Content uploaded by Koki Ho
Author content
All content in this area was uploaded by Koki Ho on Feb 05, 2020
Content may be subject to copyright.
1
Integrated In-Situ Resource Utilization System Design and Logistics for Mars Exploration
Hao Chena, Tristan Sarton du Jonchaya, Linyi Houb, and Koki Hoa*
a Daniel Guggenheim School of Aerospace Engineering, Georgia Institute of Technology, 270 Ferst Dr, Atlanta, GA
30313, USA, kokiho@gatech.edu
b Department of Aerospace Engineering, University of Illinois at Urbana-Champaign, 104 S. Wright St, Urbana, IL
61801, USA
* Corresponding Author
Abstract
This paper develops an interdisciplinary space architecture optimization framework to analyze the tradeoff on in-
situ resource utilization options, identify technology gaps, evaluate the benefits of in-situ resource utilization, and
optimize the design of infrastructure for Mars human space exploration scenarios and mission profiles. It performs
trade studies from the perspective of space logistics, which takes into account the interplanetary transportation,
infrastructure deployment, in-situ resource utilization system operation, and logistics of the produced resources. Our
method considers space architecture design and operation from the subsystem level to capture the coupling between
in-situ resource utilization technologies and in-space architecture elements for space resource logistics. A case study
involving a multi-mission human Mars exploration campaign is performed to evaluate the effectiveness of existing
and proposed in-situ resource utilization technology concepts and system designs. The results can provide us with a
better understanding of the benefits and costs of different in-situ resource utilization technologies for interplanetary
space transportation. A sensitivity analysis is also conducted to understand the impacts of lunar and near-Earth-object’s
in-situ resource utilization systems on Mars missions. The results of this analysis can help decision-makers determine
and optimize the roadmap for in-situ resource utilization technology development.
Keywords: Space logistics, Human space exploration, In-situ resource utilization
Nomenclature
= set of arcs
= cost coefficient
= mission demand
= total ISRU resource demand
= commodity transformation matrix
= subsystem index
= concurrency constraint matrix
= node index
= specific impulse
= node index
= optimization objective
= battery/fuel cell design matrix
= storage length
= subsystem mass
= ISRU hourly productivity
= set of nodes
= subsystem power
= power system output power
= subsystem power demand
= infrastructure daily operating length
= length of a solar day
= length of daytime per solar day
= subsystem daily operating length
= total ISRU system size
= time step index
= set of time steps
= time of flight
= spacecraft index
= set of spacecraft
= change of velocity
= set of time windows
= commodity variable
= ISRU daily productivity
= energy storage efficiency
Abbreviations
ISRU = in-situ resource utilization
FSPS = fission surface power system
PV = photovoltaic
RTG = radioisotope thermoelectric generator
RWGS = reverse water gas shift reaction
SR = Sabatier reaction
SOCE = solid oxide CO2 electrolyzer
MRE = molten regolith electrolysis
HR = hydrogen reduction
CR = carbothermal reduction
SWE = soil/water extraction
DWE = direct water electrolysis
ES = Earth
TLI = trans-lunar injection
NRHO = near rectilinear halo orbit
PLLO = polar low lunar orbit
LSP = lunar south pole
NEO = near-Earth object
LMO = low Mars orbit
SLS = space launch system
2
EUS = exploration upper stage
CLV = commercial launch vehicle
CUS = commercial upper stage
T/V = transfer vehicle
LLAM = lunar lander ascent module
LLDM = lunar lander descent module
IMLEO = initial mass in low-Earth orbit
1. Introduction
As the interest grows in deep space exploration, large-
scale space mission planning and space architecture
design have become increasingly important. Both NASA
[1] and SpaceX [2] have announced their exploration
plans to Mars by the 2030s. To build an affordable and
sustainable interplanetary space transportation system to
Mars, in-situ resource utilization (ISRU) systems and
propellant depots are two critical space infrastructures.
They can produce and store space resources in space,
especially spacecraft propellant, to support space
transportation and reduce mission costs. Studies on
campaign-level space mission planning have shown the
effectiveness of ISRU and propellant depots for space
transportation [3-7].
Multiple optimization frameworks have been
proposed to perform efficient space transportation
planning leveraging mission interdependencies through
heuristic methods [8], simulations [9], the graph theory
[10], or network flow models [3-7]. However, these
methods either do not consider ISRU infrastructure
design as part of the trade space or do not take into
account ISRU subsystem-level interactions and trade-
offs in the optimization. These interactions can directly
impact the ISRU system operation mechanisms and
system performances. For example, nuclear power
systems, such as the fission surface power system (FSPS)
and the large-scale radioisotope thermoelectric generator
(RTG), can support ISRU plant continuously regardless
of day and night. While the photovoltaic (PV) power
system can only support the infrastructure during the
daytime if no power storage system is deployed at the
same time. Moreover, it also suffers from radiation
degradation and dust power loss on the Martian surface.
Furthermore, some ISRU processes can share the same
subsystems, which makes their infrastructure design and
deployment more efficient. For example, the reverse
water gas shift reaction (RWGS) and Sabatier reaction
(SR) processes have the same reactant (i.e., and )
and both produce as one of the products. Their
Martian atmosphere acquisition subsystem and
storage subsystem can be designed and deployed
together.
On the other hand, several testbeds have been built by
NASA [11] and Lockheed Martin [12] to evaluate the
performance of the hydrogen reduction reaction plant in
oxygen production. Integrated prototypes were also
developed for carbothermic reduction [13] and molten
regolith electrolysis [14] for production demonstration.
Besides the soil-based ISRU systems, Meyen [15]
performed a thorough analysis of the Mars atmosphere-
based ISRU experiment, also known as MOXIE. This
system will be implemented on the Mars 2020 Rover for
an on-site test.
Numerous studies also have been done focusing on
the Martian surface transportation, resource
identification and utilization assessment, and surface
mission scenario analysis. Smirnova proposed a Mars
surface transportation vehicle concept that guaranteed
reliable flights in the Martian atmosphere [16]. Chamitoff
et al. developed a powerful software tool for Martian
resource identification and landing site optimization [17].
In addition, Kading et al. presented a manned Mars
mission based on additive manufacturing techniques and
in situ materials [18]. Do et al. conducted a detailed
assessment of the Mars One mission plan [19].
However, these studies mainly focused on the
standalone performance of surface systems after
deployment. Our research, on the contrary, proposes an
integrated ISRU design and logistics framework from an
innovative perspective to take into account the
relationships between surface operations and space
mission planning from the subsystem-level. These
relationships, which are ignored in previous literature,
can directly influence mission planning solutions and
ISRU infrastructure designs. For example, in space
missions with frequent landing and surface operations,
only a small storage subsystem is needed because most
of the propellant is used right after production; whereas
for space missions with low-frequency time windows
because of rocket launch pad availabilities or mission
demand requirements, a large storage subsystem for
ISRU infrastructure is necessary.
To effectively analyze ISRU system performances,
identify technology gaps, and evaluate the actual benefits
of ISRU systems to human exploration to Mars, this
paper proposes an interdisciplinary space architecture
optimization framework that takes into account ISRU
subsystem-level trade-offs and the infrastructure
deployment.
There are three contributions achieved in this paper.
First, the proposed architecture optimization framework
enables effective space resource logistics optimization
for future human exploration to Mars considering the
synergistic effects of ISRU technologies, infrastructure
deployment, and logistics after resource productions.
Second, multiple soil-based and atmosphere-based ISRU
infrastructure sizing models are established based on
exiting ISRU design concepts and prototypes. These
models make it possible to perform a qualitative
performance comparison between different ISRU
technologies. Finally, a detailed Mars exploration case
study is developed based on the NASA Artemis lunar
exploration architecture [20]. It is conducted to analyze
3
the impact of trajectory selections in Earth-Moon-Mars
resource logistics and ISRU technology selections
leveraging the proposed interdisciplinary space
architecture optimization framework and developed
ISRU sizing models.
Our method provides an important step forward in
system-level architecture design and evaluation for future
large-scale human space explorations. It is also
particularly useful to identify the level of resource
information we need to design ISRU hardware and plan
space missions including the in-space transportation and
landing site selection.
The remainder of this paper is organized as follows.
Section 2 briefly introduces the network-based space
logistics optimization method used for space mission
planning. Section 3 describes the ISRU subsystem
definition and infrastructure sizing models. It also covers
the detailed analyses to implement power demand and
supply relationships in the space mission planning
formulation. In Sec. 4, model details and assumptions for
human exploration to Mars are established, including
candidate space transportation trajectories, the
transportation network model, Martian ISRU operation
environment assumptions, mission demand assumptions,
and a preliminary cost model. The results and discussions
are shown in Sec. 5. Finally, In Sec. 6, we conclude this
paper and discuss future works.
2. Space mission planning formulation
This section introduces the network-based space
logistics optimization method for space mission
planning. It considers the space mission transportation
model as commodity flows along arcs in a network. In
this network, nodes denote orbits or planets; arcs
represent trajectories; crew, payload, propellant, and
spacecraft are all considered as commodities flowing
along arcs. An example of the Earth-Moon-Mars
transportation network model is shown in Fig. 1.
Fig. 1 Earth-Moon-Mars transportation network [7]
Consider a set of arcs , which contains a
set of available spacecraft (index, ), a set of nodes
(indices, ), and a set of time steps (index, ). To
compile the space mission planning formulation, we need
to define the following variables and parameters.
= commodity flow variables. It can be discrete
or continuous variables depending on the
commodities.
= mission supply and demand vector. Supplies are
positive and demands are negative.
= mission cost coefficient.
= time of flight.
= Commodity transformation matrix.
= Concurrency constraint matrix.
= Mission time windows.
Based on the aforementioned notations, we can
express the network-based space logistics optimization
formulation as follows [6].
Minimize:
Subject to:
2.1 Objective function
Equation (1a) is the objective function. It can be
mission cost or other mission performance measurements
depending on the definition of .
2.2 Mass balance constraint
Equation (1b) is the mass balance constraint that
makes sure the commodity inflows to a node is always
larger than or equal to the summation of the commodity
outflows and the mission demands. In this constraint, the
second term, , represents commodity
transformation during spaceflights or after infrastructure
deployment, including propellant burning, crew
consumptions, and ISRU resource production.
2.3 Concurrency constraint
Equation (1c) is the concurrency constraint that limits
commodity flow bounds, mainly defined by the
spacecraft payload capacity, the spacecraft propellant
capacity, and the ISRU storage capacity. This type of
constraint is also used to guarantee the non-negativity of
the commodity inflows (i.e., ) is the
number of total concurrency constraint types.
4
2.4 Time window constraint
Equation (1d) is the time window constraint. It
guarantees that only when time windows are open,
commodity flows are permitted. is the number of
commodity types.
The above is a brief introduction of the space mission
planning formulation through the generalized multi-
commodity flow model [3-7]. For detailed constraint
descriptions and constraint parameter settings, please
refer to Ref. [6].
3. ISRU system modeling
This section first shows the integrated ISRU
infrastructure model and modularized ISRU subsystems.
Then, in Sec. 3.2, we develop a series of ISRU models by
extrapolating the data from existing ISRU design
concepts and prototypes. Finally, Sec. 3.3 introduces the
ISRU operation models considering power system
demands and landing site environments.
3.1 Integrated ISRU Infrastructure Model
There are six subsystems considered in this ISRU
infrastructure model. The first subsystem is reactors,
which is the core of ISRU plants. They conduct chemical
processes to transform reactants into valuable products,
such as water (), oxygen (), hydrogen (), and
methane (). The sizing parameter of a reactor is the
ISRU resource hourly productivity , in the unit of kg/hr.
The mass and power consumption sizing models of a
reactor can be written as Eqs. (2) and (3). Note that, there
are some ISRU reactors that use rigid solar concentrators
to provide thermal energy to the chemical reactions, such
as the integrated carbothermic reduction system [13]
developed by Orbitec Inc. and NASA. For those reactors
that use a special power source, we consider the power
architecture as part of the reactor in the sizing model and
the reactor does not require external power input
anymore.
To obtain raw materials as ISRU reactants, we need
the second subsystem, excavator or acquisition systems,
to collect soil/regolith for the soil-based ISRU system or
for the Martian atmosphere-based ISRU system. We
introduce the excavation rate
, in the unit of
kg/hr, as an intermediate variable to decouple the
excavation schedule and the reactor operating time. It is
a function of the ISRU resource productivity , written
as
. Based on the excavation rate
, the sizing models for excavator/acquisition
subsystems can be written as Eqs. (4) and (5). The
excavation complexity, difficulty, and site specificity
vary depending on the target raw materials attributes.
Using Mars exploration as an example, 95% of the
Martian atmosphere is made up of . It is everywhere
on Mars. Therefore, the reactant excavation of
atmosphere-based ISRU is not a constraint for mission
landing site selection on Mars. For the Martian soil-based
water ISRU system, landing site selection can directly
impact the ISRU performance. Granular regolith is a type
of garden variety soil, which contains 1-3% water
concentration. It is easy to excavate and is found in most
places on Mars [21]. Gypsum/sulfate on Mars has a
higher water concentration, 5-10%. However, it is a
harder material that may require a rock excavator and
crushing. Its locations are also limited to the equatorial
region and mid-latitude area [21]. There is also
subsurface ice on Mars, which requires drilling devices
to collect it. The landing site for Martian icy regolith is
highly selective. The design of excavator/acquisition
subsystems is determined by the trade-off between the
landing site and ISRU technology selections.
The third subsystem is separators that are used to
separate products from the reactor exhaust gas. The
performance and sizing of separators are directly relevant
to reactor types, exhaust gas components, and operating
environments. Therefore, the sizing parameter of a
separator is the same as the reactor, which is the ISRU
resource productivity . The mass and power
consumption sizing models of separators can be written
as Eqs. (6) and (7).
The fourth subsystem is a hopper/feed/secondary
subsystem, which is supporting structures for other ISRU
subsystems. Therefore, its sizing is directly determined
by other subsystem sizing results. To make the ISRU
infrastructure design model consistent, we also use the
ISRU resource productivity as the sizing parameter. Its
mass and power consumption sizing models can be
written as Eqs. (8) and (9).
After resources are produced, it requires storage
subsystems to temporarily store the resources before they
are consumed. The capacity of storage subsystems is
determined by the maximum amount of produced
resources to be stored during space missions. We can
define the storage length variable , in the unit of days,
then the storage subsystem capacity should be equal to
, where is the daily operating length (hr/day) and
is the ISRU hourly productivity (kg/hr). Note that, is
not the same as the total space mission duration. It is the
time between two space resource logistics missions.
Frequent logistics missions reduce the value of , which
decreases the capacity requirement on storage
subsystems. However, frequent missions also increase
5
the propellant consumption and operation complexity
during spaceflights. On the other hand, occasional space
missions enlarge the value of that leads to less
propellant consumption in space transportation but
requires larger storage subsystems to be deployed. This
is a trade-off between the space logistics operation and
the ISRU production operation. Based on the storage
capacity , its mass and power consumption sizing
models can be written as Eqs. (10) and (11).
The last subsystem is the power subsystem, which is
one of the most important subsystems in ISRU trade
studies. The design and technology selections of a power
subsystem are relevant to landing site choice, space
mission planning, ISRU operation mechanism, and ISRU
infrastructure sizing. There are mainly two categories of
power sources considered in this research. The first
power source is nuclear power, including FSPS and RTG.
This type of power system works continuously regardless
of the operating environment. Therefore, it is relatively
easier to perform trade studies and system sizing
analysis. The second power source is solar power, which
mainly includes the PV power system whose
performance is highly site-sensitive. For example, at the
0-20˚ N latitude region of Mars, the peak solar irradiance
after shadowing is 450 W/ m2 and the period of high
activity for solar arrays is 6-7 hr/sol [22]. In this region,
solar arrays can work through the whole year after
implementing dust mitigation technologies. In the
northern polar area of Mars, the peak solar irradiance
after shadowing is 150 W/ m2 and the period of high
activity for solar arrays is 24.6 hr/sol [22]. However, the
exploration mission can only last for about 90 days
during summer in this area because of the limited solar
source for the rest of the year. Moreover, if the PV power
system is the main power source and ISRU systems are
planned to work during the night, additional energy
storage systems need to be deployed. They can be
batteries or fuel cells. Different attributes of power
systems and operation environments make the trade
studies more complex, especially considering their
interaction with space logistics mission planning. In Sec.
3.3, we will discuss how to integrate power system trade
studies into the space mission planning framework. The
design parameter of the power subsystem is the total
power demand of all other ISRU subsystems. Then, the
mass sizing model can be written as Eq. (12).
Based on the aforementioned ISRU subsystem sizing
models and their dependencies, we can combine them
together and obtain an integrated ISRU modeling flow
chart as shown in Fig. 2. The inputs of this integrated
model are available ISRU technologies, potential power
sources and the requirement of ISRU daily productivity
. The outputs are designated ISRU subsystem
infrastructure designs and technology selections.
Fig. 2 Integrated ISRU modeling flow chart
3.2 Available ISRU resources
In this section, we discuss available ISRU resources
and proposed ISRU sizing models. The interested
resources to be produced by ISRU include water (H2O),
oxygen (O2), methane (CH4), and hydrogen (H2). Table 1
shows available resources at different locations and
corresponding ISRU technologies to extract these
resources. These ISRU technologies are listed as follows.
Mars atmosphere based ISRU:
⚫ Solid Oxide CO2 Electrolyzer (SOCE)
⚫ Reverse Water Gas Shift Reaction (RWGS)
⚫ Sabatier Reaction (SR)
⚫ SR/RWGS Hybrid ISRU (SR/RWGS)
Lunar soil-based ISRU:
⚫ Molten Regolith Electrolysis (MRE)
⚫ Hydrogen Reduction (HR)
⚫ Carbothermal Reduction (CR)
General ISRU available for the Moon, asteroid,
and Mars:
⚫ Soil/Water Extraction (SWE)
⚫ Direct Water Electrolysis (DWE)
Table 1 Available resources and corresponding ISRU technologies
Moon
Mars
Asteroids
Water
(H2O)
Resource: Icy Regolith in
Permanently Shadowed Regions
(PRS) [ISRU: SWE]
Resource: Hydrated
Soils/Minerals/Subsurface Icy
Soils [ISRU: SWE]
Resource:
Subsurface Regolith
[ISRU: SWE]
6
Resource: Minerals containing
iron oxides [ISRU: HR (needs
H2)]
Resource: CO2 in the atmosphere
(~96%) [ISRU: RWGS (needs H2);
SR (needs H2)]
Oxygen
(O2)
Resource: Icy Regolith in
Permanently Shadowed Regions
(PRS) [ISRU: SWE/DWE]
Resource: Minerals in Lunar
Regolith [ISRU: MRE;
CR/MR/DWE; HR/DWE]
Resource: Hydrated
Soils/Minerals/Subsurface Icy
Soils [ISRU: SWE/DWE]
Resource: CO2 in the atmosphere
(~96%) [ISRU: SOCE;
RWGS/DWE; SR/DWE]
Resource: Minerals
in Regolith [ISRU:
SWE/DWE]
Methane
(CH4)
Resource: Minerals in Lunar
Regolith [ISRU: CR/MR]
Resource: CO2 in the atmosphere
(~96%) [ISRU: SOCE/MR (needs
H2); RWGS/DWE/MR; SR]
TBD
Hydrogen
(H2)
Resource: Icy Regolith in
Permanently Shadowed Regions
(PRS) [ISRU: SWE/DWE]
Resource: Hydrated
Soils/Minerals/Subsurface Icy
Soils [ISRU: SWE/DWE]
Resource:
Subsurface Regolith
[ISRU: SWE/DWE]
Table 2 ISRU infrastructure design models
System
Chemistry reactions
Reference
product
Specific power,
kW
Specific mass, kg
Reactor
Solid Oxide Electrolyzer (SOCE)
, kg/hr
8.64 [23]
184.23 [23]
Reverse Water Gas Shift Reaction
(RWGS)
, kg/hr
2.4 [24]
102.3 [23, 24]
Sabatier Reaction (SR)
, kg/hr
0.68 [23]
72.2 [24]
SR/RWGS Hybrid ISRU (SR/RWGS)
, kg/hr
4.2 [25]
275.5 [25]
Soil/Water extraction (SWE)
, kg/hr
@3%: 13.7 [26]
@5.6%: 10 [25]
@8%: 7 [24]
@3%: 357 [23]
@5.6%: 279 [26]
@8%: 195 [26]
Direct water electrolysis (DWE)
, kg/hr
5.83 [23]
83.3 [23]
Molten regolith electrolysis (MRE)
, kg/hr
26.94 [14]
197.58 [14]
Hydrogen reduction (HR)
, kg/hr
0* [13, 27]
@equator: 228 [27]
@pole: 482 [27]
Carbothermal reduction (CR)
, kg/hr
0* [13,27]
520.5 [27]
Extraction/Acquisition system
Acquisition system for
— —
, kg/hr
1.8 [23]
139.8 [23]
Excavator for soil @3%
— —
Soil, kg/hr
0.004 [28]
0.38 [28]
Excavator for soil @5.6%
— —
Soil, kg/hr
0.027 [26]
23 [26]
Excavator for soil @8%
— —
Soil, kg/hr
0.027 [26]
23 [26]
Storage system
storage
— —
, kg
0.0017 [23]
0.194 [23]
storage
— —
, kg
0.0267 [29]
3.33 [29]
storage
— —
, kg
0 [29]
40 [29]
storage
— —
, kg
0.0073 [23]
1.67 [23]
Power system
Photovoltaic (PV) power system
— —
Power, kW
— —
6.8 (@ 1 AU) [30]
Energy storage system: battery
— —
Energy, kWh
— —
4 [31]
Energy storage system: fuel cell
— —
Energy, kWh
— —
2 [31]
Fission surface power system (FSPS)
— —
Power, kW
— —
150 [32]
Radioisotope power system (RPS)
— —
Power, kW
— —
124 [33,34]
7
*HR and CR reactors are integrated with rigid solar concentrators, which are their main thermal energy source
According to existing ISRU design concepts and
prototypes, we develop a series of ISRU infrastructure
sizing models, as shown in Table 2. The specific mass
and power are the measurements of infrastructure sizing.
For reactors, they mean the required system mass and
power demand to generate 1 kg reference product per
hour. For storage systems, they mean the necessary
system mass to store 1 kg reference product. For power
systems, it means the required mass to generate 1 kW
power or store 1 kWh energy. This table also shows that
for some ISRU technologies, the landing site
environment is also a big impact factor on ISRU system
performances. For example, the performances of the
SWE reactor and soil excavator are different depending
on the soil moisture. Moreover, due to the difference in
lunar regolith composition, the performance of HR
reactors is also different when deployed in the lunar
equator area compared with the polar region.
Furthermore, the HR and CR reactors use rigid solar
concentrators as their energy source, which is considered
as part of the reactors. Thus, the nominal external power
demands of both HR and CR reactors are considered as
zero.
3.3 ISRU Power System Trade Studies
To support ISRU subsystems, we need a power
subsystem to provide enough power supply during their
operations. The ISRU design model proposed in Sec. 3.2
considers the hourly productivity as the core design
variable. However, to decouple the long-time horizon
mission planning from the complex ISRU internal
operation tradeoffs, the daily ISRU productivity is
considered as the ISRU performance criteria in space
logistics optimization. ISRU daily productivity is directly
determined by the power system design. In this section,
we introduce the equations to consider ISRU power
system design trade studies in space logistics
optimization. These equations all can be categorized as
concurrency constraints as shown in Eq. (1c).
3.3.1 Nuclear power system
The nuclear power system is one of the most common
power sources in human space missions. We know that
the ISRU infrastructures are designed based on the hourly
productivity . However, in space logistics, the
transportation system cares about ISRU daily
productivity . Now, we denote the length of a solar day
at the ISRU landing site as , the daytime length in a
solar day as , and the length of ISRU operation per
solar day as . Then, the ISRU daily productivity using
the nuclear power system can be written as follows:
In Eq. (13), to maximize the ISRU daily productivity
considering fixed hourly productivity, which means a
fixed ISRU infrastructure design, we need to operate the
ISRU for the entire solar day (i.e., ). This is to
operate the ISRU system continuously throughout the
space mission. Therefore, for nuclear power systems, if
the ISRU power input requirement is satisfied at any
specific time, the power demand of the ISRU is always
satisfied throughout the entire mission. We can define a
power demand vector for ISRU subsystems and a
power supply vector for power plants. Then, define
the ISRU infrastructure commodity flow variable as
and the power system commodity flow variable as .
We suppose that node is a surface node available for
ISRU deployment. The nuclear power supply constraint
can be expressed as Eq. (14), which has a similar format
to the concurrency constraint in space logistics.
3.3.2 PV power system
Aside from the nuclear power system, the PV power
system is another widely used power source in space
explorations. Different from nuclear power, the operating
environments, especially the daytime length and the solar
irradiance, have a significant impact on PV power system
performances. Solar arrays can only work during the
daytime. Moreover, we also need to take into account
energy storage systems, such as batteries and fuel cells,
that extend the ISRU daily operation length at the cost of
extra mission transportation and infrastructure
deployment.
Assume that we have a set of ISRU subsystems that
have power demands . The daily
operation length of each subsystem is denoted by . We
know that we have a PV power system that can provide
power during the daytime whose length is . Then,
there are two scenarios regarding different ISRU
subsystem operation lengths: 1) ; 2) .
For the first case (), given a certain level of
mission demand and mission length, we want to
minimize the size of ISRU infrastructures while the ISRU
daily productivity (i.e.,
) remains
constant. Therefore, we want to find the maximum ISRU
daily operation length for subsystem . If there exists
any ISRU subsystem that only work during the daytime
and the operation length is shorter than the daytime
length, we can always extend the operation length to
to achieve higher daily productivity because the solar
arrays work throughout the entire daytime every solar
day.
For the second case ( ), because the solar
arrays can only work during the daytime, we need an
energy storage system to support ISRU subsystems
during the night. As shown in Fig. 3, the extra energy
produced by the solar arrays during the daytime (i.e., area
) needs to be storage and then consumed at night (i.e.,
8
area ). It can involve energy loss during the energy
storage system charging and discharging. Therefore, we
define an energy storage efficiency parameter, . Then,
we get .
Fig. 3 PV Power Supply Analysis
Based on the power supply and demand profile in Fig.
3, we can write the energy conservation equation as Eq.
(15), which means that the total energy produced by the
solar arrays is equal to the summation of the energy
consumed by ISRU subsystems during the daytime and
the night.
We can rewrite the energy conservation equation as
follow, where the left-hand side represents the total ISRU
subsystem power demand during the daytime and the
night while the right-hand side represents the total power
generation of the PV power systems during the daytime.
According to the energy conservation equation shown
in Eq. (16), we define an ISRU subsystem operation
matrix , which is a diagonal matrix. If there are types
of ISRU subsystems, it is a matrix.
Then, we can write the solar power supply constraint
for ISRU architecture design trade studies as follows:
Note that if for all subsystems
, which means ISRU systems are only
operated during the daytime, Eq. (17) can be simplified
into Eq. (14). Moreover, based on the analysis for the first
case , we know that the design space of is
.
Besides the power supply constraint, we also need an
energy storage system capacity constraint. We first
define a battery/fuel cell design matrix . Then the
energy storage constraint can be expressed as Eq. (18),
where the left-hand side is the total energy to be stored
for ISRU night operations while the right-hand side is the
total battery/fuel cell energy storage capacity.
3.3.3 Optimal daily operation length
Now, we have got all the necessary constraints for
ISRU power system trade studies in space logistics.
However, in Eq. (17), if is a design variable in the
formulation, then the term
becomes quadratic. If
ISRU infrastructure design models are nonlinear, then we
can solve the problem nonlinearly and this quadratic term
is not an issue. But if ISRU infrastructure design models
are all linear, then this quadratic term is the only
nonlinear term throughout the entire formulation. It
becomes valuable to perform further analysis on the
optimal daily operation length before the optimization to
see whether we can eliminate this quadratic term and
make the entire problem to be mixed-integer linear
programming.
For ISRU subsystem , if the power system is a
nuclear power system, we know that the optimal
operation mechanism is to operate the ISRU subsystems
continuously, which means . In this situation,
becomes a constant that only depends on the
operating environment of landing sites. If the power
system is the PV power system, our previous analysis
shows that the optimal daily operation length appears
in the range of . Assume that we have a set of
ISRU linear infrastructure models for ISRU subsystem :
These models are all linear functions with zero
intercepts, which is consistent with the infrastructure
design models as proposed in Table 2. We suppose that
the total mission demands for ISRU resources are and
9
the total mission length is solar days. Then, the
required hourly productivity of the ISRU system is
. We can rewrite the ISRU linear models as Eq. (20).
For the specific mission demand and mission length,
the daily operation length is the only variable in Eq.
(20). We want to find the optimal daily operation length
that minimizes the total system sizes.
By substituting Eq. (20) into Eq. (21) and after some
manipulations, we can get,
We take a derivative of with respect to ,
then we get,
In Eq. (23), the numerator of the right-hand side is
always a constant after we know the ISRU infrastructure
design model and ISRU landing site environment.
Therefore, we can say that the system size is
monotonically increasing or decreasing with respect to
the ISRU daily operation length . It means that
or depending on the actual ISRU design
models and operating environments. Then, the entire
problem can be solved linearly.
In summary, if we only consider linear ISRU
infrastructure design models, for nuclear power systems,
the optimal daily operation length,
For PV power systems, the optimal daily operation
length, or
where the actual value of can be determined in
advance after we know the ISRU design models and the
potential landing site operation environments.
4. Human exploration to Mars: modeling and
assumptions
In this section, a case study involving a multi-mission
human Mars exploration campaign is developed based on
the NASA Artemis lunar exploration architecture [20].
This study considers ISRU infrastructure sizing, space
transportation planning, mission demand deployment,
and space resource logistics concurrently. A preliminary
cost model is used as the mission performance
measurement.
The remainder of this section is organized as follows.
Section 4.1 briefly introduces the transportation network
model for the Earth-Moon-Mars transportation system
and spacecraft models. Section 4.2 describes the mission
demand and supply. Section 4.3 introduces mission
operation assumptions and potential landing site
environments for the Moon, Mars and the asteroid.
Finally, Sec. 4.4 lists the cost model to evaluate the
performance of space mission planning.
4.1 Mission scenario
This subsection introduces the mission scenario
settings for human exploration to Mars. The mission
scenario is established based on the NASA Artemis lunar
exploration architecture [20]. The transportation network
model is shown in Fig. 4. It is a network with eight nodes:
ES = Earth
TLI = Trans-lunar injection
NRHO = Near rectilinear halo orbit
PLLO = Polar low lunar orbit
LSP = Lunar south pole
NEO = Near-Earth object
LMO = Low Mars orbit
Mars = Mars
The trajectory and time of flight (TOF) of each arc
are also shown in Fig. 4. We assume that the aeroshell is
40% of the total vehicle mass. After spacecraft land on
the Martian or Earth surface, the aeroshell cannot be
reused again.
There are seven different transportation
vehicles/spacecraft considered in the logistics:
⚫ SLS/EUS: exploration upper stage (mated with Space
Launch System Block-1B);
⚫ CLV/CUS: commercial upper stage (mated with
Commercial Launch Vehicle);
⚫ Mars T/V: Mars transfer vehicle;
⚫ Orion: Orion command and service module;
⚫ T/V: transfer vehicle;
⚫ LLAM: lunar lander ascent module;
⚫ LLDM: lunar lander descent module.
Each vehicle has its own designated service arcs in the
network. For example, the T/V delivers LLDM and
LLAM between NRHO and PLLO or helps the propellant
transportation between TLI and NRHO. Except for the
situation when the T/V is transported from the Earth, the
flight of T/V along other arcs is not permitted. Moreover,
we also assume that when enough propellant is
supported, LLDM and LLAM can also flight back from
PLLO to NRHO without the help of T/V. In this figure,
all dash lines represent human-rated flights and solid
lines represent non-human-rated flights. Note that there
10
is an underlying assumption in this logistics model that
the spacecraft capabilities of the components can be
additively combined. In reality, the interoperability
between spacecraft can be significantly more complex.
This assumption is also extended to spacecraft, which
means the vehicle that is piggybacking on another
spacecraft is treated as a payload of another spacecraft
and thus does not consume its own propellant.
To simplify the analysis, the spacecraft design is not
considered as part of the trade study in space logistics
optimization. Instead, spacecraft with fixed design
parameters are considered for space transportation. Most
of the design parameters come from existing literature.
Some spacecraft parameters that are not available are
extrapolated based on available spacecraft design
models. The spacecraft designs for this case study are
listed in Table 3. Note that, launch vehicles are not
considered as part of space logistics transportation. Only
launch cost and payload capacities are considered for
launch vehicles in mission planning.
Fig. 4 Earth-Moon-Mars transportation network model
Table 3 Spacecraft design parameters
Parameter
Assumed value
SLS/EUS
Propellant type
— —
Propellant capacity, kg
— —
Structure mass, kg
— —
Payload capacity, kg
37,000 [35,36]
Propellant , s
— —
Propellant component mass ratio
— —
CLV/CUS
Propellant type
— —
Propellant capacity, kg
— —
Structure mass, kg
— —
Payload capacity, kg
18,000 [36]
Propellant , s
— —
Propellant component mass ratio
— —
Orion
Propellant type
MON/MMH [37]
Propellant capacity, kg
8,915 [4]
Structure mass, kg
16,572 [37]
Payload capacity, kg
46,147*
Propellant , s
316 [37]
Propellant component mass ratio
— —
T/V
Propellant type
LH2/LOX
Propellant capacity, kg
7,000**
Structure mass, kg
1,194**
Payload capacity, kg
34,864*
Propellant , s
420
Propellant component mass ratio
=5.5:1
Mars T/V
Propellant type
LH2/LOX
Propellant capacity, kg
7,0000**
Structure mass, kg
9,216**
Payload capacity, kg
75,000*
Propellant , s
420
Propellant component mass ratio
=5.5:1
LLAM
Propellant type
LH2/LOX
Propellant capacity, kg
4,800**
Structure mass, kg
3,969**
Payload capacity, kg
20,756*
Propellant , s
420
Propellant component mass ratio
=5.5:1
LLDM
Propellant type
LH2/LOX
Propellant capacity, kg
15,200**
Structure mass, kg
2,367**
Payload capacity, kg
136,148*
Propellant , s
420
Propellant component mass ratio
=5.5:1
* The payload capacity is calculated based on orbital
mechanics according to other spacecraft design
parameters and their designated service arcs. Note that,
this is just an upper bound. The actual payload capacity
of a spacecraft along a specific arc is subject to the
constraints from mission demand and supply, propellant
capacity, and orbital mechanics concurrently.
11
** These design parameters (i.e., propellant capacity and
structure mass) are designed based on a data-based
spacecraft model developed by Taylor [38].
4.2 Mission demand and supply
The transportation mission demand and supply are
assumed based on NASA DRA 5.0. The Earth-Mars
transportation time window opens every 780 days. For
each time window, there is a mission demand for
delivering 51,700 kg payload or habitat to Mars from
Earth. We assume that there is a setup phase before
regular cargo transportation missions. In this setup phase,
there is one time window available to deploy ISRU
infrastructures on the Moon, the NEO, or Mars. There are
two different mission scenarios considered in this Mars
transportation mission. The first scenario is a cargo
transportation mission, which is a one-way mission to
deploy a certain amount of payload to Mars. Another
mission scenario is a human exploration mission, which
is a round-trip. Besides the cargo transportation demand,
after landing on Mars for 500 days, the crew and crew
cabin will come back to the Earth in the human mission.
The total mission of crew and crew cabin is assumed as
20,000 kg. The mission demand and supply of Mars
transportation missions are summarized in Table 4.
Table 4 Demand and supply of Mars transportation
Payload Type
Node
Demand
Time, day
Supply,
kg
Cargo Mission (One way)
Payload
Earth
780*
+51,700
Payload
Mars
980*
-51,700
Payload, propellant,
ISRU plant, ISRU
maintenance spares
Earth
All the
time
+∞
Human Exploration (Round trip)
Payload
Earth
780*
+51,700
Payload
Mars
980*
-51,700
Crew & crew cabin
Mars
1,480*
+20,000
Crew & crew cabin
Earth
1,680*
-20,000
Payload, propellant,
ISRU plant, ISRU
maintenance spares
Earth
All the
time
+∞
* These demands or supplies will repeat every 780 days
4.3 Mission operation assumptions
This section introduces mission operation
assumptions for space mission planning, including
mission time windows, ISRU maintenance requirements,
landing site environment assumptions, power system
degradations, crew consumptions, etc. All these
assumptions and parameters are listed in Table 5. For
mission time windows, we define that they are open for a
few time steps after each rocket launch opportunity.
Then, all space flights are prohibited when time windows
are closed. Moreover, during the space mission, the ISRU
infrastructure requires maintenance. The mass of
necessary maintenance spares is 10% of the ISRU system
mass every year. We also assume that the asteroid (i.e.,
near-Earth object (NEO)) has exactly the same
environment as the Moon. The purpose of considering
NEO ISRU is to analyze the impact of ISRU plant
location on the interplanetary transportation system.
Table 5 Mission operation and landing site environment
parameters and assumptions
Parameter
Assumed value
Mission Operation
Rocket launch interval, day
780
ISRU maintenance, system
mass/yr
10% [6]
PV radiation degradation, /sol
0.014% [39]
Battery charging efficiency
95% [40,41]
Fuel cell energy efficiency
60% [31]
RPS degradation rate, /yr
1.9% [33]
Food consumption rate,
kg/day/person
1.015 [6]
Water consumption rate,
kg/day/person
5.31 [6]
Oxygen consumption rate,
kg/day/person
0.84 [6]
Mars Landing Site Environment
Mars solar irradiance (5-
20°N), kW/m2
0.45 [22]
Regular dust power loss on
Mars
5% [42]
Incident energy loss in dust
storms
65% [42]
Solar day length, hr/sol
24.6 [22]
Period of operation, , hr/sol
7 [22]
System mass contingency
20% [1]
Lunar Landing Site Environment
Lunar landing site
South Pole [20]
Solar irradiance (@ 1 AU),
kW/m2
1.36 [42]
Synodic day length, day
29.5
Illumination conditions, day
27 [43]
Water ice concentration in the
regolith
5.6% [44]
NEO Landing Site Environment
Solar irradiance (@ 1 AU),
kW/m2
1.36 [42]
Synodic day length, day
29.5
Illumination conditions, day
27 [43]
Water ice concentration in the
regolith
5.6% [44]
4.4 Mission cost model
To measure and analyze the impact of ISRU systems
in space exploration, a cost model is needed to provide
an intuitive interpretation of space mission performances.
12
Due to the limited data sources, we propose a preliminary
cost model that is developed based on space cost
estimation tools and past literature. For the high-fidelity
cost model, further analysis is needed.
We assume that the total cost of space architecture is
made up of system development cost, manufacturing
cost/purchase price, and operation cost. The preliminary
cost model is shown in Table 6.
Table 6 Preliminary cost model for Mars transportation mission
Commodities
Development cost
Manufacturing cost
/Purchase price
Operation cost
SLS/EUS
— —
— —
$27,027/kg payload [36, 45]
CLV/CUS
— —
— —
$5,555/kg payload [36, 45]
Mars T/V
— —
$252M [46]
$1M/flight [47]
Orion
— —
$635M [46]
$1M/flight [47]
T/V
— —
$94M [46]
$1M/flight [47]
LLAM
— —
$211M [46]
$1M/flight [47]
LLDM
— —
$430M [46]
$1M/flight [47]
H2O tank (1000kg*)
— —
$1.09M [48]
— —
H2 tank (1000kg*)
— —
$4.78M [48]
— —
O2 tank (1000kg*)
— —
$1.34M [48]
— —
DWE reactor
— —
$1,480/kg [49]
$3,000/kg system/year [48]
SWE reactor
$10,000/kg [50]
— —
$3,000/kg system/year [48]
MRE reactor
$10,000/kg [50]
— —
$3,000/kg system/year [48]
HR reactor
$10,000/kg [50]
— —
$3,000/kg system/year [48]
CR reactor
$10,000/kg [50]
— —
$3,000/kg system/year [48]
Soil excavator
$10,000/kg [50]
— —
$3,000/kg system/year [48]
SOCE reactor
$10,000/kg [50]
— —
$3,000/kg system/year [48]
RWGS reactor
$10,000/kg [50]
— —
$3,000/kg system/year [48]
SR reactor
$10,000/kg [50]
— —
$3,000/kg system/year [48]
SR/RWGS
$10,000/kg [50]
— —
$3,000/kg system/year [48]
Gas acquisition
$10,000/kg [50]
— —
$3,000/kg system/year [48]
PV (solar panels)
— —
$15,773/kg [51]
— —
FSPS (Kilopower)
$13,333/kg [32]
— —
— —
Batteries
— —
$1,000/kg
— —
Maintenance spares
— —
— —
$2,000/kg
H2O
— —
$0.0004/kg
— —
H2
— —
$5.97/kg [47]
— —
O2
— —
$0.15/kg [52]
— —
*The manufacturing cost is defined for the tank with a structure mass of 1,000 kg
5. Human exploration to Mars: results and analysis
Now we have compiled a mission scenario for human
exploration to Mars, considering both interplanetary
cargo transportation and human exploration. This section
shows the mission planning results for two different
mission scenarios. Sensitivity analysis is also performed
to analyze the impact of power system selections, ISRU
technology selections, and lunar ISRU to Mars
transportation missions. The problem is solved using the
Gurobi 8.1 solver through Python on an i9-9900k,
3.6GHz platform with 32GB RAM. The detailed analysis
and discussion of this human lunar exploration campaign
case study are shown in the following parts. As a baseline
mission scenario assumption, the FSPS is selected as the
default stationary power supply system on the lunar and
Martian surface; The PV power system and energy
storage system are used as the default power sources in
space.
5.1 ISRU technology selections
The ISRU power system comparisons between the
PV system and the FSPS considering cargo
transportation (i.e., one-way mission) and human
exploration (i.e., round-trip mission) missions are shown
in Fig. 5 and Fig. 6, respectively.
First, we can find that when considering the cargo
transportation mission, the ISRU plant using the PV
system has limited benefit to the space mission; whereas
the ISRU plant using FSPS can provide a good mission
cost saving. In the human exploration mission, which is
a round-trip. Both the PV system and FSPS can provide
benefits to space transportation. However, the FSPS
provides a significantly larger mission cost saving by
deploying a similar amount of ISRU system as in the
13
mission planning considering the PV system. This result
shows that the FSPS is a better choice than the PV system
under current mission demand and system design model
assumptions.
Fig. 5 Power system comparison in a cargo
transportation mission
Fig. 6 Power system comparison in a human exploration
mission
By fixing the ISRU power system to the FSPS, we
perform a sensitivity analysis on ISRU technology
selection for the human exploration mission as shown in
Fig. 7. Note that, since only the LH2/LOX propulsion
system is considered as the spacecraft propulsion system
in the spacecraft design, some CH4 related Martian ISRU
is not taken into account in the analysis. From the result,
we can find a hybrid ISRU system using SWE and HR
has the best performance. It is slightly better than SWE
ISRU independently, which extracts water from Martian
soil and generates oxygen and hydrogen through water
electrolysis. Both the SWE/HR and SWE systems are
much better than other ISRU systems. Note that, even
though considering a hybrid ISRU system may be
economically more beneficial, technology development,
plant deployment, and system operations can be
significantly more complicated. Therefore, in reality, the
actual mission planning decision-making process also
needs to take into account other factors through
comprehensive evaluations. This result shows that the
proposed space infrastructure design framework is not
only able to perform ISRU technology comparison but
also able to consider the synergistic effect of
technologies.
Fig. 7 ISRU technology selection in human space
exploration mission, FSPS
5.2 Impact of lunar ISRU
In all results discussed above, ISRU systems are all
deployed on Mars. No ISRU architecture is deployed
either on the lunar surface or on the NEO. To analyze the
impact of lunar ISRU on Mars transportation, we run
multiple cases considering lunar ISRU deployment in the
Mars transportation mission. In all the following
numerical experiments, we consider a three-mission
human exploration campaign to Mars. FSPS is the only
power system considered in mission planning. We
compare the total mission cost with respect to different
amounts of lunar ISRU deployed.
First, we consider monetary mission cost as the
mission planning metric using our proposed preliminary
cost model. The result is shown in Fig. 8, where we also
take into account the sensitivity analysis on spacecraft
structure mass. We assume that because of the
technology development in structure and materials
science, the spacecraft structure mass may be reduced
significantly in future human exploration. In this
analysis, we consider the spacecraft structure mass to be
100%, 80%, and 60% of the original baseline spacecraft
dry mass from the sizing model. This result shows that
the deployment of lunar ISRU is not beneficial to the
Mars transportation mission in this considered scenario.
The reason is not that ISRU cannot provide propellant
support to the Mars transportation mission. It is that the
ISRU deployment cost is higher than the benefit it can
provide, which means the lunar ISRU deployment cost
cannot be paid off later. Fig. 9 compares the total mission
0123456
0
1
2
3
4
5
6
Mission Cost, $B
Number of Missions
Total Cost w/ ISRU (PV)
Total Cost w/ ISRU (FSPS)
Total Cost w/o ISRU
0
5000
10000
15000
20000
25000
30000
35000
40000
ISRU Structure (PV)
ISRU Structure (FSPS)
Structure mass, [kg]
(ISRU system)
0 1 2 3 4 5
0
2
4
6
8
10
Mission Cost, $B
Number of Missions
Total Cost w/ ISRU (PV)
Total Cost w/ ISRU (FSPS)
Total Cost w/o ISRU
0
20000
40000
60000
80000
100000
ISRU Structure (PV)
ISRU Structure (FSPS)
Structure mass, [kg]
(ISRU system)
0 1 2 3 4 5
0
2
4
6
8
10
Mission Cost, $B
Number of Missions
SWE+HR
CR
HR
SWE
RWGS
SOCE
No ISRU
14
cost with or without ISRU system deployment in advance
before the Mars transportation mission (e.g., for other
lunar missions). If we do not consider the ISRU
deployment cost, which represents the case that we have
already deployed ISRU in advance during other missions,
the lunar ISRU can provide propellant to support the
Mars transportation mission and reduce the mission cost.
Fig. 8 Lunar ISRU impact in human exploration to
Mars, monetary cost model
Fig. 9 Lunar ISRU impact on human exploration to
Mars w/ or w/o ISRU deployment in advance
Also, the cost metric can be a key factor for the
evaluation of the value of ISRU. To evaluate the
influence of mission planning metrics, we conduct the
same sensitivity analysis with respect to the amount of
lunar ISRU deployment but using initial mass in low-
Earth orbit (IMLEO) as the mission objective. The result
in Fig. 10 shows that when using IMLEO as the mission
cost metric, lunar ISRU can be beneficial to Mars
transportation missions even when considering the ISRU
deployment cost. The main difference comes from the
fact that all commodities are treated in the same weight
when implementing the IMLEO metric. For example,
based on our cost model, the LH2/LOX propellant is less
than $1 per kg; whereas the ISRU plant structure is
$10,000 per kg. It is significantly more expensive to get
1 kg ISRU plant in LEO than the propellant when the
monetary cost model is considered. This result shows that
the analysis results depend on the cost model and
assumptions.
Fig. 10 Lunar ISRU impact in human exploration to
Mars, IMLEO
5.3 Impact of NEO ISRU
The impact of the ISRU plant deployment location is
shown in Fig. 11, where we compare the monetary
mission cost for the cases in which the ISRU is deployed
on the lunar surface or the NEO. Note that, in both
scenarios, the Martian ISRU is also deployed and it is not
influenced by the amount of lunar or NEO ISRU
deployed.
The result shows the ISRU deployment on the NEO
is slightly better than the deployment on the lunar surface
because the NEO is closer to the TLI orbit, which is the
beginning point of the Mars transportation mission.
However, the ISRU deployment cost on the NEO is still
not paid off later. The space mission cost considering
NEO ISRU deployment is always higher than not
deploying any ISRU on the NEO.
The mission cost comparisons between the scenario
with or without ISRU plant deployment in advance is
shown in Fig. 12. It shows the impact of ISRU
deployment cost on space transportation. We can find
that if there is ISRU system deployment in advance
before the Mars transportation, which means the ISRU
deployment cost is not considered as part of the current
mission cost, the NEO ISRU is more effective than the
lunar ISRU in the considered scenario. Moreover, this
result also shows that no matter whether it is lunar ISRU
or NEO ISRU, after the ISRU system mass reaches a
certain level (i.e., 10MT~20MT in our scenarios),
increasing ISRU system size does not provide higher
benefit to the Mars transportation mission.
010000 20000 30000 40000 50000
3.5
4.0
4.5
5.0 Total Cost_100% S/C Structure
Total Cost_80% S/C Structure
Total Cost_60% S/C Structure
Mission Cost, $B
Mass of lunar ISRU, kg
010000 20000 30000 40000 50000
200
250
300
350
400 IMLEO_100% S/C Structure
IMLEO_80% S/C Structure
IMLEO_60% S/C Structure
IMLEO, MT
Mass of lunar ISRU, kg
15
Fig. 11 Lunar or NEO ISRU impact on human
exploration to Mars, monetary cost model
Fig. 12 Lunar/NEO ISRU impact on human exploration
to Mars w/ or w/o ISRU deployment in advance
6. Conclusions
This paper proposes an interdisciplinary space
infrastructure optimization framework for space resource
logistics optimization in future human exploration to
Mars. Multiple in-situ resource utilization infrastructure
design models and cost models are developed based on
existing design concepts and prototypes. A Mars
exploration case study is established based on the NASA
Artemis lunar exploration architecture to evaluate the
performance of the proposed method and analyze the
performance of in-situ resource utilization infrastructures
with different technology selections. Although the
numerical results can vary depending on the assumptions,
the analysis shows that the proposed interdisciplinary
space infrastructure optimization framework can perform
in-situ resource utilization technology comparison and
take into account the synergistic effect of technologies. It
is also able to conduct strategic analysis considering in-
situ resource utilization system deployment locations,
landing site environments, available technology options,
and mission scenarios.
Future research can focus on improving the
computational efficiency of the proposed method for
large-scale long-term space campaign design. A high-
fidelity cost model is necessary to perform in-situ
resource utilization infrastructure trade studies more
accurately. Moreover, the current interdisciplinary
infrastructure optimization method only takes into
account deterministic mission scenarios for space
logistics. Studies focusing on decision-making processes
under stochastics mission operation environments are
also important.
Acknowledgments
This material is partially based upon work supported
by the funding from NASA NextSTEP program
(80NSSC18P3418) awarded to the University of Illinois
at Urbana-Champaign, where this work was initiated.
Any opinions, findings, and conclusions or
recommendations expressed in this material are those of
the authors and do not necessarily reflect the views of
NASA. The authors would like to thank Hang Woon Lee
and Onalli Gunasekara for their reviews and thoughtful
suggestions for improvement.
References
[1] B. G. Drake, Human Exploration of Mars Design
Reference Architecture 5.0, NASA/SP2009-566,
NASA Johnson Space Center, Houston, TX, 2009.
[2] E. Musk, Making Humans a Multi-Planetary Species,
New Space, 5(2) (2017) 46-61.
[3] T. Ishimatsu, O. L. de Weck, J. A. Hoffman, Y.
Ohkami, and R. Shishko, Generalized
Multicommodity Network Flow Model for the Earth-
Moon–Mars Logistics System, J. Spacecr. Rockets,
53(1) (2016) 25-38.
[4] K. Ho, O. L. de Weck, J. A. Hoffman, and R. Shishko,
Dynamic Modeling and Optimization for Space
Logistics Using Time-Expanded Networks, Acta
Astronaut., 105(2) (2014) 428-443.
[5] K. Ho, O. L. de Weck, J. A. Hoffman, and R. Shishko,
Campaign-level dynamic network modelling for
spaceflight logistics for the flexible path concept,
Acta Astronaut., 123 (2016) 51-61.
[6] H. Chen and K. Ho, Integrated Space Logistics
Mission Planning and Spacecraft Design with Mixed-
Integer Nonlinear Programming, J. Spacecr. Rockets,
55(2) (2018) 365-381.
[7] Chen, H., Lee, H., and Ho, K., Space Transportation
System and Mission Planning for Regular
Interplanetary Missions, J. Spacecr. Rockets, 56(1)
(2019) 12-20.
[8] C. Taylor, M. Song, D. Klabjan, O. L. de Weck, and
D. Simchi-Levi, A Mathematical Model for
Interplanetary Logistics, Logistics Spectrum, 41(1)
(2007) 23-33.
Lunar ISRU_100% S/C Structure
Lunar ISRU_80% S/C Structure
Lunar ISRU_60% S/C Structure
NEO ISRU_100% S/C Structure
NEO ISRU_80% S/C Structure
NEO ISRU_60% S/C Structure
010000 20000 30000 40000 50000
3.5
4.0
4.5
5.0
Mission Cost, $B
Mass of lunar or NEO ISRU, kg
Total Cost_Lunar ISRU to be deployed
Total Cost_Lunar ISRU already deployed
Total Cost_NEO ISRU to be deployed
Total Cost_NEO ISRU already deployed
010000 20000 30000 40000 50000
3.0
3.5
4.0
4.5
5.0
Mission Cost, $B
Mass of Lunar or NEO ISRU, kg
16
[9] P. T. Grogan, H. K. Yue, and O. L. de Weck, Space
Logistics Modeling and Simulation Analysis using
SpaceNet: Four Application Case, AIAA 2011-7346,
AIAA SPACE 2011 Conference & Exposition, Long
Beach, California, 2011, 27-29 September.
[10] D. C. Arney and A. W. Wilhite, Modeling Space
System Architectures with Graph Theory, J. Spacecr.
Rockets, 51(5) (2014) 1413-1429.
[11] K. A. Lee, L. Oryshchyn, A. Paz, M. Reddington,
and T. M. Simon, The ROxygen Project: Outpost-
scale Lunar Oxygen Production System Development
at Johnson Space Center, J. Aerosp. Eng., 26(1)
(2013) 67-73.
[12] D. L. Clark, B. W. Keller, and J. A. Kirkland, Field
Test Results of the PILOT Hydrogen Reduction
Reactor, AIAA 2009-6475, AIAA Space 2009
Conference and Exposition, Pasadena, CA, 2009, 14-
17 September.
[13] R. J. Gustafson, B. C. White, and M. J. Fidler, 2010
Field Demonstration of the Solar Carbothermal
Regolith Reduction Process to Produce Oxygen,
AIAA 2011-434, 49th AIAA Aerospace Sciences
Meeting including the New Horizons Forum and
Aerospace Exposition, Aerospace Sciences Meetings,
Orlando, FL, 2011, 04-07 January.
[14] S. S. Schreiner, Molten Regolith Electrolysis
Reactor Modeling and Optimization of In-Situ
Resource Utilization Systems, M.S. Thesis,
Aeronautics and Astronautics Dept., MIT,
Cambridge, MA, 2015.
[15] F. E. Meyen, System modeling, design, and control
of the Mars Oxygen In-Situ Resource Utilization
Experiment (MOXIE) and implications for
atmospheric ISRU processing plants, Ph.D.
dissertation, Aeronautics and Astronautics Dept.,
MIT, Cambridge, MA, 2017.
[16] M. Smirnova, Mars Transportation Vehicle
Concept, Acta Astronaut., 103 (2014) 250-256.
[17] G. Chamitoff, G. James, D. Barker, and A.
Dershowitz, Martian Resource Locations:
Identification and Optimization, Acta Astronaut., 56
(2005) 756-769.
[18] B. Kading, and J. Straub, Utilizing In-Situ Resources
and 3D Printing Structures for a Manned Mars
Mission, Acta Astronaut., 107 (2015) 317-326.
[19] S. Do, A. Owens, K. Ho, S. Schreiner, and O. de
Weck, An Independent Assessment of the Technical
Feasibility of the Mars One Mission Plan – Updated
Analysis, Acta Astronaut., 120 (2016) 192-228.
[20] Forward to the Moon: NASA’s Strategic Plan for
Lunar Exploration, 23 May 2019,
https://www.nasa.gov/sites/default/files/atoms/files/a
merica_to_the_moon_2024_artemis_20190523.pdf,
(accessed 15.06.19).
[21] G. Sanders, Current NASA Plans for Mars In Situ
Resource Utilization, NASA Johnson Space Center,
Houston, TX, 20 February 2018.
[22] Advanced Radioisotope Power Systems Report, D-
20757, Joe Parrish NASA's Office of Space Science,
March 2001.
[23] G. B. Sanders, A. Paz, L. Oryshchyn, K. Araghi,
A. Muscatello, D. L. Linne, J. E. Kleinhenz, and T.
Peters, Mars ISRU for Production of Mission Critical
Consumables - Options, Recent Studies, and Current
State of the Art, AIAA 2015-4458, AIAA SPACE
2015 Conference and Exposition, Pasadena, CA,
2015, 31 August-02 September.
[24] L. Schrenk, “Development of an In-Situ
Resource Utilization (ISRU) Module for the Mission
Analysis Environment HabNet,” M.S. thesis, MIT,
Cambridge, MA, 2015.
[25] Zubrin, R. M., Muscatello, A. C., and Berggren, M.,
Integrated Mars In Situ Propellant Production
System, J. Aerospace Eng., 26 (1) (2013) 43-56.
[26] G. B. Sanders, In-Situ Resource Utilization on Mars
– Update from DRA 5.0 Study, AIAA 2010-799, 48th
AIAA Aerospace Sciences Meeting Including the
New Horizons Forum and Aerospace Exposition,
Orlando, FL, 2010, 04-07 January.
[27] Santiago-Maldonado, E. and Linne, D. L., ISRU
System Model Tool: From Excavation to Oxygen
Production, Space Resources Roundtable IX, Golden,
CO, 2007, 25-27 October.
[28] R. P. Mueller, J. D. Smith, J. M. Schuler, and A. J.
Nick, Design of an Excavation Robot: Regolith
Advanced Surface Systems Operations Robot
(RASSOR) 2.0, Earth and Space 2016, 2016.
[29] D. Rapp, J. Andringa, R. Easter, J.H. Smith, T.J.
Wilson, D.L. Clark, and K. Payne, Preliminary
system analysis of in situ resource utilization for Mars
human exploration, 2005 IEEE Aerospace
Conference, Big Sky, MT, Mar. 2005.
[30] “Solar Power Technologies for Future Planetary
Science Missions,” JPL D-101316, Jet Propulsion
Laboratory, La Cañada Flintridge, CA, Dec. 2017.
[31] “Energy Storage Technologies for Future Planetary
Science Missions,” JPL D-101146, Jet Propulsion
Laboratory, La Cañada Flintridge, CA, Dec. 2017.
[32] “Kilopower Media Event presentation slides (May 2,
2018)”, NASA,
https://www.nasa.gov/directorates/spacetech/kilopo
wer, (accessed 15.11.2018).
[33] Zakrajsek, J. F., Woerner D. F., and Fleurial J.-P.,
NASA Special Session: Next-Generation
Radioisotope Thermoelectric Generator (RTG)
Discussion, NASA,
https://rps.nasa.gov/resources/69/next-generation-
radioisotope-thermoelectric-generator-presentation,
(accessed 15.11.2018).
17
[34] Schmidt, G. R., Sutliff, T. J., and Dudzinski, L. A.,
“Radioisotope Power: A Key Technology for Deep
Space Exploration,” IAC 2009, 60th International
Astronautical Congress 2009, 8, (2009) 6612-6633.
[35] The Great Escape: SLS Provides Power for
Missions to the Moon,
https://www.nasa.gov/exploration/systems/sls/to-
the-moon.html, (accessed 15.06.2019).
[36] NASA chief explains why agency won’t buy a
bunch of Falcon Heavy rockets,
https://arstechnica.com/science/2018/03/nasa-chief-
explains-why-agency-wont-buy-a-bunch-of-falcon-
heavy-rockets/, (accessed 15.06.2019).
[37] Orion Quick Facts,
https://www.nasa.gov/sites/default/files/fs-2014-08-
004-jsc-orion_quickfacts-web.pdf, (accessed
15.06.2019).
[38] Taylor, C., “Integrated Transportation System
Design Optimization,” Ph.D. Dissertation,
Aeronautics and Astronautics Dept., Massachusetts
Inst. of Technology, Cambridge, MA, 2007.
[39] Landis, G. A., Kerslake, T. W., Jenkins, P., and
Scheiman, D., Mars solar power, 2nd International
Energy Conversion Engineering Conference, 1
(2004) 376-385.
[40] Toman, M., Cipin, R., Cervinka, D., Vorel, P., and
Prochazka, P., Li-ion Battery Charging Efficiency,
ECS Transactions, 74(1) (2016) 37-43.
[41] Parvini, Y. and Vahidi, A., Maximizing Charging
Efficiency of Lithium-Ion and Lead-Acid Batteries
Using Optimal Control Theory, 2015 American
Control Conference, Vol. 2015-July, pp. 317-322,
Chicago, IL, Jul. 2015, 01-03 July.
[42] McNatt, J., Landis G., and Fincannon, J, Design of
Photovoltaic Power System for a Precursor Mission
for Human Exploration of Mars, 2016 IEEE 43rd
Photovoltaic Specialists Conference (PVSC), 1343-
1347, Portland, OR, 2016, 05-10 June.
[43] Gläser, P., Scholten, F., De Rosa, D., Marco Figuera,
R. Oberst, J., Mazarico, E., Neumann, G.A., and
Robinson, M.S., Illumination Conditions at the Lunar
South Pole Using High Resolution Digital Terrain
Models from LOLA, Icarus, 243 (2014) 78-90.
[44] Colaprete1, A., Schultz, P., Heldmann, J., Wooden,
D., Shirley, M., Ennico, K., Hermalyn, B., Marshall,
W., Ricco, A., Elphic, R. C., Goldstein, D., Summy,
D., Bart, G. D., Asphaug, E., Korycansky, D., Landis,
D., and Sollitt, L., Detection of Water in the LCROSS
Ejecta Plume, Science, 330(6003) (2010) 463-468.
[45] NASA Making Its Moon Mission Commercial
Could Signal a Paradigm Shift for Deep-Space
Travel,
https://www.theverge.com/2019/3/15/18265417/nasa
-space-launch-system-orion-em-1-commercial-
space-tugs-docking-assembly, (accessed 16.06.2019).
[46] Arney, D. C., and Wilhite, A. W., Rapid Cost
Estimation for Space Exploration Systems, AIAA
2012-5183, AIAA SPACE 2012 Conference &
Exposition, Pasadena, CA, 2012, 11-13 September.
[47] Sowers, G., Transportation Architecture for Cislunar
Space, Dec. 2015,
https://www.mtu.edu/ece/department/faculty/full-
time/zekavat/pdfs/ssp2015-sowers.pdf, (accessed
16.06.2019).
[48] Transportation Network Design Optimization for
Cis-Lunar Space Economy (Phase II), Final Report to
United Launch Alliance, 2017.
https://drive.google.com/file/d/0B0rBRZHrNWuGZ
DBCLXlmWTZtWDA/view, (accessed 16.06.2019).
[49] Hydrogen Generator, 500mL/min.
https://www.grainger.com/product/PARKER-
Hydrogen-Generator-39T172, (accessed 16.06.2019).
[50] Transportation Enabling a Robust Cislunar Space
Economy, ULA, Space Resources Roundtable,
Planetary & Terrestrial Mining Sciences Symposium,
Jun. 2016.
[51] Solar Panels, https://www.cubesatshop.com/wp-
content/uploads/2016/06/ISIS-Solar-Panels-
Brochure-v1.pdf, (accessed 16.06.2019).
[52] Standard Prices for Aerospace Energy Category
Items, Mar. 2018,
https://www.dla.mil/Portals/104/Documents/Energy/
Standard%20Prices/Aerospace%20Prices/E_2018Oc
t1AerospaceStandardPrices_180910.pdf?ver=2018-
09-10-132709-097, (accessed 16.06.2019).
