[AR] Re: 500,000 tons per year to GEO (off topic)

  • From: Keith Henson <hkeithhenson@xxxxxxxxx>
  • To: arocket@xxxxxxxxxxxxx
  • Date: Thu, 10 Apr 2014 12:35:06 -0700

On Thu, Apr 10, 2014 at 5:48 AM, Peter Fairbrother
<zenadsl6186@xxxxxxxxx> wrote:
> On 04/04/14 17:01, Keith Henson wrote:
>>
snip

>> Please go into details.  I am not welded to Skylon and lasers.  I am
>> well aware of the problems involved.  If you have better ideas, I will
>> gladly adopt them.
>
> I rather thought that I had, but:
>
> [] The Mission:
>
> Looking at 500,000 tons per annum to GEO, in the form of powersat parts,
> they will have to be assembled in space. Assembly in space is much cheaper
> and easier when done in LEO rather than in GEO, so we will assemble the
> parts in LEO.
>
> That means a stopoff in LEO

I agree.  9 km/s to LEO plus 4.1 km/s to GEO makes the payload
fraction low if not less than zero for any exhaust velocity that seems
at all practical.  Even the laser boosted Skylon stages in LEO.

> rather than going straight to GTO, which is
> useful for all sorts of other reasons, like exactly where the assembly takes
> place (see below), stage recovery, and so on.

LEO assembly has significant advantages, I agree.  That was the
approach taken by Boeing's studies back in the 70s.  There is a good
illustration of one being built here:
http://ssi.org/assets/images/Boeing-large.jpg

The idea was to fly them out to GEO using their own power.  Took about
6 months using all the power and electric thrusters.  Most of the
higher exhaust velocity is wasted in the 2 to 1 increase in delta V
you need for spiral orbits compared to Hohmann minimum energy transfer
orbits.  But that wasn't what made Boeing abandon the idea.  It turned
out that the space junk that existed even in those days was enough to
be sure that huge damage would happen to every one of them on the trip
out.  I have not worked the numbers for this, it may be worth
revisiting.  But it's not encouraging.  You can't fold them up if you
are using the power because the smaller size is exactly compensated by
the longer trip on partial power.

> Supposing a payload of 7.5 tons to LEO (and 5 tons to GEO), that's 100,000
> launches per year or 12 per hour. We need three launch sites each launching
> every 15 minutes, and six locations in LEO where assembly takes place.
>
> We need to get the payloads to the assembly areas on first orbit, otherwise
> we will have a huge mess. That means the assembly areas pretty much have to
> be in equatorial LEO, and the launch sites located on the equator.

I agree completely.  On the equator and water to the east.  That
complicates downrange recovery.  There are not many such locations.

> [] Hardware:
>
> Two stages to LEO is much more sensible than one, or than three or more. One
> stage is too expensive and technically challenging, with more than 2 stages
> recovery for reuse becomes very difficult. So, we use two stages.

Agreed.  That's been worked out in some detail by Solar High group.

https://drive.google.com/file/d/0B5iotdmmTJQsZ0VpWndsc0c3ZFBGMW1NS3pxMUxwS2FTZUo0/edit?usp=sharing

The trouble with this analysis is on slide 14 where they estimate the
transportation contribution to the cost of electric power at "less
than 7 cents per kWh."  The total cost of power looks to be 10 cents
per kWh or a bit more, and at that price you can't make a business
case because other sources, like nuclear, are cheaper.

> [] The first stage booster:
>
> The first stage doesn't do very much in terms of delta-V - one reason for
> this is fast turnaround. If the first stage can land where it took off from
> it can turn around much faster than if it lands elsewhere. If the first
> stage adds a lot of delta-V then it is going to land several thousand miles
> away from the launch site, so instead it basically just goes up and down.
>
> Another reason for the performance of the first stage being low is
> reliability and robustness - we can build it a lot stronger if the dry mass
> is larger and the required performance is lower. Compared to Skylon, the dry
> mass of the first stage will be somewhere between 60% and 80% higher.
>
> Yet another reason why the first stage does much less than half the LEO
> delta-V is that it operates in atmosphere, where engines are less efficient
> and Isp is lower. The first stage main engines burn LOX and kero, not LOX
> and LH2 - much cheaper.

They agree with you on the fuel choice.

But the consequences of little velocity contribution from the first
stage puts a high burden on the second.  Burning LOX/LH2, if you need
to gain 8 km/s with the second stage, then 1/e^(8/4.5) is 17%.  How do
you want to split this between payload and structure?  On the other
hand, it does get you away from the problems of taking a 747 scale
aircraft supersonic.

> The first stage has wings to increase cross-range so it can add a bit of
> sideways delta-V as well as going up and down while still landing at the
> place it took off from. This extra delta_V makes the job of the second stage
> easier and increases the payload.
>
> With wings comes wheels and horizontal takeoff and landing. We add some jet
> engines so we can do go-arounds, land at alternative airports, and so on,
> increasing safety and reliability - also the jet engines increase cross
> range, and give a bit of extra height and delta-V before main rocket engine
> burn.

The RE people agree with the wings.  The lift takes the gravity loss
till they run out of air.

But wheels (and brakes) don't come cheap.  The problem of stopping a
Skylon on a rejected takeoff just short of rotation ate almost the
whole payload till they figured out water cooled brakes.

> The first stage's main rocket engines now start at 30,000 feet, and can be
> optimised more towards vacuum, making them more efficient.
>
> Summing up, the first stage booster is basically a jet aircraft with rocket
> engines stuck on it. That means we can have pilots on board too. It has a
> take-off mass of about 400-460 tons, somewhere between a Boeing 747-400 and
> an Airbus 380, and a second stage payload of 50 to 60 tons, with a LEO
> payload of 7.5 to 10 tons.

Skylon has a 300 ton takeoff mass and gets 15 tons (payload) to LEO.
For 7% structure and 10% payload, the second stage would be 100 tons
on SSME performance.  That seems awful optimistic to me.

You could consider Paul Allen's project.
https://en.wikipedia.org/wiki/Stratolaunch_Systems#Orbital_Sciences_Corporation

This 544 metric ton puppy is planned to get 6.1 metric tons to orbit.
I don't recall seeing a cost estimate.

Using a Skylon airframe lightened to 230 tons and SSME engines in the
place of the exotic SABRE, would get the orbital stage back.  Reaction
Engines has looked (if only briefly) at using the Stratolaunch vehicle
for ferry operations.  It would have to be made somewhat taller to
accommodate the Skylon's dimensions.

However, I doubt this approach would reduce the cost per kg to orbit.
Worth a more compete analysis though.

Keith

> At nominal flight rate and two-hour turnaround there are eight booster
> stages operating from each of three airfields, making 24 operating boosters
> in total. A fleet of around 60 boosters would be needed to sustain this.
>
>
> [] The second stage:
>
> The second stage is powered by a LOX/LH2 engine. It is carried inside the
> first stage until that reaches 60 km high, so aerodynamic loads are very
> low.
>
> It consists of two LH2 tanks, an engine/electronics/rcs module and an
> optional cargo module all attached to a central semi-structural LOX tank.
> Payload can also be strapped directly to the LOX tank supports without using
> the cargo module.
>
>
> One version masses 50 tons when fuelled, has a dry mass of 5 tons and a LEO
> payload of 7.5 tons at an Isp of 445s. The disassembled stage without cargo
> module fits into 3 standard shipping containers.
>
> The tanks are reused in orbit or deorbited; the engine modules are returned
> in batches under a heatshield and parachute for reuse.
>
> Depending on turnaround time, a total of about 5,000 second stage engine
> modules will be needed.
>
>
>
> [] Assembly personnel:
>
> There is also an alternative second stage for passengers, which re-enters
> and lands at a runway. It is launched from the same piloted booster.
>
>
>
> [] LEO to GEO:
>
> There are several possibilities. If we have power sat parts which can supply
> power for an electric drive we can do something with them, so that the
> amount of mass needed to get from LEO to GEO is reduced. Perhaps a small
> booster stage (or refuel and reuse one of the second stages) to get some
> initial velocity to clear any near-earth debris cloud and decrease transit
> time, then use the electric drive, or something like that. I won't mention
> tethers.. :)
>
>
> [] Conclusion:
>
> If done in a newspace SpaceX-type manner it should cost somewhere around $28
> billion in capital over 4 years for hardware, and about $9 billion per annum
> operating expenses.
>
> Somewhere around $20 per lb. Probably $80 per lb if done in an oldspace
> manner.
>
>
> That's about as cheap as I can conceive, for a requirement of 500,000 tons
> per annum to LEO, with little new technology or risk of failure. I don't
> think there is any potential new tech which would be cheaper, even if
> higher-risk.
>
>
> -- Peter Fairbrother
>

Other related posts:

  1. Home
  2. arocket
  3. Archive
  4. 04-2014