Reusable Rockets

[Dr.Strangelove]

Kind of a lull between SpaceX launches, so let’s talk about something a little different.

SpaceX’s competitors have seen the writing on the wall–that cheaper rockets through reusability are coming sooner or later. There won’t be an order of magnitude price drop immediately, but if SpaceX can cut their already low prices by 10 or 20%, they’ll have big problems.

So both ULA and ArianeSpace have announced their own programs. ULA is an American company that does US government launches; Ariane is a European group that does commercial launches.

ULA’s system is called Vulcan:

ULA unveils its future with the Vulcan rocket family – Spaceflight Now

And a cartoon of the planned system.

Basically, they plan on saving the engines only. The engine module detaches after first-stage burnout, inflates a “hypercone” (also called a ballute [balloon+parachute]), and then finally an ordinary parachute. A helicopter in the area catches the descending module in flight and lands it on a nearby ship. They bring it home and the refurbishment and integration process starts.

Ariane’s strategy is not too far off. Called Adeline, it is also a detachable engine module:

Airbus thinks it can trump SpaceX's reusable rockets | Engadget

However, instead of parachutes, it has wings, landing gear, and propeller engines. It flies to the nearest airfield and lands like a normal plane.

As much as I like to see competition, both of these techniques seem disastrous to me. The first thing is that they limit the possible upside of reusability. Yes, the engines are the most expensive part, but at the same time the engines are part of an integrated system, and to split that system ends up throwing away a lot of that value.

I think there are also serious practical issues. Looking at Adeline specifically, they show a nice shiny nosecone on the craft. Where are the holes for fuel, oxidizer, hydraulic fluid, helium lines, power, signalling, and so on? These things (in particular the fuel lines) are not something that you can easily pull apart and then close a little hatch. They will have some very serious engineering difficulties making it all work.

ULA will have a slightly easier problem in that the interface will not have to double as an aerodynamic nosecone, but just the same it’s a hard problem.

Admittedly, the Shuttle did solve some of these issues already. The external tank piped propellant to the orbiter and could cleanly separate. However, we know that Shuttle was very expensive to refurbish, so maybe that’s not the best thing to target.

At any rate, I am excited to see these ideas come to fruition, but just the same I think the SpaceX reusability model is vastly superior. Their refurbishment costs will not be negligible, but the fact that the stage comes down in one whole piece will be very beneficial, I think.

[AK84]

Why would they reduce costs anymore then the “reusable” SRB from the Shuttles ended up doing.

SpaceX’s technical demonstrations may succeed…but why do you think their attempts at reuseable would reduce costs would not hit realworld issues as the STS did?

I admit, part of my skepticism is that Musk is clearly a snakeoil salesman, but even the otherwise excellent SpaceX tech crew really has not delivered on promises made.

[Ale]

The Adeline seems rather sketchy.
Personally, and pulling things out of my rear end, if I’d go with an approach like those (reusable engines, disposable tank) I’d go with at least two engine pods attached to a main tank; not only it would make it easier to design them for reentry, it would allow modular expansion of the system.

[Ravenman]

There’s a good part of me that wonders how many customers are going to sign up their precious multi-hundred million dollar payloads onto used rockets.

If I get a new rocket straight off the factory floor, do I pay a premium? If I get engines that have been used five times, how much of a discount is that?

[AK84]

Have they given any estimate as to time and costs for refurbishment of “used” stages? What about predicted reliability after multiple use (this has insurance issues, premiums could be much much higher).

There is no point in reducing costs, if said costs will simply reappear under different and newer heads.

[Ale]

@Ravenman

There’s a good part of me that wonders how many customers are going to sign up their precious multi-hundred million dollar payloads onto used rockets.

If I get a new rocket straight off the factory floor, do I pay a premium? If I get engines that have been used five times, how much of a discount is that?

Actually that is not necessarily a point against reusable systems, there’s something in engineering called the Bathtub Curve.
Things are more likely to fail either when they are new or approaching the end of their lives, and more reliable in between; so let’s say an engine that has proven to work with no issues before can be assumed to be more reliable than a brand new, untested one.

[Ravenman]

@Ale

Actually that is not necessarily a point against reusable systems, there’s something in engineering called the Bathtub Curve.
Things are more likely to fail either when they are new or approaching the end of their lives, and more reliable in between; so let’s say an engine that has proven to work with no issues before can be assumed to be more reliable than a brand new, untested one.

I’ve heard of infant mortality, but I didn’t think about it in this context. Good point.

[Dr.Strangelove]

@Ravenman

I’ve heard of infant mortality, but I didn’t think about it in this context. Good point.

One way of looking at it: would you get into a commercial aircraft that had never flown before?

It’ll be some time before reusability programs get to this point, but I think they’ll get there eventually.

[rbroome]

@Dr.Strangelove

Kind of a lull between SpaceX launches, so let’s talk about something a little different.

SpaceX’s competitors have seen the writing on the wall–that cheaper rockets through reusability are coming sooner or later. There won’t be an order of magnitude price drop immediately, but if SpaceX can cut their already low prices by 10 or 20%, they’ll have big problems.<snip>
Basically, they plan on saving the engines only. The engine module detaches after first-stage burnout, inflates a “hypercone” (also called a ballute [balloon+parachute]), and then finally an ordinary parachute. A helicopter in the area catches the descending module in flight and lands it on a nearby ship. They bring it home and the refurbishment and integration process starts.

Ariane’s strategy is not too far off. Called Adeline, it is also a detachable engine module:
Airbus thinks it can trump SpaceX's reusable rockets

However, instead of parachutes, it has wings, landing gear, and propeller engines. It flies to the nearest airfield and lands like a normal plane.

As much as I like to see competition, both of these techniques seem disastrous to me. The first thing is that they limit the possible upside of reusability. Yes, the engines are the most expensive part, but at the same time the engines are part of an integrated system, and to split that system ends up throwing away a lot of that value.

I think there are also serious practical issues. Looking at Adeline specifically, they show a nice shiny nosecone on the craft. Where are the holes for fuel, oxidizer, hydraulic fluid, helium lines, power, signalling, and so on? These things (in particular the fuel lines) are not something that you can easily pull apart and then close a little hatch. They will have some very serious engineering difficulties making it all work.

ULA will have a slightly easier problem in that the interface will not have to double as an aerodynamic nosecone, but just the same it’s a hard problem.

Admittedly, the Shuttle did solve some of these issues already. The external tank piped propellant to the orbiter and could cleanly separate. However, we know that Shuttle was very expensive to refurbish, so maybe that’s not the best thing to target.

At any rate, I am excited to see these ideas come to fruition, but just the same I think the SpaceX reusability model is vastly superior. Their refurbishment costs will not be negligible, but the fact that the stage comes down in one whole piece will be very beneficial, I think.

I am not a rocket scientist, but I have talked to a few!
Not about this subject though.

Stranger needs to chime in, though he has made his opinion clear in previous posts.

There are 2 decision-making factors involved here:

1. How much weight will it cost to make them reusable?
2. Will the cost-savings outweigh the loss of payload capacity inherent in adding a reusable capability?

The details of how one might accomplish this task is engineering, hard but doable.

As has been pointed out by other posters, the reliability and turn-around time is a critical factor in realizing the cost-savings. Note that Space-X claims to want to turn around a booster in a day. Why the furry? I suspect it is an important part of his cost benefit. I can’t imagine them accomplishing that. It takes them weeks to get a new booster ready for launch.

A critical factor that makes rocketry different from almost any other engineering effort is that all the equipment is designed to operate at the absolute maximum capability. No one adds extra fuel just in case, or runs the engines at less than maximum temperature. If the metal handle a certain stress, the component is designed to just meet that stress. At least up to now. Space-X is explicitly changing long practice when they leave enough fuel in the first stage to bring it down safely. So the question is will future rockets keep this bleeding edge design philosophy? Re-usability will require a fundamental change in how rockets are designed.

Elon Musk is shaking up the industry by essentially overbuilding boosters so that there is excess capacity. In the past any “excess” booster capacity is quickly used by the payload.
Will users will be willing to pay, in lost weight in the payload, for excess booster capacity? Elon Musk thinks so. Time will tell.

[Dr.Strangelove]

To be clear, in this thread I’m mostly interested in talking about ULA’s and Ariane’s approaches. We’ve had a few threads on the SpaceX scheme in the past and I don’t think there’s much more to be said until they actually land one.

I’m definitely interested in Stranger’s opinion. As I recall, though, in past threads some people raised the possibility of a detachable propulsion module, and Stranger also dismissed the idea. I don’t want to put words in his mouth, though, so maybe he’ll chime in.

It seems to me that ULA and Ariane are taking a very old-school attitude toward reusability. Specifically, they target maximum performance at any expense. A detachable module has a smaller payload penalty than a fully-reusable stage, so they choose to go that route. As you say, the SpaceX approach is to simply build a larger booster. It doesn’t cost 30% more to launch a rocket that’s 30% bigger. The fuel is free and the materials are cheap. Much of the overhead simply doesn’t scale at all with rocket size.

[Dr.Strangelove]

Ahh, I found the thread.

In it, Ale asked:

[QUOTE=Ale]
What I had in mind was the lower portion of the first stage (engines, avionics, pumps, etc… the pricey bits) to land on… land, not on water. That way, by separating it from the main tanks it should be easier to steer it to a landing area, deploy parachutes at a relatively low altitude (to keep it from drifting too much) and then touch down on solid ground.
[/QUOTE]

And Stranger On A Train replied:

@Stranger On A Train

It appears that you think the “engines, avionics, pumps, etc… the pricey bits” are kind of modular parts that just plug into the rest of the stage. But actually, on a conventional rocket system the stage is built around these and they’re integral to the structure of the stage. In the case of the Falcon 9v1.1 (which despite the incremental “revision number” is actually an almost completely new vehicle compared to the original Falcon 9) the engines are assembled into a structure called the Octoweb, which also contains the engine controllers and thrust vectoring system, all of the prevalve and propellant plumbing, et cetera, as well as providing the main thrust support for the stage structure, while avionics for guidance and navigation as well as telemetry are all forward. It isn’t as if they could just cut the feed lines and fly away with the Octoweb assembly any more than you could just push a button and have your engine drive away from your car. In order to make a separable propulsion assembly you’d basically have to build it as another stage with all of the additional inert mass and complexity that requires, plus a crossfeed system to transfer propellants across the separation plane. And don’t dismiss the cost of the tanks and stage structure. These are not simple tanks like the propane tank in your back yard (if you have one). It is a high strength structure optimized for the lightest possible tank, containing a number of complex components such as valves and health management sensors. Inside the tanks are smaller composite overwrapped pressure vessels (COPV) which store helium and nitrogen under high pressure. The propellant tanks themselves are welded together using stir friction welding, which provides a very strong joint but is a complex and highly tolerance sensitive process requiring dedicated tooling which limits the rate of production.

Aerodynamic decelerators (parachute/parawing/ballute) have been the traditional means for controlled recovery because they don’t require any additional consumables or high rate active guidance; for instance, the defunct Rocketplane Kistler K-1 developed and tested parachute recovery systems for their two stage fully reusable vehicle. However, they’re complex systems with a lot of flexible elements (which engineers hate because of the unpredictability); they’re limited in the size and mass of object that can be brought down at a specified rate; and the structure that they are being used to recover has to be designed to withstand the impact load. Vehicles which use other systems have to carry propellant and some form of propulsion, or supporting a glide landing profile. The Roton used a deployable rotor which would decelerate by autorotating (and was never really proven out). The STS Orbiter Vehicle (“Shuttle”) had a large compound delta wing allowing it to glide into a landing which was very costly in terms of the impact on useable payload mass. The DC-X and Blue Origin New Goddard use a powered landing profile which requires carrying extra propellant.

So it appears he mostly shares my opinion. Actually, I suspect my opinion is based on internalizing his past statements.

At any rate, the idea of a detachable module is not impossible, but I think the view that it is basically a stage in and of itself is the right one. Despite appearances, it is not a simple thing.

[Precambrianmollusc]

@Ale

Actually that is not necessarily a point against reusable systems, there’s something in engineering called the Bathtub Curve.
Things are more likely to fail either when they are new or approaching the end of their lives, and more reliable in between; so let’s say an engine that has proven to work with no issues before can be assumed to be more reliable than a brand new, untested one.

True , but how far does one launch drive the engine down the aging range?
Shock, vibration and heat are good ways to age a component and it could be that the bath tub curve flat area is just one launch cycle. There is going to have to be a lot of work to prove that the aging failure ramp up has been pushed multiple launches out.

[Dr.Strangelove]

@Ale

it would allow modular expansion of the system.

Except for solid booster strap-ons, rockets don’t really go for modular. The Falcon Heavy may look like 3 Falcon 9s, but it’s really an entirely new rocket (with the same engines). You pretty much have to design these things in from the beginning.

Side pods does seem to help a tiny bit with the issue of clean separation, but they’re restricted by their single main engine (the Vulcain 2). They don’t have two smaller engines that they could put in there.

This is definitely one area where the “lots of small engines” approach pays off (in other ways it’s a negative). They could have had more design freedom with smaller engines, but they’re restricted by only having one that must be along the main axis. Their detachable propulsion module pretty much has to look the way it does.

[Ale]

@Dr.Strangelove

Except for solid booster strap-ons, rockets don’t really go for modular. The Falcon Heavy may look like 3 Falcon 9s, but it’s really an entirely new rocket (with the same engines). You pretty much have to design these things in from the beginning.

Side pods does seem to help a tiny bit with the issue of clean separation, but they’re restricted by their single main engine (the Vulcain 2). They don’t have two smaller engines that they could put in there.

This is definitely one area where the “lots of small engines” approach pays off (in other ways it’s a negative). They could have had more design freedom with smaller engines, but they’re restricted by only having one that must be along the main axis. Their detachable propulsion module pretty much has to look the way it does.

I was picturing something like the strap-on boosters on Soyuz rockets, but without a central sustainer engine.

Actually, if the engines and tank separate in flight (just brainstorming here), it would be possible to also recover the tank via a relatively simple splashdown with parachutes because it would be less prone to damage from immersion than the engines would be.

[Gorsnak]

@rbroome

Will users will be willing to pay, in lost weight in the payload, for excess booster capacity? Elon Musk thinks so. Time will tell.

Users don’t face that choice. Users pay SpaceX (or ESA or whoever) to place object X in orbit Y. Why would they care whether SpaceX uses an oversized rocket with a recoverable first stage rather than a just-barely-big-enough rocket, so long as their satellite ends up in the desired orbit? What Elon Musk thinks is that he can offer the service of placing object X in orbit Y for less money if he uses the excess capacity of his first stage to recover it rather than to increase the rocket payload. The customer doesn’t need to even have an opinion on that question.

[billfish678]

@Dr.Strangelove

It seems to me that ULA and Ariane are taking a very old-school attitude toward reusability. Specifically, they target maximum performance at any expense. A detachable module has a smaller payload penalty than a fully-reusable stage, so they choose to go that route. As you say, the SpaceX approach is to simply build a larger booster. It doesn’t cost 30% more to launch a rocket that’s 30% bigger. The fuel is free and the materials are cheap. Much of the overhead simply doesn’t scale at all with rocket size.

Which brings up another approach known as “Big Dumb Boosters”. Where the performance/tolerances are brought down significantly. You just make the rocket big(er) enough to compensate. As far as I know the approach has not ever really made it past the paper studies stage to any degree.

[psychobunny]

I only opened this thread because of the title/username combo. Needless to say, I was disappointed.

[Dr.Strangelove]

@psychobunny

I only opened this thread because of the title/username combo. Needless to say, I was disappointed.

It’s quite a challenge to reuse a nuclear-tipped ICBM.

[rbroome]

@Gorsnak

Users don’t face that choice. Users pay SpaceX (or ESA or whoever) to place object X in orbit Y. Why would they care whether SpaceX uses an oversized rocket with a recoverable first stage rather than a just-barely-big-enough rocket, so long as their satellite ends up in the desired orbit? What Elon Musk thinks is that he can offer the service of placing object X in orbit Y for less money if he uses the excess capacity of his first stage to recover it rather than to increase the rocket payload. The customer doesn’t need to even have an opinion on that question.

I understand your point, but I am not certain I agree with it. Putting a payload on a rocket is more than just hoisting it into the cargo bay. Getting a payload into orbit is more than just getting a big enough rocket. Fairings, vibration, orbital parameters and many factors I don’t understand drive the decision on what rocket will work. Rockets can be too big as well as too small for a given payload and orbit. And my original point remains-if a given booster can lift more payload to a given orbit, the customer will probably want to add fuel or additional sensors or power supplies. If any launch company wants to compete, they will need to offer as many kgs to orbit for the lowest price. Giving up kgs to save money is a trade-off, SpaceX thinks their trade-off is going to work. Legacy companies think it is better to offer more kgs for the price. Whether the benefits of saving just the pricey bits is a better approach-I suspect it will depend on turn-around. SpaceX wants the fastest turnaround and doesn’t have time in their schedule to hook the engines back up to new fuel tanks. I am not sure what the hurry is. But I suspect it is a significant part of their cost analysis.

[Dr.Strangelove]

@billfish678

Which brings up another approach known as “Big Dumb Boosters”. Where the performance/tolerances are brought down significantly. You just make the rocket big(er) enough to compensate. As far as I know the approach has not ever really made it past the paper studies stage to any degree.

The Sea Dragon is a great example of that. The problem is that you need giant payloads. It would be a sensible approach if we were building a lunar colony or big Mars mission, but it’s way too big for satellites and the like.

It’s the same story as so many other otherwise promising approaches to launch–you can get the $/kg figure to be low, but it requires too much capital expense and/or too high a launch rate to justify the technology. Various ground-based launchers (electromagnetic guns, etc.) fall into the same trap.

But it’s also why I’m pretty optimistic about these new reusability approaches (SpaceX’s in particular). They are fairly minor tweaks on existing systems, and allow incremental testing and development. You don’t have to spend $10B before the first rocket even flies. If the technology fails in some way, you don’t lose the payload, just the stage. It’s a much more risk-tolerant strategy.