Jump to content
Enthalpy

Expansion Cycke Rocket Engines

Recommended Posts

The expander cycle pumps oxygen to the chamber, and hydrogen to a higher pressure into the cooling jacket where it evaporates. Lukewarm hydrogen passes through a turbine and flows in the chamber. It gives the RL-10 and heirs good performance, simplicity and an easy start.

Operation with methane was demonstrated. Though, an engine burns fewer methane moles with more oxygen moles; I propose instead to pass the oxygen through the jacket and the turbine, which enables higher pressures hence improves the performance.

 

post-53915-0-75933100-1395692928.png

 

I compare with 400K before the turbine, 74% and 79% efficient pumps and turbines, injectors keeping 88% of the pressure - and no drop in the cooling jacket. The Vinci wastes much pressure there, but splitting the flow in shorter parallel sections should improve that. I've taken no shaft power margin neither. The nozzle expands to 7kPa. For oxygen/methane, the mass ratio is 3.64:1.

  • Methane can expand from 402bar to 148bar in the turbine; this puts 130bar in the chamber to achieve isp=373s.
  • Oxygen can expand from 473bar to 174bar, for 153bar in the chamber and isp=375s.
  • Both. Evaporate at different sections of the chamber, or at different chambers, pass through distinct turbines. Separate shafts would bring nothing, but if sharing power, this achieves 540bar / 199bar / 175bar for isp=378s.

Heat available from the jacket limits the thrust to upper stages. Vinci's oversized chamber achieves 180kN with hydrogen; heavier and more compact methane and oxygen bring it to half a meganewton per chamber of the same size. Cooling wih methane favours thrust, oxygen efficiency.

A cryogenic propellant helps the expander cycle start; expanding oxygen permits storable fuels, which are safer than methane. Here cyclopropyl-azetidine has carefully hand-estimated Hf=+150kJ/mol and roughly 900kg/m3, but it could be diazetidine, diazaspiropentane, and methylated instead of cyclopropylate: choose for flash point. The denser fuel permits 569bar / 209bar / 184bar for isp=375s, as good as methane and safer. Since the fuel isn't heated, something like trivinylamine may be stable enough, and it achieves isp=377s.

Marc Schaefer, aka Enthalpy


==================================================================

 

The bleeding variant of the expander cycle evaporates a fraction of a propellant in the cooling jacket, extracts power in a turbine, and ejects this fraction. The bigger pressure ratio across the turbine extracts more work from the heat, so more propellants can be pumped. This cycle powers the 137kN LE-5B, but was said too weak for Ariane 6's lower stage.

 

post-53915-0-70403100-1395693165.png

 

Though, this cycle can power a strong engine, if pumping methane rather than bulky hydrogen, as on the left diagram.

Vinci's long cooling jacket provides 21MW heat for 180kN thrust, so 40MW from a roughly D0.5m*h0.5m chamber bring 43kg/s methane from liquid 112K to gaseous 300K (yes, supercritical). Expansion from 80bar to 4bar lets the 79% efficient turbine extract 10MW plus the pumping power for the bled methane, which expands from 181K to 0.7bar and about 513m/s in a secondary nozzle to push 22kN, say as a roll actuator.

74% efficient pumps bring 672kg/s of 333:100 oxygen and methane to 80bar chamber pressure plus 11bar drop across the injectors or the jacket. Expansion to 0.7bar and 3135m/s (D~1.6m) pushes 2.1MN. The bleeder and four chambers with one turbopump push 8.4MN with isp=2969m/s=303s in vacuum, and nearly 7.6MN with isp=2681m/s=273s at sea level. The turbine and pumps have compatible speeds. As strong as the RD-171, more flammable and less efficient, but easier.

Burning the bled methane with extra oxygen worsens. Less thrust and more pressure improves the isp a bit. Heating the bled methane at the nozzle's bottom, as Vulcain 2 does, would also improve; better if feasible, inject the bled methane deep in the nozzle.

----------

On the right diagram, the jacket heats 8.9kg/s hydrogen from liquid 20K to gaseous 300K and 20bar, then expansion to 4bar provides 11MW to pump 675kg/s of 252:100 oxygen and here cyclopropane to 100+14bar. The bleeder nozzle expands to 1538m/s and pushes 14kN, helping four chambers to push 8.6MN with isp=3178m/s=324s in vacuum. The tanks volume ratio is 100:67:30; >20bar would save hydrogen.

Again, thrust can be traded for pressure, and reheating or injecting the bled hydrogen improves.

With two stages on one or two shafts, the turbine still spins at ~630m/s, too much for centrifugal pumps - but axial pumps maybe, with a single stage resembling a booster pump.

Marc Schaefer, aka Enthalpy


==================================================================

 

A bleeding expander cycle for hydrogen can be strong after all. Here, several chambers heat hydrogen by 40MW each, their common two-stages hydrogen turbine is 70% efficient, the hydrogen pump 70%, the oxygen pump 74%, the injectors or cooling jacket leave 88% of the pressure.

 

post-53915-0-17278700-1395693384.png

 

On the left diagram, 10.7kg/s hydrogen per chamber at 100+14bar are heated from liquid 20K to 250K. Expansion from 100bar to 5bar provides 1.45MJ/kg work, of which pumping this hydrogen flux leaves 1.22MJ/kg. From 5bar and 149K (losses), the D=1.3m 0.5bar nozzle obtains 1422m/s.

Pumping 620:100 oxygen:hydrogen to 100+14bar takes 43.6kJ/kg, allowing 299kg/s per chamber to accelerate to 3998m/s through the 0.5bar nozzle and push 1.20MN.

The fluxes add to isp=3909m/s=399s in vacuum and 3459m/s=353s at sea level, nearing an RS-68. Obtain 4.8MN from 4 chambers, 7.3MN from 6: enough for a first stage.

----------

The right diagram re-heats hydrogen between the two turbine stages. Cooling the wall at the junction is challenging. From the same heat, it extracts some work more, and bleeds less hydrogen. This gains 2s isp and 4% thrust.

----------

The turbine speed fits a single-stage hydrogen centrifugal pump. An oxygen pump for the same shaft speed may look like an inducer, here from the M-1; symmetrize for throughput and balance.

 

post-53915-0-25466700-1395693432.png

 

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

From the limited heat power, an expander cycle can pump more propellant to a lower pressure, and push stronger. I compare with Pmdeta, and 40MW per chamber to heat oxygen, as above.

Because evaporation takes heat, one better keeps the same flow through the cooling jacket to obtain the same temperature, and injects the additional flow directly in the chamber. Under identical assumptions (pressure /2.7 across the turbine, pumps 74%, turbine 79%, injectors and jacket 88%, expansion to 0.7bar in vacuum, mix detuned by 5m/s for thrust), I get:

  • 165kg/s propellants at 100bar, isp=319s, and 516kN per chamber
  • 206kg/s at 80bar, isp=314s (-5s), 634kN (+23%) per chamber

As the added flux doesn't pass the turbine, it needs only the injection pressure. Two pump stages save then power (and fit other rotation speeds better).

  • 273kg/s at 80bar, 840kN (+63%) per chamber

post-53915-0-01090000-1396188304.png

 

Because a stronger engine has a bigger chamber, the improvement is quicker than this comparison at constant cooling power.

Marc Schaefer, aka Enthalpy

 

====================================================================

 

This launcher obtains 4.5MN at lift-off from a six-chambers engine that expands a fraction of the oxygen and injects the rest in the 80bar chamber directly. The other stages have four chambers and expand all oxygen to obtain 105bar in the chambers. Details on the sketch, click for full size:

 

post-53915-0-90640600-1396188211_thumb.png

 

The fairing and the lower stages' structural tanks are vertical extrusions, as I describe there

http://www.scienceforums.net/topic/60359-extruded-rocket-structure/

the 1-2 interstage breaks at 12MN and 13MN*m, the rest is stronger. The tanks weigh 17 and 19kg per ton of dense propellants, details here, click for full size:

 

post-53915-0-33351000-1396188516_thumb.png

 

The 2-3 interstage is a welded hexagonal truss of aluminium tube, as is the structure of the third stage. The oxygen tank there is a balloon of brazed steel insulated with foam and multilayer sheet, and is hold by polymer belts. The Pmdeta tank is aluminium sheet welded to the truss.

The inert masses would go to Gto in two stages, but Leo plus an optional third stage is more flexible and efficient. Storing oxygen for days and weeks, the third stage can go from 34.3° Leo to Gso (4432m/s), Mars or Jupiter transfer, Moon orbit and departure from there.

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

This launcher uses hydrogen expansion cycles at both stages: a Vinci in the lighter Esc-B I describe there
http://www.scienceforums.net/topic/60359-extruded-rocket-structure/page-2#entry761740
and the bleeder expander I describe there
http://www.scienceforums.net/topic/82346-expansion-cycke-rocket-engines/?p=797705

post-53915-0-89581000-1396224672_thumb.png

The upper stage starts with 28.2t at 3000m/s below Leo. The Vinci gives it 0.65G, more than the 0.35G for Delta IV M+(5,2), so my Esc-B design can be more filled, or it can be shrinked, and the payload improves. The first stage can also nearly double with a four-chamber engine, or get side boosters.

The first engine expands from 100 to 0.3bar in 2*D1,8m for 4022m/s and 2*1.60MN in vacuum, 3908m/s and 2*1.34MN at sea level. It must throttle to 50% or 40%. Jettisoned nozzle inserts, to limit the expansion to 0.8bar in the atmosphere, and larger nozzles, would gain much - or passive flow separators.

The first stage has extruded tanks of AA6005A, described there
http://www.scienceforums.net/topic/60359-extruded-rocket-structure/

  • The interstage takes t1=t2=1mm a=60°
  • The oxygen tank, t1=2.1mm t2=1mm a=45°
  • The hydrogen tank, t1=1.6mm t2=1mm a=45°

The upper and lower heads are of 1.2mm and 1.4mm AA7022, the middle of 4mm AA6005A. With 10mm foam, the tanks weigh 6119kg or 35kg per ton of propellants. Thinner extrusions of harder alloy would improve, magnesium maybe.

Add 3.5t for the engine, 0.9t for one-third of the fairing and the Esc-B shell, 2t undetailed items.

Small dry masses let two stages reach Gso, Moon orbit and leave, and even a Jupiter transfer marginally.

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

If an engine flows the oxygen instead of the fuel in the jacket to cool the engine wall, for instance if this oxygen rotates the turbine, then the engine can burn ethylene or other fuels unsuited for high temperature.

Ethylene is flammable (detonation speed between methane and acetylene) but it brings performance second to cubane only. Here a comparison at the pressures estimated on 30 March, 2014 for a first and a second stage.

post-53915-0-10628400-1411088357.png

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

The Press reports that a studied variant of Ariane 6 has a reuseable first stage, stronger to fly with 0 or 2 boosters where the Vulcain-pushed stage would take 2 or 4. This reused first stage would burn methane with oxygen. My two cents worth of comments:

  • Cyclopropane is as efficient as methane at equal pressure but denser, so the turbopump achieves more pressure.
  • Strained amines that don't burn at room temperature lose only 4s to cyclopropane and methane.
  • Pmdeta must leave the jacket cleaner than kerosene does, cis-pinane maybe. Both are better available, more efficient, and safe.
  • Cooling the jacket with oxygen isn't done now but it would leave the engine clean for reuse. It permits ethylene that outperforms cyclopropane and methane by 2s, and in an expander cycle, safe fuels.

I'd already have proposed reuseable pressure-fed boosters for Ariane 6 if data about mass and dV were public
http://www.scienceforums.net/topic/65217-rocket-boosters-sail-back/
with oxygen and Pmdeta, sailing back using the paraglider as a kite.

Share this post


Link to post
Share on other sites

Cooling jackets waste pressure from the propellant flowing through. How much is unclear: seemingly 20% at the RL10-B2, losing 2s specific impulse, at the Vinci possibly more.

The azimutal flow direction is short and broad, as opposed to the axial one, that's how I want to save propellant pressure, and several inlets and outlets in 360° would improve further. Though, the fluid layer near the surface must be replenished quickly for good heat transfer, and a slower fluid may need help for that, for instance by successive jumps that achieve a vortex pair in each channel.

post-53915-0-19218800-1462053188.png

----------

Ariane 6 could use an eight-chamber Vinci at the main stage, with a common set of turbopumps and actuators, and D=1.3m nozzles on a circle in the D=5.4m body. This matches Vulcain's thrust (8*171kN vs 1359kN), gains 14s (Isp=443s vs 429s) and hopefully dry mass, especially if the uncooled nozzle section is of niobium instead of SiC - and must be cheaper.

The D=1.3m nozzles expand to 0.07bar hence can't start at ground level. The P120 push enough for Ariane 62 and 64, and the Vinci can start in flight. A hypothetical Ariane 60 would stop the nozzles at the cooled D=0.7m section, where the expansion to 0.34bar can be stable in the air and provides Isp = 4025m/s = 411s and 8*159kN in vacuum, and at sea level Isp = 3037m/s = 310s and 8*120kN. Inducing voluntarily a clean flow separation, for instance with the means I describe there
http://saposjoint.net/Forum/viewtopic.php?f=66&t=2411
would enable the D=1.3m nozzles and improve the sea-level performance to Isp = 3270m/s = 334s and 8*129kN to lift-off 88t, enough to put 7t in Leo.

Ariane 6 has four places for strap-on solids (to reuse the Ariane 4 launch pad?) but these could be liquids too
http://www.scienceforums.net/topic/65217-rocket-boosters-sail-back/#entry915939
or combinations with smaller solids to fine-tune the performance. For instance the Zefiro 23 works at sea level, its stronger successor supposedly too.

Marc Schaefer, aka Enthalpy

Edited by Enthalpy

Share this post


Link to post
Share on other sites

Heard recently about the new company Isar Aerospace developing the Spectrum launcher
Isar Aerospace - Spectrum launcher
whose general design is a close copy of Falcon 9, with 1 Aquila engine at the second stage, 9 Aquila at the first stage, liquid oxygen and an unspecified "mix of light hydrocarbons", to put 1000kg in Leo or 700kg in Sso. The launcher is meant for constellations.

As they hire design engineers for basic motor components, the design might still evolve. So here are my 2 cents.

Electric pumps (in this discussion) reach easily Sso with two stages. They're easier to restart, simpler to develop, maybe cheaper, and turbines can replace the electric motors later.

A common set of actuators can move the 9 first-stage engines. This avoids collisions and can be lighter and cheaper. A (turbo-) pump common to the 9 engines can be cheaper. But it goes against design reuse and redundancy.
scienceforums

Spectrum is as sleek as Falcon-9 but the next big thing, the sunheat engine, needs wide fairings
scienceforums and more

Thoughts about igniters there, for instance with a Diesel hotplug
nasaspaceflight

A new engine and launcher can burn amines. Easier to produce than synthetic hydrocarbon homologues, some are industrial compounds, gain 3s, denser. Comparison of launcher sizes there
chemicalforums
Examples of very safe and of more efficient amines there
chemicalforums - chemicalforums - chemicalforums

Heavier hydrocarbons (pinane, farnesane etc) or amines (Pmdeta etc) aren't easily flammable.
chemicalforums
I would not replace them with dangerous methane for its small gain. Spiropentane and cyclopropane are denser and more efficient. Unstrained light alkanes are both dangerous and no better than heavy alkanes. Some strained heavy alkanes seem reasonable to produce, safe and 5s better than "kerosene"
chemicalforums
Cooling the engine with oxygen would enable ethylene and other unsaturated fuels
chemicalforums
that gain 2s-6s over methane and are denser.

Marc Schaefer, aka Enthalpy

 

Share this post


Link to post
Share on other sites

I believe electric pumps are the easiest to restart and to develop fo a small two-stage launcher, but if Spectrum keeps a gas generator cycle, options vary.

Many gas generator burn the same propellants as the main chamber. It only seems simpler:

  • Some other means must start the turbopump.
  • A re-starting means must survive the normal operation.
  • Spectrum targets many orbits per launch. Solid pyro hardly re-starts the second stage many times, liquid pyros are dangerous or difficult. Or the manoeuvres demand a separate attitude and orbit control.
  •  The turbine demands a strongly de-tuned mix of propellants, which soots or is very oxidizing. Adding water to a tuned mix adds complexity, as on the Viking engine.

A separate monopropellant to feed the turbine is simpler and efficient.

  • I propose to store the monopropellant under pressure in some vessel, probably of graphite fibres. Lighter than a pump for this volume.
  • Then, a mere valve on the monopropellant starts the turbopump.
  • Some monopropellant can also ignite the main chamber, once decomposed by a catalyst. Again mere valves.
  • If desired, that monopropellant alone can control the attitude and orbit of the upper stage.

Soyuz' RD-108 pumps the monopropellant, but its design is older than graphite fibres and good steel. The opposite choice is obvious today.

If keeping the dangerous hydrogen peroxide, 86% mass concentration provides 650°C survivable by a nickel alloy turbine, good to ignite hydrocarbons, and its expansion from 100bar to arbitrary 5bar brings 1240m/s that fit a single turbine stage while 5bar provide the roll control to a single-chamber upper stage.

Monopropellants outperform bipropellants detuned to limit the temperature at the turbine, as their gas molecules are lighter and attain a higher speed, providing more energy per mass unit.

Hydrazine and MMH are more dangerous than peroxide, I exclude them.

But maybe some amine mixes that contain guanidine or an aminoguanidine might replace hydrogen peroxide? See there
scienceforums
Methylamine decomposes over catalysts, but the reaction is unstable. Hotter reactions from guanidines hopefully proceed decidedly.

Mixes of ammonium dinitramide, maybe with water, are considered too
scienceforums
they "just" need some mix that decomposes only where desired, say over a catalyst, not in the tank.

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

A gas generator cycle, as Spectrum's website claims to use, achieves about 80bar in the main chamber (for detuned kerosene+oxygen. Peroxide is more efficient). An expander cycle exceeds 94kN per chamber, as Vinci shows. The unmodified RL-10 has already run on methane. With a dense fuel, an expander cycle achieves a stronger chamber pressure than a gas generator cycle, and it and wastes no propellants at the turbine. It restarts naturally.

I've already put figures on a cycle where oxygen cools the chamber and powers the turbine
scienceforums
The unit thrust exceeds much Aquila's 94kN meant for the Spectrum launcher, so the chamber can have a more usual aspect ratio than Vinci has.

I keep uniform 184bar in the chamber as already computed for dense fuels (153bar with methane) but expand to 0.7bar in vacuum for the first stage. Cooling with oxygen enables safer or more efficient fuels:

Flam  Fuel                      Isp
=========================================
 !!   Bicyclobutane             348
 !!   Ethylene                  345  <<<
 !!   Spiropentane              344
good  Dicyclopropyldiazetidine  344
 !!   Cyclopropane              343
 !!   Azetidine                 341
GOOD  Diazetidylcyclopropane    339
GOOD  Trispiroundecane          339
 !!   Methane 153bar            339
GOOD  Pmdeta                    338  <<<
good  Beta-pinene               336
GOOD  Farnesane                 335
good  Cis-pinane                335
=========================================

The very safe, very cheap and mass-produced Pmdeta is as efficient as methane, while flammable fuels beat it. Dicyclopropyldiazetidine may be unhealthy.

Denser fuels also shrink the launcher and save dry mass
chemicalforums

Production methods are known for some fuels, I suggested others elsewhere. Beware I'm no chemist.

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

I am a chemist; but not a rocket scientist.
Can you explain why a molecule like PMDETA would be a good fuel?

It's got a lot of nitrogen in it which is "carried along for the ride".

It's loosely equivalent to adding water to the fuel. It adds weight, but not energy.

There may be times when that's a good thing, but rocketry isn't one of them.

Share this post


Link to post
Share on other sites

As an igniter, decomposed peroxide could get a bit of fuel in its injection recess at the main chamber to amplify the heat, then possibly some more oxidiser and fuel.

I have no plan to evaluate the best chamber pressure when a monopropellant powers the turbine(s). 108bar is for sooting "kerosene" and oxygen at Merlin 1D+. Does a peroxide or guanidine gas generator outperform an oxygen expansion cycle? Two turbine stages, on one or two shafts, bring little more power.

I see no excellent reason to want more pressure in the igniter monopropellant tank than in the main chamber after ignition. However, the igniter monopropellant can have a tank separate from the turbine monopropellant.

A monopropellant safer than peroxide would be nice, but has research already achieved safe fast decomposition? How much experimentation will a new launcher company do? Peroxide is well known, even if the 3% pharmacy concentration gives wrong impressions.

==========

Propellants, pumps, chambers... are much optimized for performance, but at the first stage, an over-expanded nozzle brings more with less effort.

Take an engine that pushes 94kN in vacuum by expanding O2 and Pmdeta from 100bar to 0.7bar in a D=0.34m nozzle. Sea level pressure lets lose 9kN.

Hardware that limits the expansion to 1bar at sea level gains very little thrust, but it allows wider nozzles that would be destructive or inefficient at sea level. D=0.60m fits on a 2.0m circle (easier with a common set of actuators), for instance under a 2.5m body that matches one fairing and eases a wider one. Then, expansion from 100bar to 0.17bar rather than 0.7bar gains 29s at once. Acting over 2/3 of the first-stage time, it's as much as 10s Isp over the whole flight. Raising the chamber pressure from 100bar to 150bar everywhere would bring the same improvement from a much harder effort.

The RD-0120 nozzle insert looks reasonable and has flown. Literature gives other designs. I described some at saposjoint.net, now closed.

In a 2.5m body, the second stage's nozzle too can be wider. Again for 94kN in vacuum by expanding O2 and Pmdeta from 100bar: D=0,76m achieves 9.5kPa and Isp=362s (minus the secondary flux) while D=1.6m achieves 1.5kPa and Isp=387s, 25s Isp gained over more than half of the flight.

Plansee and others sell niobium for the uncooled light nozzle extensions. The interstage truss can use aluminium tubes. There
scienceforums
I described a light and discarded shell around the truss.

Zenit has a toroidal tank to save height, and rolls to guide the separated interstage around the wide nozzle. Good ideas should spread.

Marc Schaefer, aka Enthalpy

Hi JC, thanks for your interest! I come back very soon.

Share this post


Link to post
Share on other sites
4 hours ago, John Cuthber said:

[...] Can you explain why a molecule like PMDETA would be a good fuel?

It's got a lot of nitrogen in it which is "carried along for the ride". [...] It adds weight, but not energy. [...]

In an amine reactant, nitrogen is present with relatively weak C-N and N-H bonds, sometimes N-N, while in the product it's the strong triple bond N#N. So nitrogen isn't a spectator, it's an actor of the combustion.

The comparison is too close to tell by qualitative arguments. If taking real heats of formation for the fuels, the combustion of amines gives some 20m/s bigger ejection speed than the alkanes homologues. For hydrazines it's rather 80m/s bigger. Replacing an unstrained alkane by a string of cyclopropyls gains only 40m/s.

Amines are easier to produce than alkanes. They often burn more smoothly. They may stink and emit more NOx. Hydrazines are being phased out. Amines need less oxygen, rarely an advantage. They are denser.

Share this post


Link to post
Share on other sites

The production of ammonia (and other amines) is exothermic.

You would get more heat in your rocket  by burning hydrogen than by reacting it with nitrogen, letting the heat produced by that dissipate (in the ammonia factory) , and then burning the ammonia in the rocket.

10 hours ago, Enthalpy said:

Amines are easier to produce than alkanes.

Not really.

You can drill a hole in the ground and get alkanes. (admittedly, you can piss in a pot and get ammonia)

Until the Haber Bosch process came on stream, it was difficult and expensive to make ammonia- a state of affairs that hampered food production.

The production of ammonia uses  about 1 or 2 % of the human race's energy consumption.

That energy is generally obtained by burning alkanes.

The hydrogen for the industrial production of ammonia is derived, on the whole, from alkanes.

10 hours ago, Enthalpy said:

They are denser.



If you look at this chart of energy densities

https://en.wikipedia.org/wiki/Energy_density#/media/File:Energy_density.svg

you will see why they use hydrogen- it's all the way over to the right 
The alkane fuels (LPG, petrol, diesel etc ) are all pretty much lined up at about 45 MJ/Kg

and ammonia and hydrazine- the only amines on the chart- are well over to the left.

You could make kerosene denser by adding sand to it, but that wouldn't make it a better rocket fuel.

Share this post


Link to post
Share on other sites
11 hours ago, Enthalpy said:

The comparison is too close to tell by qualitative arguments. If taking real heats of formation for the fuels, the combustion of amines gives some 20m/s bigger ejection speed than the alkanes homologues. For hydrazines it's rather 80m/s bigger. Replacing an unstrained alkane by a string of cyclopropyls gains only 40m/s.

Surely it's momentum you want not velocity per se.

I haven't computed exhaust momenta, have you calculated these for comparison ?

Share this post


Link to post
Share on other sites

Thoughts about a catalytic decomposition chamber for a monopropellant, hydrogen peroxide or other. Strongly inspired by the stone-old RD-108 design
lpre.de
see there after the chambers and the turbopump. Google Translate may help.

Rather than loose coarse grains of catalyst in a sieve, I'd have a honeycomb winding of wire sintered together. It's extremely robust: I used some, of stainless steel of several diameters, as a filter for air at 200m/s. Maybe it's a refractory metal or alloy here, or a ceramic. Fine wire gets faster hot and offers a big surface to deposit the catalyst on, maybe iron subsequently oxidized, or some transition metal.

Catalyst.png.5a27e3607f50fa69daeab58aa88a39b3.png

Radial flow gives a wider shorter path to reduce the pressure drop and the forces on the catalyst. It spares the heat losses at most chamber's walls. Optionally, two gas outlets as sketched shrink hence lighten the chamber.

A helix at the gas flow can let the gas rotate to eject residues of liquid and foam on the catalyst. Then the gas outlets can be scrolls.

Seemingly better: the helix or helices could rotate freely. The gas accepts smaller outlets, and the helices eject the liquid better. Maybe ceramic ball bearings can operate under these conditions.

Marc Schaefer, aka Enthalpy

Hi JC and Studiot, I'll be back!

Share this post


Link to post
Share on other sites

Figures, drawings and opinions about a body wider than the 1.8m announced for the Spectrum launcher. At arbitrary 100bar chamber pressure, and including some 7s lost in the peroxide gas generator - a expansion cycle may be more effective.

Mass and speed estimates tell me payloads much heavier than 1000kg from 53t at lift-off, with nine 1st stage nozzles reduced now to D=0.5m for 0.25bar on a D=1.8m circle, with D=0.27m inserts at sea level. The D=1.8m fairing wouldn't serve then.

WideBody.png.72a94694f39d6d2d5ac1d19e9baf6081.png

My sunheat engine is the next big thing
scienceforums
and already 700kg hydrogen fit in D=2.5m as depicted but not properly in D=1.8m. But deep space missions could even carry 1200kg hydrogen, which needs a fairing wider than D=2.5m. By the way, the sunheat engine needs trickle hydrogen at the launch pad under the fairing, sorry for that.

Strap-on boosters can multiply the lift-off mass by 2.1 with little development, and then the fairing must be wider than D=2.5m and longer.

The two stages reach marginally a transfer to Gso, Moon, Mars, Venus - hence the bigger tanks at 2nd stage. A true 3rd stage starting from Leo reaches nicely Gso, the lunar surface, transfers to Mars, Venus and Jupiter. Then the launcher gets taller.

My hectares solar sail makes impressive missions
scienceforums
and it fits nicely in Spectrum's mass capability, but it needs a long fairing for the telescopic booms.

Longer and wider bodies and fairings, heavier payloads increase the bending loads that dimension small launchers. I feel a D=1.8m body too difficult here.

==========

If ending 0.2mm thin, a nozzle extension of niobium from D=0.8m to D=1.6m weighs 15kg and gains 150kg in orbit. It needs a light interstage. A four-chamber engine is a different option.

Light interstages are easily made of a truss. A truss can also hold superinsulated light balloon tanks as I described
scienceforums
which I believe Ariane 6 has adopted for the upper stage, and SpaceX seemingly too for the BFR.

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites
Posted (edited)
1 hour ago, Enthalpy said:

My sunheat engine is the next big thing

You put a lot of time and effort into these threads, Do you ever do anything with these ideas of yours, like apply for patents or seek funding?

Edited by Curious layman

Share this post


Link to post
Share on other sites
On 12/30/2020 at 1:13 PM, studiot said:

Surely it's momentum you want not velocity per se.  [...]

Hi Studiot, thanks for your interest!

Where the propellant makes a small fraction of a vehicle's mass, the momentum is fully relevant. There, more propellant brings you further or faster. But a launcher's mass is essentially propellant, plus very little tank and engine and payload mass. More propellant pushing on more propellant wouldn't increase the speed. This suggests that momentum doesn't suffice.

As the rocket's mass (or rather the stage's mass) varies much while it uses propellant, the proper equation is
DeltaV = Ejectionspeed * Log (Initalmass / Finalmass), called Tsiolkovski's equation
and because the nearest orbit costs around 9500m/s but propellants achieve 3000-4600m/s ejection speed, launchers need a big ratio Initalmass / Finalmass.

This implies that all dead mass must be small in a launcher, and even then, they consume a strong part of the final mass, leaving less for the payload. Stages improve that by throwing away much dead mass, but even good designs used in a favourable mission waste easily 1/4 of the final mass for each stage.

Put together, the ejection speed (or divided by g=9.806m/s, the specific impulse Isp in seconds - plus subtleties if a secondary flux is lost) is all-important for a launcher. Already the Log would let 1% more Isp gain 3% on the final mass, but on the payload mass it acts even more. So the ejection speed rules, and designers make big efforts to improve it.

Share this post


Link to post
Share on other sites
22 minutes ago, Enthalpy said:

Hi Studiot, thanks for your interest!

Where the propellant makes a small fraction of a vehicle's mass, the momentum is fully relevant. There, more propellant brings you further or faster. But a launcher's mass is essentially propellant, plus very little tank and engine and payload mass. More propellant pushing on more propellant wouldn't increase the speed. This suggests that momentum doesn't suffice.

As the rocket's mass (or rather the stage's mass) varies much while it uses propellant, the proper equation is
DeltaV = Ejectionspeed * Log (Initalmass / Finalmass), called Tsiolkovski's equation
and because the nearest orbit costs around 9500m/s but propellants achieve 3000-4600m/s ejection speed, launchers need a big ratio Initalmass / Finalmass.

This implies that all dead mass must be small in a launcher, and even then, they consume a strong part of the final mass, leaving less for the payload. Stages improve that by throwing away much dead mass, but even good designs used in a favourable mission waste easily 1/4 of the final mass for each stage.

Put together, the ejection speed (or divided by g=9.806m/s, the specific impulse Isp in seconds - plus subtleties if a secondary flux is lost) is all-important for a launcher. Already the Log would let 1% more Isp gain 3% on the final mass, but on the payload mass it acts even more. So the ejection speed rules, and designers make big efforts to improve it.

Thanks for the informative reply. +1

Share this post


Link to post
Share on other sites
On 12/30/2020 at 12:56 PM, John Cuthber said:

[...] You would get more heat in your rocket  by burning hydrogen than by reacting it with nitrogen, letting the heat produced by that dissipate (in the ammonia factory) , and then burning the ammonia in the rocket. [...]

Launchers don't try to maximize the energy efficiency, because their energy input is cheap.

Take a 60M$ launch that puts 6t payload in transfer to geosynchronous orbit and burns 600t propellants, among them 200t fuel. At 5$/kg, the fuel costs 1M$, not a driving factor. But if a better fuel achieves 10% more payload in GTO, the launch sells for 6M$ more, which affords 30$/kg fuel more. Better: you can keep cheap fuel at the first stage, and pay a better fuel for the 40t of the second stage or the 8t of the third stage, which justify more expensive fuels.

This needs easily mass-produced compounds, but things like cis-pinane (byproduct of the paper industry) or permethylated DETA are easier to purchase and possibly cheaper than rocket "kerosene", a special compound different from aeroplanes kerosene, heavily processed. If they bring 2-3s more specific impulse, to me the choice is clear.

On 12/30/2020 at 12:56 PM, John Cuthber said:

[...] If you look at this chart of energy densities
https://en.wikipedia.org/wiki/Energy_density#/media/File:Energy_density.svg
you will see why they use hydrogen - it's all the way over to the right.
The alkane fuels (LPG, petrol, diesel etc ) are all pretty much lined up at about 45 MJ/Kg and ammonia and hydrazine- the only amines on the chart- are well over to the left.

You could make kerosene denser by adding sand to it, but that wouldn't make it a better rocket fuel.

This comparison holds when oxygen is available for free from the air. In a launcher, it must be carried and accelerated by the exhaust. So the advantage of hydrogen is less huge in a launcher, while nitrogen produces heat by making N2 without any oxygen. The reaction enthalpy per mass unit of products is a good indicator for rocket propellants, though not sufficient.

Hydrogen is the most efficient realistic rocket fuel, but it's difficult to store and use. Besides its difficulties known from chemistry or engineering, it also needs big tanks that add to the dead mass of a stage. Its pumps often demand 2 or 3 stages, whose flux straighteners are much more complicated than a scroll for 1 stage; bearings cooled by liquid hydrogen are nothing obvious neither. The turbines often demand 2 stages with vanes rather than a scroll, and temperature lets put the turbine hang outside the actively cooled bearings. Precooling of all components, and good insulation, is a big step more difficult with hydrogen than with methane or oxygen. This lets SpaceX use "kerosene" or methane even for difficult missions, and stay away from hydrogen.

But if a fuel is dense, 1 pump stage can suffice, and "kerosene" or denser booze can even share the shaft and turbine with the oxygen pump, saving much money. A fuel liquid at room temperature also simplifies the engine and tanks.

Besides the engine simplicity, denser propellants can also receive a higher pressure to enter the chamber, as the pumping power is already optimized. This gives spiropentane and cyclopropane a higher exhaust speed than methane, while at identical expansion ratio in the nozzle, methane would be as good. Lighter tanks make the rest: the launcher is lighter and smaller.

==========

By using measured heats of formation of simple amines and running a rocket propellant software to compare with alkane homologues, one observes 3s advantage for amines. But I'd like to find arguments more convincing than that.

Share this post


Link to post
Share on other sites
19 hours ago, Curious layman said:

You put a lot of time and effort into these threads, Do you ever do anything with these ideas of yours, like apply for patents or seek funding?

Hi Curious Layman, thanks for your interest!

A patent costs a fortune (application +annual fees *many countries) and is only a proof you can show to a judge. If negotiating with a small company, an inventor has some chances to win a trial, so he can get paid for an invention or prevent its use. Against a multinational, procedure ruins an individual inventor before he gets the first cent. So big companies just steal the idea, neglect any patent, and save the money. I know this concretely: at a multinational employer, two colleagues and nevertheless friends worked full time to analyse the weaknesses of external patents and find ways to circumvent them, while a complete division sought how to come best out of trials, and more people sought how to resolve the cases outside legal procedures.

You may hear and read about individual inventors getting rich from their ideas and patents, but this is truly rare. Nearly all lose a bunch of money in patents before giving up.

==========

Creating a company: I tried long ago, after building the Sara satellite with friends in a club. It failed, and we got discouraged.

More recently, I considered building the sunheat engine in Luxembourg with the help of a governmental funding. Thinking more at it, I saw basic flaws in the activity model: development costs <1M€, but the in-space proof costs 6M€ and may very well go wrong, while the first customers probably need 5 years to decide and longer to pay. I'd have put much work in a company that would fail. That's why I finally described on this forum how to develop the sunheat engine.

Basically, I describe for everyone ideas I don't plan to develop by myself, mainly because they cost too much. I keep very few for myself, like some music instruments.

Yes, I'd like to live from my inventing and engineering activity, for instance at a company. France makes that more difficult. I could work once outside France, it was in Bavaria.

Share this post


Link to post
Share on other sites

I estimate payloads for Spectrum double what Isar Aerospace claim. Comparison confirms it:
spacex.com
the engines and tanks are 1/10th of Falcon, so a decent performance is 2.3t in Leo, not 1t. Make the engines half as strong to avoid competition against Vega? Note the hydrogen tank I drew in a D=2.5m fairing is for 1t payload, so the diameters should remain as the tanks shrink.

==========

The igniters could use the same decomposition chamber as the turbines do, if they draw little gas.

==========

I estimate an optimum chamber pressure for O2 and Pmdeta with 86% peroxide gas generator. Pessimistic assumptions taken from the old RD-170:

  • Injectors leave 88% pressure in the main chamber
  • Cooling jacket wastes no pressure
  • Turbine 79% efficient
  • Pumps 74% efficient
  • Gas expansion from 20bar to 1bar, hence light pressure tanks.
  • Gas gets 1240m/s, single turbine stage.

The best chamber pressure is then 155bar, but 125bar wastes only 0.5s and 110bar 1.0s. The lost peroxide wastes 14s, so if an expander cycle achieves over 64bar, it beats the gas generator cycle. But I proposed a good way to restart the gas generator cycle.

I suggested multiple inlets and outlet to reduce the pressure drop in a cooling jacket
scienceforums
provided that cooling is efficient under the manifolds.

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

Isar Aerospace tells Spectrum's fuel is a "mixture of light hydrocarbons", and at least on astronautics forums, a recurrent frenzy is Mapp and Dmapp, an equilibrium mixture of mainly propyne and allene, optionally diluted in propane. Seducing heat of formation.

Just in case: Mapp and Dmapp can't cool the walls of a combustion chamber. Allene (propadiene) polymerizes very easily, heat and pressure would do it even with a stabilizer. The resulting polymer insulates the wall until it bursts. That's why rocket fuels are painstakingly purified of all multiple bonds, which distinguishes rocket "kerosene" from Diesel oil. Propyne (methylacetylene) is known to detonate without oxygen above +95°C, which will occur at some location and time in an engine, for instance when shutting it off. Worse: liquid propyne is known to detonate too.

Even if cooling the walls with the oxygen (which would be new), I wouldn't trust Dmapp. Ethylene isn't a huge polymerization and detonation hazard like allene and propyne are.

==========

I suggested a pressure-fed launcher "Tronador" for 400kg in Leo there
scienceforums
multiply each dimension by 1.36 and you put 1000kg in Leo.

It must be the cheapest launcher for 1000kg, by very far. Two stages as well, slightly heavier but not bigger, no pumps at all, thick strong steel. "Only" combustion, fluids, and difficult enough steering. Perhaps even reusable. Strap-on boosters can modulate the payload and may serve at other launchers. A third stage is easily scaled from the others. I described lighter graphite tanks elsewhere, but for 1000kg, steel is intuitively cheaper.

That would be a more reasonable first project for a new company lacking experience. Remember: SpaceX hired experienced engineers from old companies.

To simplify further, each stage can have a fixed main nozzle plus four verniers as Soyouz has at the RD-108. Then, each vernier needs only one rotation axis, where the propellants can arrive just with seal rings. One big worry less. Four main nozzles would do that too, but at the last stage, the verniers tune also the orbit finely, orient the launcher and its payload, and change slightly the orbit between different satellites. I described igniters elsewhere, for instance using a Diesel glowplug.

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

An oxygen expansion cycle achieves some 140bar too. Saving the monopropellant flux, it gains 13s over a gas generator cycle. But can it restart as reliably as the pressurized monopropellant gas generator does?

The same 19MW as from elongated Vinci's cooling jacket would heat 27kg/s oxygen from 90K to 280K. This pushes 145kN, but the Aquila engine announces 94kN, and 40kN would suffice, so the chamber's aspect can be normal, further shrunk by the hotter flame. Heating the oxygen more would increase the pressure, but the engine is more difficult to start, and throttling may overheat it.

2:1 oxygen pressure drop extracts 47kJ/kg enthalpy. About 2.7:1 would optimize the chamber pressure, but I didn't iterate. Oxygen isn't far above its critical point, so more data would improve over mine at 100bar only. 79% and 74% efficient turbine and pumps bring 253:100 oxygen:Pmdeta to 340bar and 160bar. The cooling jacket with many vertical inlets and outlets shall drop to 320bar, the turbine to 160bar, and the injectors to 140bar.

==========

Single staged pumps and turbine are compatible with common ~230m/s azimuthal speed. D=47mm and D=39mm centrifugal pump impellers would suffice for the 94kN engine, but the axial turbine needs more area preferably at lower speed, and D~30mm pump inlets would need booster pumps to avoid cavitation.

Centripetal turbines offer a good outlet area, doubled if symmetric. Usual at car turbochargers, absent from rocket turbopumps. Centrifugal force is no worry here.

Bigger rotors seem better, like D=64mm at the turbine, and hopefully axial pumps looking like a feeder without the usual centrifugal stage. Easier to machine. D=70mm inlets need only 4m/s, and tanks can provide the 0.1bar above vapour pressure.

Throttling and shutoff impose high-temperature alloys.

==========

The pressure achieved by the expansion cycle depends on the relative volumes of oxygen and fuel rather than their sum for the gas generator cycle, so the fuel comparison differs slightly. I believe bicyclobutane, spiropentane and diazetidylcyclopropane aren't still mass-produced. Diazetidylcyclopropane, pmdeta and farnesane are less flammable.

+5s  138bar  Bicyclobutane
+2s  132bar  Ethylene
+2s  139bar  Spiropentane
+1s  141bar  Cyclopropane
Ref  131bar  Methane
-3s  143bar  Diazetidylcyclopropane
-5s  140bar  Pmdeta
-7s  144bar  Farnesane, ramified alkanes

Isp is not everything. Big heavy methane tanks make the launcher heavier and bigger.

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

Here's a direct visual comparison for a launcher, pressure-fed or pumped by a gas generator cycle.

Masses and volumes are directly scaled, neglecting subtleties, /1.8 from the Jan 04, 2020 drawing here, and *2.5 from the "Tronador"
scienceforums
scienceforums
so both put 1t in Leo and achieve Sso in two stages. Both have a D=2.3m body and a D=2.5m fairing.

PressureVsPumps.png.8e43dc3fed01fcb9b54f973e9713f03d.png

Smaller chamber pressure and heavier tanks make the pressure-fed launcher one and a half times as heavy and tall as the pumped one. I claim it's nevertheless a better "spectrum" launcher, because it's not as extremely difficult for a new company, and it's cheaper to launch.

  • 4t low-tech tanks of Maraging steel shall cost 100k€. The light tanks cost probably more.
  • 5t more Pmdeta add for instance 20k€.
  • 0.3t helium cost roughly 6k€. The amount is available.

10 turbopumps cost obviously much more than the 100 or 150k€. Reliability also means money. These figures tell that only big launchers can pay technology by saving mass.

A pressure-fed launcher being as sturdy as a boat, reusing it after splashdown is credible with or without sailing back, provided this saves money
scienceforums

Variants:

  • The expansion cycle differs little from the gas generator. Safety, reliable restart?
  • Diazetidylcyclopropane saves some mass over Pmdeta, it's a detail. Production?
  • Graphite tanks do save good mass over steel. Cost?
  • Electric pumps are excellent, nearing gas generator performance, much simpler, reliable.

Marc Schaefer, aka Enthalpy

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.