FacebookTwitterDiggStumbleuponGoogle BookmarksRedditTechnoratiLinkedin

(ClamShell) A low input power, high temperature reactor

Scritto da Robert Greenyer on .

Details on this experiment family will be recorded in this blog. For now please keep a track of the live document below and make comments, questions and suggestions as normal.

 

Add comment


Security code
Refresh

Comments   

 
+1 #18 GlowFish 2016-10-24 07:14
Have there been any further developments with the experiment?
Quote
 
 
0 #17 Stephen Cooke 2016-10-10 13:47
Could this be useful Bob?

It's probably a way out there idea but... I wonder if corrugated borophene or graphene could play a useful role. I recently read the following article about the use of borophene for flexible electronics:

spacedaily.com/.../...

But i suppose another quality of this material apart form its flexibility would be its large surface area. If a graphene layer is useful I wonder if corrugated graphene can also bring benefits.

I wondered else where if the electrical properties and other useful characteristics of Boron Nitride Nano Tubes or or Boron Carbide Nano tubes could bring some benefits:

disqus.com/.../...

if so perhaps the borophine equivalent could also bring some benefits too. Does Boro Carbide also produce graphene like mono sheets I wonder?

The following is probably way way out there but i saw it reading about the above and it makes me wonder, however it would require someone with better theoretical knowledge than me to know if it is relevant or not:

journals.aps.org/.../...

Its interesting that there appears to be a spatial dependance of the fermi velocity in this corrugated material could this be useful in some way? or is it complicating things too much?

This is interesting too:

arxiv.org/pdf/0904.1520.pdf

nature.com/.../nmat2710.html

Could Moire fringes and dislocations play a role in enhancing Hydrogen absorption?

en.wikipedia.org/.../...

Just an idea

The following book looks interesting too I'm not sure how easy it is to get:

books.google.nl/.../
Quote
 
 
+2 #16 GlowFish 2016-09-30 11:48
Will there be some sort of link to something like Google Drive with the data from the clamshell experiment similar to the GlowStick series?
Quote
 
 
0 #15 Fredrick Hibbs 2016-09-28 15:03
Quoting Bob Higgins:

I added to the live doc above a description of the thermal simulation and the updated results."

@Bob Higgins Please may I stress the need for calorimetry (to convince all the skeptics) rather than relying on a complex mathematical simulation. All it would require is a sealed perspex box around the whole apparatus with a measured flow of ambient air going into the box and a measured delta temperature of the air coming out versus the air going in. Small adjustments could easily be made for the ambient air inlet pressure & humidity.

A highly reflective surface on the inside surface of the blocks would also help increase the COP by reducing the heat input required for any given operating temperature.
Quote
 
 
0 #14 Rene Vega 2016-09-22 22:08
Quoting Bob Higgins:
@Rene Vega
* * * it will not be possible to do with the Ni+LAH fuel which is electrically conductive. The fuel will short out the electric field. * * * * Doing this would require a dielectric seal at the plumbing end of the tube (presently a metallic seal).


It's not 0 ohms. Whatever resistance it is will be almost 7 times greater at 1200C. If you are amenable, would you measure the resistance sans metallic seal with the fuel loaded before proceeding. I'd do it myself but lack the facilities to handle LAH safely at this time. Thanks.
Quote
 
 
+1 #13 Bob Higgins 2016-09-22 22:01
@Rigel,
I presume the MFMP online video you were watching was of the "!Bang" experiment - obviously so named after the fact. That occurred during a time when we believed the high pressure H2 liberated by the LAH (hundreds of bar) was necessary for the reaction. Since then we have discovered that the experiments that produced excess heat were the ones that leaked, and the actual H2 pressure needed was below 1 bar absolute (a slight vacuum). At this lower pressure there is far less H2 present and less danger. Still it makes sense to be safe. I will take your advice and re-double my efforts to protect against bad outcomes.
Quote
 
 
0 #12 Rigel 2016-09-22 17:59
[quote name="Bob Higgins"]@Rigel
I am not sure what you mean about the null load. The entire system is being thermally modeled in SPICE using a reactor with a simulated dummy fuel. The model is better now and is extended to over 1250C core temperature.

The top brick was not designed to relieve explosions, though that is what would happen. The heater coil is heavy iron wire and that may channel any explosion axially. At my operating point, I am protected from that by 1/4" aluminum plate. /quote]

Bob, I was watching MFMP online time ago and saw the pop. You are mixing chemicals at high temperature. I would test your blast shield plate. If you do get a run-away. That said, I have had 2 small accidents. My metal shield held up, but one small piece got by where I did not have protection. The missus made me move my lab out into the cold winter garage. Nothing is more sobering than an explosion.
Quote
 
 
0 #11 Bob Higgins 2016-09-22 17:13
@Rene Vega
While it certainly would be possible to establish an RF electric field in the dummy alumina fuel that way, it will not be possible to do with the Ni+LAH fuel which is electrically conductive. The fuel will short out the electric field, but you could drive an RF current through the fuel that could provide an RF magnetic field. It is a complete unknown what effect such a current would have. Doing this would require a dielectric seal at the plumbing end of the tube (presently a metallic seal).
Quote
 
 
0 #10 Rene Vega 2016-09-22 05:58
Given the arrangement of a vacuum pipette inside the stainless steel beer tube, it might be possible to inject RF into the fuel like this:.

drive.google.com/.../
Quote
 
 
0 #9 GlowFish 2016-09-21 19:10
@Bob Higgins

Thanks for the update. I did wonder what programs you were using. Another small request. Could you add a text label next to the resistors and caps to specify what aspect they are modelling. For example, I assume that V1 is your ambient temperature, C1 your heater thermal capacity, R7 is your heater to ambient transfer,R6 your heater to fuel transfer, C6 your fuel heat capacity etc (Or something like that).
Quote
 
 
+2 #8 Bob Higgins 2016-09-21 18:17
@Glowfish
I added to the live doc above a description of the thermal simulation and the updated results. Have a look at the end. "My kingdom for an optimizer!"
Quote
 
 
0 #7 Rene Vega 2016-09-20 20:53
Quoting bob:
Bob:
If the bricks are easy to machine/drill it might be interesting to see if you could adapt this setup as a conduction calorimeter. * * *.


I asked the same question yesterday in another forum. His response:
"Bob Higgins > Rene • a day ago
2. The long term intent is to make the insulator 2" diameter on the outside which will match the diameter of the aluminum supports. Then this whole racked cassette will be inserted into a flow calorimeter. Almost anything I do to introduce variable heat outflow will make it hard to maintain calorimetry without the flow calorimeter. The flow calorimeter design I have already has variable coupling built in that will not invalidate the calorimetry.
Quote
 
 
0 #6 GlowFish 2016-09-20 17:58
A suggestion. Can a description or block diagram of the model used to simulate the process be added to the writeup (Sources of heat, heat loss vectors etc.)? It may give the community an opportunity to comment or provide suggestions regarding the model.
Quote
 
 
0 #5 bob 2016-09-20 17:03
Bob:
If the bricks are easy to machine/drill it might be interesting to see if you could adapt this setup as a conduction calorimeter. While the symmetry isn't perfect (ie. reactor is cylindrical while bricks are rectangular) it still might work. The assumption is that almost all of the heat from the apparatus is dissipated by thermal conduction through the brick material. Heat flux is known to be linear with delta T in conduction. Two axial (ie. drill longitudinal holes) located thermocouples at two different radial distances from the central chamber (ie. separated by a thin layer of brick material) might yield a linear relationship between delta T and input power during calibrations. Worth a try.

Might even work better if the bricks were machined into cylindrical geometry to be symmetrical with reactor tube.
Quote
 
 
+1 #4 Bob Higgins 2016-09-20 14:57
@Rigel
I am not sure what you mean about the null load. The entire system is being thermally modeled in SPICE using a reactor with a simulated dummy fuel. The model is better now and is extended to over 1250C core temperature.

The top brick was not designed to relieve explosions, though that is what would happen. The heater coil is heavy iron wire and that may channel any explosion axially. At my operating point, I am protected from that by 1/4" aluminum plate.

There is no plan to ship these at the moment. Let's see what happens when it is tested. It would not be hard for someone else to reproduce - the K-26 insulating brick is as easy to machine as a block of plaster. These are readily available on eBay for about $6 each.
Quote
 
 
0 #3 Rene Vega 2016-09-19 23:12
Bob,
In the run-up null test graph, is the timescale in 1000 sec units? Is the thermal lag at top of brick measurements then around 200 seconds?
Quote
 
 
0 #2 Rigel 2016-09-19 23:03
Not sure any of the design constraints. But I understand the need to make it with COTS/cheap parts though :-)
Is the null load accounted for? Does a top brick gravity design account for a way to channel possible explosions? *The lighter brick on top would be in line of sight. Canter it so it can blow out the back brick- not the top. Due to the way MFMP works I can imagine you shipping these things, do you expect to do that at some point? Then it would need to be modular and have sub-assemblies.
Quote
 
 
0 #1 Rene Vega 2016-09-19 18:03
I like the clamshell design, and yes, raising the efficiency of the bulk heating is important to reduce the losses that mask excessive heat.

1. Is the heater winding a double helix? Doing that would significantly reduce magnetic field effects from the heater process, and, perhaps, permit answering the question whether magnetic fields quench the reaction;

2. For temperature control, has BobH considered adding a couple of grooves in the bricks parallel to the reactor in which he can place heat exchanger pipes? This might be a better way than lifting the bricks and blow air to draw heat out when the reaction starts to go nonlinear. Furthermore it could be the basis of some calorimetry based measurements;

3. Does the design have the means to vary the reflection of photons back into the reactor instead of absorbing them into the bricks, terrahertz comes to mind.

-Rene
Quote
 

Here is your generous contributions so far towards our $500,000 target, thanks everyone! : $45,020   Please Donate
See the current state of our booked costs here