Performance Monoprop
Nov 10th, 2008 by johnhare
guest blogger john hare
In the other posts, I have tried somewhat to conform to accepted practice. In this one I suggest that maybe it is time to try something really new. I’m going with the monoprop again because it is the simplest configuration I could find to make the concept reasonably coherent.
One problem with rocket pumps and rocket technology in general is that it is so strongly bound by tradition. When so much destructive power and money is at stake, it really pays to be conservative. Even then, a lot of hardware and money is lost very year because things don’t go right for one reason or another. Sometimes the most conservative thing to do is take another look at the way things are done.
Running rockets at higher pressure is well known to improve performance. It also makes for a smaller engine envelope and smaller lighter tanks ending with a vehicle with less total destructive power because it is smaller. I believe it is possible to integrate the turbine, impeller, and injection system into a single unit that is lighter than any one of them in a conventional engine. In the process, we can get an engine that operates at very high pressures.
It starts with why turbines and pump impellers have speed limits. Turbines are subject to centrifugal and thermal stress while the actual work is done by converting the torque force applied to the blades through the hub to the shaft and then to the impeller. The impellers have centrifugal stress and the torque force of pushing the propellants. There are two main wheels with torque stress applied in cantilever form in opposite directions. The turbine pushing one way by hot gasses and the impeller resisting the other way with inert fluid. All of this power must be transmitted though a shaft and sometimes a gearbox.
Turbines seem to have a tip speed limit of around 2,000 feet per second, though my references may not include faster ones. LOX impellers have a tip limit of about 900 feet per second with methane toward 1,500 and hydrogen around 2,000. Hydrogen is so fluffy that several impellers are staged to get the really high pressures. LOX has a velocity head of about 12,600 feet at 900 f/s while LH2 has a velocity head of over 62,000 feet at 2,000 fps. At those head heights, LOX has a pressure of well over 6,000 psi while LH2 is under 2,000 psi. The density of the heavier propellant stresses the impeller more, but realizes much more pressure at much less speed. If the LOX could be pumped through an impeller at 2,000 feet per second tip speed, it would realize a pressure in the 30,000 psi range.
Several of us have discussed running some of the propellants through the turbine blades to cool them so that higher turbine inlet temperatures could be used. That would get more turbine power with less pressure drop so that main combustion chamber pressure would be higher for the same pump pressure. What happens if all of the propellant is run through the turbine blades? The turbine is now the impeller.The opposing torques cancel out across the same metal wall to eliminate torque applied to the hubs of either. Power transmission though a drive shaft is eliminated. The remaining stresses are centrifugal and thermal. Thermal stresses are much reduced because the entire propellant flow is regeneratively cooling the turbine blades. Centrifugal stress alone remains the same problem. 2,000 feet per second tip speed is attained in some turbomachines that also deal with the problems we just eliminated. The machinery can now physically tolerate pumping hydrogen peroxide to pressures well over 30,000 psi. The new limits are efficiency and tolerance of the other components of the rocket system.
By moving the turbine/impeller from the middle of the sphere to near the bottom just above the throat, fluid exiting the turbine/impeller tips will flow up the chamber walls in a very fast spiral. The chamber now serves as the volute with the sphere wall on one side and the high pressure gas of the chamber as the other. Aerospike volute if you will. While it will take some serious work to establish how much of the potential pressure is achieved, 50% efficient pressure recovery gives 15,000 psi in the upper combustion chamber. While there is probably some pressure limit short of this, the turbomachinery isn’t it.
With the peroxide monopropellant here, a method of fast decomposition is needed. A liquid catalyst can be injected in the chamber just above the turbine/impeller to mix with the peroxide and be glued to the chamber walls by the same centrifugal force as the liquid peroxide. The chamber walls need to be catalyst coated also. One thing that concerns me is reaction rate of peroxide decomposition. While chemical reaction rates rise with pressure in a biprop to keep L* similar, I don’t know how it works in a catalyst reacted monoprop.
The liquid will stay against the chamber walls until it converges near the top when impinging remaining liquid peroxide sheets will be forced down by the splash plate in the impingement area. Unused catalyst and liquid peroxide will try to spin back onto the side walls in vortex seperation while the gas products will move down the chamber and out through the turbine. The entire flow is a strong vortex that makes it possible to eliminate the turbine nozzles required by more vertical flows. If it requires a 50% pressure drop to drive the turbine, then a main chamber pressure of 7,500 psi remains.
A 2 inch diameter throat will give on the order of 30k thrust. Over 200 lb/sec of peroxide will require an inlet of about 3 inches diameter to keep liquid flow rates somewhat reasonable. The 12 inch diameter sphere is large enough to handle all the gear and plumbing and still have an L* of well over 200. At nominal pressures, the carbon composite shell is only stressed to 90ksi, which leaves considerable margin. The sphere structure would be a bit under 40 pounds with about 15 more for pump and plumbing. Expansion nozzle is an unknown mass.
On paper, this engine should have a T/W of over 300, with an Isp as high as monoprop peroxide is ever likely to get. The assumptions of acceptable material stress, propellant flow velocity, and required L* would shift the T/W the distance of the errors in my estimates. An acceptable L* of 50 would quadruple the T/W, and allowable shell stress of 180ksi would double it again. Or it could go the other way to give a useless engine. I don’t think it safe to assume T/W of 300 without serious experimental and theoretical confirmation.
To work out the mechanics of the configuration in hardware, a smaller engine operating at lower pressure would be useful. A good prototype static test engine target would be a 250 lb thrust engine at 1,000 psi. A half inch diameter throat in a five inch sphere would be a good start. The turbine/impeller/injector could be a two blade unit very close in size and shape to that on a model airplane. The shaping would be that of a wind turbine as opposed to propeller with hollow hub and blades with open ends. Mass would be a minimum gauge issue of perhaps a pound. I believe cost would be below that of pressure fed with the elimination of exterior pressurization.
This is a very quick sketch of the biprop concept that really interests me. On paper, 
it should be able to reach over 10,000 psi at the throat. In reality, no way. It is not a question of if I got it right, but rather where did I go wrong.
I still have my reservations about placing moving parts inside of the thrust chamber. It seems to me that the fluid flow passing through the turbine blade will be unsteady, turbulent, and probably still chemically reacting and/or combusting. Thus, the exhaust products will be imparting some rather violent forces on the turbine blades. If that doesn’t tear them apart, then you also have to consider that the spinning blades will acoustically couple back to the fluid which will most likely effect the rate of combustion within the chamber. There may even be some non-trivial interactions with the fuel being delivered through the blades.
I’m not saying that it can’t be done, but I can just imagine that it would be a nightmare to properly design the system to be both robust and efficient. Considering that we have only just recently arrived at the point where we can even begin to accurately model the flow environment inside of a simple thrust chamber, I think it might be sometime before the tools are available to properly design a rocket engine with this much complexity.
Something about the liquid catalyst is bothering me, can’t quite put my finger on it. How does the biprop version work?
Eric,
You may well be right that this is just too complicated a thing to model. I think modeling would have from the jet engine side of the house. The turbines have only the nozzles between them and the combustion chamber.
If it can be done, it would be a serious enhancement to rocket engine performance. I think early work on this will have to be small prototypes to establish an idea of possibilities.
Jsuros,
The liquid catalyst would have to be pressurized to well over 15,000 psi. That might be it.
The simple biprop version is a variant on the earlier aerospike we discussed. The lower end would need to be derated to a point that it is possible to reach with a kerosene impeller. Perhaps a 5,000 psi combustion chamber in that one.
The one that really gets my attention is more complicated and doesn’t address the issues Eric raised. I’m working on it.
Can you clear one thing up John? You say in the post that the swirl on the flow means you don’t need turbine nozzles (and there aren’t any nozzles on the drawing), but in the reply to Eric’s comment you say that the turbines have only the nozzles between them and the combustion chamber.
Tim,
I didn’t say it right. Jet engines have only the nozzles between the burner and turbine. This one uses the swirl for some of the angle instead of nozzles. I see the turbine blading resembling wind turbines more than jet or rocket blades. Under some conditions wind turbines have tip speeds a small multiple of wind speed with no swirl in the upstream air. With the entire flow driving the blades at maximum temperature, less preparation of the incoming flow should be required.
This sounds similar to the full-flow preburners/turbopumps of the Integrated Powerhead Demonstrator.
With the entire flow driving the blades at maximum temperature, less preparation of the incoming flow should be required.
Actually, I disagree. I think the swirling flow is exactly what’s going to tear apart your turbines in the end. Consider that in both jet engines and turbo-pumps, the spinning turbines are not located in the combustion chamber. The energy released during combustion must be redirected into linear momentum by nozzles or vanes of some sort. By the time the fluid reaches the turbine blades, the flow has most of its momentum directed along the axis of the turbine. At the very least, I think you are going to get very poor performance out of your turbine and your rocket. At worst, I think it will most likely eat itself when spun up to high RPM’s.
At the moment, the only way I can imagine this working is to separate the turbine from the combustion chamber by a nozzle. The resulting stream of supersonic gas will likely be much more efficient at spinning the turbine. The axis of the turbine can still be linked to a centrifugal pump located inside the thrust chamber. You may be able to design the pump to be robust enough to withstand the thrust chamber environment.
This then raises the question of what happens to the fluid downstream of the turbine? Will it be able to supply sufficient thrust? Would you want to have a secondary thrust/combustion chamber operating at a slightly reduced pressure and fed by the same kind of pump as in the first chamber?
I’m not qualified to judge the merits of your concept. I do, however, like the idea of developing more efficient, higher-performance engines, and I do understand how higher chamber pressures are a key.
Jim,
The idea is similar, execution seems a bit different. The articles I checked out suggest that my emphasis on very hot operation is misplaced with full flow. The machinery in my concept should be so relatively light that there is margin there to correct the problems pointed out here.
Eric,
This is a concept in progress of sorts. You and the others have given good criticism and ideas. I think that the turbopump core is valid, while my layout was not. I am going to think through a means of getting performance possibilities of the pump with a layout that has less baggage.
One thing that comes to mind is splitting the flow from the tips to get just enough fuel in the upper chamber to drive the pumps while the rest cools the engine before reaching the lower chamber injection system. 1,000 psi in a very light chamber would still be quite valuable now.
The Integrated Powerhead Demonstrator mentioned by the earlier poster sounded familiar to me. Out of curiosity, I looked it up. I guess I had heard of this project from a couple of years ago, but I don’t know that I ever knew the details. It uses a full-flow staged combustion cycle (FFSCC) where all of the fuel and oxidizer pass through two separate turbines. A small amount of oxidizer and fuel are exchanged to enable enough combustion to power each pump. The remaining fuel, oxidizer and resulting combustion products are then fed into the main thrust chamber. The main benefit of the FFSCC is that the turbines are kept cooler due to the greater amount of mass being passed through relative to the heat generated by the combustion.
It turns out that the concept I was trying to describe in my last post is alot closer to the staged combustion cycle (SCC), where the fuel and oxidizer are pumped directly into the main combustion chamber along with the combustion products from the turbo pump. The main advantage is that it provides higher chamber pressures and incurs less losses due to dumping the turbo-pump exhaust elsewhere. However, one of the main drawbacks is that it is still a very harsh environment for the turbines.
I doubt that they use the serial layout that I was considering, but at least we know that this approach has enough merit that people are already looking into it. The Wiki-pedia article on staged combustion cycles has a nice diagrams of both of these rocket cycles.
I’ll state the obvious: advances in material science will play a mjor role in dealing with the stresses on moving parts, like turbine blades.
Oh, and thanks for the links Eric!
I will now go talk to a bunch of gazzilionaires about funding an aggressive program of engine development.
Or not.
Article on the staged combustion cycle powerhead thingie:
http://www.pw.utc.com/vgn-ext-templating/v/index.jsp?vgnextoid=2e35288d1c83c010VgnVCM1000000881000aRCRD&prid=29c256fe452de010VgnVCM100000c45a529f____
Also, at storming media:
http://www.stormingmedia.us/01/0197/A019793.html
I’m trying to track down where this program is now.
A staged combustion cycle seems like a good fit for thrust augmentation; the turbine exhaust could go to the augmenter injector (not sure of the terminology here) rather than the combustion chamber. Since the augmenter is at much lower pressure than the combustion chamber, the turbine can be more efficient.
After checking out the integrated powerhead demonstrator links, I think we might be on to something. With the suggested performance of the IPD with cool turbines, there should be more performance with cooled turbines that allow much hotter turbine inlet temperatures. I called it the turbothroat engine at one time.
Tim,
You may have just described the gas generator engine cycle. You do seem to be on an idea here, by running much more than the normal amount of drive gas to the augmenter injector, and using the dual turbopump in the IPD, there would be gas-gas injection in the augmenter. With a very short L* in the augmenter, you get the effect that Jon mentioned with a much lower required tech level. And main chamber pressures could be extremely high.
As the engine throttles down, gas generator drive requirements drop more rapidly than the total thrust, with the result that the augmenter losses are very low. I’ll have to play with this one, trying to figure out how to eliminate augmenter flow completely on throttle down.
Yeah, somewhere along the line I got it into my head that a staged combustion cycle was just a gas ganerator cycle with the turbine exhausting into the combustion chamber, rather than sending all the fuel through the turbine.
I think that thrust augmentation gives you a lot more options when it comes to engine cycle, something that hasn’t really been mentioned much when people talk about it (at least here). For instance, on an LH2/LOX gas generator cycle engine, it might be possible to substitute a larger turbine and run it LOX rich. The turbine would be cooler, and by exhausting it into the augmenter injector, provide a oxidiser for augmentation and able to use a much lower exhaust pressure. Fuel (and possibly some oxidiser, if there’s a limit to how LOX rich you can run the gas generator turbine) could be provided by an expander cycle cooling the lower nozzle. I have to wonder if it is possible to modify an RS-68 to run on a cycle like this. I should also point out that this is the first vaguely workable setup I have been able to think of, and there are probably more (and better) options.
John pointed me to last year’s thrust augmented engine:
http://selenianboondocks.com/2007/11/random-thought-thrust-augmented-aj26-60/
and the thrust augmented nozzle:
http://selenianboondocks.com/2007/11/thrust-augmented-nozzles/
The nozzle one — essentially a rocket “afterburner” — is a special favorite of mine because the idea popped into my head recently and I tried to articulate (not well) in a comment on another post. John rescued me by pointing to the link above.
I bring them up as refreshers for the discussion (assuming this thread is still live) because I like the idea of wedding the powerhead concept to the augmented nozzle idea. It would seem feasible (for one thing, they’re both Aerojet projects). I think the thrust augmented nozzle is a simple way to introduce some of the benefits of tri-propellant effciency, as well as providing an inadvertent “dual expansion” function. Also, it would, in theory, preclude the need for an additional outside boost (from strap-ons, for instance).
The “dual expansion” notion comes from the fact — if I read the piece right — that the engine nozzle would have to be overexpanded to accommodate fuel injection and “afterburner” combustion, so that when the augmentation phase is shut down, the launch vehicle is at a point where the expanded nozzle is a plus.
Now I did note that Aerojet did the TAN tests with low-pressure engines, and cited the concept as well-suited for such engines.
http://selenianboondocks.com/2007/11/thrust-augmented-nozzles/ (point #3 in Jon’s post)
That begs the question: can there be even greater benefits with a high-pressure engine, or are their drawbacks to the idea in such a scenario?
Hmmmm . . . .
The consensus in the comments from last year imply that it’s (the TAN) best suited for lower pressure engines, if only because that is an advantage of the concept.
So, would marrying the staged combustion cycle powerhead and the TAN be overkill?
Roderick,
No, it benefits staged combustion engines as well. It’s just my bias is to take the benefits and use them to avoid having to do staged combustion. I don’t like having to deal with that high of pressure in the plumbing.
~Jon
The high pressure staged combustion system is one of the places I agree while disagreeing with Jon, or maybe vice verse. The benefits and problems of staged combustion are both undeniable. I believe it is possible to minimize the high pressure plumbing problems with a layout that eliminates the high pressure plumbing.
We just had several construction contracts close at once, so my writing will be minimal for a while. 70 hour weeks do that, but they also generate the solid rocket fuel at the base of all our decisions. Enough of the square green rocket fuel could convince even MSS to experiment with crazy ideas.
When time again permits, I hope to start closing some of the scattered concepts into a system and throw it out there to the wolves.
Thanks for the feedback, gentlemen!
I appreciate the fact that exisiting high-pressure engines are complex, and that their are operational issues with complexity. I do think that combining the TAN with a high pressure engine — especially a cryogenic one, with RP or other hydrocarbon for the TAN phase — would be the kind of game changer that would permit all-rocket SSTO.