The Space Reviewin association with SpaceNews
Waiting for Launch ebook

Falcon 9 launch
Smaller, most cost-efficient vehicles like the Falcon 9 (above) may offer a less expensive, more robust alternative for human space exploration architectures than a heavy-lift rocket. (credit: Chris Thompson/SpaceX)

Human spaceflight for less: the case for smaller launch vehicles, revisited

Bookmark and Share

Access to orbit is the common problem shared by the entire space industry. In particular, for human spaceflight to low Earth orbit and beyond, access is the main bottleneck between current space activities and sustained, meaningful space development. The lack of cheap, reliable space launch is felt by human and robotic space programs alike. At best, its absence limits what can practically be accomplished with current space budgets, and at worst, its absence masks the true economic potential of space by preventing some activities in the first place. The access problem is significant—and nowhere, it appears, is there more turmoil regarding the future of space access than the debate over NASA’s next launch system.

In considering a new rocket for NASA’s (as yet unspecified) future missions, it is worth asking: what is necessary and sufficient for enabling reliable, affordable, and increased human space activities?

While the private sector has quietly (or not so quietly) been working to address the issues of affordable and reliable access, others have struggled to address the issue at all. While NASA for its part has increasingly been embracing and assisting private initiatives in developing cheaper launch systems, there remain contingents in the agency and especially in Congress that continue to dismiss existing and emerging commercial capabilities, and who remain fixated on the belief that a heavy-lift launch vehicle (HLV) is the right and only way for human space exploration to occur. Decades of studies have called for the development of such a rocket—from the first President Bush’s Space Exploration Initiative (SEI) to the second President Bush’s Vision for Space Exploration (at least through the now defunct Constellation implementation of the Vision). But none have come to fruition since Apollo.

The latest attempt at reviving heavy-lift is a congressional demand that NASA must develop a heavy-lift launcher by 2016 (which, as the Orlando Sentinel noted two weeks ago, will be “made of recycled parts of the shuttle”). Notwithstanding the fact that Congress has not authorized sufficient funds for the completion of such a vehicle, this latest attempt at forcing a large launcher into NASA’s plans will consume at least $10 billion over the next few years, and—if history is any indication—will likely result in nothing more than another paper rocket. As Lou Friedman put it here two weeks ago, “the situation in the United States with respect to [space access] is no different than if we had a space czar whose motive was to keep the country grounded. Why does it seem like we can never get a rocket policy for civil space exploration right?” (see “The dangers of a rocket to nowhere”, The Space Review, May 23, 2011)

The new Space Launch System (also pejoratively termed the “Senate Launch System”) has the political benefit of sending billions of dollars to former shuttle contractors, and preserving some NASA shuttle jobs. But aside from being a jobs program, SLS can be expected to accomplish little. In the best case, it will probably fail entirely, and in so doing will merely be wasteful; but in the worst case, there is the possibility it might succeed, and lock NASA into using 1970s technology for the indefinite future, while also marginalizing the involvement of commercial launch providers. Under such conditions, a “post-shuttle era” would never really come.

In considering a new rocket for NASA’s (as yet unspecified) future missions, it is worth asking: what is necessary and sufficient for enabling reliable, affordable, and increased human space activities? Is there actually a sound engineering or economic case to be made for a new heavy lift launch system? Or can we accomplish just as much or more with the rockets we already have?

Is heavy lift necessary?

I’ve presented a case against heavy-lift launch vehicles before in this publication (see “The case for smaller launch vehicles in human space exploration (part 1)” and “The case for smaller launch vehicles in human space exploration (part 2)”, The Space Review, January 3 and 9, 2006). Aside from the political benefit of a shuttle-derived HLV (where the word “benefit” is used dubiously), the typical (non-nostalgic) arguments presented in favor of heavy-lift boosters usually revolve around payload efficiency and simplicity.

Heavy-lift proponents argue that HLVs are more efficient in terms of the cost per kilogram of payload delivered to orbit (since larger launch vehicles require less mass per unit payload). In terms of the marginal cost per kilogram, this should technically be correct—but only if there are no large fixed or capital costs to amortize. Unfortunately, because larger rockets tend to require significant capital investments, they also tend to have large development costs that must be remunerated over the life of the vehicle. Heavy-lift boosters also require large assembly, integration, and launch infrastructure, as well as large full-time support staff. These represent extremely large fixed costs, which also must be amortized over the vehicle’s use. This is the key issue: because the heavy-lift rocket will typically have a low flight rate (likely on the order of once per year), the HLV will have to pass its entire operating costs into the price of a small number of launches, in addition to a large fraction of its development cost. Thus, the net cost per kilogram will tend to be quite high.

Orbital assembly has arguably become a mature spaceflight capability—a capability it would be a shame to lose, since any future program of exploration will eventually outgrow any particular launch vehicle regardless the size.

Indeed, it is incorrect to apply economies of scale to the size of a rocket. Instead, economies of scale are actually realized much more powerfully by increased flight rate. A smaller launch vehicle, with lower development costs and lower recurring costs, will reliably be cheaper on a cost per kilogram basis than a heavy-lift booster delivering the same payload, because the flight rate will be higher. If a prospective HLV were to enjoy a sufficiently high flight rate that its cost/flight approaches the marginal cost of the vehicle, then efficiencies of scale could be realized; but no one can envision a time in the future where this kind of HLV demand will exist. For large capital investments, high utilization is the key to reduced cost, and is also the key to operational experience, which also reduces cost and increases reliability even further.

The size of the workforce required to support a heavy-lift booster is problematic for other reasons. With a large launch vehicle, manufactured and assembled by a large number of people at a large number of facilities across the country, there are a lot of people involved and a lot of exchanges between them to manage. This is a bad idea if you actually care about having an efficient, cost-effective operation. Every facility-to-facility exchange, every piece of hardware shipped intra-program increases the risk of something going wrong—a risk that usually demands increased management oversight and documentation to mitigate. In the interests of designing for cost, ideally a program should minimize the size of the team doing the work, locate them as centrally as practical to expedite and maximize clear communication, and minimize the burden of managing exchanges and interfaces. This is one of the enabling philosophies of small, low cost spacecraft. This also appears similar to the philosophies of SpaceX.

What about the simplicity of a large rocket? While there is certainly some aesthetic appeal to launching a big spacecraft with only one big launch vehicle, the cost of developing such a booster in the first place makes the mission design costly and problematic to begin with (as the Constellation program most recently experienced). And though HLV proponents often argue that larger boosters can minimize or eliminate orbital assembly—as though it were a bad thing—orbital assembly is in fact something that NASA has become quite good at. Indeed, the International Space Station has illustrated that very large structures can be assembled in LEO with great effectiveness. Orbital assembly has arguably become a mature spaceflight capability—a capability it would be a shame to lose, since any future program of exploration will eventually outgrow any particular launch vehicle regardless the size. This is true for the same reason that every other form of transportation outgrows the capability of any single vehicle. (Indeed, this is a fate that almost befell Apollo.) At some point it becomes silly to just keep building bigger and bigger transports.

Heavy-lift is not necessary, and even if we had it, we could reasonably choose not to use it, in favor of diverse portfolio of cheaper, smaller, simpler vehicles. A program that requires what only a single rocket can provide puts all its eggs in one basket, and is correspondingly fragile: the program will be delayed if the rocket is delayed; grounded if the rocket is grounded; and perhaps lost entirely if the rocket fails.

Are existing launchers sufficient?

Having argued that HLVs aren’t necessary, the complementary question is whether or not smaller launch vehicles are sufficient. This author contends that the answer is unequivocally yes.

Programs of both human spaceflight and human space exploration can readily be accomplished with existing or near-term launch vehicles, including (but not limited to) the United Launch Alliance Atlas 5 and Delta 4, SpaceX Falcon 9, as well as other launchers on the horizon such as the Taurus 2 and Falcon Heavy. While different vehicles are better for different types of missions (crew or cargo delivery, for example), the key advantage of using rockets that already exist (or are currently being developed by the private sector) is that the initial costs of any particular program can be substantially reduced. As well, the demand for a large number of flights can only be expected to increase competition and drive prices down, if competitively procured in the first place.

Heavy-lift is not necessary, and even if we had it, we could reasonably choose not to use it, in favor of diverse portfolio of cheaper, smaller, simpler vehicles.

Heavy lift certainly isn’t necessary for delivering hardware to low-Earth orbit, and existing launch vehicles are certainly equal to the task. SpaceX has demonstrated this with their inaugural Falcon 9 and Dragon flight. Orbital will hopefully follow suit with Taurus 2, designed to provide low-cost commercial resupply of the International Space Station. United Launch Alliance (and the companies that comprise that joint venture, Boeing and Lockheed Martin) been launching EELVs for many years now, and several initiatives funded under CCDev aim to develop new low-cost commercial crew transportation systems. While none of these have yet delivered a person to orbit or cargo to the ISS, there is certainly more cause for optimism here than with heavy lift. Several of the above launch vehicles are mature and flying. NASA, in contrast, hasn’t successfully built a launch vehicle in decades, let alone with a spec written by Congress.

Even for exploration missions, such as to near Earth asteroids, the Moon, or Mars, smaller launchers are similarly equal to the task, with the proviso that at least orbital rendezvous and docking is necessary. Fortunately, NASA has been doing rendezvous and docking for decades, and at this point can comfortably consider it something they’re good at.

An illustrative example of near-term exploration missions using existing launch vehicles, recently put forward by Alan Wilhite, Doug Stanley, Dale Arney and Chris Jones, presents a space exploration program that in no way requires SLS or anything comparable. Instead, this study and others like it rely on the use of propellant depots, which are perhaps the most enabling technology that can be developed for decoupling launch vehicle performance from exploration missions. A propellant depot doesn’t care how it gets refueled or by whom, and allows exploration spacecraft to be similarly indifferent. Propellant depots can be restocked continuously and asynchronously to the spacecraft that need them, thus allowing propellant launch to proceed on a parallel path to any particular mission. As a further benefit, the existence of propellant depots would allow spacecraft to be launched with their own tanks partially or completely empty, thus further reducing the performance requirements of both the spacecraft and its launch vehicle. Many propellant depots in the literature are designed to be launched with a single existing rocket themselves.

In spite of the advantages offered by propellant depots, they are not preconditions for undertaking exploration with smaller launch vehicles. The Mars for Less mission (with full disclosure of author bias) illustrates one approach for undertaking a Mars Direct-style mission with existing EELVs, using a set of four discrete propulsion stages assembled in low Earth orbit. Missions assembled and launched this way don’t need any on-orbit propellant transfer, so depots aren’t required at all, albeit at the expense of some efficiency and flexibility. A four-stage propulsion system of the sort specified in Mars for Less could be used to throw some 45–55 tonnes onto a Trans-Mars Injection (TMI) depending on the opportunity, and could also be used to deliver comparable payloads from LEO to Lunar orbit. This sort of approach is less effective in decoupling propellant launch from the mission, but could allow crewed expeditions to begin effectively right now, completely side-stepping the need for new launch vehicles or on-orbit infrastructure.

It is true that missions requiring a larger number of smaller launches can be expected to have launch failures; but in a mission with multiple launches and multiple launch providers, a single failure is less critical. With the exception of the booster delivering the crew to orbit, failure in any single launch can be offset by delivering a replacement component to orbit. And while a failing launch system may be shut down for investigation, a program that uses multiple launch providers needn’t suffer the same fate. On the other hand, a mission predicated on a single heavy-lift booster may find itself indefinitely delayed—or even permanently grounded—as a result of only one failed launch.

Even the most ardent supporters of using HLVs for human spaceflight, such as Bob Zubrin, have recently begun to acquiesce on using smaller launch vehicles (see “A transorbital railroad to Mars”, The Space Review, May 23, 2011.) Mission designs like Zubrin’s Mars Direct, after all, are predicated on aggressive minimalism, forward equipment deployment and supply caching. It makes sense to go a step further and eliminate the HLV bottleneck, which is otherwise incongruous with the design philosophy. Zubrin has recently proposed an ultra-minimalist Mars mission using the SpaceX Falcon Heavy, which will be capable of delivering 53 tonnes to low-Earth orbit. (While this can’t fairly be considered a small launch vehicle, it can certainly be considered an economical one: the advertised price range for a Falcon H is $85–125 million per launch, which translates into a game-changing $1,600–2,400 per kilogram.

Not having a heavy-lift vehicle doesn’t mean not having a robust and capable human space exploration program, and the benefit of using existing or near-term launch vehicles extends beyond the reduced or eliminated up-front development cost.

While Zubrin’s proposal for a two-person crewed Mars mission using three Falcon Heavy launches is a bit tight (and arguably doesn’t quite close), his full-scale four-person Mars Direct mission, consisting of two vehicles per complete expedition, could certainly be accomplished using multiple Falcon Heavy launches. A Mars Direct-style mission could be undertaken using only three Falcon Heavy launches per Mars-bound payload: the first launch would deliver the payload itself, while the subsequent two would deliver two high-performance hydrogen/oxygen propulsion stages, which would be mated to the payload and ignited successively to send the spacecraft Mars-bound. This system could throw between 45 and 55 tonnes trans-Mars, again depending on the opportunity, which would be sufficient to undertake Mars Direct with some (needed) margin. Assuming three launches per Mars-bound spacecraft, and two payloads sent to Mars roughly every two years, the average launch costs would be $375 million per complete expedition, using the upper end of the price range quoted by SpaceX. For perspective, this is about one third the cost of a single shuttle launch—a small price to pay for a continuing program of exploration. The same sort of dual-stage approach could be used to deliver comparable payloads to lunar orbit, for a more near term (and probably more realistic) return to the Moon program.

It should be noted that this sort of approach does share one disadvantage with HLV-based mission designs: it assumes, and relies upon, the capabilities of a single rocket that doesn’t yet exist. But at advertised prices, that single vehicle is hard to dismiss casually—and there is certainly cause to believe in SpaceX, which has already privately developed two launch vehicles, processing and launch facilities, and a crew vehicle for less than $800 million (compared to the approximately $9 billion spent on the defunct and flightless Ares 1, with no end in sight when the program was ultimately terminated). A compromise approach using propellant depots could be used to deal other launchers back into the game, which is probably a more robust approach regardless.

Existing or near-term commercial launch vehicles are more than sufficient for human missions in LEO and beyond. Not having a heavy-lift vehicle doesn’t mean not having a robust and capable human space exploration program, and the benefit of using existing or near-term launch vehicles extends beyond the reduced or eliminated up-front development cost. By undertaking space exploration with smaller launch vehicles, NASA could serve as an anchor tenant in the launch market, providing a demand that should encourage new providers, increase competition, and drive prices down further, to the benefit of both manned and unmanned spaceflight.

Roles and responsibilities

The US is currently rocket rich, but its space program isn’t comparably so. In 2004, President Bush’s Commission on Implementation of United States Space Exploration Policy recommended that “NASA’s role must be limited to only those areas where there is irrefutable demonstration that only government can perform the proposed activity.” Launch vehicles do not qualify—indeed, the opposite is ostensibly true in this case—and Congress certainly shouldn’t waste further billions trying to force NASA into developing a new one, for no apparent reason than maintaining jobs in the post-shuttle era (and thus preventing there from really being a post-shuttle era).

If there is a market for large, Saturn V-class heavy-lift vehicles, they will be developed. Indeed, it is notable that SpaceX is, with Falcon Heavy, in fact betting on the viability of a larger rocket—though in every important way, hopefully a fundamentally different sort of larger rocket. There is reason to be cautiously optimistic here, since SpaceX appears to approach vehicle development in a fundamentally different, cost-effective way.

In the final analysis, the argument here isn’t really against heavy-lift launch rockets, but against unaffordable or unneeded ones. The forthcoming SLS is an example of both: a rocket whose requirements are written more by politicians than engineers, developed more for political reasons than technical or economic ones, and stands in marked contrast to what the private sector is doing and what NASA could be doing more of.

If the United States actually cares about developing space—not just exploring it or studying it, but developing it in earnest, with the end goal of having a large number of people living and working in space—it would mean being able to launch crew and cargo economically.

Notwithstanding SLS, it has been exciting to watch NASA increasingly embrace commercial providers in recent years. Turning LEO transportation over to commercial vehicles would ideally allow NASA to focus on enabling technologies for missions beyond Earth orbit, for which the requirements are more challenging and several key issues remain unresolved. But “enabling technologies” should not include Senate Launch Systems, the pursuit of which will continue to cannibalize funds that could otherwise be spent addressing bigger challenges (such as advanced spacesuit technologies; high-closure life support systems; advanced space power systems; and entry, descent and landing of large payloads at Mars.) In-house launch vehicle development has had an extremely high opportunity cost for NASA, and they would better serve the cause of exploration by working on something else.

But in this regard, the agency is beholden to Congress. If the United States actually cares about developing space—not just exploring it or studying it, but developing it in earnest, with the end goal of having a large number of people living and working in space—it would mean being able to launch crew and cargo economically. The way to accomplish this is more activity and more competition, with as much commercial involvement as possible. A heavy-lift “Senate Launch System” is not consistent with these objectives, which really just affirms what we already know: that space development is not actually that important to Congress. But hopefully, at the behest of commercial efforts, a day will come when human space activities will flourish regardless of what’s important to Congress.



Enter your email address below to be notified when new articles are published:

ISPCS 2014