Design and Multi-objective Optimization of Extraplanetary surface habitats

Postdoctoral Research

Valentina Sumini

MIT | Dept. of Civil and Environmental Engineering

Digital Structures Research Group


1. Introduction 

A renewed interest in space exploration, as evidenced by the recent funding that NASA received for sending humans to Mars by 2030, leads to new challenges in architecture and structural engineering. Space architecture is deeply interdisciplinary and connects different fields of research such as aerospace engineering, architecture, design, space science, medicine, psychology and art. It combines together the accuracy of technical systems, human needs for working and living, the interface design for the relationship between humans, and the built and natural environment. In addition to traditional knowledge of planning and building processes, special knowledge is needed regarding how to design for humans in extreme environment and how to do so creatively, narrowing down to every specific detail of the construction system. 

Unlike structural engineering for the built environment on Earth, there are virtually zero rules of thumb or design precedents to draw on for construction on Mars and on the Moon. The basic requirements of future space habitat structures are defined by their ability to protect their occupants and provide usable space to live and work, but how best to meet these requirements remains an open question. There is exciting potential to shape this discussion with fundamental structural engineering principles and forward-looking material and fabrication strategies.  

The idea of Moon colonization originated far before the age of actual space exploration, as the Moon is the only Earth's natural satellite. Recent discoveries of considerable amounts of water close to the Lunar poles as well as the need to optimize space exploration by exploiting Moon bases and thus reducing the amount of fuel required for take-off (since Lunar gravity is far lower than the Earth's) makes this opportunity more concrete and appealing. However the establishment of a manned human colony on the Moon (or on Mars) will need some form of infrastructure to shelter the astronauts and scientific instrumentations from a very harsh environment. 

The habitat structural design problem has to be formulated as a multi-objective optimization of the structural form and cross-section variables where the goals are to provide minimum transportation and construction costs, minimize the probability of loss due to radiation and micrometeorite events, and provide an internal layout which allows for a high level of crew habitability and operational efficiency (respecting the functional diagrams of the habitat).  Direct trade-offs exist between these objectives allowing designers to choose between them based on different prioritization goals and strategies. Therefore, designing a space architecture implies the development of new computational design methods that respond to both functional, structural and physical requirements, offering new ways to support future space exploration. 

At the moment, my research at Massachusetts Institute of Technology is applied to both Moon and Mars environments: a permanent human outpost on the Moon, an emergency astronaut module,  a city on Mars of 10,000 people (competition held by Mars City Design) and a commercially enabled LEO / Mars Habitable Module built around common Node Modules (competition held by RASC-AL). All the design teams are multidisciplinary and involves architects, aerospace, civil and management engineers from different MIT Department and researchgroups. 


2. Structural requirements for space exploration habitats 

Structural systems for space habitats must be designed for four main loading types: internal pressure, reduced gravity, thermo-elastic loads and micrometeoroid impact.  

The most significant of these loads is pressurization.  Due to the absence of atmosphere on the Moon (and only a minimal atmosphere on Mars), a pressure differential of up to 100 kPa (0.99 atm) across the habitat enclosure is required to sustain Earth level pressures inside, resulting in outward pressures on the structure that are several orders of magnitude greater than conventional structural loads due to gravity (one sixth on the Moon and one third on Mars), wind, etc. Consequently, the structure will be mainly subjected to tensile stress instead the compressive forces typical in Earth structures dominated by gravity loading.  

According to NASA research, it is possible to safely reduce the internal pressure to values lower than typical on Earth: values as low as 55 kPa and 52.4 kPa are recommended for the Moon and Mars respectively for normal operations.  However, these recommended lower pressures require oxygen volume concentrations slightly greater than 30% (for Mars and Lunar surface habitats, 32%) which is the maximum non-metallic materials flammability certification level used by current operational human space flight programs. It is relevant to note that these recommendations for surface habitats must be studied more deeply prior to development of requirements for those elements.    

In any case, the range of possible internal pressures for surface habitats is very high.  For highly pressurized structures, inflatable or pneumatic membranes are a compelling solution because then can be easily transported and use little material. 


3. Inflatable exploration habitats 

My ongoing research aims to explore form finding strategies for deep space exploration habitats considering the Moon and Mars as target surfaces. A new methodology for space shelter form finding has been analyzed, in order to optimize the location of different functions as well as to respond to the high pressure differentials required in these environment. Currently, the internal layout is designed through an iterative process that integrates several factors (mass, volume, cost, schedule, technology level, and maintainability) and relies on adjacent matrices and bubble diagrams. Instead, the proposed algorithm automatically finds the optimal allocation of the required functions inside the habitat in the 3-dimensional space, without using a manual iterative process 

The potential impact of this work relates to the possibility to design in real-time the final layout of the habitat by simply defining the linkages between functions and subsystems. This method could be applied at different scales of the habitat, from the urban scale to the architectural one, and to even more complex systems. However, increasing the complexity could slow down the simulation process. 

Figure   : Sphere diagram connectivity diagram.   

Figure : Sphere diagram connectivity diagram. 

Figure : Sphere packing and metaball design.   

Figure : Sphere packing and metaball design. 

Moreover, because the obtained functional diagram readily translated into a structural finite element model, it was possible to evaluate its structural performance.  

While the metaball surface effectively encloses the functional regions of the habitat, its geometry is not responsive to the forces caused internal pressurization.  From a structural point of view, the presence of anticlastic areas on the metaball surface generate both tension and compression stresses that don’t are hard to resist with an inflatable membrane structure. 

As a consequence, another form finding simulation has to be performed, applying an equal internal pressure inside the metaball configuration while maintaining the volume constant, again using Kangaroo2 for form finding.  The output of the process, shown in Figure, shows a new geometrical configuration that reflects the need of having only tensile stress within the inflatable membrane. 

Figure : Inflation of the metaball design with constant volume.   

Figure : Inflation of the metaball design with constant volume. 

The structural analysisof the membrane has been tested through a finite element model, implemented in Karamba, in which the load is represented by the internal pressurization at Earth levels (100 kPa). The material used for the inflatable structure is Kevlar (DuPont), a common material for inflatable structures designed for space by NASA and Bigelow Aerospace.  The Kevlar elastic module and density are respectively equal to 8460 kN/cm2 and 1440 kg/m3 .  The boundary conditions of the system are defined considering that the inflatable structures will be located and partially dig into the ground, to improve the anchoring system. 

In Figure, the tensile stress distribution of the final membrane is shown. The areas that have higher stress are the ones in which a sharper change in the curvature of the surface is evident. The structural model has been evaluated with three different loading conditions: internal pressure, internal pressure with lunar gravity (1/6 g) and internal pressure with Mars gravity (1/3 g). The results are very similar in terms of stress distribution and displacements, highlighting that the reduced gravity is negligible when designing for space habitats that have a differential pressure of about 1 atm.   

Figure  : Tensile stress distribution inside the inflatable structure.  The colors show material utilization as a percentage of the total allowable stress for Kevlar.   

Figure: Tensile stress distribution inside the inflatable structure.  The colors show material utilization as a percentage of the total allowable stress for Kevlar. 

It has been possible to demonstrate that reduced gravity loading is negligible when designing for space habitats that have a differential pressure of about 100 kPa. Therefore, the internal pressurization is the main load to consider. Future research could expand this study considering also other types of loads, such as the micrometeoroid impact, and the airlock systems. 


4. Inflatable Emergency Shelter for Lunar Applications 

The personal shelter for one astronaut is to use in an unexpected emergency situation occurring during surface exploration. This personal emergency shelter will inflate immediately around the astronaut, to protect him or her from the elements while they wait for help to come. Potential use cases include: disorientation, unexpected severe weather, personal injury, or puncture to the spacesuit or other equipment. Unlike the 2-person overnight habitat designed by Jeffrey Hoffman, this personal emergency shelter system is lightweight, designed to be worn as a backpack, and only deployed as a short term shelter in case of an emergency. It is independent from a lunar roving vehicle and designed to be worn on the back of an astronaut who is outfitted with the counter-pressure second-skin Biosuit, which weighs approximately 20 pounds. 

Figure  :   Inflatable emergency shelter for astronaut.      [Team: Valentina Sumini and Meghan Maupin. Faculty advisor: Prof. Caitlin Mueller]

Figure: Inflatable emergency shelter for astronaut. 

[Team: Valentina Sumini and Meghan Maupin. Faculty advisor: Prof. Caitlin Mueller]

In order to serve the needs of the astronaut, the emergency shelter has to meet specific Material Requirements for Habitats, as defined by NASA. The personal emergency shelter needs to be pressurized to an internal pressure of 8.3 psi at 32% oxygen; have an airlock for entry and exit; and offer protection from wind, radiation, cold, and contaminated regolith. The shelter must deploy autonomously and rapidly, within 30 seconds, to ensure that the astronaut is protected in case of a spacesuit failure. The shelter must also have a communication and beacon system, as well as room for storage in case the astronaut needs to take off his or her spacesuit. In order for the astronaut to be able to carry this shelter on his or her back, the shelter must weight less than 138kg on the moon or 50 pounds of Earth equivalent and fold up into a volume of approximately 3 ft x 1.5 ft x 1 ft to wear comfortably on the back, without interfering with the helmet 

The architectural form of our personal emergency shelter is an arch-based prismatic shape. The highest point of the arch is 6 feet and the area is 38 square feet. The structural ribs of the shelter are also inflatable in an integrated, simple system that packs easily and deploys rapidly. The entire structure folds into a triangular backpack shape that is approximately 1 foot wide by 3 feet tall, to fit easily on an astronaut’s back. 


Once we had the basic geometric model of our shelter in Rhino, we modeled the inflation in Kangaroo. Then in Karamba, we analyzed the stress distribution with an internal pressure of 1 atmosphere, and subsequently iterated on our design to include additional ribbing elements where the compressive stresses were present. From there, we built physical models in paper to test the folding mechanisms and form factor of the backpack. 


5. Design and Structural Optimization of a city on Mars 

The Mars City project, called Redwood Forest, is located in an unusual circular depression where we will create a bright, green and water-rich habitat to nurture 10,000 people. Mimicking the structure of trees, the habitat will exist both above and below ground. Within the root network, residents will have their private spaces protected from the harsh radiation, meteoroid impact and thermal environment. The root network will house most of the machines that process, store and distribute resources vital to everyday life. The public spaces will exist above ground, in enclosed structures which filter daylight down to the root network. The main transportation network will be an underground thoroughfare modeled after rhizomes present in various plant species. The radiation protection will be helped by the inclusion into the shield of a layer as water reservoir. The regolith, that was dug out for the initial root system, will be used as a catalyst to start our water production and extract other mineral resources for construction. The same water could be used for shielding against radiation. On the project a multi-objective material optimization has been performed for designing an earth independent structural system that uses in-situ resources. Moreover, the form finding of the main habitat and its structural system has been implementedthrough different structural optimization algorithm such as the force density method. 

This project resulted to be among the top 3 finalists of the international Mars City Design competition. 

Figure  :   Mars City project: Redwood Forest.                                                            [Team: Valentina Sumini, Sam Wald, Alpha Arsano, Mark Tam, Meghan Maupin, George Lordos, Alejandro Trujillo, Matthew Moraguez, John Stillman. Faculty advisor: Prof. Caitlin Mueller]

Figure: Mars City project: Redwood Forest.  

                                                     [Team: Valentina Sumini, Sam Wald, Alpha Arsano, Mark Tam, Meghan Maupin, George Lordos, Alejandro Trujillo, Matthew Moraguez, John Stillman. Faculty advisor: Prof. Caitlin Mueller]


My future research on space habitats will relate to the further development and assessment of new computation design methods in order to facilitate the interaction between different disciplines in the design process. 

The requirements of future space habitat structures can be defined by their ability to protect their occupants and provide usable space to live and work in an extreme, isolated environment. Due to the high cost of transporting resources off of Earth’s surface, recent efforts have focused on developing increasingly Earth-independent structural designs which use local regolith-based materials as a possible solution for long-term extraterrestrial sustainability. Many options for construction methods, technologies, materials, and architectures exist for surface habitats. With a focus on sustainability and Earth-independence, my proposal looks at architectures that use regolith-based materials (concrete with carbon fiber polymer reinforcement or plastic fibers). The research approach is to formulate the structural design problem as a multi-objective optimization of the habitat structural system. The objectives that apply to the habitat geometry and shell thickness include 1) minimum transportation and construction costs, and 2) minimum probability of loss due to radiation and micrometeorite events. The multi-objective optimization applies Pareto optimization to determine which design elements or options afford the greatest efficiency.   The candidate design solutions will be examined based on their architectural effectiveness.