Author Topic: RP-1, methane, impulse density Q&A  (Read 66370 times)

Offline R7

  • Propulsophile
  • Senior Member
  • *****
  • Posts: 2738
    • Don't worry.. we can still be fans of OSC and SNC
  • Liked: 950
  • Likes Given: 662
Re: RP-1, methane, impulse density
« Reply #60 on: 06/09/2015 09:27 AM »
Ethylene may be great once you get it inside the combustion chamber but will it behave properly when driven thru turbopumps and coolant channels? Thermal stability, tendency to polymerize and all that? It's a major chemical feedstock due to capability to form chains with molecular weight up to millions to make plastics and whatnot. For this reason a colony will probably establish ethylene production anyway.

What's the route from basic Martian resources to synthesize ethylene?
AD∑ASTRA∑ASTRORVM∑GRATIA

Online Steven Pietrobon

  • Member
  • Senior Member
  • *****
  • Posts: 16375
  • Adelaide, Australia
    • Steven Pietrobon's Space Archive
  • Liked: 5256
  • Likes Given: 657
Re: RP-1, methane, impulse density
« Reply #61 on: 06/15/2015 06:00 AM »
Ethylene may be great once you get it inside the combustion chamber but will it behave properly when driven thru turbopumps and coolant channels? Thermal stability, tendency to polymerize and all that?

You need a catalyst for ethylene to polymerise, so the liquid may still be stable under heat and pressure. According to

http://www.chemguide.co.uk/mechanisms/freerad/polym.html

you need 200 C and 2000 atm of pressure. Typical staged combustion engines are only 200 atm, so this is probably well below what could cause a problem.

Quote
What's the route from basic Martian resources to synthesize ethylene?

Here is one possible way. First we have Sabatier

2CO2 + 8H2 → 2CH4 + 4H2O

Followed by water electrolysis

6H2O → 6H2 + 3O2

Followed by oxidative coupling of Methane

http://siluria.com/Technology/Oxidative_Coupling_of_Methane

2CH4  +  O2   →   C2H4 +  2H2O

The overall result is

2CO2 + 2H2 → C2H4 + 2O2

However, we need 2.37O2 for each C2H4, which means we need to generate an additional 0.37O2, which can be done using the reverse water reaction:

CO2 + H2 → CO + H2O

and then electrolysis

H2O → H2 + 0.5O2

That's a lot more efficient then directly electrolysing CO2 into CO and O2. My initial reaction is that the extra complexity is probably not worth it. However, the reduced amount of H2 from Earth plus the higher Isp and density of ethylox might possibly make it worthwhile.

This web site

http://www.marspedia.org/index.php?title=Reverse_Water-Gas_Shift_Reaction

says that you can use the CO from the reverse water reaction to also make methane and methanol, from which you can make ethylene.

There is also some research on CO2+H2 to C2H4, but I don't think it has been industrialised yet.
« Last Edit: 06/15/2015 06:38 AM by Steven Pietrobon »
Akin's Laws of Spacecraft Design #1:  Engineering is done with numbers.  Analysis without numbers is only an opinion.

Offline Proponent

  • Senior Member
  • *****
  • Posts: 5478
  • Liked: 991
  • Likes Given: 653
Re: RP-1, methane, impulse density
« Reply #62 on: 08/25/2015 06:03 AM »
Methane may have an advantage in fuel-rich staged combustion, but I'm still looking for an aha moment to explain its current popularity in other cycles.

The attached plot shows the viscosities as functions of temperatures of methane, ethane and propane as solid curves in brown, red and orange, respectively (mneumonic: express the number of carbon atoms with the resistor color code).  Dotted red and orange lines show ethylene (ethene) and propylene (propene).  The solid line across the top shows dodecane, standing in for RP-1, at 20oC; RP-1 itself would probably be a little more viscous still.  All data are from NIST.

At lox temperatures, propane is molasses in January, if you'll pardon the hemispherism.  For the most part, though, it's less viscous than RP-1, which is often used in cooling.  So, is methane's very low viscosity, even at lox temperatures, a significant advantage, or is it just a marginal benefit?

Suppose you did want to use lox-propane with a common-bulkhead tank.  The propane near the bulkhead could be pretty sludgey, but how big a problem would that be?

Offline Proponent

  • Senior Member
  • *****
  • Posts: 5478
  • Liked: 991
  • Likes Given: 653
Re: RP-1, methane, impulse density
« Reply #63 on: 03/09/2016 02:15 PM »
When I previously posted in this thread, I was under the misapprehension that Blue Origin's BE-4 ran a fuel-rich staged-combustion cycle.  In fact, it's oxidizer-rich (fact sheet attached).  While reasons for SpaceX's Raptor to burn methane rather than a slightly heavier hydrocarbon might include:

  • 1. Relative cleanliness of combustion products on the fuel-rich side of Raptor's full-flow cycle;
  • 2. High mass-specific heat capacity, which appears to partially offset methane's low density;
  • 3. Relative ease of ISRU methane production on Mars compared to heavier hydrocarbons (though this argument does not make make much sense to me); and
  • 4. Proximity of methane's boiling point to oxygen's, which allows common-bulkhead tanks;

  • only the last of these applies to the BE-4.  Even then, one could point out that propane is liquid at lox temperatures and typical ullage pressures.  However, it will be rather viscous.  Also, the use of a sub-cooled propellant makes autogenous pressurization trickier (do we know whether autogenous pressurization is planned for Vulcan, MCT or modified Falcon stages?).

    I suspect the choice of methane comes is an example of the fact that, in the paraphrased words of a forum member we're very lucky to have, "amateurs argue about engines; professionals argue about pressurization systems."

    Offline Robotbeat

    • Senior Member
    • *****
    • Posts: 28382
    • Minnesota
    • Liked: 8253
    • Likes Given: 5444
    Re: RP-1, methane, impulse density
    « Reply #64 on: 03/10/2016 02:52 AM »
    Try subcooled propylene as a propellant. Even better than ethylene.
    Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

    To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

    Offline Rei

    • Full Member
    • ****
    • Posts: 471
    • Iceland
    • Liked: 188
    • Likes Given: 58
    Re: RP-1, methane, impulse density
    « Reply #65 on: 03/11/2016 09:39 PM »
  • 3. Relative ease of ISRU methane production on Mars compared to heavier hydrocarbons (though this argument does not make make much sense to me); and

  • What about it doesn't make sense to you?  The methane fraction that comes out is much higher than that of other hydrocarbons.  You can customize it to some extent by controlling the process parameters - catalyst selection, heat, pressure, etc - but in general you get far more methane than other hydrocarbons.

    Now, if you want higher hydrocarbons and not the methane, you can always convert the methane (and other undesirable fractions) to syngas and then route it back in as a feedstock.  But it's added complexity, greater energy consumption, and lower throughput.

    Offline Proponent

    • Senior Member
    • *****
    • Posts: 5478
    • Liked: 991
    • Likes Given: 653
    Re: RP-1, methane, impulse density
    « Reply #66 on: 03/12/2016 07:43 AM »
    That the synthesis of methane is easier than that of heavier hydrocarbons makes perfect sense.  What I doubt is that plans for Martian ISRU in the future much influenced SpaceX's choice of propellants for the near term (Raptor).

    Offline Robotbeat

    • Senior Member
    • *****
    • Posts: 28382
    • Minnesota
    • Liked: 8253
    • Likes Given: 5444
    Re: RP-1, methane, impulse density
    « Reply #67 on: 03/16/2016 07:25 AM »
    That the synthesis of methane is easier than that of heavier hydrocarbons makes perfect sense.  What I doubt is that plans for Martian ISRU in the future much influenced SpaceX's choice of propellants for the near term (Raptor).
    Then you doubt incorrectly.
    Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

    To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

    Offline Nilof

    • Full Member
    • ****
    • Posts: 914
    • Liked: 376
    • Likes Given: 540
    Re: RP-1, methane, impulse density
    « Reply #68 on: 03/16/2016 11:27 AM »
    Another reason that Musk has mentioned at some point is that LNG/LOX is among the cheapest propellant combinations available. Sure, propellant costs are a tiny fraction of launch costs now, but in an architecture where you're aiming for 500 grand per seat to go to Mars the current market price of alternative propellants such as Ethylene suddenly becomes a rather significant portion of the total cost if you use it on every stage.

    So I think that Methane was chosen partly because it happens to be one of few propellants in the intersection of (high performance propellants that can easily be made on Mars) and (propellants with a reasonably low market price on Earth).
    For a variable Isp spacecraft running at constant power and constant acceleration, the mass ratio is linear in delta-v.   Δv = ve0(MR-1). Or equivalently: Δv = vef PMF. Also, this is energy-optimal for a fixed delta-v and mass ratio.

    Offline sevenperforce

    • Full Member
    • ****
    • Posts: 873
    • Liked: 250
    • Likes Given: 267
    Re: RP-1, methane, impulse density
    « Reply #69 on: 03/16/2016 03:26 PM »
    Try subcooled propylene as a propellant. Even better than ethylene.
    Wikipedia's propellant table lists ethylene as having a noticeably better specific impulse than methane. Of course, it has much higher density.

    It doesn't list propylene, though...and isn't propylene more likely to coke? I haven't been able to find anything more than 347 s for the vacuum isp of a propylene/LOX engine.

    For increased impulse density, what about using an H2O2/LOX blend during launch and then switching to LOX-only as you go higher up? Peroxide has the advantage of self-pumping, and it has a notably higher density than LOX with only a meager hit in specific impulse. It would also really improve T/W ratio at launch, reducing gravity drag.

    Offline Proponent

    • Senior Member
    • *****
    • Posts: 5478
    • Liked: 991
    • Likes Given: 653
    Re: RP-1, methane, impulse density
    « Reply #70 on: 03/16/2016 04:29 PM »
    Robotbeat, Niloff: what I'm thinking is that lots of people are keen on lox-methane these days, like BO, ULA and Firefly (last I heard Firefly had switched to lox-RP-1, but methane was originally baselined), and they're not all focused on Mars.  On the BE-4 info sheet, BO says this about its choice of LNG:
    Quote from: Blue Origin
    Liquefied natural gas enhances affordability and reusability

    Liquefied natural gas is commercially available, affordable, and highly efficient for spaceflight. Unlike other rocket fuels, such as kerosene, liquefied natural gas can be used to pressurize a rocketís propellant tanks. This is called autogenous pressurization and eliminates the need for costly and complex pressurization systems, like helium. Liquefied natural gas also leaves no soot byproducts as kerosene does, simplifying engine reuse.
    .

    Another factor is that even if SpaceX succeeds wildly, it's going to launch one heckuva lot of mass from the the surface of the Earth before it produces its first liter of ISRU methane.  Since there's a lot of hardware to be developed on the way to that first liter of ISRU methane anyway, I would think it would make sense to optimize for Earth launch for the time being and then tweak propulsion systems as needed later for ISRU methane.

    So, I could believe that if there were several propellant combinations that were approximately equally attractive, SpaceX would probably choose the one that's best for Mars.  But if there were other fuels nearly as attractive as methane, why are none of the less-Mars-obsessed players pursuing them?

    Offline Rei

    • Full Member
    • ****
    • Posts: 471
    • Iceland
    • Liked: 188
    • Likes Given: 58
    Re: RP-1, methane, impulse density
    « Reply #71 on: 03/16/2016 06:29 PM »
    You need a catalyst for ethylene to polymerise, so the liquid may still be stable under heat and pressure. According to

    http://www.chemguide.co.uk/mechanisms/freerad/polym.html

    you need 200 C and 2000 atm of pressure. Typical staged combustion engines are only 200 atm, so this is probably well below what could cause a problem.

    There's a wide range of catalytic processes for ethylene; some operate at near atmospheric pressure.  The reaction rate is positively correlated with temperature.  I'm not sure how much of a problem it would be without a catalyst but at very high temperatures.  Probably not too mucy.  Polyethylene is considered to be a good fuel in its own right in some systems - for example, hybrid rockets.  It readily melts/vaporizes, like paraffin wax, so gets whipped into a spray in the flow which gives it a large reactive surface area (very important in hybrids)

    As for how it's produced: The production process to polyethylene first involves the sabatier reaction to produce methane, a second (or combined) step to produce simple alkenes, a filter stage to extract ethylene from the rest of the stream, an oxidation stage for the rest of the stream to turn it into syngas to feed back into the original process, a couple purification steps on the ethylene stream, and then a polymerization stage.  There's as mentioned a variety of processes involved in polymerization... widely varied pressures, but generally fairly low temperatures required.  The problem with going too hot is that it catalyzes too fast and gunks up your reactor; since polymerization is exothermic, you can even (particularly in gas-phase polymerization systems) get runaway polymerization that basically turns your catalyst bed into a big block of plastic  ;)   Most processes (although not gas phase) end up with the catalyst embedded in the output product, which may or may not be extracted, and to varying degrees, and with varying levels of consumables involved.  Even in gas-phase reactors, you still need to renew/replace your catalyst over time.

    Rather than just filtering out only ethylene from the sabatier output stream, though, better would be to distill multiple products out, because each is useful.  For example, propylene reacts with ammonia (Haber process) and oxygen in the SOHIO process to produce a stream that you wash with sulfuric acid to produce a bit of N2 waste gas, some NH3SO4, some HCN, some CH3CN, but mainly acrylonitrile.  Polymerizing *that* yields PAN, which can be gel spun to strong PAN fiber.  This in turn can be first oxidized, then carbonized, followed up by a sulfuric acid surface treatment, to produce carbon fiber.   Of course, fibers alone, regardless of their composition, are only part of the story, you then generally want to produce them into weaves, and you still need resins to make composite parts... and manufacture would be quite difficult if you had to work in a space suit...

    Offline Proponent

    • Senior Member
    • *****
    • Posts: 5478
    • Liked: 991
    • Likes Given: 653
    Re: RP-1, methane, impulse density
    « Reply #72 on: 03/17/2016 06:56 AM »
    I presume, then, the suitability of propylene, ethylene or other unsaturated hydrocarbons depends on the details.  They're probably not such good coolants, and it sounds like fuel-rich or full-flow staged combustion might be a disaster.  On the other hand, if cooling and pumping requirements are not placed on them, then they could be pretty attractive.  The one actual use of propylene I'm aware of is in some of Garvey Spacecraft Corporation's rockets (hopefully leading to nanosat launchers), which I believe are pressure fed -- don't know about the cooling.  Do we know of any actual uses of ethylene (or other uses of propylene)?

    Offline Rei

    • Full Member
    • ****
    • Posts: 471
    • Iceland
    • Liked: 188
    • Likes Given: 58
    Re: RP-1, methane, impulse density
    « Reply #73 on: 03/18/2016 08:02 AM »
    It depends on what your primary needs in a coolant are.  Like most rocket propellants, they have high specific heats.  Yes, my main concerns would be anywhere that they're being heated in small channels where there's no oxidizer present.  That said, there's all sorts of things one can do to prevent polymerization.  For example, the process is impurity-sensitive; in the event of runaway polymerization the solution is often simply to "poison" the reaction by injecting carbon monoxide or other hydrocarbons.  I'm sure a reasonable approach could be developed.

    I'm not aware of any full-scale liquid fuel systems that burn ethylene or propylene, and haven't even read about experimental scale ones, although there might have been some.  I have however read about experimental scale usage of polyethylene as a hybrid rocket fuel.  Aluminized paraffin wax and aluminized polyethylene appear to be the best hybrid rocket fuels out there: cheap, stable, easy to cast, easy to melt/vaporize, burn readily with LOX, and low viscosity upon heating, so easy to whip into a fast-burning mist.  You get significantly higher ISP and propellant density with aluminized paraffin or polyethylene burned with LOX than you do with LOX/RP-1 (mainly due to the aluminum).  A lot of hybrid work had previously focused on polybutadiene as the fuel, which is very mature in its use as a solid fuel binder - but hybrids are not solids, and polybutadiene isn't as readily mobilized and thus gets a significantly lower burn rate, thus requiring significantly more combustion channels, thus leading to less burn stability.

    Offline Rei

    • Full Member
    • ****
    • Posts: 471
    • Iceland
    • Liked: 188
    • Likes Given: 58
    Re: RP-1, methane, impulse density
    « Reply #74 on: 03/18/2016 08:08 AM »
    Hmm, a thought just occurred to me, wherein cryochilled propane's viscosity could be an advantage.  Are you familiar with the research on metalized gel propellants?  The concept is to add gelling agents like fumed silica to allow you to suspend metal dusts like aluminum powder in the liquid rocket fuel.  Aluminum combustion gives off a great deal of energy for its mass, providing additional heat to the exhaust stream.  But you know, gelling basically means "increasing the viscosity".  With cryochilled propane, you already have increased viscosity vs. "runny" fuels like RP-1 (whether it's sufficient to suspend aluminum particles without gelling agents, that I can't say).
    « Last Edit: 03/18/2016 08:09 AM by Rei »

    Offline notsorandom

    • Full Member
    • ****
    • Posts: 1716
    • Ohio
    • Liked: 408
    • Likes Given: 90
    Re: RP-1, methane, impulse density
    « Reply #75 on: 03/18/2016 12:12 PM »
    Robotbeat, Niloff: what I'm thinking is that lots of people are keen on lox-methane these days, like BO, ULA and Firefly (last I heard Firefly had switched to lox-RP-1, but methane was originally baselined), and they're not all focused on Mars.  On the BE-4 info sheet, BO says this about its choice of LNG:
    Quote from: Blue Origin
    Liquefied natural gas enhances affordability and reusability

    Liquefied natural gas is commercially available, affordable, and highly efficient for spaceflight. Unlike other rocket fuels, such as kerosene, liquefied natural gas can be used to pressurize a rocketís propellant tanks. This is called autogenous pressurization and eliminates the need for costly and complex pressurization systems, like helium. Liquefied natural gas also leaves no soot byproducts as kerosene does, simplifying engine reuse.
    .

    Another factor is that even if SpaceX succeeds wildly, it's going to launch one heckuva lot of mass from the the surface of the Earth before it produces its first liter of ISRU methane.  Since there's a lot of hardware to be developed on the way to that first liter of ISRU methane anyway, I would think it would make sense to optimize for Earth launch for the time being and then tweak propulsion systems as needed later for ISRU methane.

    So, I could believe that if there were several propellant combinations that were approximately equally attractive, SpaceX would probably choose the one that's best for Mars.  But if there were other fuels nearly as attractive as methane, why are none of the less-Mars-obsessed players pursuing them?
    Methane seems to have a lot of attractive qualities and its drawbacks are managed easily. Density and IPS are important but there are other factors as well that are important. Things like pressurization, thermal capacity, viscosity, and coking. It could be that Methane is the ultimate compromise. Other propellant combinations offer better performance in some of those categories but worse in others. That it is easy to make on Mars may just be the cherry on top.

    Offline hkultala

    • Full Member
    • ****
    • Posts: 857
    • Liked: 329
    • Likes Given: 283
    Re: RP-1, methane, impulse density
    « Reply #76 on: 04/14/2016 09:01 PM »
    Very generally on the top of specific impulse and impulse density, I was thinking about optimal mixture ratios.  If oxidizer and fuel have different densities, then impulse density will peak at a mixture ratio corresponding to a higher propellant bulk density than where the specific impulse peaks.  The larger the difference in the densities of oxidizer and fuel, the larger will tend to be the difference in the mixture ratios of the two peaks.  Consider lox-hydrogen.  The attached figure shows specific impulse1 and impulse density as a function of both bulk density (lower horizontal axis) mixture ratio (O/F: upper horizontal axis)2.

    Obviously specific impulse peaks at a density of about 317 kg/m3 (O/F=4.8), while impulse density peaks at  633 kg/m3 (O/F=17.8).  Lox-hydrogen stages usually operate at mixture ratios of about 5 or 6.  You can imagine situations where you'd want to go significantly higher than that.  But you'd never want to go to a bulk density lower than that of maximum specific impulse or higher than that of maximum impulse density.  Anyway, there's nothing very profound about this, but I thought I'd mention it, because it hadn't occurred to me before.


    1. Specifically, these values are 95% of the ideal vacuum values calculated with RPA Lite; the densities of both propellants correspond to those at their respective normal boiling points.
    2. Note, though, that the mixture-ratio axis is non-linear, because bulk density is not a linear function of mixture ratio (though specific volume, the reciprocal of density, is).

    Good point. For perfect burning, mixture ratio should be 16. But increasing the amount of hydrogen by factor of 3 is the normal way of doing hydrolox engines, only less than 1/3 of all the hydrogen is burning.

    But optimizing for isp by using a lot of extra hydrogen also makes the impulse density very bad as the extra hydrogen still consumes a lot of space.

    Most air-lit/upper stage engines(RL-10,J-2) use mixture ratio 5.5 and ground-lit/sustainer or booster engines use mixture ratio of about 6(RS-25, RS-68).

    It would seem that RS-68 uses too small mixture ratio;

    Raising the mixture ratio from 6 to 7 might decrease isp by about 2%, but it would increase impulse density (and thrust) of same-sized engine by 10%. However, for a 5.5km/s rocket stage and ~400s engine, the 2% isp drop means that only about 3% more propellant is needed. So Delta IV could manage with ~7 % smaller tanks and lighter engines if the mixture ratio would b 7 instead of 6.

    Raising it to 8 instead would decrease isp about 4%, and impulse density and engine T/W would increase by about 17%. 4% of lost isp would cost only about 6% fuel for the 5.5 km/s stage, so about 10% smaller tanks and lighter engines would do.

    So what am I missing here? Why did they make RS-68 to use se low mixture ratio?


    Going much more 8 does not seem to make much sense as after 8 isp start to decrease by a lot and the impulse density increase starts to get much smaller after about 7.5.




    « Last Edit: 04/14/2016 09:01 PM by hkultala »

    Online TrevorMonty

    Re: RP-1, methane, impulse density
    « Reply #77 on: 04/14/2016 09:31 PM »
    This from Jeff Bezo interview. I imagine SpaceX are thinking along same lines hence Raptor. NB BE4 uses LNG (typically 95% methane) so it is cheaper than pure methane.

    http://www.geekwire.com/2016/interview-jeff-bezos/

     Itís why we are using liquid natural gas [for the BE-4 engine], because our goal is to make spaceflight so cheap that the cost of the fuel actually matters.

    Right now, the cost of the propellant is minimal compared to throwing the hardware away. But once we can get to a place where we are not throwing the hardware away, and we have real reusability, then we want to be using a fuel that is very low-cost. And nothing is lower-cost than liquid natural gas.



    Sent from my ALCATEL ONE TOUCH 6030X using Tapatalk


    Offline Proponent

    • Senior Member
    • *****
    • Posts: 5478
    • Liked: 991
    • Likes Given: 653
    Re: RP-1, methane, impulse density
    « Reply #78 on: 04/14/2016 10:12 PM »
    To further develop my understanding of the trade between specific impulse and density, I've done a little thought experiment on ground-launch stages.

    In 1996, John Whitehead wrote a cute little paper about SSTO mass budgets (4th attachment to this post.  For a few different propellant combinations, he used the rocket equation to calculate the mass ratios needed for a delta-V of 10 km/s (i.e., Earth to LEO with losses).  Then he estimated the masses of engines, tanks, pressurants and residual propellants (the last two can be larger than you expect) as a fraction of burn-out mass.  Let's call the part of the burn-out mass that's not devoted to those four things the available mass (I'm open to suggestions for a better term).  Many subsystems will have to be crammed into this so-called available mass:  landing gear, if any, avionics, etc., etc.  The idea, though, is that the masses of such subsystems will be approximately independent of the propellants chosen.  Hence, the vehicle with the highest available mass fraction should have the largest payload fraction as well.

    The name of the game, then, is to choose the propellant combination that maximizes the available mass.

    About the same time as Whitehead's paper, Bruce Dunn presented an analysis in a similar spirit (3rd attachment).  Although Dunn's assumptions were perhaps a bit more ad hoc, he covered a wider range of propellants.

    I've made a similar calculation similar to Whitehead's.  There are just two major differences.  Firstly, Whitehead assumes that the lift-off thrust-to-weight ratio of the engine is a linear function of propellant density and is 100 for lox/RP-1 and 50 for lox/hydrogen.  In contrast, I assume the ratio is proportional to the impulse density of the propellants (it seems to me this makes more sense; any comments?), taking a value of 123 for lox/RP-1 at a typical mixture ratio (essentially, the NK-33 or the AJ-26).

    The second significant difference is that rather than assuming a particular mixture ratio, I adjust the mixture ratio for maximal performance.

    Otherwise, to oversimplify slightly, I use pretty much the same assumptions:  10 km/s of delta-V, tanks weigh 10 kg/m3, pressurants and residuals are each 0.25% of the initial propellant load.  Specific impulses come from RPA Lite 1.2.8 and are scaled by 0.95 from ideal vacuum values.  Chamber pressure is 20 MPa and the area expansion ratio is 40:1.  For the time being, propellants are assumed to be at the lower of room temperature and the normal boiling points.

    Have a look at the first plot attached.  It shows specific impulses delivered by various fuels1 burned with oxygen as a function of propellant bulk density.  Also shown as grey curves are contours of constant "available mass."  These contours are easily calculated, since all that's required in Whitehead's model is a specific impulse and a propellant density.  The first table, below, gives optimal figures for each of 30 propellant combinations.

    Hydrogen does poorly.  If the mixture ratio is allowed to vary during flight in an optimal way, the available mass fraction with hydrogen as a fuel increases2 by about 0.026.  Other fuels don't benefit much from mixture-ratio variation, so the this enough to boost hydrogen to the middle of the table.  But, the substantially larger mass of hydrogen tanks arising from the need to insulate them has been neglected.  Taking this into account would knock hydrogen right back to the bottom of the table.

    Speaking of the table, a couple of columns may not be self-explanatory:

    * Mix:  Linear function of the mixture ratio, being zero for maximum Isp and unity for maximum impulse density.
    * T/W:  Thrust-to-weight ratio of the engine at lift-off (giving the a ratio of 1.3 for the vehicle).
    * Den exp: slope of the log Isp-log(bulk density) curve at the optimum; shows the relative importance of density compared to Isp.

    People often obsess about maximizing specific impulse.  The Mix column shows that's not generally what you want to do.

    The "Den exp" column shows the relative sensitivity of available mass fraction to density as opposed to specific impulse.  For the better performing propellant combinations, it's about 0.23, meaning that a the figure of merit is approximately:

        (specific impulse)(bulk density)0.23

    for an SSTO.  This is, of course, somewhat model dependent, but it happens to be about the same as what I estimated from Dunn's results some time ago.

    OK, so, what about hydrogen peroxide, with its high density?  Please have a look at the second plot.  This time I've left hydrogen out so as to make the hydrocarbons more visible.  As you easily see, peroxide's density does not raise bulk density enough to make up for its lower specific impulse.  Bruce Dunn told us that a long time ago, but I find it educational to see it graphically.  I also looked at nitric acid, which is even denser (1510 kg/m3) than peroxide (1460 kg/m3).  It, however, suffers from lower specific impulse and lower bulk density than you might expect: the fact that it contains quite a bit of free oxygen means that mixture ratios with nitric acid tend to be low.

    If we consider a delta-V of just 4 km/s -- see the third plot and second table -- peroxide looks much better.   As you'll see from the table, the figure of merit at this delta-V, which could correspond to a first stage or a martian SSTO, is something like:

        (specific impulse)(bulk density)0.4 ,

    Finally, consider a very low delta-V, like 40 m/s, as shown in the final plot.  In this case, impulse density reigns, and peroxide is the run-away winner.   The associated table shows that the figure of merit is very close to

        (specific impulse)(bulk density) ,

    i.e., impulse density, which is just what you expect when delta-V is small compared to exhaust velocity.  Note, though, that we do have to go to very low delta-V's before impulse density dominates.

    All of the above is applies to ground-lit stages.  For upper stages, mass will be more important, since the stage's propellant must be accelerated by lower stages.  Hence, the density exponent in the figure of merit will tend to be smaller.



    1.  Except for JP-5 (the composition of which I don't know), the color of each curve is the number of carbon atoms, modulo 10, in each fuel's principal chemical component (e.g., 1 for methane, 2 for ethane and ethylene) expressed in the resistor color code.  Solid lines are used for saturated hydrocarbons (alkanes).  The two alkenes, ethylene and proplylene, are shown with dashed lines.

    2.  If a different mixture ratio is allowed for each successive 1% of the total propellant volume, the ratio ranges from 17.8 (633 kg/m3) at lift-off to 5.7 (350 kg/m3) at burn-out.  A variable mixture-ratio program helps in two ways.  Firstly, it simply helps with the rocket equation by allowing more impulse to be packed in at the beginning, where mass doesn't matter so much, while going for higher specific impulse at later times.  Secondly, it increases the lift-off thrust-to-weight ratio of the engine, allowing for a smaller engine.

    EDIT:  Added "bulk" to very-low-delta-V figure of merit.
    « Last Edit: 04/15/2016 08:01 AM by Proponent »

    Offline Proponent

    • Senior Member
    • *****
    • Posts: 5478
    • Liked: 991
    • Likes Given: 653
    Re: RP-1, methane, impulse density
    « Reply #79 on: 04/14/2016 10:20 PM »
    Good point. For perfect burning, mixture ratio should be 16.

    You mean 8, right?

    Quote
    It would seem that RS-68 uses too small mixture ratio;

    Maybe it's partly to increase the thrust-to-weight ratio.

    Tags: