Comments on: ULA Propellant Depot Paper (Updated)

By: Andrew Swallow

Andrew Swallow — Mon, 16 Mar 2009 00:52:50 +0000

What are you using for station keeping and how much acceleration does it produce?

If the depot works better when rotated does the rocket’s tank also work better when rotated? Joint rotation is possible.

By: Jonathan Goff

Jonathan Goff — Thu, 13 Nov 2008 19:57:34 +0000

I just pulled the 30 minutes number out of a hat, but using ullage pressure is definitely the way to go compared to a pump. Hold on, lets run some numbers…

So, 4 tonnes of LOX per minute gives you 67kg/s. I’m not sure what size of ducts they use for transfer, but say we’re talking a 6″ duct (pretty reasonable for something that big, if not on the small side). That gives a cross sectional area of about .018 m^2. 67kg/s of LOX is about 59L/s of flow…or about 3.3m/s.

Of course, I really don’t have any good way of knowing what sort of flow speed you want to keep things to to avoid geysering or other issues. But I could see those kind of flows being driven by ullage pressure.

Either way though, even if you do it over 2 hours instead of 30 minutes, you’re still only talking something like 650-700lb of settling propellant for 240,000lb transferred. Even if you let it take 8 hours, you’re still at barely 1% of the total transferred propellant load. And that’s if you have to leave the RCS running constantly. At 8 hours, you’re talking flow rates of 4kg/s. That’s about the LOX flow rate on Pixel’s engines.

I just really don’t see that as being a big problem.

~Jon

By: Will McLean

Will McLean — Thu, 13 Nov 2008 17:45:38 +0000

Jon:
I don’t think it’s as simple as you say in 9 and 11. A 30 minute transfer of, say 120 tonnes of propellant is 4 tonnes a minute. That’s roughly a firehose rate of propellant flow, and I don’t see you getting that just from ullage pressure. On the other hand, if you move the propellant the way Centaur actually does it, with a pump, that’s a lot of vibration, and you need to power the pump. Maybe you can keep the tank settled at low acceleration levels with the pump running, but that has yet to be demonstrated. Likewise with propellant going into the EDS tank at 4 tonnes a minute, is it all going to stay neatly where you want it?

By: Jonathan Goff

Jonathan Goff — Thu, 13 Nov 2008 06:59:25 +0000

Comga,
Pretty simple to estimate, really. Say the depot plus EDS stack is about 400,000lb at the start of the transfer (rough guess without looking back on the numbers). If you want 10 microgees, that equals about 4lbf. If you’re using cold gas that you’ve allowed to warm to near ambient, say you can get an Isp of about 40s (I think that’s reasonable for GOX). That would imply 1lb of GOX for every 10s. If it takes a full 30 minutes to do the transfer, that would be 1800 * 4 /40 = 180lb of GOX. Which is really nothing compared to a 400,000lb vehicle. If you needed more like 1E-4 g, you might be better of using a 40lb GOX/GH2 RCS engine. You’d get similar propellant consumption because even though you’re producing a lot more thrust you’re going to get a much higher Isp….

And the reality is you probably don’t need to provide continual thrust. You’re just trying to provide a net acceleration sufficient to keep the sump from getting uncovered. That can be done with pulses of RCS followed by coasting. Of course, if you’re doing a typical settled transfer, the receiving tank is going to be continuously venting anyway, so if you’re clever about things, there’s a good chance you can get most of your settling propellant from the actual transfill boiloff. If I’m understanding this correctly. Frank’s the genius on these things.

~Jon

By: Comga

Comga — Thu, 13 Nov 2008 05:53:32 +0000

Wow. It is great to get such a thorough explanation from Frank. His arguments seem rock solid. In this case, as in many, “better” would indeed be the enemy of “good enough”. Petals make great sense as does his cautious approach.

Will. with his “pseudogravity equals about 0.5 g”, is not considering your comment about using the RCS do do the axial settling. My question would be how much RCS propellant would it take to create 1E-5 to 1E-2 g for the full depot plus the EDS to drain the LOX. How does one estimate how long that would take?

By: Jonathan Goff

Jonathan Goff — Thu, 13 Nov 2008 05:51:21 +0000

Will,
Ah…I think I see what you’re missing. The propulsive settling isn’t done by firing the main engine. It’s done by venting some of the boiloff gases through a propulsive vent. You only need thrust levels in the low high single digit to low double digit lbf range. Even a brief pulse (a couple of seconds) at that microgee level is enough to overcome the net disturbing forces and cause the propellants to shift over to the outlet side. Occasional pulses every certain number of seconds can keep the stuff settled during the transfer. And then you just transfer by venting one tank and using the ullage pressure in the other tank to force the propellants through.

It’s actually pretty simple, and it’s basically the process they use for relighting the RL-10 each time they go to do a relight. IOW, it’s a technique they’ve demonstrated hundreds of times over the years.

~Jon

By: Will McLean

Will McLean — Thu, 13 Nov 2008 03:37:46 +0000

The missing element seems to be the actual propellant transfer. Settling seems to have been demonstrated at relatively low levels of psuedogravity. Once a RL-10 has ignited, propellant transfer occurs, but once the engine has iginited psuedogravity equals about .5 g or more.

Cryogenic propellant transfer at levels below about .5 g seems to be the the process that has still not been demonstrated. Settling the propellants with the pump silent is all well and good, but getting propellants flowing with a pump running at very low g levels is a different question.

Will

By: Jonathan Goff

Jonathan Goff — Wed, 12 Nov 2008 20:22:18 +0000

Will,
I think their plan was actually to despin the tank before docking/propellant transfer and switching to axial propulsive settling. Part of the experiment they’re doing is to see if they can spin the tank up and spin it down. The RCS requirements to spin down, and then to accelerate the fluid enough to settle it is really miniscule if you can afford to give it a decent amount of time.

~Jon

By: Will McLean

Will McLean — Wed, 12 Nov 2008 20:09:33 +0000

I see one difficulty with the concept as outlined. If you rotate a cylindical tank about its long axis to settle the propellant, you effectively don’t have a low point in the tank to collect the propellant, since the cylinder wall is equidistant from the axis. It would be like trying to drain a tank in gravity with a perfectly flat floor, worsened by the low level of psuedogravity involved. Getting the last 10-5% into the pipe may be problematic, and that’s a lot of propellant unavailable when you’re using enough to refill an EDS stage.

I can see a number of solutions, such as using a cluster of tanks around a common axis, or ballasting the tank so the the axis of the tank is slightly offset from the axis of rotation. I would like to see these issues explicitly addressed.

By: Jonathan Goff

Jonathan Goff — Mon, 10 Nov 2008 21:18:11 +0000

Frank,
Thanks for the extra information! Your approach makes a lot of sense. Especially the part about not wanting to cast a pall on the whole technology field by taking too many risks up front.

~Jon