Archived

This topic is now archived and is closed to further replies.

Please be aware that the content of this thread may be outdated and no longer applicable.

Gurgel

Heat transfer in liquids

Recommended Posts

Gurgel    1655

I was wondering whether heat transfer in liquids (or gases) depends on the amount in a tile. This makes a difference, e.g. when submerging a tepidizer that is surrounded by metal-tiles to transfer the heat out.

My intuition from other observations is that it does not matter, except that with more liquid you need to heat up that additional liquid as well, but heat transfer is the same. Some experiments in sandbox were not fully conclusive, but seem to indicate the same. Can anybody confirm this? Thanks.

Share this post


Link to post
Share on other sites
MrKnister    9

As far as i understand the game just converts heat into DTUs and then transfers them from tile to tile using the rate of the lower number from the two tiles, so basically (as far as i know) it doesn't matter how much liquid you surround your tepidizer with, it will absorb the 4064 DTU/s and heat up by a certain amount determined by the used liquid, wich it will then transfer to the surrounding tiles.

Please correct me if i'm wrong

Share this post


Link to post
Share on other sites
reccurentz    4

Heat transfer is affected by temperature difference. Temperature of material is affected by mass and SHC. So yes, mass has role to play; how much depends on your use-case. But most of the time it can be ignored.

It is like thinking difference between one time expense vs operating expense in a world that doesn't have depreciating assets. Over time, the one time expense becomes negligible.

Edit:

To answer your question directly, less mass is better for transfer mediums if conductivity is the same. However, it is not like you lose that energy if you have more mass. It acts like a battery and having more buffer can be beneficial if your heat generation comes in burst.

Share this post


Link to post
Share on other sites
Coolthulhu    1045

For some objects, there was a square root of mass multiplier on heat transfer. At least when most heat transfer calculations were done on C#/unity side of things.

Share this post


Link to post
Share on other sites
KittenIsAGeek    1557

The mass of a liquid (i.e. amount on a tile) will translate to how much thermal energy the liquid can take before causing a change in temperature.  The rate that the heat moves isn't directly affected by mass, but instead by thermal conductivity.  However, the mass will limit the total amount of heat that can move.

Heat only "moves" when there is a difference in temperature.  The greater the difference in temperature, the faster the heat moves from one to the other.  But, if you have a small mass, then each unit of heat (DTU) that transfers to it is going to affect its temperature more dramatically than a large mass would.  As it gets closer to equilibrium with the hot source, the amount of DTU that can transfer becomes reduced.

Thus, in a way, a smaller amount of mass does reduce the amount of heat that can transfer, because it will more quickly reach equilibrium with the heat source, reducing the amount of DTU that can actually move.

Share this post


Link to post
Share on other sites
Gurgel    1655

I think I may have not been too clear. I do not need an general explanation how heat transfer works. The situation is (simplified)

       large hot mass A -> liquid connection 1 tile high, 1 tile long -> large cold mass B

As far as I can determine, the "thickness" (within 1 tile) of the liquid, e.g. whether I have 1kg of water or 900kg of water in the tile, has no impact on the heat transferred per time form A to B (after the liquid has heated up).  

Share this post


Link to post
Share on other sites
psusi    259

The TC for liquids is given in units of (DTU/(m*s))/°C.  There is no mass in there, so no, the mass of the liquid in the tile doesn't affect the conductivity.

Share this post


Link to post
Share on other sites
sheaker    60

I think that when I was using CO2 cooling vent line with 1kg packages then transfer were occurrin fast. When the CO2 packages were veeeery small then there was barely no heat transfer. For example 1kg package went from -28 to 33 deg C while while passing through radiator but 300mcg package went from -33 to just -31 in the same radiator. Am I wrong?

Share this post


Link to post
Share on other sites
psusi    259
6 minutes ago, sheaker said:

I think that when I was using CO2 cooling vent line with 1kg packages then transfer were occurrin fast. When the CO2 packages were veeeery small then there was barely no heat transfer. For example 1kg package went from -28 to 33 deg C while while passing through radiator but 300mcg package went from -33 to just -31 in the same radiator. Am I wrong?

That's SHC, not TC.  More mass holds more heat, but does not transfer it any faster.  Less mass holds less heat, but transfers at the same speed, so it's temperature shifts faster.

Share this post


Link to post
Share on other sites
nakomaru    1628

Transfer is worse if the mass is too small in a few places of the calculation.

  1. The obvious heat transfer is ΔQ = Δt・k・ΔT・SurfaceArea (e.g. 25・25 for liquid・liquid)
  2. If this is bigger than (Q2-Q1)/8, then (Q2-Q1)/8 is used instead. Q = SHC * m * T ← mass limited rate
  3. If either of these are less than 0.0001 J, Q = 0. ← mass/element or phase/element limited ΔT clamp

Transfer is worse if the mass is too small in a variety of ways depending on your design. Some ways, because a small mass gives a low thermal capacity for your heat:

  1. If you have an intermittent load, your heater can only work during the load. If you had a higher capacity, your heater could work for more time.
  2. A higher capacity can maintain a stable T easier, which means it could keep the ΔT between your medium higher.

Transfer is worse if the mass is too big in a variety of ways depending on your design. For example:

  1. You have a stable hot source, but the medium is not often used. Or it's used in a variety of temperatures. So it takes more work to heat it up to your target temperature.

Share this post


Link to post
Share on other sites
Gurgel    1655
1 hour ago, nakomaru said:

Transfer is worse if the mass is too small in a few places of the calculation.

  1. The obvious heat transfer is ΔQ = Δt・k・ΔT・SurfaceArea (e.g. 25・25 for liquid・liquid)
  2. If this is bigger than (Q2-Q1)/8, then (Q2-Q1)/8 is used instead. Q = SHC * m * T ← mass limited rate
  3. If either of these are less than 0.0001 J, Q = 0. ← mass/element or phase/element limited ΔT clamp

I don't care about mass of the liquid at the moment (or rather, I can factor it in easily).

But I was unaware of the surface area modifier. It explains some things I saw in experiments, as does the clamping. 

Share this post


Link to post
Share on other sites
7 hours ago, nakomaru said:
  1. The obvious heat transfer is ΔQ = Δt・k・ΔT・SurfaceArea (e.g. 25・25 for liquid・liquid

My tests show a ratio 25. Liquid lead to liquid lead ~ solid thermium to solid thermium.
Also, different elements due to the mass and heat capacity of the elements can have the same thermal energy. For example, the heat transfer from a hot diamond (~ 600 ° C) to a steel door (200 ° C) is very low due to thermal clamping.

(google translate, sorry)

My calculations in file.

ThermalClumping.xlsx

Share this post


Link to post
Share on other sites
Gurgel    1655
6 hours ago, Alexander Block said:

My tests show a ratio 25. Liquid lead to liquid lead ~ solid thermium to solid thermium.
Also, different elements due to the mass and heat capacity of the elements can have the same thermal energy. For example, the heat transfer from a hot diamond (~ 600 ° C) to a steel door (200 ° C) is very low due to thermal clamping.

(google translate, sorry)

If I read the Wiki article right (it is not too clear) the transfer factors ("Surface Area") are for cell-to-cell (not insulated, they are special):

  • liquid-liquid: 625
  • gas-solid: 25
  • all others: 1

As the nice spreadsheet by Aledander Block shows, you need to clamp as well for the final number.

Definitely a remnant from an earlier time, they would not add something this complicated today ;-)

Share this post


Link to post
Share on other sites
TheMule    193

In my tests on shiftplates, with high mass/SHC (1000kg of supercoolant), TC is basicly ignored. Dirt is much, much better than diamond. The amount of heat is basicly proportional to SHC, not TC. I suspect because of the clamping here:

On 4/15/2020 at 12:52 PM, nakomaru said:

If this is bigger than (Q2-Q1)/8, then (Q2-Q1)/8 is used instead. Q = SHC * m * T ← mass limited rate

It's hard to do, and some strange things happen (heat might be created) but under specific conditions plastic and ceramic shiftplates outperform even dirt (they have higher SHC).

So my take is that heat transferred depends on TC, until you hit the limit that depends on SHC and mass. Since (for shiftplates) mass is always 160kg, it becomes a game of pure SHC. The funny part is that usually materials with high TC have low SHC and viceversa. When you approach the limit low TC (but high SHC) materials start catching up and even outperforming high TC ones.

I haven't explored further, expecially the borderline zone. For sure, for high temp differences and mass, dirt (high SHC) wins (I don't consider plastic an option, it melts way to easily, and ceramic is inconsistent). That's 1t of high SHC liquid per tile. For low thermal masses (like 20kg/t of steam), diamond (high TC) is best. For high pressures, my bet is on aluminum. That's what my tests were about. I've tried getting to the math of it but it's hard to keep up with high volume of heat transferred.

TL;TR, in order to equilize temp inside a pool or to extract heat from it, I'd use dirt. To help transfer heat from buildings (e.g. ATs) to steam, I'd use diamond. Note than in most tests gold ranked poorly. Never use gold for shitplates unless it's a very low pressure environment (about 500g steam/1000g H2.

Next thing I have to test is radiant pipes/SHC.

Share this post


Link to post
Share on other sites

I am very ashamed of the syntax errors in my file. I don't know English and hurried then:-(.
625 - this is a very large ratio. It literally makes no sense due thermal clamping.
In my experiment, liquid lead (500 kg per cell) was equal to a solid termium (100 kg per cell). Their heat capacity is approximately is equals.
But I entered the values in my file, and saw that the heat transfer of the thermium by clamping, not thermal conductivity. Liquid lead with a coefficient of 625 does not even make sense to check.

This is weird and seems wrong. Does it make sense to write a bug report? Linking to a wiki for developers is weird (it's not an official document). It is difficult to build an experimental setup and prove that heat transfer is wrong. But anyone can make sure that the steel door conducts heat from the diamond worse than aluminum.

image.png

Share this post


Link to post
Share on other sites
psusi    259
25 minutes ago, Alexander Block said:

But anyone can make sure that the steel door conducts heat from the diamond worse than aluminum.

Wait, you mean an aluminum door does not conduct heat better than a steel one?  Actually in game?

Share this post


Link to post
Share on other sites

An aluminum door is in some cases better than a steel door. Thermal transfer limited by trermal conductivity and thermal clamping.

Thermal claming means, that heat transfer per tick cannot be more than 1/8 of the difference of thermal energy. Trermal energy = mass * temperature(K) * heat capacity. Some material combinations have the same energy at different temperatures. These combinations will conduct heat very poorly.
E.g. diamond on 600 С (100 kg per cell) and steel door (200 kg per cell) on 190C
Diamond energy=(600+273)*100000*0.516=45 046 800 DTU
Steel door energy=(190+273)*200000*0.49=45 374 000 DTU

 

Share this post


Link to post
Share on other sites
psusi    259
1 hour ago, Alexander Block said:

An aluminum door is in some cases better than a steel door. Thermal transfer limited by trermal conductivity and thermal clamping.

Thermal claming means, that heat transfer per tick cannot be more than 1/8 of the difference of thermal energy. Trermal energy = mass * temperature(K) * heat capacity. Some material combinations have the same energy at different temperatures. These combinations will conduct heat very poorly.
E.g. diamond on 600 С (100 kg per cell) and steel door (200 kg per cell) on 190C
Diamond energy=(600+273)*100000*0.516=45 046 800 DTU
Steel door energy=(190+273)*200000*0.49=45 374 000 DTU

 

Oh wow, that's fubar.  I'd lead the bug report with that example then.  Just show such a setup and how it barely conducts heat.  Why is there such a limit?  What is it supposed to do?

Share this post


Link to post
Share on other sites