Subject: Space-tech Digest #156 Contents: fiberglass, composites (3 msgs) NEA's (near earth asteroids) (1 msg) Turbopump fed boosters (10 msgs) ------------------------------------------------------------ Date: Fri, 2 Jul 93 22:16 PDT To: space-tech@cs.cmu.edu Subject: Composite pressure vessels From: Bruce_Dunn@mindlink.bc.ca (Bruce Dunn) Paul Dietz writes: > I was wondering if someone could tell me about fiberglass. > I noticed that E-glass fibers (which cost $1.65/kg in bulk) > have a tensile strength/density ratio some 8 times higher > than the high strength steel Bruce is proposing for P2. > How much of this strength actually ends up in fiberglass/epoxy > composite? from "Composites in Future Motor Hardware - a Research View", by N.J. Parratt, in the book "Solid Rocket Motor Technology", AGARD Conference Proceedings 259. "If oriented fibers are used to resist the stresses, however, the weight of structure needed to produce a vessel of given volume is theoretically the same, whether it be cylindrical or spherical." In contrast, metal pressure vessels are more efficient if they are spherical, rather than cylindrical. A sphere suffers uniform stress in all directions (called biaxial 1:1 in this reference). However, a cylinder has twice as much hoop stress as axial stress (biaxial 2:1). In effect, for a cylinder part of the strength of the metal goes unused. For a fiber composite cylinder however, only as much fiber as needed is oriented along the direction of each stress. Parratt notes that glass reinforced composites are difficult to use for solid rockets, as the case is not particularly stiff and there is considerable strain at the working stresses necessary to make use of the strength of the fiber. This tends to separate the propellant from the case, leading to all sorts of problems. This would not be a problem for liquid propellant tanks, although the structure must allow for the increase in tank size as the tanks are pressurized. The relative merits of metal and composites for pressure vessels depends on whether the shape is mostly spherical, or mostly cylindrical. Parratt makes some comparisons between steel at approximately 1400 MPa strength, and composites for pressure vessels for solid rockets. He assumes 50% fiber by volume, althouh he notes that 60% can often be obtained for wound or molded components. He suggests that the design stress in a composite pressure vessel be limited to 50% of its "theoretical" strength, to allow for variations in materials, "scale effects and statistical risk", and "some tolerance of small holes and local damage which is necessary in practice". Using this safety factor (but presumably a much less generous safety factor for metal) he gives a table of the relative masses of cylindrical cases and "end plates" (the latter stated to have biaxial 1:1 stress, and presumably approximating hemispheres). Relative mass: Material End Plates (ie. spheres) Cylinderical Cases Steel 1.00 1.00 Glass composite 1.16 0.87 Kevlar 0.6 0.45 Carbon fiber seems to perform about the same as Kevlar. The specific strength of the fibers is not as great, but the fiber is denser, allowing a higher percentage of the composite to be made up of fiber (rather than resin). Glass fiber on this basis does not look attractive for spherical pressure vessels (as are most tanks on the P2), and is only marginally advantageous for cylinders. Kevlar wins everywhere, but of course will be much more expensive. J.E. Gordon in one of his books relates how he was involved in the design of the glass composite case for one of the early solid propellant ballistic missles. A simplistic winding stratege gives a case which stretches too much in the radial direction (increase in diameter), causing propellant cracking problems. If the windings are correctly arranged "on the bias" the radial shrinking is reduced, at the expense of more case elongation than would otherwise occur. Stretch in this direction is not apparently so critical for solid propellants. > Richard Schroeppel writes: > I must be missing something in the discussion of fiberglass, etc. for > tanks. > Is there no need for compressive strength? What part of the structure is > transmitting the engine's thrust to the payload, and for that matter, to > the tanks and fuel within? The internal pressurized fluid (gas, liquid etc.) is. When I step on the inflated edge of a Zodiac boat, my foot doesn't plunge down into the inflated tube, even though the rubberized fabric of the boat has essentially no compressive strength by itself. -- Bruce Dunn Vancouver, Canada Bruce_Dunn@mindlink.bc.ca ------------------------------ From: henry@zoo.toronto.edu Date: Sat, 3 Jul 93 18:42:17 EDT To: space-tech@cs.cmu.edu Subject: Re: fiberglass, composites >I must be missing something in the discussion of fiberglass, etc. for tanks. >Is there no need for compressive strength? What part of the structure is >transmitting the engine's thrust to the payload, and for that matter, to >the tanks and fuel within? There are two answers to this. One is that fiberglass has considerable compressive strength too. It's generally a reasonably good structural material; the Voyager round-the-world aircraft had carbon-fiber wing spars but was otherwise almost entirely fiberglass. The other is that in many designs, tank internal pressure supports some or all of the external loads. Max Hunter claims that even in a pump-fed design (which typically does require modest tank pressurization to keep the pumps from cavitating), the internal pressure is the dominant load on the tank walls. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ From: henry@zoo.toronto.edu Date: Sat, 3 Jul 93 21:46:40 EDT Subject: Re: fiberglass, composites To: space-tech@cs.cmu.edu A friend has reminded me that pressurization may be the dominant *flight* load on the tank walls, but you do have to think about pre-flight conditions. For a pump-fed design with relatively modest pressurization loads, the loads encountered during ground handling or fueling may be worse, especially on a windy day. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: 3-JUL-1993 01:22:09.64 From: "Gordon D. Pusch (708)252-3843" Subject: NEA's in _Nature_ (24 June 1993) To: SPACE-TECH@BITNET.CC.CMU.EDU Near-Earth Asteroid fans should check out two articles in a recent issue of _Nature_ (v.363, no.6431, 24-June-1993): ``Explosions of small Spacewatch object in the Earth's atmosphere'' Christopher F. Chyba, pp.701--703; ``Evidence for a near-Earth asteroid belt'' D.L.Rabinowitz, et. al., pp.704--706. The first deals with burst altitudes and yield, and claims that non-ferrous objects of diameter less that 50m typically detonate too high to be dangerous; nevertheless, the author notes that several 20-ktonne bursts should occur per year, and wonders why so few of them are noticed (esp. since [he notes] US and FSU satellites should be able to detect them :-) The second claims there is an ``excess'' of NEAs over what is expected from the ``Opik formula;'' the authors speculate that perhaps they are lunar ejecta, or perhaps terrestrial equivalents of the Trojan asteroids. If this excess is *real* (rather than a selection-effect), and if the asteroids are of non-lunar origin, perhaps they constitute a useful resource...? (Personally, I've never even been able to understand why Jupiter's L-pts collect asteroids; *terrestrial* trojans exceed my boggle threshold... I mean, it's not as if L-pts are *sticky* or something --- things don't get ``trapped'' there unless something *else* perturbs them into that orbit... ...The only two mechanisms which come to my mind mind are two-body collisions near the L-pt leaving one or more fragments trapped (statistically unlikely); or co-orbital asteroids left over from the formation of the solar-system... Anyone out there know what the ``standard model'' on trojan-origins is?) Gordon D. Pusch ------------------------------ Date: Sun, 4 Jul 93 07:27:43 -0400 From: dietz@cs.rochester.edu To: space-tech@cs.cmu.edu Subject: low tech turbopumps? In Bruce's P2, the first stage tanks have a combined mass of 31 metric tons. Much of this could be saved if a pump-fed design were chosen, due to lower tank pressure. I wonder: how much cheaper and more robust can a turbopump be made if it is made with large, crude and potentially off-the-shelf parts? A mass budget of 20-30 tons would allow a pretty hefty turbopump. The pump power required by the first stage engine is about the same as a single F-1 engine, which weighs (pumps, thrust chamber and all) 9.3 metric tons. How much do the pumps weigh on the F-1 and various other conventional engines? It would be very interesting if one could just buy off-the-shelf pumps and turbines from the chemical process industry. What sort of power/mass ratio do these have? Paul ------------------------------ Date: Sun, 4 Jul 93 08:15 PDT To: space-tech@cs.cmu.edu Subject: Turbopumps for the P2? From: Bruce_Dunn@mindlink.bc.ca (Bruce Dunn) > Paul Dietz writes: > > > In Bruce's P2, the first stage tanks have a combined mass of 31 metric > tons. Much of this could be saved if a pump-fed design were chosen, > due to lower tank pressure. I wonder: how much cheaper and more > robust can a turbopump be made if it is made with large, crude and > potentially off-the-shelf parts? A mass budget of 20-30 tons would > allow a pretty hefty turbopump. The pump power required by the first > stage engine is about the same as a single F-1 engine, which weighs > (pumps, thrust chamber and all) 9.3 metric tons. How much do the > pumps weigh on the F-1 and various other conventional engines? > > It would be very interesting if one could just buy off-the-shelf pumps > and turbines from the chemical process industry. What sort of > power/mass ratio do these have? > While commercial pumps might be available, I doubt that there would be commercial turbines available. The turbines on liquid propelled rockets are optimized for power to weight ratio, and generally pay little attention to fuel economy or useful life beyond a few minutes. I can't see any commercial application where the sorts of power needed (Megawatts) would be generated with such low efficiency. It is an interesting exercise however to try to imagine whether the P2 would be improved by pumps. Dealing with the lower stage, the tank mass could be reduced from 31 tons to say 5 tons (this is a guess, not backed up by any calculations). The missing steel can be replaced by propellant, which is a good thing as something like 1 % of the propellant mass flow (about 7 tons) must be diverted to power the pumps, lowering the specific impulse by 1% Let us assume the the required el cheapo turbopump has a mass of 10 tons. Leave the upper stage alone for the moment, to avoid confusing the calculations. The net result is that the redesigned vehicle has a 7% greater payload. The redesigned vehicle has a tank which probably costs as much as the heavy steel tank. It is light, but is now made of aluminum with ribs and stringers to keep it from buckling as it is handled in an unpressurized state. The tank pressurization system is not eliminated - it is now generating much less pressure, but must still be there in some manner to provide the NPSH (Net Positive Suction Head) pressure required by the pump inlets. The vehicle is now burdened with several large pieces of rotating machinery (a minimum of one turbine and two pumps, plus associated gear trains, plus a lubrication system) plus a gas generator to run the turbine. All of these pieces are critical to the operation of the engine, and therefore will attract quality control inspectors and paperwork at the same time they add additional failure modes and costs to the vehicle. All this is for a 7% increase in payload! Why bother? I have an alternate solution: increase the size of the P2 booster by 7% and you can have the same payload benefit without the problems. Similar arguments apply to the upper stage of the P2. As an oddball way of pointing out the complexity that turbopumps add to a vehicle, I refer to "The Design of Liquid Propellant Rocket Engines" by Huzel and Huang. This standard reference work spends 22 pages talking about design issues related to pressurized propellant feed systems, but 85 pages talking about pump fed systems. If pages are proportional to complexity, then pumps take about 4 times as much designing as a pressure fed system. -- Bruce Dunn Vancouver, Canada Bruce_Dunn@mindlink.bc.ca ------------------------------ Date: Sun, 4 Jul 93 20:47:27 -0400 From: dietz@cs.rochester.edu To: Bruce_Dunn@mindlink.bc.ca, space-tech@cs.cmu.edu Subject: Re: Turbopumps for the P2? Bruce points out that a 10 ton turbopump stage for P2 would increase payload by only 7%, which he argues is not worth the bother. It may not be, but I think the argument is flawed. If turbopumps are desirable, it would be because they could be cheaper, not because they could offer much more payload. How might a turbopump fed rocket help? (1) It could enable one to greatly reduce or eliminate the use of helium, perhaps reducing the demand on the pressurization system to the point where simple compressed gases would work. This simplification could help offset the complexity of the pump system. (2) It would enable the use of LOX instead of peroxide as oxidizer. LOX is less expensive (the raw savings on the first stage of P2 would be a bit over $1 M). More important, peroxide is listed under OSHA regulations in the same class as concentrated nitric and sulfuric acids, N2O4 and chlorine. I don't believe LOX is. (3) Higher performance could instead be used to relax design margins elsewhere in the system. For example, the steel tanks in Bruce's design have a margin of 1.2. Margins that small make me nervous. In contrast, the tanks in a pump fed design can be given a larger safety factor at lower performance penalty. The power density required of the system is not outrageously high: 20 MW in (say) 20 metric tons, or 1 kW/kg. A V-8 automotive engine (offered only for comparison, not as a serious proposal) achieves perhaps .7 kW/kg. The energy density required is only ~70 Wh/kg, which suggests that even a *battery powered* pump system might be light enough. Rocket pumps often last only a few minutes, but there is no reason why they should be designed that way (witness the low temperature turbines in expander cycle engines, with their great durability). It should be possible to fire up the pumps before launch for testing. This might be simplified by the installation of surge tanks between the pumps and the engines, so that the pumps could be tested without using the thrust chamber. This doesn't test for all vibrational effects, granted. Paul ------------------------------ Date: Tue, 6 Jul 93 23:17:32 -0400 From: dietz@cs.rochester.edu To: Bruce_Dunn@mindlink.bc.ca Subject: on 3 stages for P2... Cc: space-tech@cs.cmu.edu Bruce, I was reading a bit about TRW's 3-stage low cost launch vehicle concepts. One thing they were thinking of was making the first stage very dumb, without any thrust vectoring, even. Perhaps this would be a good idea with P2. We make a first stage in which the propellants are pressurized by gas already in the tanks. Make the tanks perhaps 1/2 to 1/3 full to start (perhaps with an open-loop heater, say a solid propellant gas generator, to warm the gas as it expands). This is fed into a dumb engine through diaphragms that could be triggered to simultaneously burst. Aside from the connection to external pressurization equipment and valves for fueling/draining the tanks, and the diaphragms, there would be no moving parts. 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. Another simple idea would be gyroscopic stabilization: stick one or more big steel flywheels in the first stage, spun up by an external compressed air source before launch. The first stage here is really crude. We may use lower cost steel, use lower pressure than your engines, and pressurize with air or nitrogen. It gets discarded after delivering perhaps 1 to 2 km/s of impulse to the stack. I think this concept may work better with a cheaper propellant than peroxide in the first stage. One that might be good is azeotropic nitric acid (68% nitric acid, 32% water). This stuff is pretty cheap, I think, about half the cost of the same mass of highly concentrated nitric acid (which is itself 4 times cheaper than 100% peroxide), and is subject to somewhat less stringent regulations. It also has much lower vapor pressure than highly concentrated nitric acid, and very little NO2 contamination. The dilution does reduce the specific impulse, but not proportionally; naively, specific impulse should go roughly as the square root of the mass fraction of the active reactants (actually, better than that, since dilution shifts the equilibrium by lowering the chamber temperature). In addition, the energy content of the propellants is delivered to the vehicle most efficiently if the exhaust velocity = - vehicle velocity (ignoring the thermal content of the exhaust plume), so diluting them to reduce the exhaust velocity may reduce the fuel cost (if the diluent is essentially free, like water). Paul ------------------------------ Date: Tue, 6 Jul 93 21:52 PDT To: space-tech@cs.cmu.edu Subject: Turbopump fed boosters From: Bruce_Dunn@mindlink.bc.ca (Bruce Dunn) Paul Dietz presents several suggestions about why a turbopump fed rocket burning kerosene and LOX might be preferable to the pressure fed P2 burning peroxide and kerosene. The case he makes is quite reasonable, and arguments of this type are no doubt one of the reasons that large pressure fed rockets have not yet been developed. Paul suggests that a safety factor of 1.2 used in the design of the P2 pressure tanks makes him nervous. I should clarify this safety factor with some recent information I have obtained. I have stated that the Ladish D6AC steel used in the Titan boosters and SRBs has a yield strength of 1340 MPa. I now find that I have mis-read a secondary reference. The steel is heat treated to an ultimate (not yield) strength of 195,000 psi (1340 MPa). The yield strength for this steel however is not much below the ultimate strength, perhaps 1250 MPa judging from the yield to ultimate strength ratio for D6AC of similar heat treatments, given in a metals handbook. For propellant tanks, the textbook of rocket design by Huzel and Huang suggests that the working stress of a tank be the ultimate strength divided by a safety factor of 1.25 (for unmanned systems). They suggest a larger safety factor for manned systems. One recent reference says that the pressure loads in the shuttle ET are figured with a safety factor of 1.25, although to be fair it is not explicitly stated whether this factor is to be applied to the yield or the ultimate strength of the wall material. From this point on, I will be using a safety factor of 1.25 applied to an ultimate strength of 1340 MPa for the P2 calculations. This makes only a very small change in delivered payload. To see what might be accomplished with turbopump-fed LOX/kerosene booster as an alternative to the P2, I fired up my spreadsheet and fed it some ballpark numbers for a two stage booster based on a single F1 engine in the first stage, and a single RS-27 (Delta) engine in the second stage. While the F1 engine is not currently available, it could be put into production again, or similar thrust RD-170 engines could be bought from Russia. It turns out that this vehicle is a nice comparison with the P2. It has a gross liftoff mass about half that of the P2, and delivers the roughly the same payload. This vehicle is about as simple as you can get with a turbopump system. Both engines have a proven record, and it is not clear that they can be made much simpler or more reliable, even with a relaxation of design requirements (ie. although relaxed design margins might allow a heavier engine, the number of parts is not likely to be reduced). Relative to the P2, the cost of the propellant is lower and the oxidizer is more innocuous (although is is a cryogen). I believe that the tanks for the pump fed booster are likely to cost more than the tanks for the P2. They are smaller (although not half the size because of the poorer density of the propellant), but are made of aluminum. I also think that the engines for the turbopump fed system are likely to much more expensive than the simple P2 engines, even though the P2 engines are larger. The P2 has a much larger tank pressurization system than the turbopump fed system - the nod here probably goes to the turbopump fed system. It should be noted however that in terms of failure points, the P2 system may be simpler than the system of the turbopump fed booster. Launch operations for the turbopump fed vehicle are somewhat more complicated than for the P2 (cryogenic oxidizer). Finally, the turbopump fed engines have many more failure points than the P2 engines. I haven't tried to make an analysis of historical launch failures, but I am under the impression that loss of thrust due to engine problems is the primary cause of mission failure. I don't remember however ever hearing of a failure due to propellant tank rupture (except in cases where the primary problem was elsewhere). Perhaps Henry could comment on the historical record. -- Bruce Dunn Vancouver, Canada Bruce_Dunn@mindlink.bc.ca ------------------------------ Date: Wed, 7 Jul 93 14:07:36 -0400 From: dietz@cs.rochester.edu To: Bruce_Dunn@mindlink.bc.ca, space-tech@cs.cmu.edu Subject: Re: Turbopump fed boosters Bruce wrote: > ooster as an alternative to the P2, I fired up my spreadsheet and fed > it some ballpark numbers for a two stage booster based on a single F1 > engine in the first stage, and a single RS-27 (Delta) engine in the > second stage. While the F1 engine is not currently available, it > could be put into production again, or similar thrust RD-170 engines > could be bought from Russia. It turns out that this vehicle is a nice > comparison with the P2. It has a gross liftoff mass about half that > of the P2, and delivers the roughly the same payload. > This vehicle is about as simple as you can get with a turbopump system. Both > engines have a proven record, and it is not clear that they can be made much > simpler or more reliable, even with a relaxation of design requirements (ie. > although relaxed design margins might allow a heavier engine, the number of > parts is not likely to be reduced). I think one could make it less costly. The F-1 engine is (partially) regeneratively cooled. It has a complex injector with thousands of holes. Its turbopumps have a power density of 36 kW/kg -- far higher than the requirements I posited for P2, although still less than the SSME fuel turbopump (160 kW/kg!). Pump-fed engines don't need to be this complex. For example, injectors can be made with orders of magnitude fewer orifices, as was demonstrated in AFPRLs Project Scorpio in the mid 60s (with LOX/LH2). This reduces injector cost to about 1/3 of the conventional approach. Other cooling schemes (ablative, film) lead to simpler engines. Development cost is a good counterargument to pump-fed engines, I admit, as is reliability in the absence of an extensive testing program. Paul ------------------------------ From: henry@zoo.toronto.edu Date: Wed, 7 Jul 93 14:19:24 EDT To: space-tech@cs.cmu.edu Subject: Re: Turbopump fed boosters >... I am under the impression that >loss of thrust due to engine problems is the primary cause of mission >failure. I don't remember however ever hearing of a failure due to >propellant tank rupture (except in cases where the primary problem was >elsewhere). Perhaps Henry could comment on the historical record. I dimly recall reading a recent study which touched on this (I'll see if I can find it at home). The dominant, although not exclusive, failure mode of liquid-fuel rockets is simple loss of thrust from an engine, a "benign" failure that a vehicle with engine-out capability could survive. If my dim memory is right, this accounts for about 2/3 of all liquid-engine failures, even if you go back into fairly early days of US launchers. (Solid-rocket failures, by contrast, are historically 100% catastrophic.) I don't remember any tank failures either, except as a side effect of other causes, typically attempts to fly sideways due to engine or guidance failures. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ From: henry@zoo.toronto.edu Date: Wed, 7 Jul 93 14:26:30 EDT To: dietz@cs.rochester.edu Subject: Re: on 3 stages for P2... Cc: space-tech@cs.cmu.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... I suppose you could plan for relatively high acceleration and build the world's largest model-rocket launch rod :-) to keep the thing stable until it gets up to fin speed, but this seems rather drastic. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ To: gwh@lurnix.COM Cc: Bruce Dunn , space-tech@cs.cmu.edu Subject: Re: Turbopump fed boosters Date: Wed, 07 Jul 93 10:51:27 -0700 From: gwh@lurnix.COM I wrote: >One degree of attitude change, which fins will allow easily, >will add up to 45-60 degrees over even really short burn times. In case nobody figures this out, that's one degree per second. -george ------------------------------ To: Bruce Dunn Cc: space-tech@cs.cmu.edu, gwh@lurnix.COM Subject: Re: Turbopump fed boosters Date: Wed, 07 Jul 93 10:44:48 -0700 From: gwh@lurnix.COM From: uunet!cs.rochester.edu!dietz >I was reading a bit about TRW's 3-stage low cost launch vehicle >concepts. One thing they were thinking of was making the first stage >very dumb, without any thrust vectoring, even. Perhaps this would be >a good idea with P2. >We make a first stage in which the propellants are pressurized by gas >already in the tanks. Make the tanks perhaps 1/2 to 1/3 full to start >(perhaps with an open-loop heater, say a solid propellant gas >generator, to warm the gas as it expands). This is fed into a dumb >engine through diaphragms that could be triggered to simultaneously >burst. Aside from the connection to external pressurization equipment >and valves for fueling/draining the tanks, and the diaphragms, there >would be no moving parts. I looked at this quite a bit when I was working on liquid and hybrid vehicles earlier; it can be done, though it's better to use full tanks and a solid propellant gas generator. The variations in performance if you just use a half-full tank with fixed pressure, along with the tank fraction penalty, are enormous. >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. Another simple idea would be >gyroscopic stabilization: stick one or more big steel flywheels in the >first stage, spun up by an external compressed air source before >launch. I don't like either of these solutions; fins allow too much course change over a burn, and flywheels are too massive and prone to really painful failure modes. One degree of attitude change, which fins will allow easily, will add up to 45-60 degrees over even really short burn times. 45-60 degrees worst case is enough to put you going down again. On the average, you wouldn't, but it's a guaranteed statistical crapshoot with every launch. And that's presuming you don't start spinning or something equally unpleasant... Flywheels would have to get really really heavy to keep the vehicle on course, and big flywheels are either very very expensive (composite flywheels) or very very heavy (higher-mass lower RPM metal flywheels). From: Bruce Dunn >Paul suggests that a safety factor of 1.2 used in the design of the P2 >Pressure tanks makes him nervous. I should clarify this safety factor with >some recent information I have obtained. I have stated that the Ladish D6AC >steel used in the Titan boosters and SRBs has a yield strength of 1340 MPa. >I now find that I have mis-read a secondary reference. The steel is heat >treated to an ultimate (not yield) strength of 195,000 psi (1340 MPa). The >yield strength for this steel however is not much below the ultimate >strength, perhaps 1250 MPa judging from the yield to ultimate strength ratio >for D6AC of similar heat treatments, given in a metals handbook. For >propellant tanks, the textbook of rocket design by Huzel and Huang suggests >that the working stress of a tank be the ultimate strength divided by a >safety factor of 1.25 (for unmanned systems). They suggest a larger safety >factor for manned systems. One recent reference says that the pressure loads >in the shuttle ET are figured with a safety factor of 1.25, although to be >fair it is not explicitly stated whether this factor is to be applied to the >yield or the ultimate strength of the wall material. From this point on, I >will be using a safety factor of 1.25 applied to an ultimate strength of 1340 >MPa for the P2 calculations. This makes only a very small change in >delivered payload. As a data point, T-1 yields at 100 KSI and fails at 110-130 KSI. I am exceedingly uncomfortable about the methodology of basing working stress on ultimate, not yield, strengths. It tends to push the working stress up to the edge of yielding; at this stress range, if you get additional vibrations or an unexpected overload, you'll start to develop cracks in most of these materials very quickly. Followed annoyingly shortly thereafter by failure... To avoid this, you have to engineer the hell out of the design to make very very sure you don't overstress anything. Naval Architecture standardly uses yield strength, and a large (1.5ish) safety margin, as the standard methodology for calculating strengths. We routinely design ships in 6 months with 15 person teams. I think that the price differential involved in qualifying higher-stress designs is an excessive hump to get over at project start, which is what's killing fast development of alternative launch vehicles. Which is why I'm not doing it that way, though my mass fractions suffer for it. I'm hoping that the lower development cost will offset the higher per-flight costs, and so far my numbers say yes 8-) -george william herbert Retro Aerospace ------------------------------ End of Space-tech Digest #156 *******************