What are all of the exact mechanics of Temp Shift plates?

[KittenIsAGeek]

I have made tempshift plates out of dirt as a heat-buffer.  800kg of dirt with a SHC of 1.48 adds a lot of thermal inertia.

[Mathgeekburch]

3 hours ago, nakomaru said:

Tempshifts from insulation conduct heat just fine when buried in solids, thanks to a large multiplier and probably non zero TC. This is one way to produce tungsten.

So this means it is worth it to convert 120 kg of isoresin to 800 kg of tungsten?
 

3 hours ago, nakomaru said:

The 200x/1x factors are parsimonious interpretations of how other buried solids and buildings work. Mathman has talked about burial in solid being 200x. I'd have do an experiment to prove them with confidence.

So does that mean all solid-to-building interactions get a 200x increase in conductivity (with exception of the solid tiles a building is on/ that are below the building) or is it something that occurs only when something is buried (entombed)?  Like do pipes get the effect when they are inside tiles, or does it only triggered when they are inside natural tiles (entombed). Side question: I haven't actually entombed a pipe before, I presume liquids stop flowing through them when entombed?

Does a tempshift plat get the 'buried'  200x multiplier just from the drywall part being adjacent to a wall, or does the actual dry wall have to be entombed?

Also is the 200x only for heating the building, or does it also effect cooling the building?

3 hours ago, nakomaru said:

(I understand it doesn't tell you about the scaling,

I think the scaling is the most important part to confirm.

The test is to confirm it would be a little bit tricky, as having higher mass and thus higher heat mass means it loses less temperature for heat transferred, which can mean its transfer of heat doesn't slow as much, making it appear to have a higher 'effective conductivity' (there is probably a better term for it, but I think you understand what I mean) even when it doesn't.

You need to make sure heat loss of the medium during the experiment doesn't result in much change in the medium temperature.

I think the base way to do it would be to make three 5x5 hollow square of insulated tiles, fill the first one with an ungodly amount of super hot chlorine gas (its low conductivity should make it easier to track other factors such as Chot), fill the second one with double the ungodly amount of the same temperature. Fill the 3rd with 4 times the ungodly amount of the first.  

Place a tempshift plate in the center of each.  As long as the heat energy required to melt the tempshift plate isn't a large fraction of the "usable" heat energy of the gas, it should melt about twice as fast in the second chamber, and 4 times faster in the third. If Chot didn't depend on mass, we would expect to see very little increase between the chambers.

If the tempshift plate melts too fast for you to discern the time required to melt it; use a temp shift plate that takes more heat energy to melt and those that have a lower conductivity.


 

4 hours ago, nakomaru said:

The 200x/1x factors are parsimonious interpretations of how other buried solids and buildings work.

Secondary question.  What are mesh tiles and airflow tiles considered in heat equations?  I know liquids can exist inside mesh tiles, so does that mean the mesh tile is a liquid when liquid is occupying it for the purposes of heat equations, or is it still considered a solid, is it consider two mediums in the same spot?

I don't know if gases occupy the position of airflow tiles, I really only used them for ventilation and never tried seeing if I could actually store gases in them.  But if gases do occupy them, same question as I had for mesh tiles.

Also mention any additional interesting heat transfer interaction with mesh and airflow tiles.

[mathmanican]

1 hour ago, Mathgeekburch said:

 snip....

Tons of great questions. Each one sounds like a great experiment to run. I haven't done these tests, but it would not surprise me if @Angpaur or @wachunga have already done all these tests.  I head to this thread, whenever I have heat transfer questions. It's a simplfied version of the one by @Yothiel that includes lots more details for fringe cases. 

 

[Mathgeekburch]

1 hour ago, mathmanican said:

I head to this thread, whenever I have heat transfer questions. It's a simplfied version of the one by @Yothiel that includes lots more details for fringe cases. 

I have actually already looked at that forum before to learn about heat equations.  Actually is how I found Yothiel's thread.  Unfortunately I have doubt on whether all the equations are still true now since the thread is from 2018.  I feel like ONI uses geometric mean instead of the arithmetic mean although I haven't tested it yet.

Also with the Chot, the building equation in that thread would only be true if you are cooling a building, and not for heating buildings.  


Also some people are telling buried buildings might have an additional 200x heat transfer multiplier.  Which does change up a lot, particularly for shift plates and pipes.  

[nakomaru]

19 hours ago, Mathgeekburch said:

So this means it is worth it to convert 120 kg of isoresin to 800 kg of tungsten?

It is to most people.

19 hours ago, Mathgeekburch said:

So does that mean all solid-to-building interactions get a 200x increase in conductivity (with exception of the solid tiles a building is on/ that are below the building) or is it something that occurs only when something is buried (entombed)?  Like do pipes get the effect when they are inside tiles, or does it only triggered when they are inside natural tiles (entombed). Side question: I haven't actually entombed a pipe before, I presume liquids stop flowing through them when entombed?

Does a tempshift plat get the 'buried'  200x multiplier just from the drywall part being adjacent to a wall, or does the actual dry wall have to be entombed?

By buried I mean the medium for the tile that the object is in is a solid. For example, debris trapped in a door, buried in sand, or being carried through conveyer rails in a solid. The conveyer rails themselves in a solid. A pipe that is pathing through a solid (only for the pipe↔solid, not pipe↔contents). A bridge that touches solid on any of its 3 tiles. A tempshift plate that touches solid on any of its 8 outer tiles (or if you bury the 9th in e.g. regolith). An oil well that touches solids on any of its 8 tiles that it doesn't overlap but occupies thermally. (It is a 4x2 building that touches a 4x4 space.) I believe all of these cases share the same solid-buried multiplier, but I will have to prove it.

19 hours ago, Mathgeekburch said:

What are mesh tiles and airflow tiles considered in heat equations?

They are debris and there is always a different medium for the actual tile calculation. The heat transfer for the tile is that of the medium of the tile. If it's oxygen or water inside, it's the same as an oxygen or water tile with debris inside it (I don't know if it counts at 1/5th mass like buildings or not.) If it's a vacuum, the only valid interaction they have is with the tile below them at 1/4th factor, so this configuration can lead to perfect insulation.

On 8/25/2019 at 11:59 PM, Mathgeekburch said:

Is there any special interaction with moving/falling liquids?

If it is falling normally where it teleports into another dimension until it reaches the bottom, it is off limits thermally until it reaches the bottom and becomes a tile again. If it is beading as in the bead-pump, or forming a waterfall, it is an ordinary liquid all the way down.

[TLW]

12 hours ago, nakomaru said:

If it is falling normally where it teleports into another dimension until it reaches the bottom

I propose calling this "dripping".

[nakomaru]

1 hour ago, TLW said:

I propose calling this "dripping".

That's fair. My wording was awful. Sand and regolith also drip.

[Mathgeekburch]

If the equation for tile to tile heat transfer is 
Tile to Tile: q = klow * dT * 200
And the equation for heat transfer of a building to a solid tile occupying it (where HCB is the specific heat times the mass of material the building in  divided by 5 and divided by the area of the building which is 9 for temp plates)
Building to solid Tile: q = 200*k1 * k2 * dT * HCB/2
And the equation for heat transfer for a solid tile to a building it is occupying. (where HCT is the specific heat of the tile times the mass of material in the tile)
solid Tile to Building: q = 200*k1 * k2 * dT * HCT/2

So assuming the there isn't weirdness with clamping.  I can compare alternating diamond tempshift plate with diamond windows; with just strait diamond windows.

If I denote the heat transfer between diamond window to other diamond windows as qtt then I can write the heat transfer of the building/tile equations as

Building to Tile: q = qtt * 80 * 160 * 0.516 / 18
Tile to Building: q = qtt * 80 * 100 * 0.516 / 2

So alternating plates with windows should be much much faster heat transfer than just windows (in a vacuum)

Although should my HC be in units of kDTU/ degree C or DTU/ degree C?  Seems like it would be really fudging insane if DTU/degree C.

Hell it seems a little too insane with units in kDTU/degree.  Like 360x heat transfer if there is no clamping weirdness just doesn't seem true.  


It just seems like these equations have to be extremely outdated or incomplete.  I need to test the difference between heat transfer of alternating tempshift plates in a vacuum verse plan window tiles asap.  Only problem is I will only have access to campus computers for a week     

[wachunga]

Run tests, gather data, fit an equation to that data. Run more tests, try to disprove that equation. Share your results.

Alt- in debug advances the game one tick. The sample tool in debug gives you a cell temperature to 4 decimal places. Gathering data is trivially easy.

Cell to Cell (Tile to Tile was a potentially confusing choice of words on my part) has recently changed to geometric mean instead of lowest conductivity.

I have always advocated, and asked, for people to run their own tests. The community has refused to do so. They would rather endlessly speculate and regurgitate misinformation. If a tiny fraction of that time was spent doing their own tests, these questions would have been answered a hundred times over. 

[Mathgeekburch]

1 hour ago, wachunga said:

Run tests, gather data, fit an equation to that data. Run more tests, try to disprove that equation.

 

8 hours ago, Mathgeekburch said:

Only problem is I will only have access to campus computers for a week     

 

Edit: for clarification. I can't play ONI on campus computers, and they are all I have access to for right now.

Edit 2: this in turns means I am unable to do the test despite wanting to do them.

[Mathgeekburch]

So I got someone else to do some experiments for me, Xtraordinaire on the subreddit.  I am so thankful to him for it.  Based on the experiment outcomes, it seems that the 200x solid heat transfer rate isn't true for solid medium, or at least solid medium adjacent to the center of tempshift plates.  But that the Chot factor is true for tempshift plates. The experiments themselves were fairly messy, since I was the one coming up with them, and quite frankly I am terribly at this; I'll try coming up with better experiments once I transport my desktop to where I am at. (which might be a while).  

Link to the reddit post that contains Xtraodinaires data in the comments.https://www.reddit.com/r/Oxygennotincluded/comments/cwod6x/i_am_stuck_with_campus_computers_for_a_while_so_i/?utm_source=share&utm_medium=web2x

 

[nakomaru]

@Mathgeekburch Thanks for your dedication, and for teaching me several things. Not even being able to play but still pursing the science is a wonderful thing.

I wonder why tempshift is slightly winning in the tile to tile test. Maybe it's just C_hot. It should probably lose vs 200kg tiles if so.

If you redo the test with insulated tiles and tempshifts, you'll get much different results though. There must be some kind of multiplier for those, or a different method of clamping, or different method of averaging conductivities. Or it has a more sane multiplier for everything.

[Mathgeekburch]

8 hours ago, nakomaru said:

I wonder why tempshift is slightly winning in the tile to tile test. Maybe it's just C_hot. It should probably lose vs 200kg tiles if so.

It is probably a mix of both C_hot and k1*k2.  To my knowledge tile to tile doesn't have k1*k2; and either uses the lower conductivity or an average (I think geometric mean).  Although I haven't tested that so take it with a bag of salt.

I don't think if it was just C_hot it would be enough, as my calculation where I make the prediction of 360x conductivity was based on the 200x conductivity of solid medium to building transfer.

If I remove the 200x I get

Building to Tile: q = qtt  * 80 * 160 * 0.516 / 3600
Tile to Building: q = qtt * 80 * 100 * 0.516 / 400

If I presume the k1*k2 is also existent for tile transfer, that removes the *80 from the comparison equation to pure tiles.  And 160*0.516/3600 is less than 1; and 100*0.516/400 is also less than 1; so if it was just Chot, it should be slower unless I have the units off by a factor of a thousand, but then I should be getting crazy higher heat transfer.

If I assume that all tile to tile interactions, has only klow (or geometric mean, or arithmetic mean); then my heat transfer compared to diamond window should be what I have in the corrected equation above.  Except that gives me about 1.83 times the heat transfer for building to tile, and 10.32 times for tile to building.  Which is far higher than measured which was about 25% more heat transfer.  Like I could accept it if was 1.5x calculated and 1.25x measured, as I should be losing some heat to heating up the extra heatmass of the tempshift plate itself, but 1.83 times just doesn't seem right.  So I have no idea what the actual equation should be.


It would be nice to know if the above that puts heat transfer by building to tile at 5.63 times lower than the tile to building is anywhere close to accurate.  For all I know it could be the opposite.
 

Assuming tile-tile doesn't use Chot.  You should be able to figure out which is the bottle neck for heat transfer in the window-plate system, with a modified version of the window-plate test.  You put even more massive heat sinks on either end (like really insane heat mass, like 200T of hydrogen); allowing you to almost keep a constant temperature difference across the conductive bridge, despite heat being constantly transferred.

With this you can see what are the temperature differences between each tempshift plate and adjacent windows.

Once the system reaches a near stable transfer of heat through it, each object should be outputing as much heat energy as it is absorbing, which should be near equal across the entire conduction bridge.  As such you can compare what is something's heat transfer per dT, by looking at the dT between each object.

qin: heat entering object.  dTin: temp difference between object and the object prior.  Zin: heat transfer coefficient of the prior object to the object.
qout: heat leaving object.  dTout: temp difference between object and the object after.  Zout: heat transfer coefficient of object to object after.

qin=dTin*Zin=qout=dTout*Zout

if dTout/dTin=Zout/Zin   Which should give you the ratio of the heat transfer between different objects; and as such should allow you to know if the formula is correct.  Failure condition: this all assumes Chot doesn't apply to tile to tile transfer; if it does then that should result in the temp difference between the hot sink and the adjacent window being low; and the temp difference between the cold sink and the adjacent window being much higher.  Reason why this is a failure condition, is that we are constantly having more heat pushed into the conductive band by the hotsink, than heat being removed from the conductive band by the cold sink.  This test only works if the heat into the conductive band, is near equivalent to the heat coming out of the conductive band.


Secondary note:  If this test actually comes to be a success (IE the failure case doesn't occur, no other anomalies occur, heat sinks are large enough), you should be able to use it to find many different Z values (Z=heat transfer per dT).  Z values only exist for a relation between two specific objects and has a direction/order.  For example Z for Diamond window to Diamond Tempshift plate is different than Z for Diamond Tempshift plate to Diamond window; the order of the objects used to label the Z value matter.

This should be able give a great insight into the heat transfer equations if test of this type are successful, which is a big IF.