Benchmarking the Human Exploration of Mars Design Reference Architecture

Abstract

This research benchmarks the Mars DRA 5.0 by quantifying its mission utility in satisfying mission objectives derived from human exploration of Mars community consensus. Acolytion's Progspexion is utilized for modeling and simulation of the DRA in determination of potential mission outcomes and computation of crewed mission to Mars metrics. While establishing a baseline performance, the research determines the DRA's mission utility is limited due to a lack of its consideration during design. In addition, the research indicates the DRA simultaneously contains an overwhelming number of single points of failure with high criticality and lacks an adequate hierarchy of contingency and abort strategies to preclude loss of mission and crew fatality.

Full text

1
Benchmarking the Human Exploration of Mars Design
Reference Architecture
Ja’Mar A. Watson
Watson Institute for Scientific Engineering Research
4201 Wilson Blvd Ste 110-444
Arlington, VA 22203, USA
Email: jamar.watson@sciengresearch.org
Abstract: This research benchmarks the Mars Design Reference Architecture 5.0 by quantifying its
mission utility in satisfying mission objectives derived from human exploration of Mars community
consensus. Acolytion’s Progspexion is utilized for modeling and simulation of the DRA in determination
of potential mission outcomes and computation of crewed mission to Mars metrics. While establishing a
baseline performance, the research determines the DRA’s mission utility is limited due to a lack of its
consideration during design. In addition, the research indicates the DRA simultaneously contains an
overwhelming number of single points of failure with high criticality and lacks an adequate hierarchy of
contingency and abort strategies to preclude loss of mission and crew fatality.
Keywords: Design Reference Architecture; DRA; Mission Utility Analysis; Human Spaceflight; Mars
Exploration; Space Mission Simulation; Progspexion
Reference: Watson, J.A. (2019) ‘Benchmarking the human exploration of Mars Design Reference
Architecture’, Int. J. Space Science and Engineering, Vol. 5, No. 2, pp.138–158.
Nomenclature:
AR&D: Autonomous Rendezvous and Docking
CCF: Common Cause Failure
DAV: Descent and Ascent Vehicle
DRA: Design Reference Architecture
EDL: Entry, Descent, and Landing
EOI: Earth Orbit Insertion
ET: Earth Transit
FoM: Figures of Merit
GCR: Galactic Cosmic Radiation
ISPP: In-situ Propellant Production
ISRU: In-situ Resource Utilization
LEO: Low-Earth Orbit
LoC: Loss of Crew
LoM: Loss of Mission
LoS: Loss of System
LoV: Loss of Vehicle
MD: Mission Dynamic
MoE: Measures of Effectiveness
MOI: Mars Orbit Insertion
MoP: Measures of Performance
MT: Mars Transit
MTV: Mars Transit Vehicle
NTR: Nuclear Thermal Rocket
OCV: Orion Crew Vehicle
PPD: Post Pre-Deployment
REID: Radiation Exposure Induced Death
SHAB: Surface and Habitat Vehicle
SLS: Space Launch System
SPE: Solar Particle Event
SPoF: Single Points of Failure
TEI: Trans-Earth Injection
TMI: Trans-Mars Injection

2
1. Introduction
Human exploration of Mars is the horizon goal of human spaceflight. As such, many mission
architectures have been proposed. However, none of these propositions are substantiated because they
have not quantified their mission utility: ability to achieve mission goals and objectives. This research
establishes a benchmark for human exploration of Mars by quantifying the mission utility of its reference
mission: the Design Reference Architecture (DRA). Mission utility analysis of the DRA 5.0 [1] [2] [3] and
its nominal surface reference mission [4] is facilitated by Acolytion’s Progspexion [5]. In addition to DRA
mission utility simulation, Progspexion provides statistically guided feedback to the DRA mission design
and principles regarding the human spaceflight mission to Mars.
2. Design Reference Architecture
The DRA is classified as a conjunction-class (long stay) split mission: utilizing pre-deployment of
assets a synodic period before the crew initiates spaceflight to Mars. The pre-deployed mission systems,
an unfueled descent and ascent vehicle (DAV) and surface habitat (SHAB), require four space launch
system (SLS) launches to deposit their respected deconstructed modules into low-Earth orbit (LEO).
These four modules are then assembled into the two aforementioned mission systems via autonomous
rendezvous and docking (AR&D). Once the window for trans-Mars injection (TMI) opens, the DAV and
SHAB transit to Mars by nuclear thermal rocket (NTR) via minimum energy trajectories. After Mars
transit (MT), aerodynamic Mars orbit insertion (MOI) is performed to place the assets into a 1-sol, 250 x
33,793 kilometer, areocentric orbit. While the SHAB loiters in Mars orbit, the 40-ton payload DAV
performs entry, descent, and landing (EDL) to the Martian surface to initiate in-situ propellant production
(ISPP) at the designated landing site.
If pre-deployment is successful and the mission systems remain operational, the crew of six initiates
its journey to Mars a synodic period later. The crewed Mars transit vehicle (MTV) is also constructed in
LEO, requiring three SLS launches. Once AR&D is completed, the crew is launched separately in the
Orion crew vehicle (OCV) to rendezvous with and board the assembled MTV. Again, the mission system
begins TMI by NTR when the injection window opens; however, crewed vehicles in the DRA travel fast-
transit trajectories to Mars. After MT, the MTV performs propelled MOI utilizing the NTR to place the
crew into the 1-sol areocentric orbit. The crew then initiates the Mars exploration phase of the mission by
AR&D with the loitering SHAB. When completed, the crew is able to transfer to the SHAB to perform
EDL to the Martian surface in close proximity to the pre-deployed DAV.
Once the crew is deposited to the landing site and acclimated to the Martian environment, surface
exploration may commence. Here the crew adopts the commuter exploration strategy where the SHAB
serves as the central habitat between sorties, utilizing mobility systems to investigate regions of scientific
interest. This is performed for approximately 500 days.
Upon expiration of the allotted time for surface exploration, the crew transits to the DAV to perform
ascent back to areocentric orbit. The crew ends the Mars exploration phase of the mission by performing
AR&D of the DAV with the MTV. Once the crew transfers to the MTV and the trans-Earth injection (TEI)
window opens, the crew initiates its return to Earth. When Earth transit (ET) is completed, the crew arrives
back to the surface of the Earth by direct entry with the OCV. It is this mission profile [6] (Fig. 1) that is
used to define the human exploration of Mars DRA 5.0.

3
Fig. 1 Design Reference Architecture Mission Profile [6]
3. Methods
A benchmark for human spaceflight missions to Mars is achieved by quantifying the satisfaction of
mission objectives by the human exploration of Mars DRA 5.0. The methods of this research follow the
Holistic Methodology for Stochastic Mission Utility Analysis [7], with primary focus on applicable tasks
of defining mission objectives and their associated metrics, performing modeling and simulation of the
mission, and establishing a statistical benchmark by quantifying the mission metrics and performing
analytics of the mission utility simulation dataset.
3.1. Mission Objectives
The mission objectives (Table 1) are derived from Mars exploration community consensus. The Mars
Architecture Working Group [1], with the Mars Exploration Program Analysis Group (MEPAG) [8] [9]
providing significant contribution to the planetary science objectives, presents them. The priority of

4
objectives is consistent with their assessments with a percent of mission objective satisfaction developed
heuristically from their prioritization. It is important to note that the MEPAG agrees objectives are not
tiered; meaning, it is possible a high valued sub-objective of a lower priority objective could be of equal
or greater value than a low valued sub-objective of a higher priority objective [9]. Table 1 therefore
establishes the numerical mission utility obtainment in satisfying mission objectives during mission
simulation.
Table 1 Mission Objectives and Sub-Objectives
OBJECTIVE
DESCRIPTION
VALUE
1
Life on Mars
50%
1.1
1.2
Determine Prior Habitability
Search for Extinct Life
Determine Current Habitability
Search for Extant Life
25%
25%
2
Martian Planetology
25%
2.1
2.2
2.3
2.4
2.5
Determine Current Climate
Detail Atmosphere and Environment
Determine Geologic Record
Detail Geologic Processes
Determine Climate History
Detail Climate Changes
Characterize Planet Interior
Determine Interior Evolution
Determine Planet Evolution
Determine Planetary System Evolution
10%
5%
5%
3%
2%
3
Colonization Preparation
20%
3.1
3.2
3.3
3.4
Obtain Knowledge for Orbital Missions
Obtain Knowledge for Surface Missions
Obtain Knowledge for Missions to Moons
Obtain Knowledge for Sustained Presence
7%
7%
3%
3%
4
Ancillary Science
5%
4.1
4.2
4.3
Study Heliophysics from Mars
Study Astrophysics from Mars
Conduct Public Engagement Activities
2%
2%
1%

5
3.2. Interdependencies
Fig. 2 displays the interdependencies of the non-ancillary mission sub-objectives [9]. This identifies
how satisfaction of one sub-objective is dependent on the means of identification of another objective. By
definition, this is not equivalent to intercorrelation and does not need to be mutual between sub-objectives.
DEPENDENCY
D
E
P
E
N
D
E
N
T
1.1
1.2
2.1
2.2
2.3
2.4
2.5
3.1
3.2
3.3
3.4
1.1
X
X
X
1.2
X
X
2.1
X
X
X
X
2.2
X
X
X
2.3
X
X
2.4
X
2.5
X
3.1
X
X
3.2
X
X
X
3.3
X
3.4
Fig. 2 Mission Objectives' Interdependencies
To reference Fig. 2, a dependent sub-objective is traced across its row to determine its interdependencies.
If a cell contains an X then that column’s sub-objective is responsible for the dependency. Therefore, the
row sub-objective is dependent on X marked column sub-objectives. Mission utility in satisfying mission
objectives is tabulated by successful scientific investigations and Fig. 2 is used to update mission
intelligence using Bayesian inference following sorties [11].

6
3.3. Scientific Investigations
The primary driver of delivering the crew to the surface of Mars is to perform scientific investigations
to accomplish mission objectives. Derived from the mission sub-objectives are 30 investigations [1] [8]
[9] (Error! Reference source not found.). This establishes the mission phenomenology.
Table 2 Crewed Mars Mission Scientific Investigations
No.
Investigation
Objective
Utility
I
Identify past habitable environments and characterize its habitability
1.1
12.5
II
Identify biosignatures of prior ecosystem
1.1
12.5
III
Identify present habitable environments and characterize its
habitability
1.2
12.5
IV
Identify biosignatures of current ecosystem
1.2
12.5
V
Determine processes that control the composition, dynamics, and
exchange of the upper-atmosphere and plasma environment
2.1
5.0
VI
Determine processes that control the composition, dynamics, and
exchange of the surface and lower-atmosphere
2.1
5.0
VII
Characterize geologic environments and processes relevant to the crust
2.2
2.5
VIII
Determine age of geologic units and events on Mars
2.2
2.5
IX
Find and interpret physical and chemical records of past climates
2.3
2.5
X
Determine present escape rates of key species
2.3
2.5
XI
Identify and evaluate manifestations of crust-mantle interactions
2.4
1.5
XII
Determine age and processes of accretion, differentiation, and thermal
evolution
2.4
1.5
XIII
Determine the geologic composition and interior of the moons
2.5
1.0
XIV
Determine the material and impactor flux of Mars’ environment
2.5
1.0
XV
Characterize atmospheric effects on aerobraking and aerocapture for
missions
3.1
3.5
XVI
Determine orbit environment of high areocentric orbit that effects
missions
3.1
3.5
XVII
Characterize atmospheric effects on EDL for missions
3.2
1.4
XVIII
Determine if biohazards and ionization are present in mission surface
environments
3.2
1.4
XIX
Characterize atmospheric ISRU around mission surface environments
3.2
1.4
XX
Determine landing site hazards in a mission
3.2
1.4
XXI
Determine Aeolian dust effects on mission systems
3.2
1.4
XXII
Determine the planetary properties of the moons for missions
3.3
1.5
XXIII
Determine surface and orbital environment of the moons for missions
3.3
1.5
XXIV
Characterize aqueous environments and locale of water
3.4
1.5
XXV
Provide detail maps of potential landing sites
3.4
1.5
XXVI
Conduct solar observations from Mars’ surface
4.1
1.0
XXVII
Conduct solar observations from Mars orbit
4.1
1.0
XXVIII
Conduct laser ranging on Mars’ surface
4.2
1.0
XXIX
Conduct laser ranging from Mars orbit
4.2
1.0
XXX
Conduct publicity events with astronauts
4.3
1.0

7
3.4. Progspexion
With the DRA and the calculation of mission utility defined, Progspexion is used to perform mission
utility analysis. Progspexion is a software suite of stochastic Monte Carlo space mission simulators, whose
modeling and simulation method has been verified and validated in backtesting of the Apollo 11-17
crewed missions to the Moon [10]. To analyze the DRA, Progspexion’s Mars mission simulator, Areo,
completes numerous DRA simulations, altering the intrinsic, specified, and input stochastic parameters of
the Mars mission model, until the average mission utility converges on a value. All parameters of the
model derived from the DRA are defined by static and deterministic variables. However, in addition to
the 60+ intrinsic possibilistic determinations of mission event outcomes and contingency actions, the
following variables are specified as stochastic:
1. the synodic period of the mission – making variable the mission epoch, starting Martian season
during surface exploration, and subsequently the occurrences of dust storms during the mission
2. interplanetary transit durations, which makes dynamic the apportionment of exploration phases
in the mission
3. EDL accuracy of mission systems, altering the possible ranges at which the DAV and SHAB
are deposited from the nominal landing site
4. selection of a landing site among the 58 potential locations identified by the MEPAG’s Human
Exploration of Mars – Science Analysis Group [8]. Additionally, the following input variables
are defined as static
5. mission intelligence is set static at probable, which directly affects the initial knowledge of
Mars in the mission simulation concerning scientific investigations
6. the mission operational weather epoch is static at mid-solar cycle occurrence, making solar
particle event (SPE), galactic cosmic radiation (GCR), and radiation exposure induced death
(REID) occur at their standard space weather probabilities.
Also, inherent in Progspexion is analytics of the simulation dataset. This includes the computation of
many measures of effectiveness (MoE: metrics quantifying the effectiveness of satisfying mission
objectives), figures of merit (FoM: metrics quantifying multi-attribute/criteria effectiveness), and
measures of performance (MoP: metrics quantifying operational, systematic, technological, etc.
performance in the mission) [7], as well as tabulation of mission event occurrences and outcomes [5].
Therefore, not only is the DRA benchmarked by mission utility obtainment, but its overall mission
portfolio provides a baseline performance for the design and/or analysis of all human exploration of Mars
mission architectures and operational concepts.

8
4. Results
4.1. Simulation
Progspexion completed 1 million simulations of the DRA, which is sufficient to exhibit convergence
behavior (Fig. 3). While being run on a MacBook Pro with a 2.3 GHz quad core processor and 16GB of
RAM, Progspexion is able to complete well over 1,000 DRA simulations per second.
Fig. 3 DRA Convergence Behavior in Progspexion
It is found that the pre-deployment phase of the mission alone is exceptionally difficult - failing in 94.98%
of simulations. Fig. 4 breaks down failure contributions among the sequence of four SLS launches, two
AR&D integrations, two NTR TMIs of the integrated modules, two MTs, two aerodynamic MOI, Mars
EDL of the 40-ton payload DAV, and simultaneous SHAB loiter and DAV ISPP.

9
Fig. 4 Pre-Deployment Failure Events
Launch
AR&D
TMI
MT LoV
Aerocapture
MOI
DAV EDL
SHAB Loiter ISPP
SHAB Power SPE GCR DAV EDL
Accuracy

10
However, the mission is sequential, and the results are due to the fact that launch, AR&D, and TMI are
the initial mission events for pre-deployment, thus occur more often than subsequent phases when failures
are present. Fig. 5 shows that when the two assets successfully initiate Mars transit, the leading cause of
pre-deployment failure from then on is the aerodynamic MOI, which comprises 57.48% of post-TMI
failures.
Fig. 5 Post TMI Pre-Deployment Failure Events
MT LoV
Aerocapture
MOI
DAV EDL
SHAB Loiter ISPP
SHAB Power
SPE DAV EDL
Accuracy

11
Similarly, proceeding the MOI phase of pre-deployment, the leading cause of failure is loss of the DAV
during EDL (Fig. 6). Here it comprises 49.63% of failures, with ISPP failure contributing 27.82%.
Fig. 6 Post MOI Pre-Deployment Failure Events
Although only representing the remaining 5.02% of mission simulations, results for the post pre-
deployment (PPD) crewed portion of the mission are still obtained by isolating the runs in which pre-
deployment was successful. Of these, only 1.21% successfully saw the crew complete the mission and
return to Earth; meaning, the DRA successfully completes the mission in only 0.06% of the total
simulations. The failures for the crewed portion of the mission are shown in Fig. 7.
DAV EDL
SHAB Loiter
ISPP
SHAB Power
SPE
DAV EDL
Accuracy

12
Fig. 7 Post Pre-Deployment Mission Failure Events
Launch
LoM
Launch LoC
LEO AR&D
LEO Repair
Miss Insertion
Window
TMI
TMI LoC
TMI Stranded
MT LoC
MT LoV
MTV Habitat During
MT
MOI Abort
MOI
SHAB AR&D
SHAB Transfer
Mars EDL
MTV Loiter
DAV Damaged
ISPP
Surface
Abort
SHAB
Ascent
Ascent Abort
MTV AR&D
TEI
TEI LoC
ET LoV
MTV Habitat During
ET OCV
EDL
Earth EDL CCF
REID Human
Error LoM
Human Error LoC
GCR SHAB EDL Accuracy
DAV EDL Accuracy
Miss Return
Window

13
An abort was required in 64.88% of the missions in which the crew initiated its journey (Fig. 8), resulting
in a loss of mission (LoM).
Fig. 8 Post Pre-Deployment LoM Events
Again, these failure events are separated into post-TMI and post-MOI mission failure events, shown in
Fig. 9 and Fig. 10 respectively.
Launch
LEO AR&D
LEO Repair
Miss TMI
Window
TMI
MOI Abort
MOI Error
SHAB AR&D
SHAB Transfer ISPP
Surface Abort Loss of SHAB Human Error SHAB
EDL

14
Fig. 9 Post Pre-Deployment Post-TMI LoM Events
MOI Abort
MOI Error
SHAB AR&DDAV Propellant
Storage
Surface Abort
Loss of SHAB
SHAB Transfer
Human Error
SHAB EDL
Accuracy

15
Fig. 10 Post Pre-Deployment Post-MOI LoM Events
Propellant
Storage
Surface Abort
Loss of SHAB
Human Error SHAB EDL Range

16
However, only 69.97% of aborts were actually successful, with the remaining resulting in loss of crew
(LoC). These, along with events that directly result in LoC are displayed in Fig. 11.
Fig. 11 LoC Events
OCV
Launch
TMI
TMI Stranded
Mars Transit
Loss of MTV
During MT
MTV Habitat
During MT
SHAB
EDL
MTV Loiter
DAV Damaged
Ascent
Ascent Abort
MTV AR&D
Loss of MTV
During TEI
TEI
Loss of MTV
During ET
MTV Habitat
During ET
EOI
OCV EDL CCF
REID Human Error GCR DAV EDL
Accuracy
Miss
Return
Window

17
The following figures break down the LoC contributing events into phases. The first is for events on the
way to Mars (Fig. 12).
Fig. 12 Pre-MOI LoC Events
The second is events at Mars (Fig. 13).
Fig. 13 LoC Events at Mars
OCV
Launch
Loss of
MTV
During TMI
TMI Stranded
MT
Loss of MTV
During MT
MTV Habitat CCF Human Error GCR
SHAB EDL
MTV Loiter
DAV
Damaged
Ascent
Ascent Abort
DAV Post-
Ascent AR&D
CCF
Human Error GCR SHAB EDL
Accuracy

18
After which, displayed are Martian return events (Fig. 14).
Fig. 14 LoC Events Returning from Mars
Loss of MTV
During TEI
TEI
Loss of MTV
During ET
MTV
Habitat
EOI
OCV EDL
CCF
REID Human
Error
GCR Miss Return
Window

19
To keep these three PPD mission phases (journey to, exploration of, and return from Mars) in reference
to the mission as a whole, the sequential results are also captured. Post-TMI LoC events are shown in Fig.
15.
Fig. 15 Post-TMI LoC Events
MT
Loss of MTV
During MT
MTV Habitat
During MT
SHAB EDL
MTV
Loiter
DAV
Damaged
Ascent
Ascent Abort
MTV AR&D
Loss of MTV
During TEI
TEI
Loss of
MTV
During ET
MTV Habitat
During ET
EOI
OCV
EDL
CCF
REID
Human
Error
GCR DAV EDL
Range
Miss
Return
Window

20
Fig. 16 displays LoC events post-MOI.
Fig. 16 Post-MOI LoC Events
SHAB EDL MTV Loiter
DAV Damaged
Ascent
Ascent Abort
MTV/SHAB
AR&D
Loss of MTV
During TEI
TEI
Loss of MTV
During ET
MTV Habitat
During ET
EOI
OCV EDL
CCF
REID Human Error GCR DAV Range Miss
Return
Window

21
Fig. 17 outlines post-SHAB EDL LoC events.
Fig. 17 Post-SHAB EDL LoC Events
LoC events Post-TEI are shown in Fig. 18.
Fig. 18 Post-TEI LoC Events
Ascent Ascent Abort
MTV AR&D
Loss of MTV
During TEI
TEI
Loss of MTV
During ET
MTV Habitat
During ET
EOI
OCV EDL
CCF
REID
Human Error GCR Miss
Return
Window
Loss of MTV
During ET
MTV Habitat
During ET
EOI
OCV EDL
CCF REID
Human Error GCR

22
4.2. Analytics
Progspexion’s analytics of the simulation dataset also provides quantification of numerous crewed
mission to Mars metrics. Although Progspexion calculates a host of metrics, the major metrics impacting
the DRA’s performance are discussed in this section. The first relevant metric is that of mission utility
(satisfaction of mission objectives from Table 1), which computes to a mean of 0.22% and median of 0%
amidst an overwhelming number of mission failures. Even when considering only PPD, the DRA’s mean
mission utility is 4.47%, its median mission utility is still 0% with a median absolute deviation of 0%, and
an interquartile range of 1.4%. Although outliers, the DRA obtains an average mission utility of 77.28%
in missions which are successfully completed. Additionally, the dataset shows the DRA is capable of
achieving its maximum mission utility of 95.5% and its maximum 28 feasible investigations of the 30
(DRA cannot achieve investigation XIII since the concept does not include exploration of Phobos or
Deimos and cannot achieve investigation XV because the crewed MOI is not aerodynamic).
Computation of the risk metric (probability of LoC or LoM) further quantifies the degree of mission
failures, tabulating LoM in 97.26% of missions and LoC in 2.68% of missions. However, risk alone is
misleading with so many pre-deployment failures; therefore, the PPD risk is also calculated to determine
this metric when pre-deployment is successful and the crew launches in the mission. Once the DRA
progresses to PPD, the risk becomes 45.43% LoM and 53.33% LoC. To clarify, although there is also a
LoM in LoC outcomes, the LoM percentages are tabulated for simulations where the mission is loss but
the crew is not. This helps distinguish mission outcome factors. Therefore, absolute LoM outcomes is the
cumulative percentage from the sum of LoM and LoC values.
The resiliency metric (ability to maintain mission utility with loss of system (LoS), loss of vehicle
(LoV), and mission dynamic (MD)) shows the mission utility metric is nearly unchanged when these
perturbations are present. This, however, is a mathematical consequence of a MD (mission event altering
the mission phasespace trajectory) occurring in 99.40% of missions and a LoV or LoS occurring in 99.36%
of missions. When calculating resiliency for missions with only a MD the mission utility is also
unchanged; however, when only a LoV or LoS occurs, the mission utility reduces to 53.50% of the
average. With these mission system losses, the metric of fragility (ability to result in LoC or LoM from
LoS or LoV) demonstrates that LoM was the result in 37.08% of LoV/LoS missions and LoC in 0.94%.
Additionally, the robustness metric (ability to result in LoC or LoM from MD) shows 93.18% LoM and
0.90% LoC in MD instances. These metrics, however, are also skewed by dominance of pre-deployment
failure, which eliminates crew involvement and therefore LoC outcomes. If only considering PPD
missions with successful pre-deployment, the resiliency indicates that the average mission utility
(compared to the PPD average mission utility) decreases 1.57% with MD and the resiliency with LoS and
LoV shows a 39.60% decrease in DRA mission utility. PPD fragility is 7.97% for LoM and 27.53% for
LoC; PPD robustness alters to 26.13% LoM and 17.75% LoC. Tabulation of these metrics comprises
Table 3.

23
Table 3 Average Value of Mission Metrics
Metric
All %
PPD %
Mission Utility
0.22
4.47
Risk (LoC)
2.68
53.33
Risk (LoM)
97.26
45.43
Resiliency (LoS/V)
-53.50
-39.60
Resiliency (MD)
-0.00
-1.57
Fragility (LoC)
0.94
27.53
Fragility (LoM)
37.08
7.97
Robustness (LoC)
0.90
17.75
Robustness (LoM)
93.18
26.13
Results of temporal variability in DRA mission utility is affected by both solar and Martian seasons.
Mission utility showed little variance in missions where an SPE occurred but exhibited a 47.30% decrease
if a GCR storm occurred during the mission. A dust storm was encountered in 94.07% of missions;
however, there is no statistically significant variation in mission utility when varying the starting Martian
season during surface exploration, with autumnal, winter, vernal, and summer mission utilities being
roughly equivalent.
5. Discussion
Although the DRA is a well-constructed engineering-based solution, the research benchmarks the DRA
as having little to no mission utility with stifling risk. This result occurs due to the lack of mission utility
analysis in the DRA’s design process itself. The DRA removes concurrent consideration of primary
mission engineering metrics (MoE, such as mission utility) in its design. In fact, the metrics backing all
DRA design decisions are either secondary mission engineering metrics (FoM, such as cost and risk), or
tertiary metrics (MoP, such as initial mass to LEO and mission gear ratio). Therefore, the DRA provides
a poor benchmark because its design is not based on satisfying its own objectives.
Additionally, the mission simulation determines pre-deployment design is a limiting factor. This is
primarily due to high criticality. While not all failures affect the mission equally, the results highlight
many recurrent contributors: the first of which is the pre-MT sequence. Although it contains single points
of failure (SPoF), many of its failure events occur in LEO (Fig. 4), which can be resolved with repair
actions. However, subsequent events do not have this luxury. Results show that aerocapture failure leads
directly to pre-deployment failure and is a primary contributor (Fig. 5). Being an SPoF, it is also of high
criticality. Therefore, redundancy or contingency must be built into the DRA design if aerocapture and its
associated systems cannot be made more reliant. Due to apriori knowledge from FoM that aerodynamic
MOI is a high-risk maneuver (driver for why the crewed MOI is propulsive), its selection now requires
more justification than just trades on MoP concerning mass savings. Since its selection over other
alternatives and trades on its design did not include consideration of MoE, it has not been determined if
aerocapture’s associated decreases in the number of required launches and launch campaign duration are
worth the SPoF limitations, especially when the results also suggest that the higher mass propulsive MOI
produces a lower percentage of LoM outcomes (Fig. 9).
The DAV is of great concern to pre-deployment failure. Particularly, its 40 t payload EDL is a primary
contributor (Fig. 6). However, the choice of EDL mass in the DRA design has not been justified against
MoE. Since EDL design is highly selectable based on MoP, such a justification of EDL requirements, in

24
the presence of FoM and in terms of both division and summation of payload mass, is crucial to DRA
success. This also drives the requirement for redundant and contingent design of EDL in the mission. The
results therefore show not only is the EDL architecture a primary contributor to pre-deployment failures,
but 40t payload EDL is a high criticality SPoF.
Even if EDL of the DAV is successful, both its ISPP (Fig. 6) and propellant storage (Fig. 9) functions
lead directly to LoM if failed. Since ISPP is strictly during the pre-deployment phase, it suffers many of
the same limitations as the aforementioned contributors. However, its criticality may be mitigated with
relatively low-mass system design changes to improve its reliability. While DAV propellant storage has
the added benefit of repair action during crew surface exploration, it may not be advantageous to rely on
this capability. While pre-crewed-EDL detection keeps DAV propellant storage failure as a LoM outcome,
these results could convert to LoC if the crew attempts to repair or maintain DAV propellant storage but
is unsuccessful, as seen by the impacts of the DAV losing its ascent capability due to damage (Fig. 13).
Therefore, the DAV and its operational concept impose many limitations which must be addressed in the
DRA mission design in addition to its system design. The remaining key SPoF from the pre-deployment
failure contributions is ensuring the availability of the SHAB and its subsystems during loiter (Fig. 6).
However, even if pre-deployment can be guaranteed to be successful, there are still a host of failures
that must be addressed during the crewed portion of the mission (Fig. 7). While many contributors produce
LoM outcomes (Fig. 8), the PPD metrics indicate that successful pre-deployment will convert mission
outcomes into LoC (Fig. 11). During the crewed portion of the mission, the deployment sequence affords
the same luxury of taking place in LEO but LoM outcomes are still present with repair attempts (Fig. 8).
The conversion from LoM to LoC outcomes already becomes prevalent during MT. During pre-
deployment, a LoS or LoV leads to LoM (Fig. 5); however, an unrecoverable MD or loss of MTV during
MT leads to LoC (Fig. 12). The same is true for ET after Mars exploration (Fig. 14) and these SPoF go
beyond the MTV habitat alone (Fig. 15). Therefore, the DRA design requires significant change to ensure
the availability and reliability of the MTV since no major contingencies are available during interplanetary
transit. The results therefore indicate that mission-level redundancy of the MTV may be required to reduce
its criticality if system and sub-system design alterations are not sufficient.
As previously discussed, the propulsive approach to MOI with the crew over aerodynamic MOI suits
the mission well in avoiding SPoF. However, while these events do not directly lead to LoC, the metrics
show that most mission aborts following LoM are unsuccessful. Therefore, when abort events are
significant contributors (Fig. 9)(Fig. 10) their events convert into more LoC outcomes. Consequently, the
results show that the DRA’s operational concept must be redesigned regarding its hierarchy of
contingency and abort actions. While crewed EDL is also of high criticality (Fig. 13), the results
furthermore suggest the assessment of failure modes and effects must address the SPoF criticality during
the crew’s return to Earth (Fig. 16)(Fig. 17)(Fig. 18).
Finally, the results are indicative of GCR being a limiting factor. This does not apply to SPE since
indications and warning is partially established and its effects criticality is less than GCR. Therefore, the
DRA should be executed when the probability of encountering significant GCR is least. The results
indicate this change in epoch would not affect surface exploration since there is no statistically significant
differential in mission utility across temporal variability of Martian seasons and the crew’s subsequent
dust storm encounters. However, dust storm failure modes should be included in analysis of surface
exploration abort strategies (Fig. 10) since the metrics indicate a dust storm encounter during the mission
is common.

25
6. Conclusion
Benchmarking the DRA establishes its baseline performance for human exploration of Mars as having
little to no mission utility with stifling risk. This is due to:
1. a lack of mission utility analysis during the DRA design
2. an overwhelming number of SPoF with high criticality
3. an ineffective hierarchy of contingencies and aborts.
It is recommended human exploration of Mars underdo concept development and experimentation to re-
architect the DRA with concurrent quantification of its efficacy.
Acknowledgements
The author would like to thank Acolytion for enabling the use of Progspexion in this research and all
reviewers for their recommended edits to this article for publication.
References
[1]
National Aeronautics and Space Administration, "Human Exploration of Mars Design Reference
Architecture 5.0," NASA, NASA/SP 2009-566, Houston, TX, July 2009.
[2]
National Aeronautics and Space Administration, "Human Exploration of Mars Design Reference
Architecture 5.0 - Addendum," NASA, NASA/SP-2009-566-ADD , Houston, TX, July 2009.
[3]
National Aeronautics and Space Administration, "Human Exploration of Mars Design Reference
Architecture 5.0 - Addendum 2," NASA, NASA/SP-2009-566-ADD2, Houston, TX, March 2014.
[4]
S. J. Hoffman, "The Mars Surface Reference Mission: A Description of Human and Robotic Surface
Activities," NASA, NASA/TP-2001-209371, Houston, TX, December 2001.
[5]
J. A. Watson, "Progspexion: Simulators for the Prognostication of Space Exploration Missions,"
Acolytion LLC, Cape Canaveral, FL, 2017.
[6]
Jet Propulsion Laboratory, "Missions to Mars: Robotics and Humans Together," Johnson Space
Center, Houston, TX, Novemeber 2011.
[7]
J. A. Watson, "Holistic Methodology for Stochastic Mission Utility Analysis," Int J System of
Systems Engineering, vol. 8, no. 2, pp. 174-188, 2017. doi: 10.1504/IJSSE.2017.10009471
[8]
Mars Exploration Program Analysis Group - Human Exploration of Mars Science Analysis Group,
"Planning for the Scientific Exploration of Mars by Humans," NASA, Pasadena, CA, January 2008.
[9]
MEPAG, "Mars Science Goals, Objectives, Investigations, and Priorities," NASA, Pasadena, CA,
June 2015.
[10]
J. A. Watson, "Modeling for Concise Space Mission Utility Simulation with Apollo as Exemplar," J
Astronaut Sci.,vol. 66, no. 4, pp/ 404-418, 2017
H. Dezfuli et al, "Bayesian Inference for NASA Probabilistic Risk and Reliability Analysis,
NASA/SP-2009-569, National Aeronautics and Space Administration