Here is my introductory post for the series.
The subject is section 4, Interplanetary Communications Network.
Headline results: I believe that the cost of this system can be reduced by nearly 50% without altering the underlying performance assumptions.
Details after the break.
This study's communications design is a major factor in overall cost. Every effort should be made to reduce this cost; the headline price is $228 billion, which works out to $4,560 per colonist per year for connectivity.
The basic design using radio links for surface to orbit and optical links for orbit to orbit is sound. The performance numbers seem well-researched, as do the fault rates. The need for a relay is well-established. The amount of data considered is also reasonable: a single HD video stream for each 100 colonists plus two minutes of SD video (or its equivalent in other data) per person per day. My disagreement is that the use of small disposable satellites is an oldspace norm which unnecessarily drives cost.
What's the alternative? A large multiuser platform with human maintenance missions. ITS has more than enough dV to deliver a maintenance crew and 300 Mg of cargo from Earth to ESL-5 or from Mars to areosynchrous orbit. By delivering discrete components like transponders, processors, batteries, apertures, etc., the mass of the thrust tube and propellant masses for orbital insertion and decommissioning are eliminated. Beyond that, a shared facility offers options for cost recovery. Science missions such as large telescopes may draw funding for shared services like electrical power and maintenance. Organizations like the L5 society may choose to dock habitats to the platforms for similar reasons.
Consider for example the ESL-5 relay. Baseline costs for the node are $47.5 billion. The baseline plan requires 53 separate Falcon Heavy launches, 12 of which are contingency satellites. Each vehicle has dedicated avionics, GNC, optical apertures, structural bus, thrust tube, propulsion, etc., etc., all of which are discarded once any one system fails.
For redundancy and to minimize interference, two platforms should be deployed. The physical structures are not significantly affected by exposure to deep space, so their effective service lifetime is greater than the length of the study. By allowing one ITS launch per cycle and alternating the visited platform, each platform is visited roughly every five years. Components can be tested, repaired or replaced if necessary and certified for another 5-year period. System expansion is done by adding more components than are removed.
The baseline Falcon Heavy launch and operation costs for this location are $6.2 billion in current dollars. 47 fully-refueled ITS launches (at $46 million each, using a higher lifetime flight rate for near-Earth operation and the same $25 million operations cost) would run $2.16 billion. Total savings: $4.04 billion.
The solar panels are a major portion of that cost, $240 million for ~360 kWe on one satellite, or $12.72 billion for the relay satellite system. The project considers SAFE-series reactors in other contexts, for example a SAFE-800 design producing 240 kWe for about $1.4 million. At steady state, six satellite equivalents will be operational at each platform; 9 SAFE-800 reactors plus one spare would therefore cost $28 million initial plus $4.3 million per refuel for the L5 relay system. Reactors are assumed to have a 60-year operational life, so two complete sets of hardware (40 total reactors) must be built and deployed during the project along with eight refueling cycles. Total cost $90.4 million, assuming launch costs are covered in the per-cycle maintenance visit cost. Total savings: $12.63 billion. Even if the development process for these reactors took several billion dollars it would be a net gain.
The thrusters, propellant tanks, lines, valves, avionics, guidance and other systems are shared across the entire platform. Because no insertion or deorbit burns are necessary, the amount of stationkeeping propellant required is a small fraction of the baseline amount. I'll assume that mass and cost of these systems is instead spent on a structural truss for mounting components, on a stationkeeping system with ion thrusters or PITs to eliminate toxic fuels, and on a platform pointing system with very long lever arms and robust reaction wheels. No net change.
The electronic communication components would be unchanged, and remain the primary expense at $268.8 million each ($11.02 billion total without flight spares). However, because a failure of other critical systems does not lead to loss of use of the comm systems, we are able to eliminate the 12 extra satellites and comm sets. Human servicing missions will be able to repair, repurpose or replace as needed, while one extra set on each platform will provide short-term protection against an outage. Total savings, $3.22 billion.
This node in the communication network can be implemented as a shared facility for a savings of $19.89 billion, a total cost of $27.65 billion. The average colonist population over 100 years is half a million, so this node costs $553 per colonist per year.
Similar logic can be applied to the Earth-orbit and Mars-orbit nodes.
The Earth-orbit node would consist of three geosynchronous shared platforms. Due to the high population of GEO satellites, these could be useful places for permanent satellite servicing facilities. A rotating workforce of technicians could keep the platforms operational while also servicing other customers' hardware to offset costs. Over the life of the node, 66 satellite equivalents are required; 15 satellites are required at steady state, or 5 per platform. For simplicity these platforms should be identical to the ESL-5 platforms, though they will use one fewer reactor due to reduced power requirements. To maintain a five-year service interval, ITS flights are required every 20 months which equals 60 total flights. An alternative is to place a permanent manned maintenance facility at EML-1 and use an orbital tug to service platforms as needed; this would provide the added benefit of a one-week response time to problems.
The baseline cost of this node is $61.1 billion with 14 flight spares. Electrical savings are ($19.2b baseline - $75.6m capital - $46.44m refueling) = $19.08 billion. Launch savings are ($9.36 billion baseline - $390 million ITS/$6.5m) = $8.97 billion. Spares savings are 14 x $189.86 million = $2.66 billion. This node's total savings are $30.71 billion, total cost is $30.39 billion.
The Mars-orbit node would consist of three areosynchronous shared platforms. Due to the study's power limitations, bandwidth through each satellite is quite a bit lower than through the relay satellites and so more craft are specified. Over the life of the node, 135 satellite equivalents are required; 33 satellites are required at steady state, or 11 per platform. These would ideally be two standard platforms linked together, with 19 reactor modules. The hardware would be delivered via cargo ITS, one per cycle. This vehicle would aerocapture and then rendezvous with a central maintenance facility on Phobos (or possibly Deimos) for offloading, followed by a descent to the surface with remaining cargo and a standard return. Service missions would be dispatched from the Martian surface using one of several Mars-dedicated ITS vehicles, dock with facilities on Phobos, transit to the appropriate platform, perform maintenance and installation tasks, then return first to Phobos and then the surface. The Phobos base may eventually be permanently manned and used as a port for handling electric-propulsion cargo vehicles via docking tether.
The baseline cost of this node is $119 billion with 21 flight spares. I believe there is a potential savings from increasing the transmission power of the Mars node to match the data rates of the other nodes, but let's stick with the baseline for now. Electrical savings are ($37.44b baseline - $159.6m capital - $98.04m refueling) = $37.18 billion. Launch savings are ($18.25 billion baseline - $1,457 million ITS/$31m) = $16.79 billion. Spares savings are 21 x $189.86 million = $3.99 billion. This node's total savings are $57.96 billion, total cost is $61.04 billion.
These changes bring the total I-Comm network costs to $119.08 billion, or $2,382 per colonist per (Earth) year / $198.47 per (Earth) month. That's a combined savings of $108.92 billion, or 47.8%. I believe there may be another $30-$60 billion of potential savings in the actual comms components due to standardization and large production runs, and at the Mars end by adding more power through fewer apertures.
The use of shared facilities encourages the further development of space. Permanent crewed maintenance bases near the three nodes would revolutionize the way satellites are built, launched and operated, and would open the road to asteroid mining and the colonization of free space.