Knowledgeable reduction of my ignorance RE electric propulsion much
appreciated - thanks!
So, all other electric-propulsion things being equal, in terms of
overall efficiency of converting electrical energy in to thrust energy
out, the ionization energy requirement favors heavier atomic-weight
propellants and higher isp's. But given current ~10%-of-theoretical
ionization efficiencies, there is some scope for reducing that factor.
Good to know. I will shelve for now pondering what propellant might be
economically optimal twenty years from now - it all depends both on what
actually turns out to be cheaply available locally, and on how the
propulsion tech evolves.
One observation, however. Electrical energy is just one more economic
input to balance. It's possible the eventual economically optimum
balance will accept less than the max possible electrical efficiency in
order to reduce other system costs. Which leads to a cultural
observation I suspect we both can agree with, having both seen horrible
examples of it in action: Fixing on one parameter to optimize above all
others early in a system development process can lead to less than
optimal (and sometimes downright peculiar) results!
RE the need to reduce inner-system travel times in general (you should
take "Mars" as an easily grasped representative example for popular
consumption), it needs to be done for cost reasons. One, ships with
systems requiring less per-mission endurance will, all else equal, cost
less. Two, ships that fly more round trips over their economic lifetime
will, all else equal, cost less. Three, people will always prefer
faster transportation if they can afford it - our time also has a cost.
Now for once-in-a-lifetime national-prestige "exploration" missions,
none of these are major factors. (Though significantly faster transit
time is one fix for the still-poorly-managed risk of intrepid explorers
arriving too debilitated to explore.) But for future human expansion
out into the Solar System, speed and cost very much matter. (so, which
side did you say you were on again?.. :-D )
Henry
On 8/26/2019 7:57 PM, John Schilling wrote:
Lower efficiency is to some extent inherent in the technology. Unless you have a way to use electricity to accelerate neutral atoms or molecules, you have to ionize everything first and that's a fixed energy cost (theoretically ~10 eV/ion, in practice ~100 eV) before you get any acceleration. The more you accelerate, the less energy you're wasting. So, maybe 50% at 1500 seconds, 65% at 3000 seconds, and possibly up to 90% when you are dominated by losses in the power-conversion and acceleration mechanisms. And we're looking for less wasteful ways to do the ionization.
Xenon is definitely preferred as a propellant; Krypton will also work but cost you 5-10% in efficiency because of the higher ionization energy per unit mass (roughly the same per ion, but lighter ions. Light elements like hydrogen are problematic because it takes a *lot* of hydrogen ions to make a kilogram. Anything with oxygen is problematic because it is still too light and because you basically can't ionize oxygen without turning it into atomic oxygen in the process, and that stuff is about as corrosive as Alien blood. Also, if your way of making ions involves a cathode, that's going to electrostatically beckon "Here, oxygen. Here, oxygen oxygen oxygen! Come get your free electrons!", and then dissolve into goo.
Heavy metal vapors will work quite well if you can find a metal with a high vapor pressure at reasonable temperatures and can avoid having it plate out over your feed system, insulators, etc. I've seen preliminary work done with bismuth, zinc, and iodine. Mercury would be ideal if it weren't for the toxicity problems. But if you're in an environment where everybody has to wear a spacesuit anyway, and if I recall correctly the LCROSS measurements suggested there was quite a bit of mercury in the cold trap at the Lunar south pole.
I'm not sure we need to get Mars travel times down to weeks, but your math looks right for what it would take to do that.
John Schilling
john.schilling@xxxxxxxxxxxxxx
(661) 718-0955
On 8/25/2019 10:48 AM, Henry Vanderbilt wrote:
65% electricity-to-thrust efficiency? Not bad at all. I assumed something like 70% in coming up with my ~1 kW/kg power supply requirement for a usefully faster-than-chemrocks space transport. (More on what went into that in a moment.)
Is the lower efficiency at lower Isp inherent in the technology, or a matter of the chosen design optimization point being ~3000 seconds Isp?
And, AIUI, Hall-effect thrusters tend to use inert gases for propellant. What are the prospects, if any, for hitting similar efficiencies with cheaper and/or more common propellants - say, for example, water? Or liquid oxygen? With, perhaps, electric thruster technologies other than Hall-effect?
I ask because for a number of reasons, ~2000 seconds Isp, AKA ~20 km/s exhaust velocity, seems to me about right for a high-energy space transport intended to supplant chemical rockets for inner solar system runs. And a network of such transports would one hopes expand traffic to the point where cheap common (and dense and easily storable) propellant would be important.
Per various Jim Wertz papers (and grossly oversimplifying) Earth-Mars transfers start getting down to weeks rather than months for one-way delta V's on the order of 20 to 30 km/s. (Earth-nearby asteroid and Earth-Venus transits require somewhat less.) A basic ship Mass Ratio of around three, AKA two parts propellant to one part everything else, is reasonable (and reasonably expandable) designwise, and yields mission delta V of ~1.1 times exhaust velocity, or about 22 km/s baseline delta V. Double the propellant via auxiliary tanks for MR=5 and the mission delta V increases to ~32 km/s. Meanwhile, exhaust velocity roughly matching the lower end of the expected range of mission velocities seems to me the way to shade things to minimize the acceleration-time issue, AKA make it easier to get past "mousefart thrust" requiring months of acceleration to reach transit-in-weeks velocities.
The very generic rough system breakdown I come up with: For (semi-arbitrary) 10 MW ship electrical power available, at ~70% electric-to-thrust conversion efficiency and 20 km/s exhaust velocity, you get ~700 newtons of thrust. I assume the ability to accelerate at 1 km/s per day at basic full-propellant-no-auxiliary-tanks mass ratio of 3 as being well into the usefully "zoomy" region - at 86,400 seconds in a day, this means you need to accelerate at .011574 m/s^2, which means you can accelerate 86.4 kg at that rate per newton of thrust, which means your 10 MW power supply will propel a 60 (metric) tone ship - 40 tons of propellant, 20 tons of everything else.
1 kW/kg from your power supply then makes it 50% of "everything else", IE of the entire ship dry mass, which is definitely squeezing the rest of "everything else." But then you can back off a fair amount from that initial full-tanks acceleration of 1 km/s per day and still reach 10-15 km/s delta V in days not weeks. (I need to run those numbers again at some point, but not this morning - I have a real-world ditch to dig.) But ~1 kW/kg does give a reasonable ballpark idea of what's needed in the way of watts per unit mass from a power supply for a usefully faster space transport.
Henry
On 8/24/2019 5:47 PM, John Schilling wrote:
Roughly 3000 seconds Isp at 65% wall-plug efficiency. If you want higher Isp that's easy and the efficiency will go up a bit; if you prefer thrust to power you can go down to ~1500 seconds and still keep the efficiency above 50%.
We'd like to do better on the efficiency front, to be sure, but that's probably going to be a long slog and a few percent at a time.
John Schilling
On 8/24/2019 9:47 AM, Henry Vanderbilt wrote:
What Isp does this setup operate at? With what efficiency turning watts into thrust?
RE nuke-electric versus solar-electric, yes, solar-electric current SOTA is much closer to useful fast-transit power/mass efficiency levels than nukes. With the added assumption of remote lasers to bring solar panels up to near their typical practical upper limit of roughly 2x Earth-solar input levels whether they're here or out near Mars, from what you say, roughly one order of magnitude away in the system you describe. I tend to assume some additional economies from larger scale installations plus a certain amount of subsytem-marketing-brags cherrypicking and assumed engineering tradeoffs, in saying that current optical-electric is about a half OofM short of what's needed.
Conventional fission nukes I suspect are a dead end (for transport apps at least) that will never get near the ~1 kW/kg needed for general-purpose fast inner-system transports. The energy just shows up in way too inconvenient a form, requiring lots of mass for shielding/conversion to thermal energy, then conversion from thermal to electrical energy. Some unconventional nuke technology that does NOT produce lots of messy neutrons might get there - positron-electron? He3 fusion? The ability to convert the energy directly to high Isp thrust without an intermediate electrical step would also help. But, as best I know out here in the unclassified world, those are a LONG way off.
Solar/beam-electric strikes me as the most promising lowest-risk approach to while-we're-still-alive fast inner system transportation. I mention nukes for fairness. And you never know, someone may next year come out with the He3 1 MW Home AC Power Fusor, purchase conversion kit for thrust applications separately, warranty void if not installed by a factory-trained technician. It'd be nice! But I don't count on it.
And yes, in this business, always check the math. It eliminates huge amounts of nonsense, and occasionally leads you someplace unexpectedly useful.
Henry
On 8/23/2019 11:37 AM, John Schilling wrote:
The NASA "HERMes <https://urldefense.proofpoint.com/v2/url?u=https-3A__en.wikipedia.org_wiki_Advanced-5FElectric-5FPropulsion-5FSystem&d=DwMDaQ&c=clK7kQUTWtAVEOVIgvi0NU5BOUHhpN0H8p7CSfnc_gI&r=rPTfWqtJdrL0Ber-yr0E_hSjRXuvJH6ZmQx03u8-2as&m=lgmr_Ur1YnNMRUlLyXe3ZjZyBN8dvIL9I1VaXb7uGUo&s=IZRXE6wZhHUzf8OmvyiYCr1MBZnHLI8F9tfhz7oz3Gw&e=>" Hall thruster is targeting 100 kg mass for a complete 12.5 kW system including power processing unit but power supply. Actual input power is 13.3 kW, which using the quoted performance of the Northrop-Grumman "Ultraflex <https://urldefense.proofpoint.com/v2/url?u=https-3A__www.northropgrumman.com_Capabilities_SolarArrays_Documents_UltraFlex-5FFactsheet.pdf&d=DwMDaQ&c=clK7kQUTWtAVEOVIgvi0NU5BOUHhpN0H8p7CSfnc_gI&r=rPTfWqtJdrL0Ber-yr0E_hSjRXuvJH6ZmQx03u8-2as&m=lgmr_Ur1YnNMRUlLyXe3ZjZyBN8dvIL9I1VaXb7uGUo&s=qJWjaZ9lOLcDhYOJzQV5E5-fB2giSoxV1sWLzcRhazs&e=>" solar array would be an additional ~90 kg BOL. The HERMes numbers are conservative, IMO, and Ultraflex has flight heritage. Accounting for solar array degradation and system integration probably puts you at ~250 kg for a state-of-the-art 12.5 kW solar-electric propulsion system, or 20 kg/KW.
If you need full power at Mars orbit, that would go to ~30 kg/kW. There's room for improvement, of course, and plausible paths to same. It isn't clear that nuclear-electric power supplies will outperform solar-electric until you get somewhere past Mars; nuclear power plants are heavy beasts, and people handwaving that the Magic N-Word means as much energy as we need in as small a package as we need to fulfill the dream du jour are probably selling snake oil. Or maybe aren't selling anything and just haven't done the math. Insist on checking the math.
John Schilling
john.schilling@xxxxxxxxxxxxxx
(661) 718-0955
On 8/23/2019 10:26 AM, Henry Vanderbilt wrote:
A tangentially related thought: In looking over the years at the need for ~2000s Isp, ~ 1km/s/day delta V inner-system space transport to supplant chemical rockets, I came up with a rule-of-thumb: That to do this with nuclear-electric ships, the nuclear electrical generation part of such a system needs sustained (weeks) power output on the rough order of one kilowatt per kilogram of generating plant. Else we're back in the mousefart thrust trap.
My impression is that the current nuclear-electric SOTA might approach 100 kg per kw sustained output. Anyone have better numbers than that, or thoughts on alternative nuclear approaches?
Henry (V)
On 8/23/2019 10:14 AM, Henry Spencer wrote:
On Fri, 23 Aug 2019, William Claybaugh wrote:
The de facto Western ban on testing in the atmosphere means testing will
have to be off-planet.
And for something like NSWR, I'm not sure I can reasonably object to that. For a solid-core design that can be expected to pretty much contain all fission products, I could see testing on the ground using exhaust scrubbers, *perhaps* moving to open-air testing after containment was well verified. But for advanced systems like NSWR, where there's inherently a lot of radioactive garbage in the exhaust and so even one unscrubbed run would be bad, I'd worry about scrubber effectiveness and scrubber failure modes.
Moreover, just a little bit off-planet isn't enough -- you want to test (and operate) outside Earth's magnetosphere.
(Gosh, could this be something that the Lunar Gateway is actually good for? :-) )
This favors "known" designs that can be relied upon to produce a usable vehicle on the first try.
They may be favored, but if they can't meet the mission specs -- and the "known" designs realistically have at best modest advantages over chemical rockets -- then they're not even in the running. The objective is to be useful, not just usable. A modest gain over chemical isn't worth the very high development costs and nasty political hassles.
If reuse is a part of the plan, then some sort of refueling facility is implied, all in a nuclear safe orbit (1000 km, generally). Getting a new payload on a hot vehicle is left as a problem for the reader.
What probably makes the most sense is a concept that was seen in the late 60s: reusable nuclear tugs. They're not part of your mission-specific vehicle; they dock to your ready-to-go vehicle, boost it from LEO to (say) Mars trajectory, and then turn around and come back (immediately, using brute-force high-delta-V retrofire, not an economy orbit) to return to their base. So your payload never spends much time in their vicinity.
If you're going to have to fiddle around for a while during/after docking but before departure, you can fit substantial shielding around the tug docking interface, and have it moved aside five minutes before departure. (Launching it will cost something, but it doesn't *go* anywhere, so that cost only has to be paid once and then it's available for all future departures. Think of it as the equivalent of airport de-icing trucks.)
I also observe that the whole notion of "nuclear safe" orbits implicitly assumes that a failed nuclear-propelled vehicle would be abandoned and left to come down at random. This makes no sense -- quite apart from the side effects of doing so, it's too expensive to just abandon.
(Well, except perhaps in a spectacular one-shot program with no continuation or follow-on, which is expected to crash and burn once it meets its primary objective, and therefore isn't interested in building support infrastructure -- e.g. tow trucks for disabled vehicles -- for future use... and that doesn't make much sense either. Been there, done that, know better now.)
Henry
