 31.164 32104/28/33 7.780 01/27/34 19.157 27406/26/35 10.199 01/04/36 7.273 192

09/06/37 14.848 10/06/38 11.148 39509/27/39 12.176 09/21/40 7.264 36010/20/41 9.813 09/01/42 6.183 31611/15/43 8.969 09/18/44 7.837 308

Table 3: Minimum Earth departure trajectories from2030-2043.

2. Minimum Mars Arrival C3Earth

DepartureDate

DepartureC3

(km2/s

2)

MarsArrivalDate

ArrivalC3

(km2/s

2)

TransitTime(days)

12/13/30 12.397 09/25/31 11.870 28604/20/33 9.305 11/05/33 10.962 19907/09/35 11.910 01/24/36 6.762 199

09/24/37 32.612 06/01/38 5.838 25010/25/39 28.739 07/17/40 5.576 26610/22/41 9.859 09/04/42 6.167 31711/14/43 8.974 09/14/44 7.801 305

Table 4: Minimum Mars arrival trajectories from 2030-2043.

3. Minimum Total C3Earth

DepartureDate

DepartureC3

(km2/s

2)

MarsArrivalDate

ArrivalC3

(km2/s

2)

TransitTime(days)

12/28/30 10.445 10/08/31 12.372 28404/15/33 8.996 10/31/33 11.072 19906/25/35 10.274 01/12/36 6.989 201

08/23/37 15.754 08/18/38 8.338 36009/22/39 12.444 08/30/40 6.281 34310/20/41 9.814 09/03/42 6.171 31811/14/43 8.974 09/14/44 7.801 305Table 5: Minimum total energy trajectories from 2030-

2043.

Table 3 serves as a minimum propulsiverequirement to get to Mars during the givenopportunity. Table 4 is included as it placesa lower bound on the Mars capture andentry, descent, and landing (EDL)requirements of the payload. Table 5

attempts to balance departure and arrivalenergy in order to provide a more practicalsolution than the previous two.

ANALYSIS AND RESULTSAres V Ascent TrajectoryThe Program to Optimize SimulatedTrajectories (POST) was used to simulateascent of the DRM payloads given the Ares


5/15

5

V configuration detailed above. Thissimulation essentially provided theremaining EDS capability once the givenpayload achieved orbit. Figure 3 is anascent trajectory obtained for the DRMCargo launch with an 800 psf maximum

dynamic pressure (max-q) constraintimposed on the trajectory. This constraintwas used assuming at least sometechnology advancement above the currentSpace Shuttle max-q of 650 psf and thelargest max-q witnessed by the Saturn V at777 psf[6].

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

0 100 200 300 400 500 600 700 800 900

Time (s)

We

ight(103

lbs

)

0

100

200

300

400

500

600

700

800

Altitu

de

(10

3f

t)

Weight

Altitude

Target Altitude

SRB Separation(116 sec)

Core Separation(343 sec) MECO

(857 sec)

0

100

200

300

400

500

600

700

800

0 100 200 300 400 500 600 700 800 900

Time (s)

Dynam

icPressure

(ps

f)

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

Ve

loc

ity(

fps

)

Dynamic Pressure

Inertial Velocity

Target Velocity

Figure 3: Ares V ascent trajectory for DRM Cargo

payload to 30 x 160 nmi x 28.5o

orbit.From the trajectory above, it is noticeable

that the EDS lofts above the desired altitudeas part of the optimized trajectory. Thislofting action is caused by the particularlylow thrust-to-weight ratio (0.38 to 0.40) ofthe modified EDS design which only usesone J-2X engine. The loft itself is a result ofthe optimization in which the EDS attemptsto minimize drag losses in order toaccelerate the payload to the desired

velocity. Increasing the number of engineson the EDS would eliminate the loft andcreate a much smoother trajectory.

EDS Capability for Aerocapture PayloadThe DRM v3.0 mission assumes that each

payload uses aerocapture technology todecelerate into an elliptical Mars orbit uponarrival (Figure 4). Additionally, it assumesthat each of the nuclear TMI stages is usedonly once and is discarded after the TMImaneuver is performed. All of the trajectorycorrection maneuvers (TCMs), orbit adjustand trim maneuvers after aerocapture, andde-orbit burns for payload entry areperformed by the propulsive descent systemonboard each payload. Finally, the nuclearTMI stages in the DRM are launched

separately from the payloads, requiring sixheavy launches per mission as well as threeEarth-orbit rendezvous (EORs).

Figure 4: DRM triconic aerocapture vehicle withestimated L/D of 0.6[3].

In this study, each payload is launched ontop of the Ares V / EDS stack, thus requiringthree heavy launches per mission plus thenecessary depot resupply launches(explained in detail later). Although stackingthe DRM payloads on top of the EDScreates a very tall launch vehicle, currentfigures for the Ares V core and EDS heightscoupled with the cited payload dimensionsin the DRM v3.0 report do not violate themaximum door height of the Vehicle

Assembly Building (VAB) at KSC, as shownin Table 3. As a reference point, the SaturnV measured approximately 364 ft in heightand the Shuttle Transportation System(STS) measures 184 ft in height. It shouldbe noted that a flight stability analysis wasnot performed on this configuration toensure a stable ascent vehicle.


6/15

6

HeightElement

Element CumulativeAres V Core 212.6 ft 212.6 ftEDS 73.4 ft 286.0 ftPayload + Shroud (max) 91.9 ft 377.9 ftVAB Doors (max)[7] 456 ftClearance 78.1 ft

Table 6: Estimated height of the Ares V stack withDRM payload compared to the maximum heightof the VAB high bay doors.

In addition to the analysis performed in theDRM concerning the feasibility of a largeaerocaptured Mars payload, an additionalsource was used to confirm that thebaseline DRM aerocapture vehicle hadsufficient lift-to-drag (L/D) to aerocapture atthe entry velocities witnessed throughoutthe 2030-2040 timeframe (5.477.45 km/s).According to Braun and Powell[8], a vehiclewould require an L/D of 0.5 to capture withinthe velocity range of 6.08.5 km/s usingatmospheric guidance control throughoutthe trajectory. Given that the DRM baselineaerocapture vehicle has an L/D of 0.6, it isassumed that the vehicle can aerocaptureat velocities lower than 6.0 km/s as well asretain the capability to aerocapture atvelocities up through 7.45 km/s withatmospheric guidance. The final captureorbit around Mars is assumed to be a highlyelliptical 500 x 33,640 km elliptical orbit,

resulting in a period of 25.67 hours thesame as the Martian day.

From the above C3 requirements andreference parking orbit, the required amountof propellant to initiate transfer to Mars iscalculated. The ascent simulation performedin POST provided the amount of EDSpropellant remaining after achieving orbit.The difference between the remaining EDSpropellant and the amount required toinitiate Mars transfer is the amount of

propellant that needs to be replenished viathe propellant depot. Figure 5 plots theamount of refuel propellant required foreach DRM element at each opportunity.Three points for each payload at eachopportunity translate into the minimumdeparture C3 (min refuel mass), minimumarrival C3 (max refuel mass), and minimum

total C3 (mid refuel mass). When only oneor two points are shown, refuel mass for thetrajectories are approximately equal.

0

10

20

30

40

50

60

70

80

90

100

09/28 06/31 02/34 11/36 08/39 05/42 02/45

Earth Depature Date

RefuelPropellantMass(M

T) ERV

CargoCrew

Figure 5: Required propellant resupply mass given

payload aerocapture. The Crew payload is offsetfrom its corresponding ERV and Cargo payload

by one opportunity.Figure 5 illustrates that a propellant depotcan vary widely in capacity depending onthe departure date and the payload mass.However, given the DRM aerocapturevehicle, only the minimum departure C3case (minimum refuel propellant mass)need be considered.

Figure 6 illustrates the total refuel propellantrequired per mission (one mission equals

one ERV and Cargo departure on oneopportunity with the Crew departure oneopportunity later) considering only thelowest Earth departure energy cases. Theplot shows that, even for the least energeticmission to Mars, approximately 52.5 MT ofpropellant must be boosted into orbit overthe course of two years. These figures donot include propellant boiled off from thedepot while waiting for the EDS to launchand refuel.


7/15

7

0

10

20

30

40

50

60

70

80

90

100

2031/33 2033/35 2035/37 2037/39 2039/41 2041/43

Mission

RefuelPropellantMass(MT) Crew

Cargo

ERV

52.5 55.0

73.3

87.6

72.1

60.9

Figure 6: Total refuel propellant required for

aerocaptured payloads assuming a minimumdeparture C3 trajectory.

EDS Capability for Propulsive CapturePayloadIt should be noted that the refuel masses for

the aerocapture cases do not completely fillthe EDS i.e. the EDS has additionalcapability if provided more propellant. MoreEDS propellant could be used as analternative to aerocapture allowing thepayload to be propulsively captured into aMars parking orbit. This approach holds aset of distinct advantages:

1. Aerocapture has never beendemonstrated and a more riskymaneuver than propulsive capture.

2. Aerocapture requires additional

hardware, increasing the mass of thevehicle. Additionally, it requires theentry vehicle to have a prescribedamount of L/D, which could lead topackaging constraints both inside theaeroshell and on top of the launchvehicle.

3. NASA may not have enoughresources to support the simultaneousdevelopment of Mars surfacehardware and descent capability alongwith aerocapture technology.

For these reasons, the concept of apropulsive capture should be investigated.To simulate payloads without aerocapturehardware, each vehicle mass was reducedby about 8%, resulting in the payloadmasses in Table 7. The DRM found that theheatshield required for aerocapture andentry of these large payloads resulted in

about 16% of the total vehicle mass. It isassumed in this study that eliminatingaerocapture would reduce the heatshieldmass by about half.

MassPayload

lbm kgEarth Return Vehicle (ERV) 150,237 68,146Cargo Lander 133,952 60,760Crew Lander 123,330 55,942

Table 7: DRM payloads less aerocapture hardware.The lighter payloads result in slightly moreEDS capability once the launch stackreached LEO. However, since the EDS isnot being discarded after the TMI burn, thedescent stage cannot perform the TCMsduring transit. Thus, for this case, the EDSperforms the TMI burn, all TCMs, and the

Mars-orbit insertion (MOI) burn before beingdiscarded.

Since transit times to Mars can takeanywhere from 190 to 400 days within thetimespan investigated, propellant boiloffmust be accounted for. Propellant boiloffrates are modelled as a constant percent ofthe current propellant mass lost per day.The boiloff values for liquid oxygen andliquid hydrogen were backed out of theESAS final report and slightly improved to

account for some technologicaladvancement in cryogenic propellantstorage, resulting in an LH2 boiloff rate ofabout 0.20% per day and a LOX boiloff rateof about 0.02% per day. Additionally, TCMpropellant is estimated at 2% of the totalEDS propellant mass.

Compiling the TMI, TCM, boiloff, and MOIpropellant, a total required propellant for themission is obtained. Using the ascent dataproduced by POST for the lighter payloads,

the total amount of refuel propellant shownin Figure 7.


8/15

8

0

50

100

150

200

250

300

09/28 06/31 02/34 11/36 08/39 05/42 02/45

Earth Depature Date

RefuelPropellantMass(MT) ERV

Cargo

Crew

Figure 7: Required propellant resupply mass given

propulsive payload capture. The Crew payload isoffset from its corresponding ERV and Cargopayload by one opportunity.

Figure 7 is distinctively different from Figure5 as the amount of refuel propellant is now

strongly dependent on the Mars arrival C3 higher arrival C3 not only means more MOIpropellant, but it also means morepropellant boiled off during transit. Theminimum amount of refuel propellant,providing the smallest propellant depot,generally coincides with minimum total C3trajectory. The refuel requirements for thistrajectory are shown in Figure 8. Thesevalues also do not include propellant boiloffwhile loitering in the depot.

0

50

100

150

200

250

300

350

2031/33 2033/35 2035/37 2037/39 2039/41 2041/43

Mission

RefuelPropellantMass(MT)

Crew

Cargo

ERV

288.5

238.8221.9

257.0

202.8193.2

Figure 8: Total refuel propellant required for

propulsively caputred payloads assuming aminimum total C3 trajectory.

For the propulsive option, the first missionopportunity in the 2031/33 timeframe is nowthe worst option, which is exactly oppositeof the aerocapture results in Figure 6.However, even the most attractive

propulsive mission requires over 193 MT ofrefuel propellant.

Based purely on these results, aerocapturerequires a smaller depot and less total massin LEO. Initial mass in low-Earth orbit

(IMLEO) is generally a good indication as tohow much a given architecture will cost.This rationale favors aerocapture abovepropulsive capture.

Propellant Depot SizingSince the amount of refuel propellant canvary quite widely depending on theopportunity and the associated payloadmass, the worst-case propellant resupplyfrom each scenario is used to size anappropriate depot. Each depot is sized

using the scalable depot sizer developed byStreet[9] assuming aluminium storage tankswith insulation but no cryocoolers (i.e.propellant will boiloff while the depot is inorbit). The propellant boiloff rates were notmodified in the scalable depot sizer asthese rates apply to LEO whereas the EDSboiloff rate applies to deep space.

Since the ERV and Cargo mission bothleave Earth on the same opportunity, bothmust be launched and refuelled within a

small window of time. This means that thedepot must be sized to provide enoughpropellant for both departures. A 90-daywindow between the last commercial depotresupply and EDS refuel is assumed insizing the depot.

For the aerocapture mission, the maximumamount of refuel propellant is 76.5 MT,which occurs during the 2037 departure ofthe ERV and Cargo payloads. In order tohave that much propellant in the proper

oxidizer-to-fuel (O/F) ratio after theassumed 90-day wait period, the orbitingpropellant depot would be required to store94.5 MT of propellant. Table 8 provides thepertinent details about the propellant depot.The depots outer diameter was chosensuch that the depot would be able to fitinside the Ares V payload fairing. However,


9/15

9

the fairing would have to be lengthenedslightly in order to contain the depot.

Depot Diameter 7.8 mDepot Length 17.59 mDry Mass 13.6 MTLH2 Storage Capacity 11.9 MTLOX Storage Capacity 66.3 MT

O/F Ratio @ Capacity 5.6Total Wet Mass 94.0 MT

Table 8: Propellant depot metrics for the aerocapturearchitecture.

For the propulsive capture mission, themaximum amount of refuel propellant is 221MT, which occurs during the 2031 departureof the ERV and Cargo payloads. Table 9outlines the depot required to support thismission. This depot is also sized to fit withinthe Ares V payload shroud. However, due toits length, the depot would have to belaunched using two Ares Vs and assembledon orbit.

Depot Diameter 7.8 mDepot Length 36.22 mDry Mass 28.9 MTLH2 Storage Capacity 32.5 MTLOX Storage Capacity 190.2 MTO/F Ratio @ Capacity 5.8Total Wet Mass 265.5 MT

Table 9: Propellant depot metrics for the propulsivecapture architecture.

As stated previously, a propellant depotwould likely use commercial launch servicesto replenish its stores when necessary. Forextremely large amounts of propellant, theAres V could be used to boost a largerresupply module. A Propellant ResupplyModule (PRM) could be quite similar to theRussian Progress M spacecraft or theItalian Multi-Purpose Logistic Module(MPLM), both of which are used to ferrysupplies to the International Space Station(ISS). Assuming a conservative structural

mass fraction similar to the MPLM(mloaded/mempty = 0.31), a rough massestimate for a PRM can be obtained.

Figure 9: Multi-Purpose Logistics Module (MPLM) isanalogous to a Propellant Resupply Module(PRM) for depot resupply.

Using this sizing strategy, Table 10 detailsthe propellant resupply capability of each ofthe Delta IV launch vehicles as well as theirassumed launch costs.

Delta IV VariantCapacity to

LEO[10]Refuel

Capability

LaunchCost[11]

FY06Medium 9.1 MT 6.3 MT $140M

Medium+ (4,2) 12.3 MT 8.5 MT $155MMedium+ (5,2) 10.6 MT 7.3 MT $155MMedium+ (5,4) 13.9 MT 9.6 MT $171MHeavy 21.9 MT 15.1 MT $264MAres V 140.6 MT 97.0 MT $563MTable 10: Delta IV propellant resupply capability andassumed cost.The cost of each PRM can be roughlyobtained based purely on the estimated drymass by using the Spacecraft/Vehicle LevelCost Model (SVLCM)[12] developed by theNASA JSC Cost Estimation group. The

design, development, testing, andengineering (DDT&E) and theoretical firstunit (TFU) can be obtained using anunmanned Earth orbital spacecraft analogyin the cost model. These costs are detailedin Table 11.

ModuleSize

DryMass

DDT&EFY06

TFUFY06

Medium 2.8 MT $304.9M $84.3MMedium+ (4,2) 3.8 MT $359.7M $102.9MMedium+ (5,2) 3.3 MT $331.7M $93.3MMedium+ (5,4) 4.3 MT $384.2M $111.4M

Heavy 6.8 MT $493.9M $150.7MAres V 43.6 MT $1,374M $516.2MTable 11: Development and unit cost estimates for the

PRM associated with each launch vehicle.Realistically, only one, potentially two, sizesof PRM would be fabricated and used tosupply the depot. Since both options requirea significant amount of propellant, the Delta


10/15

10

IV Heavy PRM and the Ares V PRM aretraded in the following section.

Fixed Cost of Depot ArchitecturesEach of the propellant depot configurationswere costed using NAFCOM 2004 and

inflated to 2006 costs. Analogous systemsused to generate the cost data included theSkylab Orbital Workshop, STS ExternalTank, Centaur-D upper stage, and SaturnIV-B upper stage. This resulted in the costslisted in Table 12. All costs are $M FY2006USD. Launch costs assume one Ares V forthe aerocapture depot and two Ares Vs forthe propulsive depot. The launch cost forthe Ares V was estimated at about 25%more than the average Shuttle launch cost($450M). This cost increase from Shuttle is

derived from the fact that the Ares V will bea much larger vehicle, requiring moreprocessing and manpower to prepare andlaunch.

Aerocapture Propulsive DifferenceDDT&E $1,418.8 $3,530.4 $2,111.6TFU $110.3 $240.5 $130.2Launch $563 $1,126 $563Total $2,029 $4,897 $2,868

Table 12: Development and unit costs for propellantdepot options.

DDT&E includes the cost associate tomature the technologies associated withdepot operations; Chato[13] describes someof these technologies in his paper.

The cost to develop, test, and reliablyimplement aerocapture technology forcrewed vehicles also needs to be capturedin order to accurately compare the twooptions. This sort of technology programwould likely involve high-fidelity computersimulations, high-altitude testing, and

potentially subscale and full-scale testing atMars. Extensive testing and verification toman-rate an aerocapture vehicle wouldundoubtedly be an expensive venture. As apoint of comparison, total costs to developthe X-38 crew return vehicle were estimatedaround $2.17B FY05[14]. According to thedepot cost analysis above, there would beabout a $2.87B difference between the

hardware costs of the larger, propulsivecapture depot and the smaller, aerocapturedepot. Once again the SVLCM is usedobtain a cost estimate for a man-ratedaerocapture vehicle. Using mannedspacecraft as the analogy and the DRM-

cited dry mass of the aeroshell at 9.9 MT, aDDT&E cost of about $3B and a unit cost of$282M are obtained. The depot costdifference is virtually equivalent to the costof developing and testing a new man-ratedaerocapture vehicle.

Table 13 summarizes the total fixed costsfor both depot options. The PRM cost is theDDT&E cost associated with a Delta IVHeavy-sized PRM; the aerocapture costrepresents its associated DDT&E cost. All

costs are in $M FY2006 USD.

Aerocapture Propulsive DifferenceDepot $2,092 $4,897 $2,868PRM $494 $494 $0Aerocapture $3,195 N/A $3,195Total $5,781 $5,391 $327Table 13: Total fixed costs for both depot architecture

options.

Recurring Cost of Depot ArchitecturesSince the two depot architectures are aboutequivalent in up-front development cost, the

recurring cost of each architecture will bethe deciding factor. In order to assess eachmission on equal grounds, it is assumedthat the depot must be resupplied with therequired departure propellant prior to anEarth departure. Additionally, the suppliedpropellant must be sufficient to satisfy therefuelling requirements of the departureafter a 90-day loiter time.

Figure 10a and b compare the recurring forthe aerocapture and propulsive capture

options. These both assume that a Delta IVHeavy is used exclusively to deliver alldepot resupply propellant. The threerecurring Ares V launches are not shown asthey are the same for both options. As alarge number of PRMs need to be built, an85% learning curve is applied to the PRMunit cost.


11/15

11

$0

$1,000

$2,000$3,000

$4,000

$5,000

$6,000

$7,000

$8,000

2031/33 2033/35 2035/37 2037/39 2039/41 2041/43

Mission

Recurrin

gCost($M

FY06) Aerocapture Cost

PRM Cost

Delta IV Cost

$2,668 $2,668$3,031

$3,391$3,031

$2,668

(a) Aerocapture

$0

$1,000

$2,000

$3,000

$4,000

$5,000

$6,000

$7,000

$8,000

2031/33 2033/35 2035/37 2037/39 2039/41 2041/43

Mission

ResupplyCost($MFY06)

PRM Cost

Delta IV Cost

$7,154

$6,134

$5,450

$6,475

$5,450$5,106

(b) Propulsive Capture

Figure 10: Recurring cost (less Ares V) of both depotoptions assuming Delta IV Heavy depot resupply.

From the above graphs, it is apparent that

the propulsive capture option hasconsiderably higher resupply costs:between $2.4B and $4.5B more than theaerocapture option depending on theopportunity. With fixed costs between thetwo depot options approximately equal andthe recurring costs decidedly in favor ofaerocapture, it appears that it is worth theinvestment to pursue aerocapture.

Boiloff SensitivityThere is currently no standardized method

to model cryogenic propellant boiloff while avehicle is in space. Additionally, the choiceof the propellant boiloff rates in this studymay seem somewhat arbitrary despite theirtraceability from the ESAS Final Report. Inan effort to assess the effect that boiloff hason the propulsive capture option, asensitivity analysis was conductedassuming a departure on the 2031

opportunity for minimum total C3. The MarsCargo Lander payload (less aerocapture)was used as the reference payload. Sincethis is a large payload with a relatively largearrival C3, the effect of boiloff will be readilyapparent from the change in refuel mass.

The baseline amount of propellant resupplyis 97.8 MT.

A general range of LH2 and LOX boiloffrates can be obtained by examining otherdocuments which cite propellant boiloff.These documents and their assumed boiloffrates are presented in Table 14.

Reference LH2 Rate LOX RateBaseline 0.200 %/day 0.020 %/dayRef. [3]* ~0.363 %/day N/ARef. [15]* ~0.260 %/day ~0.047 %/dayRef. [16]* ~0.150 %/day ~0.020 %/dayRef. [17] 0.127 %/day 0.016 %/day

*Calculated based on surface area of EDS LOX andLH2 tanksTable 14: Representative boiloff rates from various

sources.

Given the data above, the LH2 boiloff rate iswidely varied from 0% per day (zero-boiloffcase) to 1.0% per day. Similarly, the LOXboiloff rate is widely varied from 0% per dayto 0.1% per day. The results of thesensitivity are shown in Figure 11, with the

Most Likely Region being bounded by theextreme values in the Table 14.

0.0

00%

0.0

14%

0.0

28%

0.0

42%

0.0

56%

0.0

70%

0.0

84%

0.0

98%

0.00%

0.12%

0.24%

0.36%

0.48%

0.60%

0.72%

0.84%

0.96%

Refuel Mass

(MT)

LOX Boiloff (%/day)

LH2Boiloff(%/day)

123-129

117-123

111-117

105-111

99-105

93-99

87-93

Baseline

"Most Likely"

Region

Figure 11: Refuel propellant mass sensitivity to boiloff

rates.


12/15

12

According to Figure 11, the amount of refuelmass can vary by up to 42 MT for this widerange of propellant boiloff rates. However,within the Most Likely Region, the refuelmass varies by 17 MT. The advantage ofimplementing a low-boiloff system over a

long storage time (approximately 300 daysin this case) becomes readily apparent fromthe sort of sensitivity.

To bring this sensitivity analysis back intothe architecture study at hand would azero-boiloff system make the propulsivecapture option desirable over theaerocapture option? This question is easilyanswered as No. Given the samedeparture conditions, the aerocapture optionwould still result in a 67 MT propellant mass

savings over the propulsive option. Thissuggests that the least expensive Marstransportation architecture is not dependenton boiloff rate but rather on the mode ofarrival.

Comparison with NTRIn this study, the idea of using a nuclearthermal departure stage was discarded fromthe start. This elimination does need some

justification before it can be readilyaccepted. To frame this trade, the EDS-

Propellant Depot combo with anaerocaptured payload will be comparedagainst a dedicated nuclear thermal TMIstage with a similar aerocapture system,shown in Figure 12.

Figure 12: DRM nuclear thermal rocket stage andaerocapture payload.

The DRM v3.0 is leveraged to obtain thenuclear thermal rocket (NTR) departurestage. Earlier analysis has shown that the

amount of propellant needed for departureis dependent on the opportunity, so therequired propellant to accomplish eachmission is once again calculated to ensurethat the DRM NTR stage can be usedwithout major design changes. The DRM

sizes the NTR for the fast-transferopportunity in 2009, resulting in a maximumpropellant capacity of 54 MT of LH2.Applying losses for ullage, start-up losses,residuals, and 90 days of boiloff, each NTRwould have just over 47 MT of usablepropellant.

Analysis for an aerocaptured payloaddemonstrates propellant requirementsranging from 42 MT to 49 MT using an NTRdeparture stage. The assumed NTR is

unable to push the ERV on three of theopportunities investigated. The NTR couldpossibly achieve these three burns if theboiloff time was reduced to only 41 days.Alternatively, the NTR stage could belengthened such that its maximumpropellant capacity was around 56 MT. Asseen in the boiloff sensitivity, the propellantmass variability due to boiloff is quite large.This kind of propellant mass uncertaintyprecludes a detailed sizing of the NTR sothe nuclear stage remains as sized in the

DRM, which is detailed Table 15. Theshadow shield mentioned in the table isused to absorb and deflect radiation fromthe nuclear reactor on the crewed flightonly. The concept of operations involvingthe NTR also remains the same i.e. eachNTR stage is expended after the TMI burnis executed.

Propellant LH2Thrust 45,000 lbfIsp 960 sDry Mass 23,400 kg

Max Propellant 53,729 kgShadow Shield Mass 3,200 kgStage Length 28 mStage Diameter 8.4 m

Table 15: DRM v3.0 baseline NTR departure stage.Unfortunately, do to the radiation emitted bythe nuclear reactor, the system cannot bestarted until it reaches orbit, meaning the


13/15

13

NTR stage cannot function as an upperstage similar to the EDS. Additionally, theVAB cannot support an Ares V which stacksthe Mars payload and the NTR stage on topof an upper stage used to achieve thecorrect Earth parking orbit. This immediately

forces the departure stage to be launchedseparately from the payload, creating anadditional three launches per mission. Anascent trajectory was analyzed for an AresV launch of just the TMI stage. The Ares Vdemonstrated enough capability to achievethe reference 30 x 160 nmi orbit using onlythe solid rocket boosters and the Ares Vcore. Additionally, each of the individualMars payloads can also achieve thereference orbit without the use of anadditional upper stage.

The particularly prohibitive aspect aboutspace nuclear systems is their highdevelopment costs. Most recently, this wasdemonstrated with the Prometheus Project,whose projected budget was $430MFY05Error! Reference source not found.prior to cancellation. The Nuclear Engine forRocket Vehicle Applications (NERVA)program was a program executed from1961 to 1973 to build nuclear poweredrockets and missiles. Before it was

cancelled, the program consumed $3.9BFY96[19], which translates into over $4.73BFY05. Since the cost of developing andimplementing a man-rated nuclear thermalrocket cannot be estimated with anyreasonable accuracy, an attempt is made todetermine how much it would have to costin order to be competitive with the EDS-Depot architecture. Table 16 and 18outlines the fixed and recurring costs foreach system.

Depot DDT&E $1,419Depot Unit $110Depot Launch $563PRM DDT&E $494Aerocapture DDT&E $3,195

Total Fixed Cost $5,781Aerocapture Unit (3x) $739PRM Hardware (max) $802Delta IV Heavy Launches (max) $1,850Ares V Launches (3x) $1,688

Total Recurring Cost $5,079

Overall Cost (through 1st

mission) $10,860Table 16: Depot architecture cost through first

mission.

NTR DDT&E N/ATotal Fixed Cost N/A

NTR Unit (3x) N/A

Ares V Launches (6x) $3,375Total Recurring Cost $3,375

Overall Cost (through 1st

mission) $3,375Table 17: NTR architecture cost through first mission.

Although the data in the table above isincomplete, it is possible to derive someconclusions about an NTR system versusthe EDS-Depot architecture.

First, considering recurring costs, the NTRsystem is at a disadvantage as it requiressix Ares V launches per mission. This alonecompensates for nearly 3/5 of the depotrecurring cost. The remaining unknown isthe NTR departure stage unit cost. Intuitionsays that three nuclear reactors and theirassociated vehicles will cost more than$1.7B to fabricate. Looking at the fixed cost,an NTR system would have to bedeveloped, flight tested, and certified forunder $6B. Considering that NERVAconsumed $4.7B pursing the NTR engineonly, it might be difficult to achieve ahuman-rated nuclear vehicle for under $6B.However, regardless of the fixed costsassociated with the development of eitherarchitecture, the NTR system will be moreexpensive in the long run if each stage isdiscarded as the DRM suggests.

Next, considering the near-simultaneousdepartures of the ERV and Cargo payloadseach mission, the NTR architecturerepresents a more complicated concept ofoperations (conops) than the depotalternative. Four Ares V launches and two

automated EORs must be performed withina short amount of time (to preventexcessive boiloff of the TMI propellant) aswell as two remote nuclear reactor start-ups. The propellant depot conops involvesonly two Ares V launches and twoautomated EORs (to refuel at the depot).


14/15

14

Additionally, there will be a host ofenvironmental concerns regardinglaunching three nuclear reactors (fourincluding the Mars surface power system)for each mission to Mars. These concernscould likely delay the verification process

and drive up cost and development time.Both the propellant depot and anaerocapture vehicle are not subject to thiskind of scrutiny, eliminating some possiblebureaucratic hang-ups during developmentand implementation.

Overall, it appears that the EDS-Depotoption is superior to the development of anuclear rocket. Despite the cost ambiguityof developing such technologies as NTR,intuition and operational considerations both

indicate that an NTR stage in thearchitecture would be disadvantageouswhen compared to a depot architecture.

Launch and IMLEO ComparisonCost, thus far, has been the primary driverin deciding which architecture is superior tothe other. However, cost estimation at thislevel can have a lot of variability. It may bemore beneficial to look at the threeinvestigated architectures purely from alaunch and IMLEO standpoint. The total

number of launches for the threearchitectures is shown in Figure 13assuming the Delta IV Heavy as theexclusive depot resupply vehicle.

0

5

10

15

20

25

30

2031/33 2033/35 2035/37 2037/39 2039/41 2041/43

Mission

N

um

bero

fLaunc

hes

Depot w/ AerocaptureDepot w/ Prop Capture

NTR

Figure 13: Total number of launches per mission for

each architecture.

The total IMLEO is shown below in Figure14. For the depot options, this IMLEOincludes: three Mars payloads three EDS stages with propellant

remaining from ascent all necessary propellant required to

resupply the depot (including boiloff) forthe mission

all PRM hardware mass associatedwith all resupply missions

the propellant depot dry massFor the NTR option, the IMLEO includes thethree Mars payloads and three of thenuclear TMI stages.

0

200

400

600

800

1,000

1,200

1,400

2031/33 2033/35 2035/37 2037/39 2039/41 2041/43

Mission

IMLEO(

MT)

Depot w/ AerocaptureDepot w/ Propulsive Capture

NTR

Figure 14: IMLEO for all three architectures.

The NTR system consistently has the least

number of launches and IMLEO ascompared to both depot options. However,the uncertainty associated with itsdevelopment cost and the large costincurred when each nuclear stage isdiscarded prevent a nuclear solution frombeing viable in this study.

CONCLUSIONIn order to make human excursions to Marsmore feasible, the use of existing hardwarewould be desirable. Assuming that the

NASA-proposed Ares V launch vehicle isoperational during the lunar campaign, itcan be used to send significant payloads toMars given a pre-deployed propellant depot.Three different levels of technologydevelopment were also considered propulsive Mars capture (low technologydevelopment), aerocapture at Mars(moderate to extensive), and nuclear


15/15

15

propulsion (highly extensive) to helpgauge which technology would be mosteconomic for a Mars campaign. It wasshown that propulsively capturing at Marsusing the ESAS-proposed EDS provides thelowest life-cycle cost and the lowest risk. It

was also shown that an NTR solution wouldlikely be more operationally complex andexpensive than both depot options in thelong run. Overall, the use of a propellantdepot can be a relatively inexpensive,enabling system for human-Marsexploration.

REFERENCES[1] The Vision for Space Exploration,

NASA, Feb. 2004.[2] NASAs Exploration Systems

Architecture Study Final Report, NASATM-2005-214062, Nov. 2005.

[3] Reference Mission Version 3.0Addendum to the Human Exploration ofMars: The Reference Mission of theNASA Mars Exploration Study Team,NASA EX13-98-036, Jun. 1998.

[4] JAQAR Space Engineering.[http://www.jaqar.com.Accessed 09/13/2006.]

[5] Troutman, P.A., Orbital Aggregation &Space Infrastructure Systems (OASIS)

Executive Summary, Oct. 2001.[6] Orloff, R.W., Apollo by the Numbers,

NASA SP-2000-4029, Oct. 2000.[7] Apollo 10 Press Kit, NASA, May 1969.[8] Braun, R.D., Powell, R.W.,

Aerodynamic Requirements of aManned Mars Aerobraking TransferVehicle, AIAA Journal of Spacecraftand Rockets, Vol. 28, No. 4, 1991, pp.361-367.

[9] Street, D., A Scalable Orbital PropellantDepot Design, M.S. Special Problems

Report, Guggenheim School ofAerospace Engineering, GeorgiaInstitute of Technology, Atlanta, GA,2006.

[10] Delta IV Technical Summary, BoeingLaunch Services, Jul. 2004.

[11] Isakowitz, S.J., Hopkins, J.B.,International Reference Guide to

Space Launch Systems, AIAA,Reston, VA, 2004.

[12] Spacecraft/Vehicle Level CostModel, Johnson Space Center,Houston, TX.[http://www1.jsc.nasa.gov/bu2/SVLCM

.html. Accessed 09/13/2006.][13] Chato, D.J., Flight Development for

Cryogenic Fluid Management inSupport of Exploration Missions, 44thAIAA Aerospace Sciences Meetingand Exhibit, Reno, NV, Jan 2006,AIAA 2006-940.

[14] X-38 Fact Sheet,[http://www.nasa.gov/centers/dryden/news/FactSheets/FS-038-DFRC.html.Accessed 09/13/2006.]

[15] Borowski, S.K., Dudzinski, L.A.,

McGuire, M.L., Vehicle and MissionDesign Options for the HumanExploration of Mars/Phobos UsingBimodal NTR and LANTRPropulsion, NASA TM-1998-208834/REV1, Dec. 2002.

[16] Borowski, S.K., Dudzinski, L.A., 2001:A Space Odyssey Revisited TheFeasibility of 24 Hour CommuterFlights to the Moon Using NTRPropulsion with LUNOX Afterburners,NASA TM-1998-208830, Dec. 1998.

[17] Future Orbital Transfer VehicleTechnology Study, Volume II Technical Report, NASA ContractorReport 3536, 1982.

[18] Project Prometheus,[http://en.wikipedia.org/wiki/Project_Prometheus. Accessed 09/13/2006.]

[19] Budget data for NERVA were obtainedfrom successive volumes of Bureau ofthe Budget/Office of Management andBudget, Budget of the United StatesGovernment, Fiscal Year 1958

through 1975, and converted toconstant 1996 dollars.

Documents

On-Orbit Propellant Resupply Options for Mars Exploration