zaterdag 21 december 2013

Lunar resources and a new Space Race

With the successful landing of Chang'e 3, the attention of the world's spaceflight enthusiasts and experts alike once again turns towards the moon. These things spawn up discussions about why we should go to the moon, whether there is economical benefit to it, or about the "new space race", as well as the usual complaining about NASA's lack of productivity or that those damn commies is taking muh spaceflight.

And to be honest, when these things happen, I can't help but feel a slight annoyance. Many of these things are based upon false premises, or people who live in the 1960's, or general ignorance of spaceflight that results in xenophobia, anti-americanism, armchair experts (I'm being a little bit of a hypocrite here, I'll admit that) schooling others like it's their job and nationalist slapfights. 

Lunar resources

Something I've seen so many times lately is that we can use lunar resources to pay off our debt, make fusion attainable, makes spaceflight more sustainable and many more things. But is this really true? Can lunar resources make the moon an economically viable source for materials needed on Earth? I don't think so. Let me show you this with an example.

Say we assume the cheapest LV in terms of cost/kg currently available in the near future, Falcon heavy. With ~2500 $/kg, this vehicle could get 53 metric tons into LEO with a cost of $135 million. The ∆V needed to return from the lunar surface is 3 km/s. A 53 ton stage with a 0.87 PMF (realistic since there have to be legs and everything attached to the stage) could get about ~35 tonnes back to Earth. This stage would land on the moon, be refueled there, and could be sent back. So how much would this cost?

The cost of this material would be $3900 per kg. But it gets worse. This assumes that the entire payload is nothing but the actual material. If you put it in a container, it becomes more expensive. Something like a Dragon capsule can return 3/4 of it's own mass back down to the surface, or 3/7 of total mass. Suddenly, you're looking at 4 tons returned to Earth, and a cost of $9000/kg. And in the case of a low density material like Helium 3, you get even lower return down to the surface. 

To make things even better, this assumes that the base put there was entirely free and that the stage could be refueled there for free. If you don't assume refueling there's no payload at all, and if you take into account the cost of a lunar mining base, which is on the order of $100 billion, you'll quickly see why lunar mining isn't very economical currently. This also assumed the cheapest vehicle to LEO available in the near future; if you went with Delta IV it would cost 6 times as much.

So, lunar resources are, for the time being, a pie in the sky dream. They are extremely expensive to get back and the ROI would take so long (if there was any profit at all) that it simply isn't worth it for a long, long time. I'm not saying they can't be useful, a prop depot in lunar orbit or a small refueling base could really ease up lunar exploration. I don't think it should be held like a good reason to go there. If lunar resources are to be seriously considered as a reason for lunar exploration, we should first get launch costs down by one or two orders of magnitude before it is a serious argument in favour of lunar resources for use on earth.

And Helium-3... How about we perfect Tritium-Deuterium fusion before we go there to pick it up? And there's always Proton-Boron fusion if neutron-less fusion is so important. It's not really necessary for attainable fusion.

A new Space Race and China's space dominance

By far my favorite. Go on any thread on reddit related to China's space program and you'll find a bunch of people preaching about the new space race, that the US should step up their game because the Chinese are overtaking "us", etc. And I usually just sit there laughing or being slightly frustrated with this.

First, why a new space race? There certainly are no signs of one. China is stepping up their game but neither the US nor Russia show any sign of doing the same. Their plans have changed little regarding exploration; the US still has to work in the direction of an asteroid, and Russia still works in the direction of the moon and Lagrange points. There is no space race going on at the moment.

What surprises me even more are all the people who wish for a new space race. And I flat out don't get it. What happened the last time we had one? The US landed on the moon. The costs, which were so high in part because of time pressure, sent Congress and the administration into shell shock.  Apollo applications was canceled and a "cheaper" LV, the space shuttle, was signed into law. That "cheap" LV got the US stuck into orbit for 30 years. Apollo was expensive, accomplished little and forced NASA into the terrible position it was in for the past few decades, until it was decided to finally axe the shuttle in favor of a conventional rocket with a normal capsule. A safer, more affordable and more flexible vehicle. Alright, Orion and Ares 1 are a bad example. Falcon 9 and Dragon, or Atlas V with CST-100 are better examples. Still, the last space race got us stuck for a long time. 

A new space race would get NASA to the surface of the moon, maybe Mars, for a few times. We would alienate a potential spaceflight partner, would cause NASA to create an expensive and unsustainable program similar to CxP, only to cause massive budget cuts once the program is over and we'll never go anywhere for many years afterwards. The other option is teaming up with China/Russia/whatever dirty commie country 'Murica is afraid of for no reason/ESA and spread costs and effort over several different nations, allowing large scale exploration for much lower cost in a more sustainable way. Which one do you think will result in more stuff getting done? 

Second thing is, why are people so afraid of China? As reddit user Ambiwlans so eloquently put it, "Since Russia collapsed all the people that need a bad guy to fear seem to have latched on to China." Since there is no big scary red Russia to fear, people start irrationally fearing the other "big scary commie country" out there, which is China. But is China really that big of a danger to the US space "dominance"? Let's compare the two, shall we:

Biggest LV: Long March 2F/G, 11.2 tons into LEO. 25 ton launcher in the works.
One crewed spacecraft, three crew, a few days on orbit life time. No new one in the works.
Budget $1.3 billion

Biggest LV: Delta IV Heavy, 28.8 tons to LEO. 53 and 70 ton launchers in the works.
No crewed spacecraft, but four in the works. One can support crew of four for three weeks.  
Budget $16.8 billion

So yeah. Another big red scare that really isn't that scary. China is not yet capable of doing anything the US hasn't long been able to do, and they don't have the means to do anything significant for the time being. They have a focus, which is a big improvement over what the US has. But should the next administration of the US decide to switch focus, all that has to be done is developing a lunar lander and the US could be back on the moon by 2025. China isn't even trying to land earlier than that. Older articles claim 2017 but these are usually very outdated.

"NASA should get off their arse and do something productive, they don't do anything unlike China"

Let me just answer with a few pictures. 

Curiosity is a lot cooler than Yutu.

Cassini, currently orbiting Saturn
New Horizons spacecraft, currently underway to Pluto
SLS tank barrels being manufactured right now
Dragon spacecraft, funded and co-developed by NASA

zaterdag 14 december 2013

Random Thought: Should we avoid SLS Block 2, or go straight for it?

In previous blogs of mine, I wrote about how I wish NASA would stick with Block 1 instead of continuing development of SLS after EM-1. The reasons for this are simple; development is expensive. If SLS Block 1 was to be NASA's new main launcher, it would save billions in development cost, while at the same time giving a very capable 90 ton launch vehicle that can take whatever we throw at it. However, I changed my mind somewhat. I have given the option of going straight toward Block 2 as our main exploration launcher, instead of sticking with Block 1 or 1A, a chance in my mind. There are many advantages to doing this, and I'll explain myself here.

Advantages of Block 2

#1: Block 2 doesn't have to be so expensive to develop. The current path baselined by NASA is expensive, but it doesn't have to be. The current one has the following order of upgrades:

1. Develop new boosters for SLS, either advanced solid or liquid; increases payload to >105 tons.j
2. Develop an in-space stage for SLS to replace the iCPS. 
3. Wait nine years, fly SLS once every one or two years. 
4. Develop a new, massive J-2X powered upper stage, redesign the core you've been flying for 15 years by adding an extra engine. Increases payload to >130 tons to LEO.

Since then, however, a few new upgrade paths have been drawn up. One of them was put forward in a paper by Boeing named "THE SPACE LAUNCH SYSTEM CAPABILITIES FOR ENABLING CREWED LUNAR AND MARS EXPLORATION" which is available on L2. The paper describes adding a new upper stage to SLS, powered by 2 J-2X engines. This variant, sometimes nicknamed by the community as Block 1C, would use normal boosters, a normal 4x RS-25 core, and a new upper stage. Just by adding a new upper stage, the "70" ton version gets 130 metric tons to LEO. Now, I have to be honest, I'm a little skeptical of this payload. Similar vehicles shown in ESAS couldn't reach the 130 goal, and ATK claimed in their Advanced Booster paper that SLS needs 5 engines to reach 130 tons to LEO. The Boeing paper also renders this "Evolved SLS" as having 5 core engines, despite claiming it has 4 engines. However, it should still be possible to evolve the core design for the follow up to Block 1 to allow for 5 engines. By making such a decision early on, it should save a lot of development cost.

Now, after some time, new boosters will be required. There are sufficient booster casings for ten SLS flights with standard 5-segment boosters, but if we ever get serious with SLS, new ones would become necessary. However, having already reached 130 tons, this would allow for cheaper, affordability focused boosters instead of the big, performance focused advanced boosters using the F-1 engine proposed by Dynetics. In fact, they could be even cheaper than the ATK advanced boosters because they won't have to be as big.

And other possible development path is by adding a 4x RL-10 upper stage to Block 1, then adding advanced boosters. No core redesign required, and although Advanced Boosters will become necessary, the new upper stage would be significantly cheaper than the J-2X upper stage, both to develop and operate.

Either of these paths would probably allow Block 2 to be ready by ~2025 instead of the current 2032.

#2: Block 2 is actually cheaper than sticking with Block 1 in the long run. While it would cost more to develop, Block 2 would cost pretty much the same on a yearly or per-flight basis as Block 1. The rocket shares tooling for the upper stage, costing very little extra. The engines, either RL-10 or J-2X aren't free but don't cost hundreds of millions either. The boosters, whether they're designed to be powerful or affordable, are supposed to be cheaper than the 5-segment ones on Block 1. The fixed costs would be slightly lower, marginal cost slightly higher, but the difference is very small either way. Assuming identical cost, SLS Block 2 costs significantly less at payloads above 70-90 tons, and anything above 140-180 tons because of requiring less flights. The use of advanced boosters compared to normal boosters also allows for lower fixed costs, though the exact costs of this is hard to estimate because of a lack of info. ATK claims a cost reduction of 40% for the boosters, but I don't know how much a booster costs, so the exact difference is hard to pin down.

There's also something I glanced over previously. For sticking with Block 1, you'll eventually need to develop an in-space cryogenic propulsion stage and new boosters too. The cost saved from sticking with Block 1 aren't very significant at all.

#3: Block 2 frees up more money for actual missions. You might be wondering what I'm smoking right now. Believe me, I'm not smoking, this might actually be true. The current plan is to operate Block 1A for nine years to perform all missions to asteroids/the moon and continue developing Block 2 in the background at the same time. By going to Block 2 directly, you save money afterwards because you don't have to keep on developing upgrades for SLS. You might put back missions compared to sticking with Block 1A, but they can be developed quicker afterwards. Considering NASA's dire lack of funding, stopping funding for development of the launcher and flying the final config as soon as possible seems like a more realistic path to other worlds than sticking with a less capable, just as expensive variant that will have to be replaced later anyway. Unless a new NASA Authorization Act is made dropping the 130 ton requirement, Block 1A would require an upgrade sooner and later, and putting it off might only cost more money in the long run.

#4: Unlike Block 1, Block 2 is actually a big improvement over other launchers. Block 1's 70-90 ton LEO payload is very good compared to current rockets, but it's not something cheaper rockets could reach for less. 130 tons to LEO, however, is a different story. I previously estimated that Falcon Heavy with a Raptor upper stage could get ~65 tons to LEO, and Atlas Phase 2 can get 75 tons into LEO according to ULA. Sticking with Block 1 would neuter the point of even having SLS, since these launchers would give similar LEO performance for less. Block 1A would give a decent increase to 105 tons, but 130 tons is sky high above all other launchers and is the only payload capacity that could somewhat justify an expensive Shuttle Derived Architecture over, for example a 70 ton inline kerolox concept combined with in-space refueling. Whether we need that capacity is a bit of a different story, but there certainly are reasons it could be useful.

There are also many mission possibilities that could be opened up by Block 2 that would require several launches on other launchers, including other SLS blocks. Assuming 130.0 tons to LEO, a Block 2 with 0.9 PMF CPS optimized for lunar missions and 462 Isp would get the following performances to other orbits:

  • 44.6 metric tons into Low Lunar Orbit
  • 56.3 tons to Trans Lunar Injection
  • 54.2 tons to the nearest Near Earth Asteroids
  • 48.9 tons to Mars ( ~7-9 month travel time, 3700 m/s)
Now assuming these values, there are a few options that open up that Block 1A would require multiple launches for. 

One is a manned lunar landing with 3-4 astronauts for a week down on the surface. A paper with presentation from Spaceworks Enterprises Inc. includes the estimates for lunar landers from L2 with several different propellants. The lander+in space stage combinations show mass estimates for single stage landers that can land from and return to LLO. The mass of a single stage methane lander, including 4 ton habitat, is 20.9 tons, and with hydrogen it's 16.8 metric tons. According to the Boeing paper I referred to earlier, the empty mass of an Orion spacecraft is 14.9 tons; even with a specific impulse of only 316 (STS OMS) and a worst case 1200 m/s ∆V to return, the Orion has a mass of 21.95 tons in LLO. 20.8+21.95= 42.85 tons into Low Lunar Orbit, which fits within the 44.6 ton capability mentioned earlier. A more optimistic case with a hydrolox lander and 900 m/s return ∆V gives a total mass of only 36.8 tons, which fits very well and gives plenty of margin. This Lander uses a habitat that is actually bigger and with more supplies than the Boeing reusable lander, which carries 3 people for 7 days. While the SEI lander is designed for 2 people for 20 days, it could support a crew of 3 or 4 people for a full week, giving the lander similar capability to the Altair lunar lander for much lower mass. A Constellation class mission in a single launch could be possible with SLS Block 2.

Another option could be a mission to a Near Earth Object mission in a single launch. The mass of the notional habitat baselined by NASA has a mass of 23.9 metric tons. It has 72 m^3 of habitable volume and can support a crew of three for a year. Judging from the diameter, this hab would be based on the ISS MPLM module which is built in Italy. Combined with a fully fueled Orion, worst case mass of 24.2 metric tons, the mass of this spacecraft would be 48.1 tons. SLS Block 2 with in-space stage would be able to get this entire stack to a trajectory of 3760 m/s, enough to visit an asteroid. In fact, this is enough to send the crew to visit Mars, though the hab would require some mods to allow for such a mission time, and it wouldn't be able to do anything useful there.

The last option is a Mars Direct type mission. The 48.9 ton capacity to Mars with a C3 of 11 km^2/s^2. According to the NASA trajectory browser this is sufficient to reach Mars in 200-300 days, depending on launch window. That same browser shows plenty of windows to return from Mars with 800 m/s of delta V, or 2.3 km/s total including escape velocity. Assuming aerocapture to reduce MOI delta V to only 100 m/s and 2300 m/s to return, as well as methane/oxygen propulsion to return, a spacecraft of 48.9 tons would have an empty mass of 25.2 tons. This is enough to hold the propellant tanks and propulsion system (2.7 tons), a crew return capsule (8.94 tons) and an inflatable habitat (13.5 tons). That might seem small for such a habitat, but a study from SEI estimated the mass of an inflatable hab for a crew of three with 60m^3 of habitable volume at just 10 tons. Such an ERV, similar to the Mars Semi Direct approach, would be much roomier than the capsule-sized ERV Zubrin proposed in the original Mars Direct, which had a mass of just 7 tons.

All in all, it seems to me that developing Block 2 immediately instead of sticking with Block 1 or 1A has many advantages and it could very well be the right path forward. Unless the 130 ton requirement is dropped, upgrading to it will be necessary, and getting the upgrading out of the way will certainly help. It will all depend on what missions SLS is supposed to perform though. For any case though, getting as much out of SLS as possible certainly seems like the right way to do so. 

maandag 25 november 2013

Delta IV for exploration

NASA's current focus is exploration, and for exploration of anything, you will need some kind of infrastructure, whether the goal is the moon, asteroids, or Mars. NASA is currently working on two vehicles that are supposed to provide the foundation of all exploration architectures for the coming decades: the Space Launch System and the Orion crew vehicle. These two vehicles are very useful for building a deep-space architecture, but they aren't guaranteed a future, especially with sequestration and other political nonsense lately putting many big NASA projects in danger. And at $1 billion and $1.4 billion annual for Orion and SLS respectively, they are some of the biggest projects around. Therefore, it's always necessary to hold some backup plan around in case Congress goes nuclear on NASA's budget. Here I present an plan that might be part of such a backup plan, that could get humans to places in a more budget constrained environment. 

Delta IV launch vehicle

The Delta IV launch vehicle, in its Heavy configuration, is the most powerful launch vehicle currently available, capable of bringing 28,790 kg to a Low Earth Orbit [1]. The launcher is very reliable, having suffered only a single failure, and is completely American and provides lots of upgrade potential. Delta IV is not exactly cheap at this moment, with a price tag of approximately $370 million[2], but the main reason for this high price tag is the very low flight rate; only once a year for Heavy, with about 3-4 launches per year for the complete Delta IV family. This low flight rate is because of lack of demand, caused by the competition of Atlas V. Increasing this flight rate by two launches and six additional cores every year would reduce the per-unit cost by a decent amount.  

The reason for picking Delta IV for this comparison was the upgrade potential and the fact that all the infrastructure for launching large amounts of cryogenic propellant to orbit are all already included; something that does not exist yet for Falcon Heavy. Delta IV is also being used for an Orion flight test next year, and is therefore likely closer to being adapted for launching Orion than Falcon heavy or Atlas V heavy is. Delta IV Heavy already exists, unlike Atlas V Heavy or Falcon heavy. Also, unlike Falcon, upgrades to Delta can be partially paid for by the Air Force. All in all, upgrading Delta IV for the job could likely be done quicker and cheaper than either Falcon Heavy or Atlas V. However, they remain strong alternatives, and in case of Congress going nuclear, it's up to NASA to decide which one is better.

These alternatives, as well as Atlas Phase 2 and others, can easily be copy-pasted in this paper mission, and I see no reason why it couldn't work. Delta IV was simply assumed here for reasons mentioned earlier.

The exploration gateway

For exploration, some infrastructure to support it will be needed. By far my favorite proposal to do exploration is Boeing's Exploration Gateway plan. This option allows for lower cost access to the lunar surface, and provides a great place to do deep space research and as a staging point to asteroids and Mars. In this post I will focus on the lunar architecture, like I usually do. It can be expanded to Mars later.

The plan I'm presenting here would not dump Orion, but rather, it would dump SLS. As much as I personally support SLS, it's one of the more likely to be canceled and it's more expensive than Orion, and potentially also more replaceable. 

Assembling the station

The Gateway station plan includes a small space station to be assembled behind the moon, at the EM-L2 point. For this plan, the space station would consist of two main smaller components which can be launched in two launches with the Block 1 SLS. The first launch would carry the Science/Power module, the second one the Node and Utility module. They would be docked at the L2 point and a third SLS brings the Orion spacecraft there to man the station and start doing science. The architecture assumes SLS can lift 27 tons to TLI; Delta IV Heavy probably can't even get half of that there in a single launch, so a slight change in architecture will be required.

ACES compared to Centaur and DCSS
In order to reach the 27 TLI goal, a dual launch is obviously required. However, even with a fully fueled DCSS in orbit, the TLI capacity falls a few tons short. In order to reach the goal, an upgrade to Delta IV will be required. The ACES (Advanced Common Evolved Stage) allows Delta IV to lift a whopping 37 tons to LEO [1]. It carried 41 tons of propellant and has an empty mass of ~3.6 tons, and uses four RL-10 engines with a specific impulse of 461.5 [3]. Using a propellant drop tank, it can refuel itself in LEO; such a drop tank could realistically carry about 34 tons of propellant, allowing the ACES to refuel itself to about 83% of maximum capacity. With a propellant load of 34 tons and 1% prop residuals, the ACES stage can send 28.8 metric tons to a TLI trajectory with a delta V of 3200 m/s. While this is not using the most accurate numbers in existence, it nonetheless provides 1.8 metric tons of margin, as well as the not-insignificant launch margin from the payloads (ranging from 1.6 to 4.6 tons of margin). 

To recap, in order to assemble and man the station, Delta IV would launch six times:
1. DIVH Launches drop tank, upper stage refuels in orbit
2. DIVH launches Science and Power module
3. (see 1)
4. DIVH launches node and Utility module
5. (see 1, 3)
6. DIVH launches an Orion spacecraft with a crew of four astronauts

The parts would rendezvous and dock at L2; Orion, the Utility and Science module all have their own propulsion system with sufficient delta V to dock.

Lunar surface missions

The Boeing HLO lander
In the Boeing plan, there are two architectures described for landing on the lunar surface. One uses the SLS upper stage as a crasher stage and a methane-oxygen lander. The other one uses a single-stage two man lander with storable propellant and a LTV to get the lander from HLO to LLO and back. The architecture that is most suited for a Delta IV based architecture is the storable propellant lander, since it does not use any parts bigger than SLS block 1 allows for. [4]

A humble change to the architecture is to use Orion's SM instead of ATV as the LTV. ATV production will stop after number 5, and there is not chance for revival. Orion's SM, which is ATV derived, will be able to operate independently from the crew module (required by law) and is European anyway. One of the big reasons to use a modified ATV is to allow for international cooperation to share the costs, but if ESA provides and maintains the Orion SMs for this mission it will do that just as well. Assuming a lander mass of 21.4 tons with 6 tons empty mass, as given in [3], an Orion SM with 316 second Isp and 4115 kg empty mass (Oxygen, nitrogen and water are not needed) can do this job of transferring the ship from HLO to LLO (571 m/s [5]) with a propellant load of 7620 kg, which fits within the 7907 kg maximum prop load Orion can hold [6]. The tanker would still be a single vehicle though; Orion SM with a fuel tank for 15.4 tons of hypergolic fuel would (barely) exceed the 28.8 ton TLI capacity of two Delta IV rockets. It's still a possibility though, but with very little margin, only ~200 kg, which NASA likely wouldn't risk. 

For lunar missions, the Utility module would use its propulsion system to transfer the station from L2 to High Lunar Orbit, where it will remain for the rest of the lunar campaign. After this, the lunar lander is launched fully fueled, followed by the LTV (small enough for a single ACES Delta IV H) and a crewed Orion spacecraft. The crew would dock with the station, transfer to the lander, and would descend to the lunar surface from where they can do a 14 day science mission with 300 kg of scientific instruments. After the mission, they ascend to LLO, where they will dock with the LTV and return to the station. At the station, the crew enters their Orion spacecraft and heads home. For follow-on missions, a tanker vehicle and Orion are sent to the station, taking in total four launches per follow-on mission. 

It would be possible to launch Orion to the station in one go by adding 6x GEM boosters to the Delta IV booster. This would increase payload to about 45 tons [1]. Using Orion and a drop tank at the same time, it could refuel to about 25 tons of propellant. Orion could be short-fueled to about 17 tons for L2 missions, in which case ACES has sufficient delta V to get Orion to the station. A single Delta IV would have about 19 ton TLI capacity, enough for Orion or any other crew or cargo vehicle currently planned. This upgrade would save 1 launch per mission on the manifest, but would require additional upgrades to Delta IV which might end up costing more.

The advantages of a Gateway station

The architecture described by Boeing, and the modifications I made, have a few significant advantages over normal exploration. First, there's technical/economical advantages: the Gateway allows the lunar lander, by far the most expensive part of this architecture, to be fully reusable and used for many times. Unlike letting the thing float in LLO, it can easily be refueled and can be repaired at the station, should maintenance be required. It is also an efficient staging point for missions to asteroids and Mars; while those will likely require bigger launchers in the 80 ton class, the groundwork for such missions could be laid with near-term launchers. 

Another advantage is the international aspect of the plan. The costs are split over several participating countries. For example, the US provides the LV, the Node and Utility Module, the Orion crew module, and commercial resupply. Russia provides the Science and Power module and does most of the lunar lander. They also provide crew access and HLV capacity later on via Angara 7, PTK NP, and Sodruzhestvo (Zenit super Heavy). Europe contributes to the lander, the LTV and provides a LV for commercial resupply. The advantages of such an approach are obvious. America can provide their part within the current exploration budget, while the high cost of a lander of ~$7 billion [7] is split over two space agencies rather than one. 

All in all, an international Gateway station program allows for a flexible, low cost exploration program supporting a wide range of missions, and Delta IV with the ACES upper stage upgrade is very well suited for laying the groundwork for an exploration program and supporting the exploration of cislunar space. 


[5] Done using the Vis-Viva equation: 232.6+(1972-1633.9)=571 m/s Delta V

donderdag 7 november 2013

Low Cost Lunar Missions; To the moon with Ariane 6

Lunar exploration has always been a huge interest of mine. The moon is our closest neighbor and could teach us tremendous amounts about the history of our solar system and our planet. It could also function as a place to gather resources to explore further into the solar system. In short, there's plenty of reason to go there, but how? Many earlier plans to got there have yielded nothing but powerpoints and pretty animations. Getting there in a low-cost manner would be critical.

(For info on the plan discussed, scroll to the bottom.)

Constellation: How not to go there

The Constellation program was initiated by NASA as part of Bush's Vision for Space Exploration policy in 2005. It had the goal to return Americans to the moon by 2020 and give America independent manned access to space by 2014 after the Shuttle's retirement in 2010. However, the way they wished to accomplish this was doomed to fail from the beginning. 
The first fatal flaw of the program was the way two different LVs were used for achieving one goal. Using two different vehicles delayed Heavy Lift capability and costs a lot of additional money for developing the extra vehicle. Even if the smaller vehicle is a lot lower in cost than the big one, which Ares 1 certainly wasn't, you still wind up having to pay more. Using two or more launches of the same launch vehicle, like is the plan with SLS, significantly reduces the costs and allows the mission to take place much earlier, and doesn't require a vehicle as big as Ares V, which by itself was way too big and expensive.

The second fatal flaw in the program is that it's very ambitious. Landing four people on the moon for at least 7 days, anywhere on the surface, is a very big requirement, which ends you up with a spacecraft and lander which are both extremely huge and very expensive to develop. It also caused Ares V to be huge; up to 188 metric tons to LEO according to some sources. Even the biggest version of SLS won't go over 130 tons, and that is still years into the future. Ares V would have used a new, bigger core, new engines, new boosters and would have nothing in common with the space shuttle. It was a completely clean-sheet design and it would have cost over 20 billion dollars to develop. The whole constellation program would've cost over $40 billion dollars just for the first lunar flights. NASA doesn't have the funds for that.

Lastly, another big flaw in the program was Ares 1 itself. The vehicle was designed from the start as "it must use a space shuttle solid rocket booster as the first stage" and that is were the problems started. Even when using the powerful RS-25 engine for the upper stage, it was underpowered and provided almost no margin for Orion. The slightest grow of Orion or performance reduction of Ares would have made the LV useless.

Early Lunar Access

ELA lander docking with EDS. Credit: NASA, Wikimedia
By far one of the better ways to do "budget moon flights" is the Early Lunar Access architecture designed in the early 1990's. It was based on the basic "Faster Cheaper Better" spirit at NASA at the time. It requires only two launches of Medium Lift Vehicles; one space shuttle and one Titan IV or Ariane 5. The total mass in LEO is only 52 tons. It would use only near-term hardware and could be ready by the year 2000, and would bring two people to the surface of the moon. While less capable than constellation, the use of smaller expendable LVs allows more missions to be done for a lower cost, more than making up for the smaller capability. The idea is great, but slightly outdated. Many modernized versions of this architecture are possible, using newer launchers like Falcon Heavy, Delta IV, Ariane 5 or SLS. This time, however, I'll focus on an architecture using Ariane 6, Europe's new, low cost launcher.

Ariane 6

Ariane 6. Credit: ESA

Ariane 6 is the successor to Ariane 5. It is expected to become operational after 2021, and cost approximately €70 million a piece, which is about $93 million dollars. It can get 6.5 tons to GTO, but its LEO performance has not been released yet. The vehicle is designed for GTO and that's where its market is at, so LEO performance isn't as important. However, it is possible to estimate the performance of the vehicle.

Using the LV performance calculator, which should be added is just an estimation tool, I've been able to model the performance of 6.5 tons to GTO using the following numbers:

Stage 1, 2 and boosters: 135 tons of propellant, 11.8 tons empty mass, 4500 kN of thrust, 280 second specific impulse. 

Stage 3: 28 tons of propellant, 4.94 tons empty mass, 180 kN of thrust, 464 second specific impulse. 

All these values are based on Vega's P80 lower stage (which forms the basis for A6's P135 main stages) and Ariane 5's upper stage information (propellant mass fraction was improved, because A6 doesn't have the same volume restrictions as A5). For LEO, a mission to a 200x200 km orbit with an inclination of 6ยบ, it has a payload capacity of 17.3 tons. This seems a little optimistic to me, so I took out a chunk of payload by including a 12.5% performance margin for error, reducing payload to 15.1 tons to LEO. This is a lot more realistic for this vehicle, and it's the LEO payload I went with for the rest of this article. 15 tons is a lot less than the 23 tons Ariane 5 ME can get into orbit, but it does so at a much lower price; only $93 million instead of the estimated $210 million for Ariane 5. That's a 32% decrease in cost per kg (and pretty close to the claimed 30% cost decrease for A6). For this reason, Ariane 6 is the vehicle I went with. That, and it's European, and the European space program is a great interest of mine.

The mission 

Launch 1: The first Ariane 6 launches the Lunar Landing and Return Vehicle. This vehicle consists of a capsule (3.7 tons fully loaded), a propulsion module carrying an engine (Aestus 2, storable propellant, 55.4 kN and 340s Isp) and landing legs and a propellant load of approximately 9.3 tons.  Total mass of the system in LEO is 15 metric tons. The hypergolic Aestus 2 engine was chosen because it allows fuel to be stored in space for much longer. The Aestus 2 is by far the most efficient upper stage engine available for this purpose right now. It is launched without a crew: to prevent another Ares 1 fiasco, it's probably better not to launch humans on a solid powered rocket without any significant margin, especially not because it doesn't have a Launch Escape System yet. This part of the ship has 3226 m/s of ∆V.

Launches 2,3 and 4: Carry small propulsion modules. They each have a mass of 15 tons and a structural index of 10%, meaning that 10% of their mass is non-propellant. Thee of them give the stack, including the LLRV ∆V, a total of 8697 m/s, which is just enough for the total round trip from LEO back to Earth. 

Soyuz can provide crew transport to the LLRV
Launch 5: Should Ariane 6 not be capable of launching humans, which it probably won't be, a 5th launch would be necessary. This would be a Soyuz, launched from Kourou Space Center in French Guyana. It would bring a crew of 2 to the loitering spacecraft in LEO. The Soyuz facilities at GSC are made so that they can easily be adapted to human launches, so this shouldn't be a big problem. Landing locations after the flight are a different story, but they aren't supposed to reenter in Soyuz anyway.

Major mission events include a 3100 m/s burn to a trans-lunar trajectory, a 2700 m/s burn to land, surface exploration by the two astronauts, an 1872 m/s burn to enter LLO and a 1025 m/s burn to return to Earth. In order to save ∆V, the spacecraft sent on a trajectory straight to the moon and does not enter LLO before landing. This saves about ~300 m/s compared to entering LLO first (only 200 m/s was assumed here). However, in order to allow some more global access to the moon, the craft does enter LLO on the return trip

The launch costs of this mission are not insignificant. 4x Ariane 6 is a total of $372 million dollars, and a Soyuz adds another 40-60 million. In the worst case it's $432 million total. However, this is still a lot lower than the cost for two SLS launches (which I previously estimated at $1.4 billion a piece, for $2.8 billion total). It also helps kick up the flight rate of Ariane 6. In order to get a low price out of the vehicle, a low price is required. It is supposed to launch 7-15 times a year, and reach the 70 million euro goal at 7 launches a year. Costs can only go down by flying the vehicle more often, and four extra launches will certainly help with this. 

Sources and other interesting information:

zaterdag 5 oktober 2013

SLS: Some more about costs

Some weeks ago I wrote a blog about SLS costs. Now, it was more meant to provide a rough idea as to what SLS would cost per launch, not a full on cost analysis; the methods I used for this were obviously flawed in several ways. In this one I'll look a NASA budget availability study from 2011 (I won't do my own analysis), as well as some more things about budget that I often see confusion about.

What exactly is "SLS"?
This is a dumb sounding question that is actually a lot more complicated than you might think. When people talk about "SLS" they often mean the complete system, not just the rocket which is what I refer to as "SLS". The current Exploration Systems Development programs contains three different main programs: SLS, MPCV, and 21st Century Ground Systems. They all have their own costs and their own purposes, with the current budget looking like roughly $1.4 billion for SLS, $1 billion for the MPCV and about $400 million for 21st CGS, totaling at 2.8 billion dollars. Now, when you want to compare the costs of SLS with other systems, it's very important with what you actually compare. If you compare SLS to a complete system like the Space Shuttle, you need to factor in complete costs. However, if you compare it with just a launch vehicle, something like Saturn V or Falcon Heavy, like is often done, you need to look only at SLS itself.

"According to NASA, SLS would cost $10 Billion per launch." No, it wouldn't.
I've seen this one far too often. The study this refers to does indeed claim the entire system would cost $41 billion through to FY2025. However, taking this as "SLS cost $10 billion" is completely wrong. First of all, it includes the entire system, not SLS. The costs for SLS itself are "only" 20.245 billion dollars. In the second case scenario, SLS can launch five times for 21.524 billion dollars. That's a 1.279 billion dollar difference for a single SLS. Not quite 10 billion per launch. Anyway, this study takes development costs into account as well as the first few flights, and with a projected total development cost of $18 billion dollars and with a low initial flight rate of 1 every two years, I'm hardly surprised cost would run up that high.

How much would it really cost according to that study? For SLS alone, if you take case #2, which is president budget+escalation, it's estimated SLS can launch once a year from 2023 on at an annual budget of $1.73 billion dollars. I should say, however, that these are estimated 2023 dollars. These numbers are estimated with an inflation rate of 2.5% per year, which means that translating it to 2013 dollars gives a modern day flight cost of $1.35 billion, for one flight per year. That includes on going development of SLS Block 2. If you take SLS as a complete system, it would cost about $2.8 billion, of which ~1.4 is for SLS, ~0.8 is for Orion and ~0.4 is for 21st CGS. This budget fits roughly within the current budget. So, SLS would, according to the study, be capable of one flight per year even under current budgets when adjusted for inflation.

Final words on cost discussion
To be honest, I think it's better to not argue too much about the cost of SLS. There are simply way too many unknowns at this point. The NASA study I talked about assumes SLS as a Space Shuttle External tank with SSME engines and Ares 1 boosters and upper stage engines. However, many design changes have been made that would affect costs in a number of ways. The core, for example, is pretty different from the STS ET. It's a lot bigger and uses more engines, for example, but on the other hand it uses easier to work with materials, new more automated tooling and new "mass produced" engines that all significantly reduce the amount of people needed. The boosters will be replaced after 2023 if the Block 1A approach is taken, and while these boosters are much more capable than the current ones, they also offer a number of ways to reduce costs, be it either cheaper and easier materials (ATK's AB), liquid kerosene propulsion (Aerojet, Dynetics) or simply using cheaper, existing engines (RS-68 boosters are still being considered). If the Block 1B approach is taken, SLS will use a Dual Use Upper Stage, which is a lot lower on development costs than the J-2X powered Large Upper Stage currently baselined.  And then there's the fact that it's just all estimates. SLS cost could end up much lower or much higher than what is currently estimated at about 1.4 billion (or the $500 million cost figure that makes no sense and is very unclear in every way). Until Block 1A or 1B has flown we simply can't know for sure what the costs for the rocket will be. The only answer for SLS costs you can give right now is "somewhere between 500 million and 3 billion."

maandag 16 september 2013

SLS Block 1 missions: Returning to the Moon using only the basic SLS

In my previous blog, I wrote about canceling the Block 2 SLS because I considered it unnecessary, claiming that Block 1 should be more than enough.  Now, to be fair, Block 1 will probably not last us long enough for a serious program because the current booster casings only allow up to 10 flights. After some time, Block 1 will run out of boosters and we'll have to make new ones, which would basically turn it into Block 1A with advanced boosters. Still, I think that the initial 90 and 105 metric ton capacities of these vehicles are more than enough, and in this article I will describe a possible lunar mission using only Block 1 capabilities.

Block 1 capabilities 

The Block 1 payload to Low Earth Orbit is usually described as 70 metric tons. However, if you have paid close attention to the project and have an eye for rocketry, you might have noticed that this isn't true. Initially, SLS was to have four different variants: Block 0, Block 1, Block 2 and Block 3. Block 0 was going to have 4 segment boosters, an External Tank length core and three RS-25 engines. This version would lift 70 tons to LEO. The Block  1 described here would have a stretched core, 5 RS-25 engines and 5 segment boosters and would be able to lift 100 metric tons to LEO. (More info on these variants, as well as other SLS alternatives, can be found here, on When NASA later started to refine the project, they found that Block 0 would only complicate development since the core would need to be completely redesigned for Block 1 and later, so they decided to take it out. Block 1, however, remained mostly unchanged, except that one of the engines was taken out because of being unnecessary. The payload decreased slightly due to this, from 100 tons to an estimated 90-95 tons. However, due to a variety of political reasons that I don't want to get involved with, NASA still has to claim 70 tons for the initial variant. If you don't believe me, this NASA document also estimates Block 1 capacity at 90 metric tons. While it's probably slightly more than that, I'll stay at 90 metric tons.

The Boeing proposed SLS third Stage
As for TLI capacity of SLS, things are a little less rosy. The Delta Cryogenic Second Stage used on the Basic SLS is very underpowered for actual BEO exploration. A bigger stage is certainly necessary if we want to get some serious exploration going, sin
ce the basic 25 metric tons is not enough. For this role, I think the scaled up DCSS proposed by Boeing in their lunar exploration architecture is sufficient. It has a wet mass of 45.6 metric tons and an empty mass of 4.7 metric tons, with a specific impulse of 462 seconds. If we assume the delta V required for TLI at 3150 m/s I get 36.13 metric tons to TLI. And this is conservative, since SLS would be able to launch the stack into a somewhat higher orbit because of it's higher lift capacity, and the TLI delta V is usually slightly lower; 3100m/s gave me 37 tons, for example. So let's assume Block 1+ASCS (Advanced SLS Cryogenic Stage, which is how I shall call it from now on) TLI capacity at 37 metric tons.

The mission architecture

The most important part of the mission is, of course, the mission architecture, which is how you will actually execute the mission. The main goal for this mission is to get people on the surface of the moon in a single launch on the basic Block 1 SLS. The primary components will be the crew vehicle and the lunar lander. For the crew vehicle, we will of course assume Orion. The lander will be of my own design. First thing the ship will need to do is enter orbit around the moon. For this task, the service module of Orion should be sufficient. Instead of entering a low lunar orbit, like Apollo did, the vehicle will enter a highly elliptical orbit to save delta V during both orbit injection and return to Earth. This will require a bigger lander, of course, but I will adjust for that later. Lander mass is 15.1 tons, as I will describe later, and I'll assume a 500 m/s injection burn and return burn.

For Orion, finding hard numbers is kind of difficult. Wikipedia has them but doesn't give a source, NASA has a fact sheet on their site but those numbers don't really match up with a realistic specific impulse for the engine. It claims a 22.78 ton full mass on orbit, 1500 m/s delta V, and 7.91 tons of propellant. If all of these values were correct the Orion main engine would need an isp of 358.5 seconds, which is impossible with current hypergolic propellants. A specific impulse of 320 seconds if far more realistic. 1500m/s for Orion is a requirement, so that means either propellant mass or total mass is incorrect. We will assume that total mass is incorrect here, though increasing propellant mass relative to the total should result in similar capabilities in the end. More recent sources indicate a total mass of 21.250 kilograms of 21.25 tons. This gives a specific impulse of 327.5, which is also what NASA claims in a SLS ConOps documents (which I can't cite here due to paywall). Still a little optimistic but certainly possible in the range of hypergolic propellants. Now, assuming a specific impulse of 327.5, a propellant mass of 7.91 tons and total mass of 21.25 tons, we can start doing some math. Total mass of Orion+lander will be 36.35 tons. Using the rocket equation, we get the following:

MassRatio=e^500/(327.5*9.81)= 1.1683923.

If we divide the total system's mass by this we get 31.11 tons, meaning Orion's SM has used a total of 5.24 tons of propellant. This leaves us with Orion having a mass in lunar orbit of 16.01 tons. With an empty mass of 13.34 tons, we get the following delta V left in Orion:

delta V= 327.5*9.81*ln(16.01/13.34) = 586.16 m/s!

This means that even if Orion has to do the LOI burn for both itself and the lander into and elliptical orbit, it still has enough delta V left to return, including a margin of over 86 m/s! That is definitely enough to enter an elliptical lunar orbit and be able to return afterward.

Next, the lunar lander. Like Apollo, it is composed of two main parts, and ascent module and a descent module. In order to increase the delta V of the system but keep the mass close to 15 tons, I used a staged approach for the Ascent Module and more efficient cryogenic fuels for the Descent Module.

For the Ascent module, we can't use efficient propellants like hydrogen. These are very volume inefficient, leading to high empty masses and poor mass ratios. Instead, I decided to use storable propellants and a dual stage approach to make up for the poor efficiency. The habitable module for a crew of two, including an engine and empty propellant tanks would have a mass of 2000 kg, and would be capable of carrying 1000 kg of propellant. The engine we used for this would be derived from Aerojet R-4D, which Boeing claims could have their Isp increased to +320 seconds easily.
The empty mass of 2000 kg is a 15% decrease from the Apollo LM, which is very much attainable due to three factors:

  1. The miniaturization of electronics means a computer weighing 100 kg during that time can fit in a pocket now. They also take up less space now, meaning that the cabin can be smaller but just as spacious for the crew.
  2. The stage has to hold less propellant, meaning that the tanks are smaller and therefore lighter
  3. Modern manufacturing technologies and lighter materials like plastics and carbon composites mean that even a direct design copy of the LM built using newer alloys and composite materials could be a lot lighter
So I'd say 15% is very attainable if not a little pessimistic. One of the Golden Spike designs for a lunar lander is basically a plastic ball, and it's pressurized and can hold two people for several hours with a mass of roughly 1500 kg, also including empty tanks and engines, so 2000 kg should be enough. It might even be capable to cram four people in there and give them a surface stay of a few hours, which is enough for visiting a base on the surface.

The STAR 48 booster
The AM also includes a booster stage. For this, I assumed the STAR 48 solid rocket motor. This rocket motor might be solid fuel, which can sound scaring to some people, but it's actually a very sensible choice. On smaller systems, solid rockets are very reliable due to few mechanical parts, they have extremely good mass ratios and they are very long storable, unlike cryogenic fuels. The Star 48 is also a flown, proven and affordable system, which makes the lander simpler and safer. It has a fueled mass of 2146 kg and an empty mass of 117 kilograms, with a specific impulse of 291. They are not capable of thrust vectoring control though, so the reaction control thrusters on the AM will have to take care controlling the craft during ascent. 

For the Descent Module, I again took the Apollo LM as an example. The Apollo DM had a wet mass of 10344 kilograms and an empty mass of 2144. Then I decreased the total mass to 10000 and increased the wet mass to 2600 kilograms to make up for the extra insulation of the cryogenic fuel, as well as a light fairing to cover up the Star 48 during landing. This gives a mass ratio of 3.85, which is plausible for a hydrogen powered stage. The engine used is the RL-10 with a specific impulse of 462 seconds. 

Lunar Lander Delta V overview:
Ascent Module, Habitat Module = 1273 m/s 
Ascent Module, Booster = 1431 m/s 
Ascent Module, Total = 2704 m/s
Descent Module Total = 3039 m/s
Lunar Lander Total = 5743 m/s
Lunar Lander Total Mass = 15146 kg

The published figures for the Apollo LM are 2500 m/s for the descent and 2200 m/s for the ascent. Because of our choice of parking orbit, an extra 400m/s is needed to go between the parking orbit and the standard Low Lunar Orbit. This means 2900 m/s and 2600 m/s are required respectively, which gives us 139 m/s of margin during descent and 104 m/s of margin during ascent.

Mission Summary

So, to recap, the mission is divided into the following steps:
  1. SLS Block 1 launches an ASCS, a lander and an Orion into LEO
  2. The ASCS performs the TLI burn after a short on-orbit checkout
  3. Orion separates from the stack, docks with the lander and pulls it away from the ASCS
  4. Orion performs the LOI burn of 500 m/s
  5. The crew transfers to the lander and separates from Orion
  6. Lander performs a 400m/s burn to circularize into a LLO
  7. Lander descends to the surface and lands
  8. Crew Performs mission, either a 1-2 day sortie or a 27 day stay in a pre-landed habitat.
  9. Crew lifts off in the AM
  10. STAR 48 burns out and is jettisoned
  11. AM liquid engine performs burn to LLO
  12. AM engine does a 400 m/s burn to rendezvous with Orion
  13. AM docks with Orion, crew transfers
  14. Lander separates from Orion (could use spare dV to crash itself into the moon)
  15. Orion performs TEI burn of 500 m/s
  16. Orion re-enters the atmosphere and lands in the ocean.

While the use of an elliptical parking orbit limits the possible surface stays either to a few hours or days or to a full lunar orbit of 27 days, it greatly reduces the mission mass, complexity and cost. While this architecture is notional and certainly not anything that could pass NASA without any significant changes, it is an architecture that proves that even the basic Block 1 SLS has the capability to do lunar missions in the same style as Apollo. I will later write an article describing a possible cargo lander, to allow this architecture to be used to set up a whole lunar base.

Some sources: