NASA Study Shows Cheaper Alternatives to SLS

[Xplor]

Muomega0
I fully agree that one wants to keep the radiator viewing deep space.  In LEO this is a challenge when half the view is Earth and the sun fills a good fraction of the sky for half the orbit.

The article you site while interesting was specifically looking at state of the art and needed development of cryocoolers to support the high temperature superconductor market.  I couldn’t tell if the 100 to 200 input SHAFT power over cooling capacity (at 20k) was for an actual unit or a desired value.  This didn’t include power converter losses, motor losses or other real application losses.

The Advanced Cryocooler Technology Development Program goal was to demonstrate 0.25 W cooling (@18k) for 150 w input power, a 1:600.

So a question, does anyone have data for actual space cryo cooler performance? 

[Xplor]

So to achieve a 1/5 the heating rate, then 1(n+1) layers of MLI are needed in between, in theory, so 4 layers of MLI. 

The simple concept is to maintain the intermediate shields at ~90K and ~20K, with fluid/plumbing completely different than the LOX/LH2 lines.

In the worse case, assuming the active system failed, the system reverts back to the passive 0.1%/day boiloff.

In other words, the active system only helps the architecture reduce boiloff with minimum extra development dollars, and has a contingency plan of *more launches*, identical the passive system *baseline* plan.

Passive or active cryo storage will require efficient thermal designs, incorporating MLI (a lot more than 4 layers) and other features.  So I don’t understand your above MLI reference?

In a passive system boiling LH2 to 20k GH2 absorbs 428 kj/kg.  But using this GH2 to vapor cool the tank, shields, plumbing, structural and electrical penetrations can absorb an additional 3,256 kj/kg, the equivalent of a huge amount of active cooling without the need for solar arrays, radiators, pumps, etc.

All space systems require some means of station keeping, especially in LEO where one has atmospheric drag that large solar arrays will only make worse.  If one achieves ZBO what do you do for this station keeping?  Loft tons of storable propellant each year?  Alternatively use the GH2 as a very efficient monopropellant.

[Xplor]

ASSUME that the LH2 boiloff maintains LOX to zero (it does not).

Why can one not use the LH2 boiloff to maintain LO2 ZBO?

So after 180 days, one has lost about 0.835 of the LH2.  Hence the LH2 tank size must be increased by ~20% (1/0.835) to arrive at the 5:1 ratio.

A 15 ft (~ 4.5 m) diameter by 31 ft (~10 m) long tank, according to ULA, holds 140 mT of LOX *or* 15 mT of LH2, or a 9.3:1 ratio.  So to achieve a 5:1 ratio, the LH2 tank is almost twice as large as the LOX tank.  If *only * LH2 has boiloff, it would have to increased ~ 20% in length to accommodate 0.1% boiloff, to a first order approximation.

So if the LOX is also has 0.1% boiloff, then the quantity of both need to be increased accordingly.

The 180 days is an *approximation*  to say that 2 years or 5 to 10 years provided by XPLOR is incorrect, unless one plans more resupply launches, at more cost.

The LH2 and LO2 should be launched over time supporting a reasonable, steady launch tempo, not all at once.  So I am not assuming that the depot/CPS would start full in LEO and then sustain boil-off for years.  A LEO depot really should be a collecting/staging point for moving propellant and hardware up the gravity well.  Any long term storage should be at a LaGrange point or other suitable high orbit location where Earth’s thermal influence is minimized.

[muomega0]

Muomega0
I fully agree that one wants to keep the radiator viewing deep space.  In LEO this is a challenge when half the view is Earth and the sun fills a good fraction of the sky for half the orbit.

The article you site while interesting was specifically looking at state of the art and needed development of cryocoolers to support the high temperature superconductor market.  I couldn’t tell if the 100 to 200 input SHAFT power over cooling capacity (at 20k) was for an actual unit or a desired value.  This didn’t include power converter losses, motor losses or other real application losses.

The Advanced Cryocooler Technology Development Program goal was to demonstrate 0.25 W cooling (@18k) for 150 w input power, a 1:600.

So a question, does anyone have data for actual space cryo cooler performance? 

And where is that NASA report btw? 

[muomega0]

ASSUME that the LH2 boiloff maintains LOX to zero (it does not).

Why can one not use the LH2 boiloff to maintain LO2 ZBO?

So after 180 days, one has lost about 0.835 of the LH2.  Hence the LH2 tank size must be increased by ~20% (1/0.835) to arrive at the 5:1 ratio.

A 15 ft (~ 4.5 m) diameter by 31 ft (~10 m) long tank, according to ULA, holds 140 mT of LOX *or* 15 mT of LH2, or a 9.3:1 ratio.  So to achieve a 5:1 ratio, the LH2 tank is almost twice as large as the LOX tank.  If *only * LH2 has boiloff, it would have to increased ~ 20% in length to accommodate 0.1% boiloff, to a first order approximation.

So if the LOX is also has 0.1% boiloff, then the quantity of both need to be increased accordingly.

The 180 days is an *approximation*  to say that 2 years or 5 to 10 years provided by XPLOR is incorrect, unless one plans more resupply launches, at more cost.

The LH2 and LO2 should be launched over time supporting a reasonable, steady launch tempo, not all at once.  So I am not assuming that the depot/CPS would start full in LEO and then sustain boil-off for years.  A LEO depot really should be a collecting/staging point for moving propellant and hardware up the gravity well.  Any long term storage should be at a LaGrange point or other suitable high orbit location where Earth’s thermal influence is minimized.

If phasing the delivery of the LOX/LH2 is shown to be optimum, then great.
I just stated my assumptions of LH2 first, so as to be clear.  If delivered as a pair, then perhaps ZBO is maintained on the LOX, and the % boiloff applies to a lesser amount of LH2, which extends the ~ 180 days to a larger number of days.   

-----
If the design is properly done, then ZBO on the LOX is probably a good first order approximation.

Note however John F's observation and inconsistency between two papers:

Quote from one ULA paper "Atlas Centaur Extendability...." : "By using the LH2 boil-off to cool the LO2 tank, LO2 boil-off can be completely eliminated."

In "A practical Affordable Cryogenic Propellant Depot Based on ULAs Flight Experience", Figure 9 shows the % boiloff in a plot of Shield half angle to shield length, an in the figure this text is shown "Values in bubbles is net LO2 boiloff (%/day)".

Now most have assumed that the LO2 in the caption was really a typo.  So anyone want to comment on famous 0.1%/day figure 9?

From the same paper:

"It is also possible, using a Thermodynamic Venting System (TVS), to use the vented hydrogen gas to remove some more heat from its own tank. This is done by running the moderate-pressure boil-off gas through a Joule-Kelvin valve, which drops both the pressure and temperature of the vented gas, allowing it to provide some extra cooling for the LH2 tank before it is used for removing heat from the other tanks and subsystems in the depot. After cooling the LO2 tank, and the connection between the LO2 tank and the electronics section of the depot, the now much warmer gaseous hydrogen can be run through a nozzle to provide thrust for settling and station-keeping."

So again, this statement suggest that linear thrust, rather than rotation, can also be used for settling, and some other boiloff for station keeping.

Since hydrogen expends at a 1:848 ratio, significant amounts cannot be stored, and thrust/reboost would have to be done *continuously*.  But a lower altitude with reboost later offers more propellant. 

[oldAtlas_Eguy]

Thus the process of boiling the LH2 is absorbing 500 W (0.0012 kg/s * 428,000 J/Kg)

I would have to agree that if someone proposed using a 1:1000 cryocooler, then it makes the design a wee bit more challenging- not impossible but more challenging.

The current state of the art for Space Solar panels is 91W/kg so for a 500kW array (enough power to fully power the cryo cooler at 1:1000) it would weigh ~5.5MT. For a 100MT propellant depot 5.5MT is only 5% of the total weight and half of the depot dry weight making the depot dry weight about 10-12MT.

Edit: Re considering tank and sunshield weights, the depot dry weight not counting the solar array its going to be about 20-25MT so with arrays and the array deployment mechanisms and heat rejection for the cryo cooler the dry weight would be about 30-40MT. It could still be put up by an FH.

[muomega0]

So to achieve a 1/5 the heating rate, then 1(n+1) layers of MLI are needed in between, in theory, so 4 layers of MLI. 

The simple concept is to maintain the intermediate shields at ~90K and ~20K, with fluid/plumbing completely different than the LOX/LH2 lines.

In the worse case, assuming the active system failed, the system reverts back to the passive 0.1%/day boiloff.

In other words, the active system only helps the architecture reduce boiloff with minimum extra development dollars, and has a contingency plan of *more launches*, identical the passive system *baseline* plan.

Passive or active cryo storage will require efficient thermal designs, incorporating MLI (a lot more than 4 layers) and other features.  So I don’t understand your above MLI reference?

In a passive system boiling LH2 to 20k GH2 absorbs 428 kj/kg.  But using this GH2 to vapor cool the tank, shields, plumbing, structural and electrical penetrations can absorb an additional 3,256 kj/kg, the equivalent of a huge amount of active cooling without the need for solar arrays, radiators, pumps, etc.

All space systems require some means of station keeping, especially in LEO where one has atmospheric drag that large solar arrays will only make worse.  If one achieves ZBO what do you do for this station keeping?  Loft tons of storable propellant each year?  Alternatively use the GH2 as a very efficient monopropellant.

Station keeping:  If one is adding power, then attitude control can be performed using power.
---

Boiloff for cooling:  But LH2 costs 4 to 5 times as much as LOX (density LOX/density LH2/ 5, where 5 is the fuel ratio), so $10000/kg becomes  $40K/kg if launched separately, with no hardware reuse.   Why bring a LH2 tank each *cheap* tanker trip, unless its more cost effective?

---
MLI design.    When a "shield" or MLI layer is added, it effectively reduces the heat load to the inner wall.  IOW, a temperature gradient is created.

So assume that the 20K cryocooler has a thermal capacity of 20 W.  If the total tank area was 100 m2, then the allowable heat rate is 0.2 W/m2. 

So if the solar flux is 1373 W/m2 and 0.2 absorbed, then the heat gain is 275 W/m2.      So theoretically, one could add 1/(n+1) layers of MLI to cut the heat load to 0.2 W/m2, but over 1000 layers of MLI is not practical.

Fortunately, some heat transfer rocket scientists have other methods to cut this heat load to save weight.

Analytically, a simple calculation can relate the JWST shield or BEO design to the conical design for LEO.

[muomega0]

Thus the process of boiling the LH2 is absorbing 500 W (0.0012 kg/s * 428,000 J/Kg)

I would have to agree that if someone proposed using a 1:1000 cryocooler, then it makes the design a wee bit more challenging- not impossible but more challenging.

The current state of the art for Space Solar panels is 91W/kg so for a 500kW array (enough power to fully power the cryo cooler at 1:1000) it would weigh ~5.5MT. For a 100MT propellant depot 5.5MT is only 5% of the total weight and half of the depot dry weight making the depot dry weight about 10-12MT.

Lost of Garbage In, Lots of Garbage Out.
----

"its never flown before"

IOW, if the TRL of technology is never raised, then the status quo is what you get.  This is not *research*, which would help even more, but simply flying technology to obtain a higher TRL,*IF* it helps the architecture.

Oh well back to reality tv, how much money did SLS get this year, and was anyone voted off the island?!

[oldAtlas_Eguy]

Thus the process of boiling the LH2 is absorbing 500 W (0.0012 kg/s * 428,000 J/Kg)

I would have to agree that if someone proposed using a 1:1000 cryocooler, then it makes the design a wee bit more challenging- not impossible but more challenging.

The current state of the art for Space Solar panels is 91W/kg so for a 500kW array (enough power to fully power the cryo cooler at 1:1000) it would weigh ~5.5MT. For a 100MT propellant depot 5.5MT is only 5% of the total weight and half of the depot dry weight making the depot dry weight about 10-12MT.

Lost of Garbage In, Lots of Garbage Out.
----

"its never flown before"

IOW, if the TRL of technology is never raised, then the status quo is what you get.  This is not *research*, which would help even more, but simply flying technology to obtain a higher TRL,*IF* it helps the architecture.

Oh well back to reality tv, how much money did SLS get this year, and was anyone voted off the island?!

Ok, GIGO.

In considering a cryo cooler using 500kW of power you also need 500kW of heat rejection or about 50MT of a system to do this. Heat rejection is why the ISS solar power and heat rejection combined system weighs so much for the power it produces. The heat rejection portion of the system is what will make a ZBO system based on a cryo cooler for a LEO depot not feasible, at least in the next 20 years. But using cryo coolers for deep space depots is feasible since they would need 1/10 the power or about 50kW and 1/10 the heat rejection to be able to achieve ZBO, a must for efficient storage for a Mars mission or even a good to have for a L1/L2 depot.

[Xplor]

Boiloff for cooling:  But LH2 costs 4 to 5 times as much as LOX (density LOX/density LH2/ 5, where 5 is the fuel ratio), so $10000/kg becomes  $40K/kg if launched separately, with no hardware reuse.   Why bring a LH2 tank each *cheap* tanker trip, unless its more cost effective?
 

Why does LH2 cost more to launch than LO2?  If a rocket can lift X mass it doesn't really care about the density, so long as it will fit within the payload fairing.

[muomega0]

Thus the process of boiling the LH2 is absorbing 500 W (0.0012 kg/s * 428,000 J/Kg)

I would have to agree that if someone proposed using a 1:1000 cryocooler, then it makes the design a wee bit more challenging- not impossible but more challenging.

The current state of the art for Space Solar panels is 91W/kg so for a 500kW array (enough power to fully power the cryo cooler at 1:1000) it would weigh ~5.5MT. For a 100MT propellant depot 5.5MT is only 5% of the total weight and half of the depot dry weight making the depot dry weight about 10-12MT.

Lost of Garbage In, Lots of Garbage Out.
----

"its never flown before"

IOW, if the TRL of technology is never raised, then the status quo is what you get.  This is not *research*, which would help even more, but simply flying technology to obtain a higher TRL,*IF* it helps the architecture.

Oh well back to reality tv, how much money did SLS get this year, and was anyone voted off the island?!

Ok, GIGO.

In considering a cryo cooler using 500kW of power you also need 500kW of heat rejection or about 50MT of a system to do this. Heat rejection is why the ISS solar power and heat rejection combined system weighs so much for the power it produces. The heat rejection portion of the system is what will make a ZBO system based on a cryo cooler for a LEO depot not feasible, at least in the next 20 years. But using cryo coolers for deep space depots is feasible since they would need 1/10 the power or about 50kW and 1/10 the heat rejection to be able to achieve ZBO, a must for efficient storage for a Mars mission or even a good to have for a L1/L2 depot.

--- Only 500 kWe, why not 12.4 MWe?!
 
Why did you start with 500 kWe, when you could have started with a bare LH2 tank in space, requiring 12.4 MWe of power plus the heat rejection system?  The calculation, by the ‘old atlas guy’ estimation would then be: 1373 W/m^2 –solar constant * 0.2 abs * 45 m^2 * 1000:1 = 12.4 MWe, without including any earth ir and albedo!
 
Again, the 500 KWe is complete garbage, which is quite a disappointing assessment from ‘ an old Atlas guy’.   Ironically,   a similar ‘argument’ was made to eliminate ZBO systems from any consideration during Constellation- yet another ‘thumb on the scale’ approach.

So the good news is ZBO in a dedicated depot is achievable in LEO with about 10 KWe of power and is by far the cheaper approach (billions) than the refueling stages relying passive TCS rather than an active.

This will be shown later.  Stay tuned...

--- NASA has two big $$$ problems

1. Choosing depot centric over SLS HLV reduces NASAs cost by the 10s of billions over 20 years.

2. The permanent depot saves billions over 20 years versus refueling stages.

--- 0.1%/day boiloff is *Lousy* and limits operations and drives costs
 
0.1%/day is *Lousy* - it limits LEO operations of a refueling stage to ~ 180 days and throws away hardware that could last for over a decade.  To state otherwise, needs, well proof:
 

LEO operations are limited to ~180 days is a more more reasonable estimate assuming 0.1%/day

because the the LH2 tank is much larger than the LOX tank, even at 5:1 ratio!  Now it must increase 20% just to achieve 180 days (which increases the boiloff rate!).


---  Why does the passive system alone not work?

Its rather simple.  Weight needs to be added to reduce the passive heat load to the shields for the depot (or tank walls for the refueling stages).

But when weight is added to the upper stage, it reduces mass fraction, killing overall architecture costs and performance, increasing mass to orbit requirements.

To justify the refueling stage approach, an "acceptable" boiloff percentage is arbitrarily defined to justify the hardware approach.

Some simple calculations can show this. 

Stay tuned for more details....


--- Advantages of a dedicated depot

Read the last page of the leaked NASA study, "Depots vs Refueling".
Internal NASA Studies Show Cheaper and Faster Alternatives to The Space Launch System

The study shows that costs can be substantially cheaper and risk significantly lower with a depot versus refueling, the exact opposite of  unsubstantiated claims.

Depot Study:
* Most expensive hardware/capability can be located on the depot to be re-used over and over again rather than expended every flight
* The expendable CPS and delivery tankers can be made as dumb/cheap as possible
* Mass of the CPS that has to be pushed through thousands of m/s and delta-V can be reduced
* All of the important and costly avionic/software/IVHM can be on the depot
* The prox-ops and rendezvous and docking systems can be on the depot, rather than the CPS
* The depot could do the last prox-ops maneuvers and even berth the tanker/CPS with an RMS
* Relieves CPS of need for active boil-off control for cis-lunar missions (...)
* Reduces risk to CPS from MMOD by reducing time in orbit prior to departure
* Reduces number of rendezvous events required to fuel CPS from many to ONE, reducing risk of collision or propellant transfer rate
* Reduces risk of LOM by decoupling propellant delivery flights FROM delivery of mission elements

Edit:  Added Links

[JohnFornaro]

Note however John F's observation and inconsistency between two papers:

Just a nit, but:  John F is the monicker of some other person who posts on this forum.  I believe that you are mentioning my previous questioning of the 0.1% boil-off meme?  It would be nice if you'd fix this, if appropriate.

Now most have assumed that the LO2 in the caption was really a typo.  So anyone want to comment on famous 0.1%/day figure 9?

What?  I have never heard of this!  I'm really fed up with these "typos".  On another thread some guy posted a calculation based on one meter.  Others commented that he really meant one centimeter.

Apparently, we are not entitled to published papers being free of typos in the most important places.  I don't believe that it's a typo.  There is no discussion in the ULA paper of hydrogen boil-off rates, and little discussion beyond the schematic notion of using O2 to ameliorate H2 boiloff.  If the published data is so completely error prone, then it is no wonder that Congress has difficulty in appropriations based on rational bases.

Of course, I would appreciate other papers on this important issue.

[oldAtlas_Eguy]

Some of the discussions related to energy input into the system (Sun, Earth as heat radiation sources) and the passive thermal shielding limiting the thermal energy reaching the inside of the tank, made me realize the boiloff is actually related to the energy IN not to the amount of propellant. A 100MT tank only partially full say 10MT would boiloff nearly the same amount of propellant in a day measured in kg that a full tank of 100MT of propellant would (basic thermodynamics Energy IN = Energy OUT). So the boiloff rates expressed in % of the propellant weight has no real meaning except in a quick evaluation of systems all with the same size tank that is full of prop of the same type of prop!

The % boiloff rates being thrown around do not specify the tank sizes (volume or surface area) or the amount of prop in the tank, without which the numbers have little meaning from an engineering standpoint.

muomega0 you are correct a lot of the info based on % boiloff is just so much GIGO, which unfortunately includes my own estimates and conclusions presented based on % boiloff.

[muomega0]

Note however John F's observation and inconsistency between two papers:

Just a nit, but:  John F is the monicker of some other person who posts on this forum.  I believe that you are mentioning my previous questioning of the 0.1% boil-off meme?  It would be nice if you'd fix this, if appropriate.

Now most have assumed that the LO2 in the caption was really a typo.  So anyone want to comment on famous 0.1%/day figure 9?

What?  I have never heard of this!  I'm really fed up with these "typos".  On another thread some guy posted a calculation based on one meter.  Others commented that he really meant one centimeter.

Apparently, we are not entitled to published papers being free of typos in the most important places.  I don't believe that it's a typo.  There is no discussion in the ULA paper of hydrogen boil-off rates, and little discussion beyond the schematic notion of using O2 to ameliorate H2 boiloff.  If the published data is so completely error prone, then it is no wonder that Congress has difficulty in appropriations based on rational bases.

Of course, I would appreciate other papers on this important issue.

The name will be edited.

--
In a way the question is already answered. 

Assume that the LOX tank has no net heat load on its surface from the environment.  In this case, the LH2 boiloff will simply be subcooling the LOX, so no boiloff.

So now assume 0.1% boiloff of LH2, calculate how much cooling can be provided at 90K, and magically spread this heat over the surface of the LOX tank to determine the allowable environmental limit on the LOX tank.  One will now have subtract, however, that amount be used for power, attitude control, reboost, *if* that is the operational mode being used.  Sure is quite a bit of work to do.

The issue returns now to weight and cost.  One can always add more passive thermal control to cut heat gain from the environment, but it weighs more.   To magically spread the cooling gas to all sides of the tank takes more tubing and conduction paths, but this *may* increase heat gain from the environment.

So *only* knowing the boiloff is not the complete answer, one needs the mass fraction of the refueling stage also, as well as the tank sizes, etc.

But if one never studies options not on the table, then only local minimums will be found.

Not a precise answer, so stay tuned

[Xplor]

 

So now assume 0.1% boiloff of LH2, calculate how much cooling can be provided at 90K, and magically spread this heat over the surface of the LOX tank to determine the allowable environmental limit on the LOX tank.  One will now have subtract, however, that amount be used for power, attitude control, reboost, *if* that is the operational mode being used.  Sure is quite a bit of work to do.

How do you propose to spread your active cooling into the LO2 tank? 
Why does it take magic to spread hydrogen vapor cooling over the LOX tank

The issue returns now to weight and cost.  One can always add more passive thermal control to cut heat gain from the environment, but it weighs more.   To magically spread the cooling gas to all sides of the tank takes more tubing and conduction paths, but this *may* increase heat gain from the environment.

Do you have any information to back up this claim?  Do you suggest a bare tank (no MLI) with active cooling is lighter than a tank with 3 layers of MLI or 40 or 200? Do you suggest that adding 10 kw of power, power conditioning unit, cryo cooler and radiator are cheaper than a Joule Thompson orifice and 100 feet of 3/8” tubing?   

The real question is: what is the appropriate balance between passive and active cooling?  The answer is application specific.