(I'm 80% sure this should be in "design" and not "theory". Let me know if it's in the wrong place).
How does one throttle and idle a direct-conversion polywell reactor? Demand for energy will vary, so how do you turn it down without turning it off?
Most other power plants use a generator which will experience increased loading torque with increased demand, so generating hotter steam or opening sluice gates will increase the power driving the generator to meet demand; this can even be regulated.
How do you do that with direct-conversion?
Throttling and idling....
-
blaisepascal
- Posts: 191
- Joined: Thu Jun 05, 2008 3:57 am
- Location: Ithaca, NY
- Contact:
Re: Throttling and idling....
Pulse width modulation. If you send the pulses to the converter at a 30 KHz rate the converter itself can respond to harmonic distortion (out to the 39th harmonic or so of the power frequency) on the line. By correcting it at the source you reduce it some at the load.blaisepascal wrote:(I'm 80% sure this should be in "design" and not "theory". Let me know if it's in the wrong place).
How does one throttle and idle a direct-conversion polywell reactor? Demand for energy will vary, so how do you turn it down without turning it off?
Most other power plants use a generator which will experience increased loading torque with increased demand, so generating hotter steam or opening sluice gates will increase the power driving the generator to meet demand; this can even be regulated.
How do you do that with direct-conversion?
Engineering is the art of making what you want from what you can get at a profit.
-
blaisepascal
- Posts: 191
- Joined: Thu Jun 05, 2008 3:57 am
- Location: Ithaca, NY
- Contact:
That's not quite what I meant...
If you have a 100MW reactor on the grid, the grid will be demanding (say) 90MW during the day and 40MW during the night. How do you manage that variable load? Do you do it by controlling the flow through the ion guns? Adjusting the voltages of the various E-Fields or the strength of the various B-Fields?
If you have a 100MW reactor on the grid, the grid will be demanding (say) 90MW during the day and 40MW during the night. How do you manage that variable load? Do you do it by controlling the flow through the ion guns? Adjusting the voltages of the various E-Fields or the strength of the various B-Fields?
I would do it with fuel load control. You can vary the amount of fuel you feed in in lots of different ways, and that in turn will vary the fusion density in the core. That will change the system dynamics and power extraction.
It is an "engineering detail". When we understand the physics, the engineering will be "easy".
Or to quote Bill Cosby - "RIGHT! What's a cubit?"
It is an "engineering detail". When we understand the physics, the engineering will be "easy".
Or to quote Bill Cosby - "RIGHT! What's a cubit?"
Without the science engineering is possible. It is just much harder. :-)
I have always thought the fuel flow was critical. And a good way of controlling the reaction.
However, given the flow through required to maintain internal pressures it may be just as easy to change power out by changing the grid voltage. i.e. since you are recycling the flow any way a change of plus or minus 20% to control power out may not make much difference in economics and control by grid would have a much faster response time.
I have always thought the fuel flow was critical. And a good way of controlling the reaction.
However, given the flow through required to maintain internal pressures it may be just as easy to change power out by changing the grid voltage. i.e. since you are recycling the flow any way a change of plus or minus 20% to control power out may not make much difference in economics and control by grid would have a much faster response time.
Engineering is the art of making what you want from what you can get at a profit.
Yes, I think several methods will be used depending on how much change is needed. Like a BWR inside a sub, you just add a little more water and you get a little more power, cut back and it the power output goes down. But if you need to shut down, you use the control rods.
I can see several methods being used to to maintain a BFR at several levels with smooth control at each one. I suspect we'll learn how to do a whole bunch of different methods once we get to play.
I can see several methods being used to to maintain a BFR at several levels with smooth control at each one. I suspect we'll learn how to do a whole bunch of different methods once we get to play.
for the purposes of meeting demand-cycle at least, i wouldnt have thought that reactor throttling would be the most efficient thing to do - would it?
surely it would be best to keep the fusor running constant state within absolutely optimal parameters and divert any excess power to say a hydro reservoir pump for reclamation later during peak demand - else diverted elsewhere on the (wider/international) electricity grid.
that is pretty much the preferred approach with conventional generation anyway - so far as i am aware.
however, throttling for other purposes - such as 'tuning' the device itself - would still seem to be a very worthwhile topic in its own right.
surely it would be best to keep the fusor running constant state within absolutely optimal parameters and divert any excess power to say a hydro reservoir pump for reclamation later during peak demand - else diverted elsewhere on the (wider/international) electricity grid.
that is pretty much the preferred approach with conventional generation anyway - so far as i am aware.
however, throttling for other purposes - such as 'tuning' the device itself - would still seem to be a very worthwhile topic in its own right.
The most useful (valuable) plants on the grid are those that can meet a varying load in short order. The BFR can be a base load device. It can also be an incremental load device.rcain wrote:for the purposes of meeting demand-cycle at least, i wouldnt have thought that reactor throttling would be the most efficient thing to do - would it?
surely it would be best to keep the fusor running constant state within absolutely optimal parameters and divert any excess power to say a hydro reservoir pump for reclamation later during peak demand - else diverted elsewhere on the (wider/international) electricity grid.
that is pretty much the preferred approach with conventional generation anyway - so far as i am aware.
however, throttling for other purposes - such as 'tuning' the device itself - would still seem to be a very worthwhile topic in its own right.
Why build pumped storage (sites are hard to find) if BFRs are cheap enough?
Engineering is the art of making what you want from what you can get at a profit.
On the other hand, the most 'valuable' may be the ones which yield greatest profit to it's operators - ie. the plant operates at greatest (net) efficiency - I just wonder if we aren't asking too much of our (future) BFR that it also be very 'turn-upable-and-downable' (as our British Gas Ad's used to say over here).MSimon wrote:The most useful (valuable) plants on the grid are those that can meet a varying load in short order. The BFR can be a base load device. It can also be an incremental load device.rcain wrote:for the purposes of meeting demand-cycle at least, i wouldnt have thought that reactor throttling would be the most efficient thing to do - would it?
surely it would be best to keep the fusor running constant state within absolutely optimal parameters and divert any excess power to say a hydro reservoir pump for reclamation later during peak demand - else diverted elsewhere on the (wider/international) electricity grid.
that is pretty much the preferred approach with conventional generation anyway - so far as i am aware.
however, throttling for other purposes - such as 'tuning' the device itself - would still seem to be a very worthwhile topic in its own right.
Why build pumped storage (sites are hard to find) if BFRs are cheap enough?
I do take your point though:
From what I understand from these pages, first generation BFR's at least are constrained to operate at relatively low power output (order of 100MW to 1GW or thereabouts, is it?), due mainly to thermal volume scaling factors. And, as you say, they are (likely) to be relatively cheap.
So, I can imagine a regional cluster of such machines operating to as to have the effect of net variable output over a more or less continuous range, when combined on the grid, though a combination of:
a) turning individual BFR plant completely on or completely off (more or less)
b) available '(fine-) throttling' envelopes across a number of BFR machines simultaneously.
c) variability available from other (extant) generating technology (eg BWR's, hydro, geo, whatever) contributing to the grid.
And as you are suggesting I think, a large network of such machines would also supply the base load nicely, even if they do have to operate fairly flatly.
I'm obviously still assuming (possibly quite wrongly), that BFR's are likely to exhibit a fairly narrow operating efficiency curve. Do any theoretical/experimental figures exist for it yet?
I was thinking particularly of its ability to sustain the B=1 WB inflation region as being fairly critical in all this - though by the argument above, even a narrow envelope becomes useful.
(As to your point about Hydro alternative, well I suppose it depends on where you live
)My theory on this "throttling" has been a little different. To borrow an example from Sci-Fi, I sort of compared the BFR to the naquadria reactor in Stargate SG-1. The reactor produced a massive energy flow that was unstable and had to power a constant draw (basicly reverse of a BFR but the same solution works). To solve the problem they built a "buffer" to hold the energy and distribute it as needed.
Back in the real world, we could build a buffer consisting of either large capacitors or large batteries (or large superconducting inductors when technology permits), whichever fits the discharge model best. Have the buffer large enough that it won't over-charge at minimum consumption. As I understand it, a B-P polywell would produce high DC voltages at relatively low currents so some sort of large-scale switch-mode power supply or power converter is going to be necessary anyway.
Back in the real world, we could build a buffer consisting of either large capacitors or large batteries (or large superconducting inductors when technology permits), whichever fits the discharge model best. Have the buffer large enough that it won't over-charge at minimum consumption. As I understand it, a B-P polywell would produce high DC voltages at relatively low currents so some sort of large-scale switch-mode power supply or power converter is going to be necessary anyway.
Re: Throttling and idling....
blaisepascal wrote:How does one throttle and idle a direct-conversion polywell reactor? Demand for energy will vary, so how do you turn it down without turning it off?
Throttling the fusion reaction itself would be very difficult. The nice little power bump you feel in a car engine doesnt exists on jet engines. They take time to spool up and down. Fusion reactors would take this concept of "lag" even further.
At the heart of a functional polywell reactor exists a fine balance of many parameters. For example it seems that we will have to keep the core hydrogen rich to minimise Bremstrallung radiation. It would appear tricky to modulate all parameters over large values.
So instead it seems polywells will be tuned to run mostly at an optimum performance range inherent from design and resemble more like a car alternator. In a car alternator, the rotor speed is directly tied to the engine speed via a pulley. The load is adjusted by varying the current fed to the stator (electromagnet) windings. Essentially the power out is controlled by how much power you put in.
p+B11 Direct Conversion Polywells will be very similar. Around the outer permimeter will be a 2Mev Grid that will convert the work done from deaccellerating the 2Mev Alpha particles to electric power. You need to put power into this grid to get power out. Modulating this grid will let one adjust power output to grid load. Simon was correct as usual ^ above. This is just my attempt to dumb it down a few notches.
Purity is Power
That is when you call in the control engineers.It would appear tricky to modulate all parameters over large values.
RoughOrderOfMagnitude for a full scale BFR says a time constant of around 10 to 100 mS.
Piece of cake. I can buy an off the shelf processor that runs at about 100 MHz for under 15 bucks. That means roughly 1 to 10 million instructions per update cycle. Even if it is 1/10th (due to typical inefficiencies) that it is still 100,000 instructions per update cycle.
Piece of cake.
Just give me the controlling algorithm.
Engineering is the art of making what you want from what you can get at a profit.
Agreed. This is what we programmers live for.MSimon wrote:That is when you call in the control engineers.It would appear tricky to modulate all parameters over large values.
RoughOrderOfMagnitude for a full scale BFR says a time constant of around 10 to 100 mS.
Piece of cake. I can buy an off the shelf processor that runs at about 100 MHz for under 15 bucks. That means roughly 1 to 10 million instructions per update cycle. Even if it is 1/10th (due to typical inefficiencies) that it is still 100,000 instructions per update cycle.
Piece of cake.
Just give me the controlling algorithm.
Now, if I just had more practical experience in FORTH. Read the book, played around a bit, can see how its very powerful and well suited to this kind of thing, but I still think in 4GL because that's what I usually program in. OTOH, I'm only 33, so my brain is only 33% full.