Thursday, May 4, 2017

Project Destiny: Habitats and Food, part 2

This is a topic post referring to Purdue University's project Destiny.
Here is my introductory post for the series.
The subject is section 7, Mars habitats. By necessity, section 8 (Food Production) is also discussed.

Due to the broad scope of this section, I've broken it into three separate posts. This post discusses living spaces. Part 1 discussed life support.

 Headline results:
No net change to costs. Per capita living space increased by 2.3x, while per capita hydroponic space about doubled. I present two alternatives to the 'cylinder farm' approach, each with pros and cons. Basalt fibers would be used as rockwool insulation to reduce heating requirements.

Details after the jump.

Living Space

 First, let's discuss living space. 131 m³ may sound like a lot, but imagine living your entire life in a 6.6x6.6 meter square that also has to grow all of your food. I believe that by allocating substantially more area to private living space, the demand for EVA recreation can be reduced. I also believe that the use of public greenspaces with high clearance is beneficial for mood and mental health. For example, have a look at the first picture in this article on The Expanse. Imagine a row of fruit trees down the middle, or one of these bamboo groves.

 I have used several heights for various posts and projects, but I will standardize on floors or levels of 3 meters gross height. That's 2.5 meters net height after the floor thickness and utility space. For certain cases like horizontal tunnels, the utility space might be allocated to deadspace caused by packing inefficiencies rather than charged against floor heights.

 A rule of thumb I have used in long-term habitat design is 100 m² per person. In this case that means a habitable volume of 300 m³ gross, 250 m³ net. That covers all types of space: private, public and work. The exact split depends somewhat on the settlement's cultural norms: some people prefer privacy and have little public space, while others prefer large communal spaces and small private areas. A reasonable balance might be 32/36/32, which means each person has just under three times as much private space as the baseline. A family of four would have about the same private floorspace as an average American home.

Construction - baseline

 The baseline plan is to cut and cover an array of 18 m dia. vertical cylinders made of cast basalt. I have to admit, this was an inspired choice. Cast basalt is solid stone, airtight, strong, impermeable and highly durable. Burying the structure is necessary for radiation protection, and the cut and cover method is fast, simple and economical.

 The effort requires a fleet of 10 dump trucks ($3.766m ea), 20 excavators ($1.688m ea) and 42 printers ($4.759m ea). Printer spares run about $21.5 million per cycle. Excavation is expected to take 203 sols, construction 353 sols and an uncertain but relatively short period for backfill.

 One concern is that under the baseline these structures will mostly be in tension rather than compression. I believe the excavation depth needs to roughly double, and here's why: Assuming the habitats operate at 100 kPa, that's 100 kN per m² of inner surface area. Gravel is approximately 1700 kg/m³ and Mars gravity is 3.711 m/s², so each meter of soil offsets 6.3 kN/m² (6.3 kPa) of pressure. Depending on the actual soil density, a habitat's upper surface would have to be about 16 meters below the surface to achieve neutral pressure.

 One minor modification to the baseline would be to use a different printer design. The current design is an X-Y gantry, which means that virtually all head movements will engage two out of three motors. A polar-coordinate design would be more efficient for printing round structures as only a single motor would engage and would run at a steady speed most of the time. This printer could use grip points printed into the structure's shell to climb along the outside during construction, allowing much taller structures to be built.

 I see two options beyond the study's approach.

Construction - slab and column

 First is a more traditional rectangular construction. The entire quarry would be excavated to appropriate depth, with footings dug to bedrock. An outer layer of locally-sourced rockwool insulation would be applied first. Basalt or concrete columns would support beams which support internal slab floors comparable to modern high-rise construction. (If desired, reinforcing steel can be omitted and the columns and beams can be formed as arches to allow fully compressive loads and extremely long service life.) The nearly flat base slab, curved outer walls and flat roof would be quite thick, perhaps 40-50 cm. The whole structure would be buried under 16-20 meters of regolith. Basalt requirements would be 145,800 m³ for the outer shell and about the same for interior structures (perhaps 65,400 m³ for floor slabs and about the same for columns). This is about 61.2% of the baseline basalt volume, meaning we can cut the power and hardware requirements of the printers by 2/5 and save ~ $427 million over the course of the project.

 This would provide a very large and convenient space, efficient use of materials and maximum utility of the pressurized volume. Without increasing the required excavated material, the resulting livable volume is increased by a factor of 2.7 to roughly 4.5 million m³. This nearly compensates for the desired increase in volume without affecting the excavation plan. Each additional 3-meter floor would add another ~250,000 m³ of volume if desired; another two floors would be necessary to hit my volume goals, which adds half a million m³ of excavation work (about 8.3% more excavation).

 It is also somewhat risky, in that a single major structural failure risks exposing the entire volume to ambient atmosphere. The city could be split into several compartments with pressure bulkheads to reduce the risk. It's worth noting that a modern high-rise building is expected to last a hundred years or more and encounters much more demanding loads, so the risk of catastrophic collapse is very low. Periodic observation and preventive maintenance of the structure should push its expected life well past the program timeline.

 Pressure loss would be a minimal risk as the soil pressure exceeds the air pressure, so the primary risk is actually CO2 infiltration. Life support can handle a certain amount of this infiltration, and the life support system itself would include an array of CO2 sensors that would detect a leak.

Construction - tunnels

 Second is to use large-diameter (~16 m) tunnel boring machines. These would be Mars-adapted: able to break down all parts into transportable pieces, using liquid CO2 as hydraulic fluid and using cutting or wearing parts that can be replaced locally.
 First, an initial downslope is excavated and lined in basalt. This serves as the egress point for roving vehicles. Next a series of parallel tunnels are cut and lined, either with precast blocks or with cast-in-place basalt (or concrete, depending on the state of industry at the time). The tunnels are outfitted with multiple decks. For example, have a look at this packing: it could provide two ~12-meter floors and a ~15-meter floor for normal use. The rest of the cross-section would carry utility lines, storage, life support and access ways. Cross-tunnel access would be provided with smaller perpendicular tunnels. The overall excavation would require 26.6 km of main tunnel for each quarry equivalent. Unless a transit system is built, individual tunnels should probably be limited to 2 km; that would mean 13 tunnels of 2,045 meters length. The completed million-person colony would be 2x10 km.

 Tunneling through solid basalt should be able to manage 2 meters per hour, 40-50 meters per day (up to 10,000 m³), or 43 days per tunnel excluding setup and teardown. If a second downslope is built at the other end of the tunnel array then the TBM can be hauled out, turned and sent back in the other direction without needing to be disassembled. The full quarry would require 554 days of TBM operation, meaning only one machine is required. With a 50% margin, two machines are necessary. If we use the authors' 2.5x factor for space hardening and $80 million for the cost of a very large TBM, that's $200 million per machine. Replacement cutting bits can be manufactured on Mars, while seals, motors and pumps would need to be sourced from Earth. The overall cost for two 16-meter TBMs and two smaller TBMs should be about $525 million, with running maintenance costs of around $115 million per cycle.

 TBMs are normally extremely heavy (thousands of tons), with the cutting head as a single large element. A Mars-adapted version would need to be much lighter, with modular plates for the cutting head and overlapping segments for the shield. While the soil pressures experienced by the machines would be miniscule compared to Earth tunnels, the cutting face must produce the same enormous force to fracture the rock; the support structure for this thrust is definitely going to be heavy and complex.

 This approach would be useful for settlements on rough terrain or shallow soil cover, particularly with solid basalt layers (most of the northern hemisphere, for example). TBMs can be operated with few people and would eliminate the need for excavating the initial area all at once. In other words, this approach allows for organic growth rather than building an entire quarry neighborhood all at once. The cross tunnels serve as refuge spaces as well as access, so long as connections between tunnels use airlocks. Because the volume figures already include considerable margin, some of the space would be left tall enough for trees; a sort of linear orchard / park area may help with psychological health and will produce luxury foods at reasonable yields.

 Further, the availability of a TBM means that access to useful locations such as the water plant can be made underground. 
 A moderate or large diameter tunnel can link the colony site with the water refinery, allowing shirtsleeves access to the pipeline and reduced heating requirements. It would also allow the use of multiple basalt pipes for redundancy and expandability, as well as reducing the demand for steel pipe from Earth.
 A similar tunnel could run to the main landing area. Coupled with pressurized towers and extendable docking ports, colonists would not need to ride in a rover to get to their rooms.
 A large central tunnel with conveyors and/or rails for material handling could be used as the central artery of industrial water extraction. Side tunnels with automated excavation equipment could feed regolith to the main tunnel, allowing for a system similar to longwall coal mining. This would be useful in hydrated mineral beds or ice beds, avoiding the need to remove overburden from the surface.

 For the truly ambitious colony, a 12-meter tunnel could be built to both poles. That's 10,606 km more or less. (Less ambitious but still useful would be a north polar tunnel; length depends on colony latitude.) With two machines in operation (one northbound, one southbound) the work would take 142 cycles, about 307 years. Spacing a series of ten construction sites along the route and building multiple sections simultaneously would yield a working tunnel by the 15th cycle, about 31 years. The benefits of this would be virtually limitless access to water ice from the north pole and solid CO2 from the south pole, a colossal dataset on subsurface geology, 1.2 billion m³ of excavated volume (and any interesting resources it contained), plus fast access to any latitude. Settlements could be developed along the route, allowing some separation of groups as insurance against pathogens.


 Here are three reasonable approaches to building the habitable volumes needed. The baseline approach can be used, though I submit that the structures need to be both taller and buried deeper. A more Earth-normal slab and pillar method can be used as well. There is also the option of tunneling, which has advantages for some types of terrain and some use cases.

 The cost differences between the approaches are trivial in comparison to other costs such as life support. The actual implementation may be a mix of all of these as the colony expands and experiments. For that reason, I will use the baseline cost estimate for construction. In order for the baseline plan to produce the necessary extra volume, all systems are run continuously. Excavation keeps working the quarry face while printers put up habs and backfill covers completed structures.


  1. I've been looking at roadheaders as alternatives to tunnel boring machines you might want to investigate them as well. It could also be interesting to bore into softer sandstone, or similar material for faster throughput. In particular, sandstone can have a fairly high void area, opening the intriguing possibility of melting the material in place and using it do create airtight lining.

  2. you might enjoy my little colony video...

  3. The fleet of heavy equipment postulated for construction seems right given your assumptions, but is nonetheless considerable. I suggest that the use of explosives for both excavation and backfill would save both time, money and most crucially person hours. Oxyliquit explosives can be manufactured locally. Invented in 1895, and used in large quantities in the 30's, they would require only a strong container, liquid oxygen, carbon black or similar compound,and magnesium as a detonator. Nitrate based explosives could also be used, but their manufacture is more demanding. Skilled explosives experts can do much, and it is a readily available skill set.