kunkmiester wrote:This has been covered a bit. There's ways to shift the voltage down reasonably efficiently. Someone else will have to go into detail on that.
On the other hand, if you can pulse your fuel feed so you get a pulsed output, you can then just run the output through the proper transformers to step down the voltage, and to get AC. That was my dad's idea, I recall someone mentioning a criticism of the idea, but I don't remember what it was.
If Direct Conversion works...
Similar comment upstream:
-
kunkmiester
- Posts: 892
- Joined: Thu Mar 12, 2009 3:51 pm
- Contact:
Thanks, Delta.
Big thing was that a giant transformer that'll take a million volts might be easier to build than a mosfet or something like that.
My brother and I once messed with how big a wire you need to get it straight out. It was something like a foot thick of copper. While someone who has a better idea of how that works might change it, simply coiling that kind of wire up into a transformer would simply be capital costs and the labor, not much engineering versus semiconductors, the idea was.
Big thing was that a giant transformer that'll take a million volts might be easier to build than a mosfet or something like that.
My brother and I once messed with how big a wire you need to get it straight out. It was something like a foot thick of copper. While someone who has a better idea of how that works might change it, simply coiling that kind of wire up into a transformer would simply be capital costs and the labor, not much engineering versus semiconductors, the idea was.
Evil is evil, no matter how small
At 1 MV, I think the biggest concern is arcing (see above picture of a 0.8 MV transformer).
For a flying Polywell, with some oscillation of direct conversion high voltage thrown in (by whatever means, POPS or ?), something else is probably needed. Maybe something like electrostatically-driven, mechanical oscillation of a plate, connected to a nonconducting pushrod, that exits the vacuum chamber through a flex-seal to drive an external linear generator (like the ones used for solar-Stirling systems... small linear oscillations of permanent magnet armatures inside coils). The frequency would have to be in a "mechanical" range, though. Maybe there's a "solid-state" equivalent for higher frequencies.
I'm talking here about "only" 10% or so of max reactor power, but still, several linear generators would probably be needed around the outside of the vacuum chamber.
For a flying Polywell, with some oscillation of direct conversion high voltage thrown in (by whatever means, POPS or ?), something else is probably needed. Maybe something like electrostatically-driven, mechanical oscillation of a plate, connected to a nonconducting pushrod, that exits the vacuum chamber through a flex-seal to drive an external linear generator (like the ones used for solar-Stirling systems... small linear oscillations of permanent magnet armatures inside coils). The frequency would have to be in a "mechanical" range, though. Maybe there's a "solid-state" equivalent for higher frequencies.
I'm talking here about "only" 10% or so of max reactor power, but still, several linear generators would probably be needed around the outside of the vacuum chamber.
I calculate that with 1.3 MVDC, a tungsten heating element capable of dissipating 3 GW at about 3200 K while weighing 2000 kg would involve 7.5 km of coiled wire with a radius of about 2 mm. Variations in temperature along the length of the wire will change this; at room temperature the values are 33 km and just under 1 mm...
[I hope I didn't make a mistake there; it's essentially Ohm's law and cylindrical geometry...]
Yes, that's 2300 A going through that wire...
Naturally this would have to be electrically insulated by something with a high breakdown field, high thermal conductivity and very good high-temperature properties... which could then in turn be protected with a strong, high-thermal-conductivity shell that wouldn't evaporate or ablate during flight...
Funny - reducing the voltage only increases the difficulty of getting the airflow to accept the energy from the heating element... No idea how hard or otherwise it is with the above harebrained scheme; my heat transfer textbook is elsewhere at the moment...
On the other hand, it would be nice not to have to worry about turbines and such (makes the combined-cycle engine easier to design and potentially more efficient), and the heat pumps for the wing radiators are going to be in the GW range for sure, even if you only ever use one or two cores while in CSR mode...
(BTW, I'm currently working with 8 cores for a total of 48 GW. I've assumed that with POPS and whatnot we can get the vacuum chambers down to 6 metres across. Gamma and X-ray shielding has been reduced to 3" of lead except for a 12" thick shadow area in the front of the cluster, yielding about 900 tonnes of shield mass for the whole thing. Subsonic won't require that much power, but VTOL will require multiple GW - even modern ultra-high-bypass turbofans would need to expend several GW to lift a nearly-3000-tonne aircraft straight up... actually, this thing might just be able to land on Mars and take off again without refueling; for 48 GW at an Isp of 800 seconds, assuming jet power efficiency of 90%, you can get about 11 MN of thrust, which will get a 2900-tonne spaceship off the surface of Mars vertically without using any of the ambient air as extra propellant... which you can start to do once you get going... okay, I'm rambling...)
[I hope I didn't make a mistake there; it's essentially Ohm's law and cylindrical geometry...]
Yes, that's 2300 A going through that wire...
Naturally this would have to be electrically insulated by something with a high breakdown field, high thermal conductivity and very good high-temperature properties... which could then in turn be protected with a strong, high-thermal-conductivity shell that wouldn't evaporate or ablate during flight...
Funny - reducing the voltage only increases the difficulty of getting the airflow to accept the energy from the heating element... No idea how hard or otherwise it is with the above harebrained scheme; my heat transfer textbook is elsewhere at the moment...
On the other hand, it would be nice not to have to worry about turbines and such (makes the combined-cycle engine easier to design and potentially more efficient), and the heat pumps for the wing radiators are going to be in the GW range for sure, even if you only ever use one or two cores while in CSR mode...
(BTW, I'm currently working with 8 cores for a total of 48 GW. I've assumed that with POPS and whatnot we can get the vacuum chambers down to 6 metres across. Gamma and X-ray shielding has been reduced to 3" of lead except for a 12" thick shadow area in the front of the cluster, yielding about 900 tonnes of shield mass for the whole thing. Subsonic won't require that much power, but VTOL will require multiple GW - even modern ultra-high-bypass turbofans would need to expend several GW to lift a nearly-3000-tonne aircraft straight up... actually, this thing might just be able to land on Mars and take off again without refueling; for 48 GW at an Isp of 800 seconds, assuming jet power efficiency of 90%, you can get about 11 MN of thrust, which will get a 2900-tonne spaceship off the surface of Mars vertically without using any of the ambient air as extra propellant... which you can start to do once you get going... okay, I'm rambling...)
Last edited by 93143 on Mon Dec 07, 2009 2:53 am, edited 3 times in total.
Re: ripple...
Newly built traction grids (like the High Speed 1 in UK and HSL in Belgium/Holland) use 50Hz line at 25kV. It makes distribution a lot easier (no need for huge rectifiers in substations) and, more importantly, it allows power to flow back into the grid during regenerative braking.Nik wrote:Not to put too fine a point on it, sure would help if you could pulse the system to some extent. Being able to 'syphon off' ripple with a transformer gives you utility power at manageable voltage...
FWIW, some trains run on 16.3 Hz...
Dumb answer here, but ever heard of APU's? They could easily run in a closed loop fuel circuit on hydrogen or any other fuel that is replenished by Polywell power once you're in orbit.Just to be clear, I'm talking about stepping down voltage for 100+ MW of power (at most hundreds of MW), used for the avionics, actuation, reactor thermal management, cabin/EE bay environmental control systems and most importantly the superconducting, electromagnetic lift fans and turbines, before the QED-ARC engine fires up for the kick into orbit, not conversion of the full Polywell output (GWs) that the power-hungry, high-voltage QED-ARC uses.
Because we can.
Re: ripple...
APUs are normally used to start propulsion engines, like the GE-90s for a 777, which combined put out about 150-160MW. APUs are at least an order of magnitude less power than needed for propulsion. If you had a lot of them, however... maybe. If you don't ignore reactor shielding mass (as I do, and 93143 does not), add more to the power needed for VTOL.Stoney3K wrote:Dumb answer here, but ever heard of APU's? They could easily run in a closed loop fuel circuit on hydrogen or any other fuel that is replenished by Polywell power once you're in orbit.
And don't forget that those pulses will probably have currents of several kA, if not more. If you start switching that, any circuit with a line inductance of more than zero will quickly ramp that up into the MV range.93143 wrote:Oh, and pulse isn't necessarily so useful, even if you could get a perfect sine wave, because the mean is still several hundred kV DC...
Only true in aviation. Since you've got a theoretically limitless supply of fuel anyway, you can get away with a huge APU, no problem. You're talking about a bigger craft than a T7 on all aspects, so that definitely warrants a bigger APU. They're called Auxilliary Power Units for a reason, otherwise they'd be called 'starter motors'.APUs are normally used to start propulsion engines, like the GE-90s for a 777, which combined put out about 150-160MW. APUs are at least an order of magnitude less power than needed for propulsion. If you had a lot of them, however... maybe. If you don't ignore reactor shielding mass (as I do, and 93143 does not), add more to the power needed for VTOL.
Because we can.
Parallel hardware...
Uh, what did I miss ? Why have multiple modular polywell hardware, then try to force all that power through one (1) conduit / busbar ?? And the waste heat through one (1) heat exchanger ??
Surely, you'd want to parallel lower-rated power-handling equipment for safety, redundancy and economy of production ??
Slightly OT: The astonishing heat-exchangers designed for RE's Skylon's Sabre engines may be useful here, too. Their power density is remarkable-- 400 MW through a 'trashcan'.....
http://www.reactionengines.co.uk/he_man.html
Surely, you'd want to parallel lower-rated power-handling equipment for safety, redundancy and economy of production ??
Slightly OT: The astonishing heat-exchangers designed for RE's Skylon's Sabre engines may be useful here, too. Their power density is remarkable-- 400 MW through a 'trashcan'.....
http://www.reactionengines.co.uk/he_man.html
Who said anything about using only one unit?
3 GW is one sixteenth of the available power. If you wanted to push full power through four engine nacelles in ARC-ZMO mode (which I hope isn't necessary), you'd need four of those elements in each one. Plus a heat exchanger upstream of them to dump pumped-up waste heat into the airflow... actually, you might need to waste a quarter or so of your available power just on the heat pumps, so make that three heating elements per engine...
Naturally if you could efficiently step down the voltage, you could run multiple shorter heating elements in parallel. Even for the high-voltage case, I'm not pretending that my numbers are optimal - but they are at least physical... you know, it would definitely help to have more than three available throttle settings for the engines - this requires more consideration, but voltage stepdown is the obvious solution if it can be done within a reasonable mass budget...
As for cooling, the wing radiators I've conceptualized would use 4 or 8 separate loops, at least, and they only have to handle waste heat from one or two cores (for all eight, the required temperature and/or wing area become ridiculous). It remains to be seen what sort of pressure drop an adequate heat exchanger design would induce...
Actually, with the number of systems (and hypersonic leading edges) that need to be cooled, this vehicle is going to be plumbed all through like a living thing...
I suspect that using hundreds of smaller systems, each in the capacity range of a 100 MW power plant, is not an optimal engineering solution - there's a reason only the very largest airliners have more than two engines...
Lest you think these power levels are preposterous, I remind you that the Saturn V first stage developed about 44 GW of jet power...
3 GW is one sixteenth of the available power. If you wanted to push full power through four engine nacelles in ARC-ZMO mode (which I hope isn't necessary), you'd need four of those elements in each one. Plus a heat exchanger upstream of them to dump pumped-up waste heat into the airflow... actually, you might need to waste a quarter or so of your available power just on the heat pumps, so make that three heating elements per engine...
Naturally if you could efficiently step down the voltage, you could run multiple shorter heating elements in parallel. Even for the high-voltage case, I'm not pretending that my numbers are optimal - but they are at least physical... you know, it would definitely help to have more than three available throttle settings for the engines - this requires more consideration, but voltage stepdown is the obvious solution if it can be done within a reasonable mass budget...
As for cooling, the wing radiators I've conceptualized would use 4 or 8 separate loops, at least, and they only have to handle waste heat from one or two cores (for all eight, the required temperature and/or wing area become ridiculous). It remains to be seen what sort of pressure drop an adequate heat exchanger design would induce...
Actually, with the number of systems (and hypersonic leading edges) that need to be cooled, this vehicle is going to be plumbed all through like a living thing...
I suspect that using hundreds of smaller systems, each in the capacity range of a 100 MW power plant, is not an optimal engineering solution - there's a reason only the very largest airliners have more than two engines...
Lest you think these power levels are preposterous, I remind you that the Saturn V first stage developed about 44 GW of jet power...
Actually I was thinking around 2,000,000 - 3,000,000 lbs of LH2, with about 300,000 lbs of payload to the lunar surface.
Unfortunately, I don't have an airframe design yet (and anything I can come up with by myself will be very low-fidelity), so the above is just handwaving.
Take a 747, which is about 200 tonnes dry, multiply it by 3 for extra systems and high-temperature materials, and double it for the wider fuselage and bigger wings, yielding 1200 tonnes, or 2,645,547.12 lbs. Add 730 tonnes for the shielding (assuming you save 200 tonnes by not bothering with the neutron moderator/absorber, and just letting the lead handle it). I did a back-of-the-eyelids calculation indicating that you might be able to fit about 1400 tonnes of LH2 in a big tank wrapped around the top of the fuselage. Assuming you need 4 km/s at 6000 seconds (airbreathing) and 4 km/s at 3000 seconds (rocket) for the ascent from maximum ARC-ZMO speed to orbit, and about 8.5 km/s at 8000 seconds and 4 km/s at 2000 seconds to get to the moon and back, and assuming you can do a high-lift unpowered reentry back down to ARC-ZMO speed (which may well be impossible for a vehicle this heavy, regardless of how big the wings are) you have a total mission mass ratio of 1.62. So the payload is around 130 tonnes to the lunar surface... not bad. That number, BTW, is extremely sensitive to the trade between Isp and gravity losses during lunar surface access. Detailed design and trajectory optimization could easily give a payload anywhere between 1000 and -1000 tonnes - hopefully there'll be enough hydrogen left over to do a partially-propulsive descent once you get back to Earth...
See? Handwaving.
A propellant depot in LEO makes this look a lot better. Another one in LMO (Phobos station maybe?), even better...
Does anyone have any idea what sort of masses we're talking about for GW-range DC voltage stepdown?
Unfortunately, I don't have an airframe design yet (and anything I can come up with by myself will be very low-fidelity), so the above is just handwaving.
Take a 747, which is about 200 tonnes dry, multiply it by 3 for extra systems and high-temperature materials, and double it for the wider fuselage and bigger wings, yielding 1200 tonnes, or 2,645,547.12 lbs. Add 730 tonnes for the shielding (assuming you save 200 tonnes by not bothering with the neutron moderator/absorber, and just letting the lead handle it). I did a back-of-the-eyelids calculation indicating that you might be able to fit about 1400 tonnes of LH2 in a big tank wrapped around the top of the fuselage. Assuming you need 4 km/s at 6000 seconds (airbreathing) and 4 km/s at 3000 seconds (rocket) for the ascent from maximum ARC-ZMO speed to orbit, and about 8.5 km/s at 8000 seconds and 4 km/s at 2000 seconds to get to the moon and back, and assuming you can do a high-lift unpowered reentry back down to ARC-ZMO speed (which may well be impossible for a vehicle this heavy, regardless of how big the wings are) you have a total mission mass ratio of 1.62. So the payload is around 130 tonnes to the lunar surface... not bad. That number, BTW, is extremely sensitive to the trade between Isp and gravity losses during lunar surface access. Detailed design and trajectory optimization could easily give a payload anywhere between 1000 and -1000 tonnes - hopefully there'll be enough hydrogen left over to do a partially-propulsive descent once you get back to Earth...
See? Handwaving.
A propellant depot in LEO makes this look a lot better. Another one in LMO (Phobos station maybe?), even better...
Does anyone have any idea what sort of masses we're talking about for GW-range DC voltage stepdown?
I don't, but just to keep the ball rolling...93143 wrote:Does anyone have any idea what sort of masses we're talking about for GW-range DC voltage stepdown?
If the alphas are really channelled along the magrid coil axes, use them to spin up six or more copies of something like this (scaled up)
http://www.arcsandsparks.com/tourbillion.html
connected to several copies of something like the generator version of this
http://www.synchrony.com/products/high- ... ators.aspx
to power several copies of the motor version of the same, connected to lift fans and turbines, to get to QED-ARC altitude.
Superconductors used where practical.
Curiously enough, if you assume Mars surface access from Phobos Station requires 7 km/s at 30,000 N·s/kg (3059 seconds) and 2 km/s at 8000 N·s/kg (816 seconds), you end up with a mission mass ratio of 1.62... hopefully using ambient air (such as it is) will help out with that a bit...
Is there an easy or cheap way to cut a high voltage in half, or something like that? Something like that rotating disk capacitor thingy? That might ease the load on the rest of the system...
I just thought of something. If you were using multiple heating elements in series in the engine (there would be multiple sets of these in parallel, of course), you could get power at a certain voltage by tapping into the line partway down the chain. Sort of like a high-power rheostat, except the rejected high-voltage energy is doing something useful... Trouble is, this doesn't work if you aren't using the heating elements... and the maximum power ratio equals the voltage ratio, so you can't get a substantial fraction of the total power down to an insubstantial fraction of the total voltage...
Didn't we, on that other thread, discuss the possibility of using a closed heat engine cycle with a REB as the heat source? Waste heat from that could be dumped into the engine along with the rest - but the heat is dumped at a pretty high temperature; what would the efficiency of the overall cycle be? How hot would the hot side have to get in order to produce enough net power? 2000 K, even in ARC-ZMO mode? More? Way more, in ARC-QED mode, which leaves us with a magnetically shielded fluid loop and a magnetohydrodynamic turbine - whoops, those don't have very good efficiency at all, do they?
I can't see that tourbillion thing having a significant efficiency. It seems like more of a curiosity to me... besides, if you had a high-voltage motor, why would you hook it to a generator hooked to a low-voltage motor? Why not simply use the resulting rotary power as is, since it's what you want anyway?
Superconducting motors WILL be necessary...
Is there an easy or cheap way to cut a high voltage in half, or something like that? Something like that rotating disk capacitor thingy? That might ease the load on the rest of the system...
I just thought of something. If you were using multiple heating elements in series in the engine (there would be multiple sets of these in parallel, of course), you could get power at a certain voltage by tapping into the line partway down the chain. Sort of like a high-power rheostat, except the rejected high-voltage energy is doing something useful... Trouble is, this doesn't work if you aren't using the heating elements... and the maximum power ratio equals the voltage ratio, so you can't get a substantial fraction of the total power down to an insubstantial fraction of the total voltage...
Didn't we, on that other thread, discuss the possibility of using a closed heat engine cycle with a REB as the heat source? Waste heat from that could be dumped into the engine along with the rest - but the heat is dumped at a pretty high temperature; what would the efficiency of the overall cycle be? How hot would the hot side have to get in order to produce enough net power? 2000 K, even in ARC-ZMO mode? More? Way more, in ARC-QED mode, which leaves us with a magnetically shielded fluid loop and a magnetohydrodynamic turbine - whoops, those don't have very good efficiency at all, do they?
I can't see that tourbillion thing having a significant efficiency. It seems like more of a curiosity to me... besides, if you had a high-voltage motor, why would you hook it to a generator hooked to a low-voltage motor? Why not simply use the resulting rotary power as is, since it's what you want anyway?
Superconducting motors WILL be necessary...
The voltage across the rotating capacitor could be designed low enough to avoid vacuum arcs (field electron emission), and then several stages could be cascaded in series to get enough voltage step-down. To get sufficient power transfer through the "transformer" a high-enough capacitance per stage is necessary. Close spacing of rotating plates in a vacuum, using plates that are thick and stiff (honeycomb/isogrid) and/or some sort of active structural control would be required. The higher the RPM, the greater the power transferred, up to a point (you need to allow 5 time constants for charge/discharge). If the RPM was low enough, maybe a dielectric with a high breakdown voltage could be used ( http://en.wikipedia.org/wiki/Dielectric_strength ), but then you have to deal with friction/wear, which gets worse with plate area, compression force, and RPM.93143 wrote:Is there an easy or cheap way to cut a high voltage in half, or something like that? Something like that rotating disk capacitor thingy? That might ease the load on the rest of the system...
The "toy" electrostatic motor linked above runs in air and is thus prone to arcing. I don't know if sufficient charge deposition could occur in a scaled-up, vacuum-enclosed variant (via thermionic or field electron emission), or what the best rotor material would be, best electrode shape, and so on. I have not seen any research done on electrostatic motors in vacuum, power levels obtained, etc. (there is some recent research from Japan on linear electrostatic motors/actuators, not in vacuum, but they probably work in one). The rotary motor in a vacuum seems to be unexplored territory, but it's likely research was done one or two centuries ago, before electromagnetic motors took the lead. Hobbyist potential...93143 wrote:I can't see that tourbillion thing having a significant efficiency. It seems like more of a curiosity to me...
That means using a lot of driveshafts if you want the redundancy of distributed lift/thrust, as opposed to routing low-voltage cables. For using four thrust pods, like you mentioned before, that might be the way to go. I get the willies when I think about a pod failure on the V-22, even though I know the rotors are cross-linked via driveshafts.93143 wrote:besides, if you had a high-voltage motor, why would you hook it to a generator hooked to a low-voltage motor? Why not simply use the resulting rotary power as is, since it's what you want anyway?