Heat exchangers

[TallDave]

Well, hopefully cooling is our major problem. That would be a nice situation to be in.

[hanelyp]

I did some calculations on this blog showing why, if the heat load was limited to 1 MW/m^2 then power could only go up as the square of the radius above 100MWth. Very bad.

Especially since costs tend to go up as r^3.

If a polywell was built at a scale where power output was limited to r^2 by heat, I'm thinking it would be built with weaker magnets, B^4R^3 = kR^2. So costs wouldn't necessarily scale as r^3 in this case.

[TallDave]

Well, r^3 is just based on the idea that things having three dimensions tend to have costs that increase as a function of their volume, all else being equal. I'm assuming magnetic field strength stays the same.

[tombo]

Which m^2 are we talking about?
The inside of the whole chamber?
Coil surface area?
The coil alpha intercept area?

Different heat loads have different appropriate areas.

So scaling could be something of a wash for equal energy flux in the core and equal heat dissipation per m^2 in the heat transfer mechanism.
(assuming superconductors otherwise resistance loss too)
Bigger certainly is better for coolant flow rates.

msimon,
Where did you find the resistivity of Cu at LN2 temperature?

[MSimon]

Cu resistivity drops .39% per deg K. Which is quite a bit actually. A 100 deg K rise raises the resistance by around 50%.

From the CRC Handbook of P&C:

micro-ohm cm Deg C

1.692 20 C 293.15 K
.163 -206.6 C 66.55 K
.014 -258.6 C 14.55K

Here is a nice graph:

http://www.physics.ubc.ca/~outreach/phy ... opper.html

It also has some interesting thermodynamic demonstrations re: heat transfer in a boiling liquid.

[charliem]

Tombo there is a graph for Cu resistivity vs temperature from 0 to 900K here:

http://hypertextbook.com/facts/2004/BridgetRitter.shtml

Scroll down and click on the graph to see the table of values.

[MSimon]

Vernier Format 2
BridgetRitter.ga3 5:38 PM 6/13/05 .
Data Set
Temperature (K) Resistivity x 10e-8 ým
X res.


1 0.002
10 0.00202
20 0.0028
40 0.0239
60 0.0971
80 0.215
100 0.348
150 0.699
200 1.046
273 1.543
293 1.678
298 1.712
300 1.725
400 2.402
500 3.09
600 3.792
700 4.514
800 5.262
900 6.041

http://hypertextbook.com/facts/2004/BridgetRitter.txt

*

[eros]

V
Temperature (K) Resistivity x 10e-8 ým
X res.


20 0.0028
40 0.0239
60 0.0971
80 0.215
293 1.678
298 1.712

Hmm, if copper 0.017ohm/mm² drops 0.0017 (~70K) then power needs drops 10x. 20K drops power needs ~600x.

So if 1m research machine with 30cm thick coils cooled to 20K (liquid hydrogen) it coils consume 0.68kw at 0.13T and give 1W fusion power.

But if so hard cooling it make sense to use superconductors. Beter B and breakeven.

What field and A/mm² is posible with YCBO?

[MSimon]

Eros,

MgB is a better bet. IMO. Lower costs - not so brittle. And if it is made with B11 its resistance to neutron flux is pretty good.

[eros]

Eros,

MgB is a better bet. IMO. Lower costs - not so brittle. And if it is made with B11 its resistance to neutron flux is pretty good.

Yes, MgB2 is best bet, but supply is limited still. And it needs He or liquid H2 cooling LN2 is not enough.

Maufacturers make YBCO rings for magnet bearings. Some teslas are maybe posible..
I mean somekind of WB-10 steady state test machine for low budget..

Power reactors needs MgB2, but I think we are not yet ready for such building. More experiments needed.

One tesla at 250mm size is maybe 50w fusion power. 5T jumps ~30kw.
It is maybe posible to do pB11 experiments, DD neutrons quite soon destroy YBCO.
But if show fusion in that size I am quite sure that funds are no more broblem.

I am not calculated other power loss, but 5T fusion power is maybe enough for short time breakeven. 5T mean at 10% rule ~5000A/mm³ which is inside YBCO limits. (but it depens, some bulk are not so good)

Ready build bearing rings can't be too expensive and they should be magnetically good. Size is ~WB6 but B much higher and last long, only cooling loss.

I see some good in idea, but what other people thinks?

[MSimon]

To get YBCO up to the Teslas we need (would like) YBCO would need to be cooled to the same temp (roughly) as MgB.

And mfg capacity is to the point where km lengths are the norm and they can be spliced.

==

Cu and LN2 would be relatively off the shelf and provide a field of about .45T (higher pulsed). I'd use a Bitter type design (flat plates with cooling holes).

Neutrons would not be a problem. Total operating time in the D-D fusion regime would be in the 10s of hour range. With two coil sets one could be operating while the spare was being annealed to remove neutron flux damage.

[eros]

To get YBCO up to the Teslas we need (would like) YBCO would need to be cooled to the same temp (roughly) as MgB.

http://aries.ucsd.edu/ARIES/MEETINGS/00 ... sld006.htm

I see still quite lot of amperes at 77K. Anyway it needs lees cooling than MgB2.

And mfg capacity is to the point where km lengths are the norm and they can be spliced.

We don't need conductor we need only magnets. Bulk rings are available.

Cu and LN2 would be relatively off the shelf and provide a field of about .45T (higher pulsed). I'd use a Bitter type design (flat plates with cooling holes).

Bitter is maybe good for super magnets, but forces are guite small in 0.45T so normal desing is enough. It is easyer to do round shape vs. bitter.
Btw can you give size what think about and calculate it power needs?

Neutrons would not be a problem. Total operating time in the D-D fusion regime would be in the 10s of hour range. With two coil sets one could be operating while the spare was being annealed to remove neutron flux damage.

Hmm, 1hr continous operation win tokamak 6-0. I am not sure anout annealing, but you think whole core heating or something? How about vacuum?
I think if got some mins to hour operation and then break vacuum and replace core (if posible due chamber radiation) then pump vacuum next week and analyze results.

Btw I just found that Nd permanent magnets are available. 1.4T fields and european modernest factory is ~200km where I live. Maybe they offer some ring samples..
1.4T don't make power reactor but give good test enviroment for other things like alpha collectors, electron keep ratios, other geometry than cube etc..

Well, on monday I must call and ask what kind of rings they have on shelf.
Next need to weld vacuum chamber (hmm, I need to fix my plasma weld things). Friend have some pumps and 75kw HV supply,
Maybe we get some results next year, only broblem is zero budget..

[MSimon]

Permanent Magnets won't work. The field lines are in the wrong place. Electromagnets are the only way.

This has been discussed extensively on this board. Look around.

Any magnet used would have to be custom designed due to the necessity of dealing with high voltage. LN2 Cu would be the quickest. The Bitter design is my preference because of mechanical stability and cooling. Mechanical stability is important if you are going to pulse them.

I did some Cu LN2 calculations at:

http://iecfusiontech.blogspot.com/

Have a look around.

I also have a spread sheet I could dig up if you were interested.

[KitemanSA]

Would someone explain to me why a Bitter magnet, which is a stack of flat rings, is better than a coil of flat plate? The number of turns and x-section can be identical, and the coil seems so much easier to build.

The bolt holes and coolant holes can be effectively identical. Why so Bitter?

[MSimon]

Would someone explain to me why a Bitter magnet, which is a stack of flat rings, is better than a coil of flat plate? The number of turns and x-section can be identical, and the coil seems so much easier to build.

The bolt holes and coolant holes can be effectively identical. Why so Bitter?

The Bitter can be made from stampings. Winding a wide flat spiral is a tougher manufacturing problem.