Subject: Space-tech Digest #157 Contents: Misc. booster discussion (7 msgs) throttling for thrust vectoring (7 msgs) ------------------------------------------------------------ Date: Wed, 7 Jul 93 21:48:02 -0400 From: dietz@cs.rochester.edu To: gwh@lurnix.COM Subject: safety factors Cc: space-tech@cs.cmu.edu On safety factor, the following description of one design guideline for the AFRPL/TRW/etc. minimum-cost design booster program may be of interest: The vehicle design must utilize conservative safety factors to simplify fabrication and acceptance testing. ICBM derived expendables typically employ design safety factors of 1.25 to achieve a minimum weight vehicle at the expense of specialized manufacturing methods and quality assurance programs. The use of aircraft design safety factor of 1.5 incurs a higher weight penalty (20%-30%) but greatly reduces costs associated with development testing, production and quality control. Similarly, conservative performance margins must be used initially to avoid expensive weight reduction efforts later in the program. The launch vehicle design must maximize the use of commercially available parts, conventional fabrication techniques, and existing test and launch support facilities. Large savings result from moving production operations out of specialized, dedicated manufacturing facilities and into highly competitive shops. (from "Beyond Percheron: Launch Vehicle Systems From the Private Sector", AAS 83-081.) This article describe very briefly some commercial minimum cost launcher concepts from Phoenix Engineering; I assume these never attracted funding to go into development. The concept would have used HY-140 steel tanks, storable propellants, and liquid-injection thrust vectoring. Three stages, not two. This paper had references to four papers on the earlier low cost concepts, but I don't have any of these. These are from the TRW/etc. "Big Dumb Booster" program. Elverum, G. W. "Scale Up to Keep Mission Costs Down", IAA 24th Int. Congr., Baku, Soviet Union, Oct. 1973. (Describes a large engine with a large thrust/element injector that was fabricated and tested for a few percent of the cost of a conventional engine test program.) Kendrick, J. B. and Dergarabedian, P., "Low Cost Launch Vehicle Study", NASA CR 106662, NASA Headquarters, May 1969. Dergarabedian, P. and Kendrick, J. B. "Comparison of Six Designs of Low Cost Launch Vehicles", American Astronautical Society, 7th Goddard Symposium "Reducing the Cost of Space Transportation", Proceedings, Mar. 1969, Vol. 21. Fritz D. E. and Sackheim, R. L. "Study of a Cost Optimized Liquid Rocket Launch Vehicle," AIAA/SAE/ASME Joint Propulsion Conference, Cleveland, June 1982, AIAA 82-1108. (Describes updated TRW study concluding that for development cost of $715 M [1982 dollars; about $1.3B 1993 dollars] one can get a booster capable of launching a "shuttle payload" [presumably, 65K lbs at that time] for recurring cost of $31.8M/launch [about $840/lb in 1993 dollars].) Wasn't Goldin at TRW? He should be pushing this stuff. Paul ------------------------------ Date: Thu, 8 Jul 93 08:47:15 EDT From: Paul Carr Subject: TRW To: space-tech@cs.cmu.edu > Wasn't Goldin at TRW? He should be pushing this stuff. Goldin was an exec at TRW before taking the NASA job. But remember, TRW are the masters of the lowball. Lowballing is a common practive, but they're especially good at it. ------------------------------ Date: Fri, 9 Jul 93 11:01:54 EDT From: John Roberts Disclaimer: Opinions expressed are those of the sender and do not reflect NIST policy or agreement. To: space-tech@cs.cmu.edu Subject: Re: modulating thrust for attitude control >Date: 09 Jul 1993 09:44:58 -0500 >From: Matthew Feulner >Subject: FWD>RE>on 3 stages for P2.. > >Has anyone thought about modulating the thrust on different engines >to achieve attitude control? In other words, if you're leaning to the >left, increase the thrust on the engines on the left, or decrease on the >right, or both. What would be the inherent problems? From a controls >point of view, you have to take into account the response of the engines >to changes in desired thrust levels (which might need studying). The >engine must allow small thrust changes from nominal (increasing >complexity to allow throttling?). The structure would have to withstand >this type of torque. Are these major obstacles, or are there others? A few other factors to consider: - Time lag between control and response. (In comparison to control surfaces?) - How constant and controllable is the thrust? - Are there any problems with rapidly and frequently changing the thrust? (The SSMEs change thrust only a few times, and in a known and tested sequence.) - With all the changes in thrust for steering, how hard will it be to make sure you end up where you wanted to go? - Unless the engines are spaced far apart, the torque supplied by differential thrust will be very low. At least three engines are needed to control pitch and yaw. If the engines are spaced far apart, then control of relative thrust of the engines has to be very precise, and failure of a single engine will badly unbalance the loading, making survival very unlikely unless there are a large number of engines. - Differential thrust won't control roll (unless engines are directed off-axis). John Roberts roberts@cmr.ncsl.nist.gov ------------------------------ Date: Fri, 9 Jul 93 17:29:55 -0400 From: dietz@cs.rochester.edu To: space-tech@cs.cmu.edu Subject: turbopumps for dumb boosters; TRW study I got a copy of the Dergarabedian/Kendrick paper, "Comparison of Six Designs of Low Cost Launch Vehicles" (Mar. 1969) -- the conference ("Reducing the Cost of Space Transportation", 7th Goddard Conference) was in our library after all. As it turns out, one of the options they looked at *was* a low performance pump fed booster. They used commercial pipeline water pumps in each of three stages; inlet pressures were 125, 100 and 100 psia, respectively. No commercial turbine would be suitable, so they used a low cost two-stage impulse turbine (one turbine type only, used in multiples for lower stages). They found the turbopump fed boost had a considerable reduction in gross weight (6 Mlb vs. 9.3 Mlb for 3 stage, 100 Klb payload), but only a very slight reduction in cost. They judged the increase in complexity and development cost to be not worth the trouble. There are some differences between the reference TRW concept and Bruce's P2: -- 3 stages, lifting 100 klb. -- fuel was UDMH/N2O4. -- HY 140 steel tanks were used; tanks in a stage share a common bulkhead. This steel has a very high fracture toughness and (I think) good elevated temperature behavior. -- Tank pressurization by Main Tank Injection (that is, injection of oxidizer into the fuel tank and vice versa); monopropellant gas generator as a second choice. -- Larger safety factor (for tanks, 1.5 safety factor based on UTS at 600 deg. F); 90% of theoretical Isp is assumed. -- Assumed a fabricated tank cost of $3/lb *in 1969 dollars*. -- Velocity required is 31,700 fps (9.66 km/s). -- Lower chamber pressures: Stage 1: 300 psia epsilon = 6 Isp = 267 sec (vac) Stage 2: 250 psia epsilon = 31 Isp = 300 sec Stage 3: 200 psia epsilon = 50 Isp = 306 sec Even with all that, the payload/gross weight ratio is still > 1%. The paper does not have details of the cost estimates for thrust vectoring or pressurization equipment. A companion paper includes a comparison of the welding of a tank from HY 140 with the welding of an aluminum tank for the S-IC stage. Great reductions in cost are possible due to lack of in-process inspection for the steel tank, as well as much larger tolerances. Flaws show up as (noncatastrophic) leaks during proof test, and can be repaired by in-place hand welding. Paul F. Dietz dietz@cs.rochester.edu ------------------------------ From: henry@zoo.toronto.edu Date: Thu, 8 Jul 93 11:17:20 EDT To: space-tech list Subject: failure mechanisms The report I remembered was From Earth To Orbit, an NRC report from last year. Turns out that it was primarily discussing engine failures, rather than whole-system failures, alas. For both ground and flight firings, LOX/LH2 engines had a benign failure rate of 2% and a catastrophic failure rate of 0%. Other liquid systems had 0.6% benign and 0.2% catastrophic. Solids have not had many failures, but they have all been catastrophic. The data was from a company report whose name I didn't get down -- didn't seem likely to be obtainable. If anyone wants the exact reference, let me know. I would suspect that those numbers exclude the official development period for the engines in question. I have racked my brains a bit further, and still can't think of any tank failures in flight that weren't provoked by something else. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ From: henry@zoo.toronto.edu Date: Thu, 8 Jul 93 17:16:55 EDT To: space-tech list Subject: engine failures, p.s. In regard to those engine failure benign/catastrophic numbers, a friend has pointed out (in another context, actually) that a significant fraction of the liquid-fuel catastrophic failures were the result of "damn the torpedoes" engineering that made no attempt to shut the engine down even when safe operation was about to become impossible -- for example, a turbopump that made no attempt to shut down if the tanks ran dry. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: 09 Jul 1993 09:44:58 -0500 From: Matthew Feulner Subject: FWD>RE>on 3 stages for P2.. To: Space Tech Content-transfer-encoding: 7BIT Mail*Link(r) SMTP FWD>RE>on 3 stages for P2... From: henry@zoo.toronto.edu > > Some kind of guidance might be needed. If the stage operates only > > in the atmosphere, then fins would provide some stabilization, at > > least after the stage is moving... > I really think you're going to need something for stabilization during > the period immediately after liftoff, when fins are not yet effective > and you're in dense and turbulent atmosphere with a heavy vehicle. > Even the V2 did; von Braun didn't put those vanes in the exhaust stream > because he thought they were pretty... Has anyone thought about modulating the thrust on different engines to achieve attitude control? In other words, if you're leaning to the left, increase the thrust on the engines on the left, or decrease on the right, or both. What would be the inherent problems? From a controls point of view, you have to take into account the response of the engines to changes in desired thrust levels (which might need studying). The engine must allow small thrust changes from nominal (increasing complexity to allow throttling?). The structure would have to withstand this type of torque. Are these major obstacles, or are there others? I ask this not purely for discussion purposes as it's being considered for our Project Olympus rocket - which I'll describe sometime in the future. Matt ------------------------------ From: henry@zoo.toronto.edu Date: Fri, 9 Jul 93 11:29:52 EDT Subject: throttling for thrust vectoring To: Space Tech >Has anyone thought about modulating the thrust on different engines >to achieve attitude control? In other words, if you're leaning to the >left, increase the thrust on the engines on the left, or decrease on the >right, or both. What would be the inherent problems? ... It's been considered; I've seen plug-nozzle concepts that must surely plan on it, since they have no *other* visible means of vectoring... As I understand it, there are three significant problems. One is that you need a substantial number of engines, especially if you want an engine-out capability (which pays off *big* in reliability) as well. The flip side of that is that if throttling one or two engines is going to have a major effect -- and you need quite a bit of control authority to cope with the worst-case situations of stratospheric winds and such -- then you need engines that can be throttled pretty deeply, which multiplies engine- development headaches by a substantial factor (not only must the engine run well at a range of settings, but it must handle throttling transients well too). The final problem is that effective control in atmospheric turbulence requires fast response, which is hard to get out of substantial engines; according to Mitch Burnside-Clapp (who's a test pilot), 1Hz response is not enough, you need circa 10Hz for good control, and that's going to be hard to get unless your individual engines are fairly small. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: 09 Jul 1993 12:41:34 -0500 From: Matthew Feulner Subject: FWD>throttling for thrust v To: Space Tech From: henry@zoo.toronto.edu > >Has anyone thought about modulating the thrust on different engines > >to achieve attitude control? In other words, if you're leaning to the > >left, increase the thrust on the engines on the left, or decrease on the > >right, or both. What would be the inherent problems? ... > As I understand it, there are three significant problems. One is that you > need a substantial number of engines, especially if you want an engine-out > capability (which pays off *big* in reliability) as well. The flip side Yep, you need a couple of engines, but aren't you in big trouble if one engine goes out no matter what? > of that is that if throttling one or two engines is going to have a major > effect -- and you need quite a bit of control authority to cope with the > worst-case situations of stratospheric winds and such -- then you need > engines that can be throttled pretty deeply, which multiplies engine- > development headaches by a substantial factor (not only must the engine > run well at a range of settings, but it must handle throttling transients This was my fear. Can someone define "quite a bit", though? An example of maximum load converted to torque on the vehicle? > well too). The final problem is that effective control in atmospheric > turbulence requires fast response, which is hard to get out of substantial > engines; according to Mitch Burnside-Clapp (who's a test pilot), 1Hz > response is not enough, you need circa 10Hz for good control, and that's > going to be hard to get unless your individual engines are fairly small. I don't know how throttling is actually implemented, but everything that happens after injection happens pretty fast (doesn't it?). So what would be the sources of delays and slow transients? Something in the pressurization system? Don't they have to worry about throttling issues for the DC-X? Matt ------------------------------ Date: 09 Jul 1993 13:28:54 -0500 From: Matthew Feulner Subject: FWD>RE>modulating thrust fo To: Space Tech From: John Roberts > A few other factors to consider: > - How constant and controllable is the thrust? > - Are there any problems with rapidly and frequently changing the thrust? > (The SSMEs change thrust only a few times, and in a known and tested > sequence.) I would think the DC-X program would answer some of these questions (at least whether it's possible) since it needs to land, although the time constants might be slower for just the vertical motion of landing. > - With all the changes in thrust for steering, how hard will it be to > make sure you end up where you wanted to go? I had hoped you could increase the thrust of one side while decreasing the thrust of the other for no net vertical effect. I guess this depends on whether you can get higher than nominal thrust, and how much higher. > - Unless the engines are spaced far apart, the torque supplied by > differential thrust will be very low. At least three engines are > needed to control pitch and yaw. If the engines are spaced far apart, > then control of relative thrust of the engines has to be very precise, > and failure of a single engine will badly unbalance the loading, making > survival very unlikely unless there are a large number of engines. I have no feel for what torques need to be appied during ascent. Failure of a single engine would seem to make survival unlikely no matter what. > - Differential thrust won't control roll (unless engines are directed > off-axis). True. Is roll control necessary on the first stage, though? Matt ------------------------------ From: henry@zoo.toronto.edu Date: Fri, 9 Jul 93 13:49:12 EDT Subject: Re: throttling for thrust v To: Space Tech >> need a substantial number of engines, especially if you want an engine-out >> capability (which pays off *big* in reliability) as well... > >Yep, you need a couple of engines, but aren't you in big trouble if one >engine goes out no matter what? No, not if you have a reasonable number of them and good design margins. Most liquid-fuel engine failures are fairly benign, and the only problem is the loss of thrust. It is increasingly popular to design launchers to survive a single benign engine failure. Actually, many multi-engine launchers can survive single failures late in flight already, if their control systems are flexible enough to revise the flight plan to cope; what's new is planning launchers to survive a single failure at any time. To do this, you need either an excess of performance to start with, or the ability to push the remaining engines to higher thrust in a pinch (it may hurt their reliability, but taking a chance on this is better than giving up), plus enough control authority to cancel out the asymmetric thrust (or enough performance margin to kill another engine to balance out the bad one). The Delta Clipper design rule (well, one of them), for example, is to fly a fully successful mission with a single engine failing at liftoff. >> ... you need quite a bit of control authority to cope with the >> worst-case situations of stratospheric winds and such... > >This was my fear. Can someone define "quite a bit", though? An example >of maximum load converted to torque on the vehicle? This depends a lot on details, because your vehicle's stability (or lack thereof) also figures into the problem, and you need to decide how big a disturbance you want your launcher to survive. I haven't heard any simple rules of thumb; has anyone else? >I don't know how throttling is actually implemented, but everything that >happens after injection happens pretty fast (doesn't it?). So what would >be the sources of delays and slow transients? ... After injection things happen fast, but before injection it's not so simple. Your pumps or pressurization system will have a finite response time, due to the inertia of the flowing fuel if nothing else -- the fuel flow won't increase instantaneously when you open the valve wider. The valve itself won't operate instantaneously. And an abrupt change in flow may trigger oscillations that will take time to damp out, especially in a complex system with lots of feedback paths. It all adds up. >... Don't they have to worry about throttling issues for the DC-X? Not for steering; the DC-X engines are gimballed. They need throttling for a vertical landing, but the response times can be a lot longer. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ From: henry@zoo.toronto.edu Date: Fri, 9 Jul 93 14:26:31 EDT Subject: Re: modulating thrust fo To: Space Tech >> - Differential thrust won't control roll... > >True. Is roll control necessary on the first stage, though? Generally speaking, yes. A slow roll is no big deal, if your guidance system is halfway intelligent. (An Atlas survived a 60RPM roll once!) It just needs to ignore the roll when figuring corrections and planned turns, and keep the thrust pointing in the right direction (which may involve rotating the vehicle-relative thrust vector to point the thrust in a constant ground-relative direction). But you must keep the roll rate down to where your guidance system can cope, and you will have stray torques, from engine misalignment if from nothing else. If the rate gets too high for the guidance system, or too high for the thrust-vectoring actuators, it's game over. The torques involved typically aren't particularly high, and a small auxiliary system can cope adequately. The forces, and the required response rate, are much less than for pitch and yaw. It's not uncommon to put a roll-control system on an upper stage and have it do roll control even while lower stages are still burning; as I recall, the Pegasus third-stage thrusters do roll control for the second-stage burn as well as the third, and I think Scout is the same way. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: 12 Jul 1993 09:39:09 -0500 From: Matthew Feulner Subject: FWD>RE>modulating thrust To: Space Tech Date: 7/10/93 12:58 PM From: John Roberts >Keeping total thrust the same while modulating the thrust of individual >engines is a fairly complex control problem. It might be simpler to >modify the overall thrust profile (i.e. by changing the total burn time). >I presume that's what the Shuttle would do if an SSME were to fail. >(I don't think they'd try to run the remaining engines at 150%.) There are two separate problems: 1) Using thrust modulation for attitude control I was thinking of on the order of 5% or so thrust changes, though I'll have to look at our aerodynamics and some kind of worst case wind profile to see what we actually need. If we can't do an increase in thrust, we will alter the attitude profile to reach our desired orbit. 2) Engine failure In our hybrid rocket, if lose an engine, we lose a big chunk of delta V since the solid part can't be used by the other engines. So we our out of luck in an engine failure, and therefore don't have to prepare a control strategy for it. Matt ------------------------------ Date: 12 Jul 1993 09:45:06 -0500 From: Matthew Feulner Subject: FWD>RE>modulating thrust To: Space Tech Date: 7/10/93 11:28 PM From: George William Herbert >With low altitude winds, you get a slightly different problem. >Right off the pad, the rocket can get blown around a lot >and bad things can happen if it gets blown too far over. >You can avoid this by careful launch planning, but you have >to design for occational gusts. In this case, motor ang >guidance reaction time compared to the vehicles lateral >ballistic quotient (mass divided by lateral surface area) >is the design rule. Can you be more specific? I can't compare seconds with kg/m^2. Thanks Matt ------------------------------ End of Space-tech Digest #157 *******************