Monday, August 02, 2004
Price to Orbit II... revision
Lt.Col. John R. London III wrote for the USAF the book length study of space costs and space cost reduction, LEO On the Cheap, which is available for free from here:
LEO On the Cheap
The author released it for general distribution because his message is vitally important - Space costs too much chiefly because it is done at the extreme edge of technical ability, and the launch vehicle and satellite makers like it that way.
Enter Elon Musk who, with the SpaceX team, has taken on board Colonel London's findings and has simplified LEO bound rockets. If SpaceX can reach their goals the price to LEO will drop to ~ $2,200/kg, instead of the ludicrous ~ $40,000 - $10,000/kg currently on offer. And the price to GTO will drop to ~ $4,400/kg or so.
SpaceX is offering 4,200 kg payloads to LEO delivered for ~ $12 million. That's $2,860/kg, but they'll get better at it after a few launches and the price should come down. The main point about their efforts is that the rockets and avionics might not be absolute marvels of engineering perfection - instead they work well enough. Incredible amounts can be spent pushing machinery and designs to their absolute limits because of the continual review and refinement process uses large numbers of staff and resources to achieve the incremental approach to such limits.
A common satellite design mistake is to try to fit the satellite into a specific mass - the cost of space launch means people want to use all the mass budget they're alloted. As a result in the final stages $100,000s are spent refining the design. Larger, cheaper launcher payload options would mean a cheaper design process. Bigger satellites, because of ease of design, are also cheaper satellites. A common satellite frame-work that can be adjusted for several roles - rather than needing total redesign each time - would make for cheaper satellites too.
And rocket design? Advanced, computer controlled rockets pushed to their absolute limit, pushing propellants at high pressure into exhaust chambers with complicated and expensive turbo-pumps, and cooling jackets... well it all adds to costs. Exotic alloys for propellant tanks and refined rocket motors that push the envelope are long labours of (very expensive) engineering love - and simpler, proven designs and components could do the job for a lot less money. No commercial rocket these days needs to push the envelope when there is so much prior experience already paid for.
Posted at 12:53 pm by Adam
Back in the '70s the Space Shuttle was supposed to be the answer to revolutionise space-travel. No more expensive throwaway rockets, but routine, reuseable space-flight cheap enough for all to be involved in some way. In reality NASA over-sold the Shuttle in a valiant attempt to avoid getting the whole manned space program shut down by Richard Nixon. To achieve the hoped for price-tag of $750/kg to LEO (1980$... more like ~ $1,200/kg today) the Shuttle would need to be flying ~ 60 flights to LEO per annum with a full load of ~ 29,500 kg each time.
So why is the Shuttle so damn expensive? For starters the whole ground support system costs $2.8 billion per annum without a single Shuttle lifting off. Developing the Shuttles cost ~ $7.5 billion and building one costs ~ $1.5 billion. Say NASA had their full complement of 5 Shuttles. If they can manage 100 flights each, they then cost ~ $90 million plus those per annum housekeeping costs ($560 million per Shuttle per year), which totals (@ 60 flights/annum) $137 million/flight before we buy fuel. Surprisingly the fuel only costs ~ $1.5 million. Go figure...
So lets call it ~ $138 million to haul our 29,500 kg to LEO. That's still $4,700/kg - I guess NASA expected cheaper operating costs and wrote off the development costs - they expected to be making fleets of Shuttles eventually which would have driven that initial cost down dramatically.
My old favourite from 1979, "The Space-Traveller's Handbook", is set in a fictional 2061 in which all the 1970s dreams have come true - and Shuttle costs a mere ~ $60/kg to fly. About 100,000 people fly on a Shuttle per year, plus who knows how many cargo flights launched via reusable Heavy Lift Vehicles. Hence development costs have long been paid for, mass production has cut costs per Shuttle and ground support costs are spread over LOTS of flights (~ +2,000/annum.) That's how flight costs could be cut down dramatically - LOTS of traffic to LEO and beyond.
The reality is different, very different. In the fiction there are several Space Colonies - all Standford Torus designs - with a space population +100,000. There is a market - if not many space markets - that make Space pay-off for Earth's investment. In our current reality Earth-oriented "services" are the only pay-off. There is NO primary industry in Space, and that's what it really needs. So what can we get in Space and sell back here for $$$ ???
Posted at 1:43 am by Adam
Here's a cool Optics news bite...
Sydney Opera House repro/reduced
...micro-lithography - if it can be sped up - enables all sorts of amazing things to be crafted at micro-scale.
Posted at 12:35 am by Adam
Mass break-down for our fictional Voyager to Mars...
CSM (fully fuelled) 30,329 kg (inc. 18,413 kg propellant)
Crew: 204 kg (3 @ 68 kg)
MEM: 24,947 kg
Multiple Docking Adapter: 6,260 kg
Airlock Module: 22,225 kg
Saturn Workshop: 35,380 kg
Instrument Unit: 2,065 kg
extra supplies (1,000 days @ 1.27 kg/day/person): 3,810 kg
[the extra supplies are arranged in lockers that form a radiation storm shelter in the core of the Workshop]
Sum Total: 125,220 kg
Earth Return sub-total: 98,862 kg
The CSM main engine puts out 9,979 kgf of thrust - 97.86 kN. As you can see the Earth Return stack masses just 98,862 kg - perhaps a couple of tons less once waste is dumped pre-orbital insertion. Hence it begins deccelerating at 1.01 m/s^2, burning its 18,413 kg of propellant which enables ~ 650 m/s dv. This is easily enough to park the Voyager in a Highly Elliptical Earth Orbit and let the CSM de-orbit. The re-entry speed is higher than an LEO return, but a bit lower than a Lunar mission return, hence it is easily handled by the CM's heat shields.
But what will it be like for the crew? The gees can get quite high - will it be tolerable after months of zero-gee? In the novel, Voyage, NASA has already made several long duration missions to a Moon-Lab, which is Skylab in Lunar Orbit. Hence they have already tested astronauts at the expected gee-load. But I wonder, of course, if some sort of gees can't be supplied by rotating the Voyager...
And tonight on Channel 10 News is a quite good article on the Mars Society's mission to Australia's Red Centre - not quite the Red Planet, but a very useful exercise. A real voyage to Mars will require a lot of parallel thinking to get carried off successfully - raw Space and Mars itself are alien, deadly environments that need our collective skills as a species to survive. Hence my analogy, parallel thinking by many minds to let those first pioneers have the best chance of survival.
The elastic space-suit that got mentioned sounds really interesting - rather than gas pressure the idea is to supply counter-pressure for breathing via elastic force of the suit itself. This allows better freedom of movement and uses less air-supply to sustain. Also there can be no explosive decompression if there is a tear. Another technology they should be working on (and probably are) is rebreather oxygen supplies, which scrub expelled air of carbon dioxide and return it to the system, rather than directly venting like SCUBA systems. With such an astronaut could carry oxygen for a day easily, rather than mere hours.
Mars also allows lighter meteroid protection. On the Moon micro-meteroids had to be dissipated by solid and padded protection in the suits, else the astronauts would be covered by bruises (or worse) fairly quickly. On Mars the atmosphere burns-up/deccelerates all those hyper-velocity pests, letting the far fewer larger meteorites through to dig those craters being explored by Spirit and Opportunity currently.
Posted at 12:18 am by Adam
Sunday, August 01, 2004
While Discovery has been my main project I haven't forgotten the other fictional option - taking a Saturn Orbital Workshop - Skylab - to Mars and back. To boost inwards to Venus from LEO takes ~ 3.5 km/s, takes 0.4 years then 0.55 years back out to Mars. To save fuel a highly elliptical orbit around Mars is needed - Mars arrival takes ~ 1.5 km/s and departure [sans 25 tons of MEM] takes 0.9 km/s for an Earth return.
The major problem in that case is choice of fuel - a Saturn IV-B uses LH2/LOX, but LH2 boils away in its tanks and needs to be vented or else the tank will rupture. Over a year plus much of the initial fuel will be gone. Passive cooling in insulated tanks needs the slow vent of cryogens to work, while active cooling takes energy and machinery, but avoids venting.
Perhaps trusty old UDMH/N2O4 is the way to go. Certainly storable indefinitely but the ~2.5 km/s dv at Mars will need ~ 130 tons fuel. Which means more propellant to boost out of LEO.
Yesterday I discovered happily that the Apollo CSM can easily brake the stack into a Highly Elliptical Earth Orbit (300 km x 200,000 km say) - a 100 ton cluster gets a dv ~ 650 m/s out of the 18,413 kg fuel in the CSM's tanks. Only about ~ 570 m/s is needed, the rest is to put the CSM into a re-entry orbit.
Friday I figured out how much propellant a Saturn V could boost to LEO in the tanks of the IV-B stage, with just an aerodynamic shroud. Approximately 96 tons useable propellant will remain in the tanks after Orbital Insertion. So not much modification would be needed to launch suitable boosters for the Mars or Titan missions.
Posted at 10:21 am by Adam
Thursday, July 29, 2004
Last entry I linked to the idea that the Shuttle Orbiter should carry a Soyuz in its cargo bay as an emergency re-entry vehicle [ERV] - in reality all that is needed is the re-entry capsule and the engine unit to give the de-orbit burn and escape maneuvering. Such a configuration is not too far off the Zond configuration for circumlunar flights...
So I wonder what does a Soyuz need to get to the Moon? Or at least a looping orbit around the Moon - a free return trajectory - which has recently been advocated for a Space Tourism option, here...
July 27 Astronotes
A Moon-bound Soyuz needs a fairly large boost to get there - it's about a ~ 3 km/s boost from LEO and a Soyuz/Zond masses ~ 5.4 tons. Here's the Soviet-era circum-lunar orbital system that successfully flew a few flights around the Moon, but only managed one successful re-entry that wouldn't have killed/injured its crew...
Lunar 1 configuration
...obviously needs some work before it becomes a tourist option. Twenty-gee ballistic burns through the atmosphere would be a battering for anyone - much better to have a controlled "double-dip" re-entry just like Apollo. However the Block D stage that is the main-engine has been launched successfully for years - unlike the explosive hiccups in its early days - and has proven reliabilty...
...all in all the system masses the 13,360 kg of the Block D, plus the 5,400 kg for the Soyuz/Zond. About ~ 18,800 kg all connected up. This also fits neatly into the Shuttle's launch bay and is well under the maximum payload mass of ~ 24,400 kg. The actual Zond circum-lunar stack massed ~ 18.2 tons at the start of its translunar burn, so a tourist version might mass about the same. The system was ~ 11 metres long so there would be plenty of room in the Shuttle Orbiter's 18.3 metre long payload bay.
Posted at 1:06 pm by Adam
Wednesday, July 28, 2004
Here's a good concept - if NASA's so afraid of going to service the Hubble Space Telescope because the astronauts have no bail-out option in case of Orbiter damage, then why not take an emergency ride home? Here's a post at Save Hubble...
Save Hubble Posts
...simply: a Soyuz fits neatly in the Shuttle's payload bay, so why not take your spare spacecraft with you? Not an insane idea at all. Two Soyuz could easily fit in the payload bay and only mass 7 tons each, which is half the Shuttle's maximum payload. They're only 7 metres long so a docking node or Hubble repair gear could fit in there with them too.
Really radical idea, since the Russians sent stripped back Soyuz around the Moon as Zonds, is to use Soyuz/Zond vehicles for crew transfer to a Moon Station... hmmm...
Posted at 2:06 pm by Adam
Stephen Baxter's Discovery recapitulates Cassini's original flight-plan to the day, but in reality this is highly unlikely. Cassini launched in 1997 while Discovery launches in 2008 - and all the planets involved would have moved into new relative positions between those two dates. Multi-planet gravitational assists that can reach Saturn do happen frequently - several flight-plans were available for Cassini if the early launch window was missed. Fortunately Cassini was launched close to on time as the later flight-plans all took substantially longer, but eventually a shorter flight configuration would reappear.
If Discovery were in Cassini's current orbit (periapsis 1.33 Saturn radii, apoapsis 150.5 radii) and then able to change its periapsis to meet Titan directly the total time from SOI would be about 18 weeks and the aero-braking speed a mere 3.18 km/s - to enter a 4 hour orbit Discovery would then need to aerobrake by a mere 1,295 m/s. The periapsis raise would cost ~ 728 m/s of dv and rocket-braking into that 4 hour orbit would take 1,326 m/s, so aerobraking is definitely better. To do so I am not so sure how the Discovery could be configured to aerobrake - it has two Apollo Command Modules hanging off a docking node in the payload bay.
An Apollo CM is 12'10" (3.91 m) wide, and certainly short enough to fit the 4.6m x 18.3 m bay. But there are Topaz reactors, Spacelab and ISS modules squeezed in there too. However the aerobraking at 3.18 km/s need not be as fiery as a full re-entry and a few dips through the atmosphere might do the job.
Posted at 1:32 pm by Adam
Tuesday, July 27, 2004
Discovery arrives at Titan...
Back to our major theme - the fictional Discovery mission to Saturn and Titan. Once Discovery crosses into Saturn's Hill Sphere it has some important tasks:
deccelerate from hyperbolic velocity
change into Titan's orbital plane
line up an intercept orbit for Titan
deccelerate into orbit around Titan
land on Titan
Cassini - and Baxter's Discovery presumably - deccelerated into a highly elliptical orbit, then maneuvered into a Titan approach orbit. Cassini's current orbital path - modified from the originally planned Titan approach to improve communication with Huygens - needs two major maneuvers:
Saturn Orbital Insertion (SOI) - about 626 m/s
Periapsis raise - about 391 m/s
The SOI was completed successfully, though Cassini plowed through substantial amounts of dust nothing fatally injured the vehicle. The periapsis raising maneuver should go smoothly as well being a shorter main engine burn than the SOI.
The fictional Discovery suffered a decompression from a ring-plane impact, but otherwise successfully arrived. Baxter's described decceleration level was actually about 10 times too high - 1/10th of a gee instead of the actual 1/100th. The Shuttle's OMS is actually only capable of about ~ 2/100th gee. Poetic license I guess.
A direct braking into a Titan approach orbit would take a dv of ~ 1,400 m/s, but Titan is unlikely to be so obligingly in the right part of its orbit to be approached so directly. And I can't currently work out where Titan really would be with respect to a real Discovery in circa 2014 AD, but I'm working on it ;-)
However a direct approach is not the best. To minimise propellant load Discovery needs to aerobrake in Titan's upper atmosphere to shed its relative velocity and a direct orbit doesn't minimise that relative velocity. An indirect approach from a higher orbit with Titan's radius as the periapsis has a better chance of minimising the relative velocity.
Cassini's current orbit would have required aerobraking at 8.32 km/s (29,950 km/h) which is higher than the Shuttle's usual aerobraking range.
A direct approach would require aerobraking at 5.38 km/s (19,400 km/h.)
An apoapsis at 210 Saturn radii with Titan at the periapsis (20.273 radii) means aerobraking at a mere 3.26 km/s (11,740 km/h.)
Shedding a mere ~ 1,375 m/s for a 4 hour parking orbit means a lower heating load. Propellant needed for the two maneuvers is also lower than the direct orbit, saving about 320 m/s of delta-v. The penalty is the +30 weeks spent getting into the right orbital positions.
Posted at 3:15 pm by Adam
Monday, July 26, 2004
In the late 1960s physicist Peter Glaser proposed to build huge solar energy collecting satellites, but no one took him seriously in the age of cheap oil and "safe" nuclear power. Then the Oil Crunch of the mid-70s hit and the US Department of Energy (DoE) and NASA took him seriously. The Solar Power Satellites (SPS) they designed were typically capable of delivering 10 gigawatts power to the ground, with about 62.9% efficiency. In space that meant 129 km^2 arrays of solar cells massing perhaps ~ 80 - 100 thousand tons, beaming power to the ground as microwaves.
Their designs made a lot of assumptions that made them difficult -
on-orbit assembly was assumed with a huge work-force and inter-orbital transportation system requiring everything haulled from the ground, cargo haulled to GEO via giant ion-drive space-tugs and faster chemical fuel tugs delivering workers
arrays were supported by rigid frameworks that had to be assembled on-orbit
power transmission was at lower microwave frequencies that minimised atmospheric absorption, but need huge transmitting antennae
the solar array had to track the Sun continuously and the antenna had to point to the power receiver continuously
huge dedicated cargo delivery aerospace vehicles had to be developed for the program to deliver equipment to LEO
Each assumption complicates design and increases cost and mass delivered to the target orbit. Geoffrey Landis, a physicist who works for NASA and writes SF, has discussed SPS alternatives extensively. He has several interesting papers at his web-page...
Landis technical papers online
...scroll down and you will find the SPS section. The first article in the list is the most recent and well worth a read. He proposes a serious rethink of all assumptions about SPS and then describes a system that is nearly viable for power supply today.
His next article on Super-Synchronous SPS provides a good description of an ultra-light weight system. Compare it to the old DoE/NASA SPS satellites - its inflatable, uses advanced concentrators, doesn't need a separate, rotating transmitter, can supply power from the beginning and only masses 1,300 tons for 1 to 2 gigawatts power supplied to the ground.
An alternative for transporting SPS is to deploy an array sub-unit at LEO and use its power for an ion/plasma drive to move it to GEO and/or L1/2. Potentially this approach will cut mass delivered to LEO to a mere 30% of trying to do it all via chemical rockets. The apparently defunct PowerSat corporation tried to patent this concept, but I believe that this article...
Electric Propulsion for SSPS
...indicates a prior useage.
Posted at 12:58 pm by Adam