This is the sixth in a series of articles on how to go about dramatically reducing space mission cost while maintaining a high level of mission utility. This article discusses methods for dramatically reducing the cost of traditional large spacecraft programs, or fatsats for short.
First, we should make the definitions clear. Many of the methods we have discussed previously are applicable primarily to smallsats, which are usually, though not necessarily, physically small spacecraft similar to a NASA Class D mission, low-priority, high-risk payloads for which many of the traditional rules and requirements do not apply. Smallsats have a greater implementation risk, but an analysis of long-term performance by NASA’s Goddard Space Flight Center has shown that the reliability of smallsats is essentially comparable to that of more traditional missions.
Fatsats are usually, though not always, physically large (but not overweight) spacecraft characterized primarily by having multiple payloads, a long mission lifetime and the requirement to obey all of the most stringent rules and requirements of long-lived, expensive spacecraft, such as a NASA Class A or flagship mission. Reducing the cost of fatsats is particularly challenging specifically because of the desire to not change any of the rules, procedures or even technology and the high mission assurance requirements. Due to the high cost, the fatsat cannot be allowed to fail, but of course that is what has created much of the high cost in the first place.
Clearly, if we don’t change anything about the fatsat except for minor modifications in the design or manufacturing, we also won’t make dramatic changes in the cost. What can we do that leaves all of the “rules of engagement” intact but still has a dramatic impact on cost? Certainly it’s a challenging problem — but perhaps not impossible.
To make the discussion more specific and a bit easier, let’s assume we have a large Earth observation spacecraft weighing 5,000 kilograms, flying at 800 kilometers, having a design life of 15 years, a resolution with the primary instrument of 0.5 meter at nadir, and a first flight unit production cost of $1 billion. Because of budget cuts and sequestration, we need a dramatic cost reduction (by a factor of two to 10), but we don’t want to change any of the traditional rules for this flagship mission. If we keep all the rules intact, then the traditional cost models will also apply. For most of the traditional cost models, the principal determining factor in the spacecraft cost is the mass, and the cost varies approximately linearly with mass over a rather wide range.
One option is to reduce the altitude from 800 kilometers to 400 kilometers. To achieve the same 0.5-meter resolution we would need a primary payload instrument that has only half the aperture and, consequently, about one-eighth the volume and mass. This will reduce the entire spacecraft mass by a factor of about eight, to 625 kilograms, and the cost to about $125 million. (One could argue that not all of the spacecraft subsystems will shrink proportionally, so we might have to reduce the altitude to, say, 350 kilometers.) We also propose to reduce the design life to only five years in order to greatly reduce the level of redundancy and therefore the mass and cost by an additional 20 percent, to 500 kilograms and $100 million, respectively. Because we have come down in altitude by a factor of two, we will need two spacecraft to cover the full swath width previously covered by one and we need three times as many to meet the full design life of 15 years. So our first-round estimate is six “trimsats” for $600 million. (They aren’t smallsats because they still follow all of the rigorous design rules of the traditional missions, except for the reduced design life.)
But multiple spacecraft offer some additional, and very real, economic advantages. Building multiple spacecraft always saves money relative to building a single one. This is expressed by a “learning curve” that takes into account both actual learning on subsequent units and also things like tooling and spares that can be amortized over more spacecraft. A typical learning curve for spacecraft is around 90 percent, which means that the average cost of all of the spacecraft is reduced to 90 percent of the previous value every time the number of spacecraft is doubled. For a $100 million first production unit cost and a 90 percent learning curve, the total cost of six spacecraft will be $457 million and the cost of the sixth unit will be $65 million. If I want to build a spare spacecraft, it will cost only $64 million additional for a total cost of $521 million for seven units. The average cost of the seven spacecraft will be about $75 million each.
There is another purely economic advantage as well. Originally, we had to spend the entire $1 billion “today,” before we got any return on our investment. But the six-spacecraft purchase is spread out over 10 years (two now, two more in five years, and the last two in 10 years). As anyone who watches the lottery will tell you, paying a winner $1 billion over 10 years costs the payer much less than paying the whole $1 billion today. If the government had lots of excess money lying around and wanted to use it right away to stimulate the economy, our $1 billion up-front expense would be a nice way to do it. But if we have to borrow that $1 billion from the Chinese at 5 percent interest and pay it back over 10 years (the same time span over which we’re paying for the six trimsats), the total cost for our one fatsat will be $1.3 billion. If we buy seven trimsats (six operational ones plus a spare), it costs us only $520 million, or 40 percent of what a single fatsat would cost under the same payment conditions. We have substantially reduced both the cost and the mission risk and transformed a potentially devastating launch failure into something with only a moderate economic impact, rather than a catastrophic mission impact. And Congress is much happier, having to fund only the $125 million (or possibly $250 million for two) right now, rather than the full $1 billion.
As we discussed previously, going with more spacecraft spread out over time has quite a few other advantages as well:
- Allows us to maintain a continuous production line and therefore knowledgeable people who know how to build the spacecraft.
- Allows the potential of introducing new technology along the way, or modifying the mission to meet changing needs.
- Allows us to adjust the launch rate to match the actual on-orbit life rather than the design life (most spacecraft outlive their design life; for example, LandSat 5 just set a record for operating 29 years with an initial design life of only three years).
- Reduces risk by having another spacecraft in the pipeline if something does go wrong.
- Being at a lower altitude also means that we have effectively solved the orbital debris problem for the trimsats, although there is a cost in terms of added propellant. The delta-V needed to maintain the 400-kilometer altitude ranges from about 5 to 50 meters per second per year, depending on the satellite and the time in the solar cycle.
An alternative to building multiple smaller spacecraft that each do the same job as our original fatsat is disaggregation — i.e., breaking up our original spacecraft along functional lines and having one, or possibly two, functions on each of many small spacecraft. (See the fourth article in this series for a more detailed discussion.) We also have the option to mix the two approaches. Some of the functions of the original spacecraft may benefit from flying low while others may do better at a higher altitude. Similarly, the optimal design life may be different for different functions. Breaking our original fatsat into various pieces allows a great deal more flexibility and robustness and has the potential to substantially reduce cost, even if we don’t change the rules of the game.
Thus far, we have applied only the approaches of flying low and reducing the mission design life to build multiple trimsats and therefore take advantage of the learning curve and spreading the costs out over time. Other than that, trimsats are following the traditional program rules. However, the cost of an individual satellite has come down from about $1.3 billion to about $75 million, the total cost has been spread out over time, and we have significantly reduced the risk by reducing the required design life, having a spare spacecraft, and reducing the amount of money at risk. (Numerically, risk is the probability of failure times the consequences of that failure.) All of this makes many of the other approaches more reasonable to consider, such as trading on requirements, considering multitier requirements, setting functional rather then technical requirements, making cost more important, providing stable funding or using multiple sources of data (particularly because new types of data may become available over time). All of this can begin to get us back to where we would like to be — creating far-lower-cost spacecraft where a launch failure or debris collision is certainly not good but is also not a threat to national security or the long-term continuity of critical science data.
The same learning curve that we used to reduce the cost of the trimsats could, of course, also be applied to the fatsats themselves to reduce cost. For example, with the assumed 90 percent learning curve, five fatsats could be bought for $3.9 billion and a spare would be an additional $650 million. If we want all five for the entire 12 years, then our initial outlay before any return on investment goes from $1 billion to $3.9 billion. (And the $3.9 billion becomes $5.1 billion when paid for over the next 10 years.) If we use them serially spread out over time, we are buying the same satellite to be used for the next 75 years. In contrast the 30 trimsats cost a total of $1.8 billion over 10 years, the first five cost $390 million, and we really do have a trimsat production line under way.
The seventh article in this series will look at personnel issues for reducing cost.
James R. Wertz is president of Microcosm Inc. He is co-author of “Reducing Space Mission Cost,” published in 1996, and has taught a graduate course at the University of Southern California on that topic since then. If you have questions, comments or suggestions, or simply want to discuss these issues, he can be reached at email@example.com. Information on the joint Microcosm/USC Reinventing Space Project can be found at www.smad.com/reinventingspace.html
Reinventing Space: Dramatically Reducing Space Mission Cost — Mission Design
Reinventing Space: Dramatically Reducing Space Mission Cost — Programmatic Approaches
Reinventing Space: Dramatically Reducing Space Mission Cost — Systems Engineering Approaches
Reinventing Space: Dramatically Reducing Space Mission Cost — Attitude