Jump to content

Heat clamping in chillers


Recommended Posts

I'm playing around with heat exchangers and came across this thread on regolith melters.

I don't want to necrobump it, but I have a few questions.

First, to clarify why @nakomaru used obsidian and iron (ore) instead of materials with higher TC. If I understand correctly the game clamps heat transfer based on the delta and using higher TC would not transfer energy quicker because it's limited by the massive temperature difference between the regolith and igneous rock.

Secondly, the convoluted route the regolith takes is because there is a little bit of heat transfer between a rail and the tile beneath it. Is that still a thing?

Spoiler

 

5.thumb.gif.da121eee5940bd11656e994c975b


 

The thread also mentions off-screen vs on-screen performance. Is that still a concern?

TY

Link to comment
Share on other sites

For testing purposes to show on the forum it is customary to use resources that is available on any map, so iron and iron ore are almost always used for metals unless higher overheating temperature demands another metal to be used. Aluminium for example is never used for demonstration purposes because it requires a map with a forest start. Copper and iron are very close to eachother but since forest starts don't have copper, iron is the preferred metal.

Obsidian has a very high melting point and is usually the material choice for showing of builds where there's extremely high temperatures. Note that if you're dealing with metal volcano output or rocket exhaust then it can melt ceramics too.

Link to comment
Share on other sites

Ceramic should only be melting in a metal volcano if you're not dumping the heat into a medium like steam. If you were just letting the molten metal build up then I could see it happening, but if you're using a steam chamber to absorb the heat and a steam turbine to convert the heat to electricity then your ceramic should never melt.

Link to comment
Share on other sites

7 hours ago, Saturnus said:

For testing purposes to show on the forum it is customary to use resources that is available on any map

I think Nakomaru choose obsidian because of its thermal properties. If I understand correctly debris on a rail obey the entity inside a cell heat transfer rules:

Entity inside of a cell:

{\displaystyle q=k_{lowest}*\Delta T*\Delta t*1000m}

source

That means that the regolith (TC 1) to tile and igneous rock (TC 2) to tile heat transfer is clamped by k_low - in this case 1. Using a material for the blocks with a higher TC would only improve the cell-to-cell transfer, not the heat movement in or out of the material on the rail.

  Still curious if the off-screen bug they talk about is still a thing.
Link to comment
Share on other sites

Exactly as @occamrazor says. Debris transfer uses the lower conductivity. Regolith has conductivity 1, so the best material for this has a TC as low as possible without going under 1, to minimize left-right transfer. Up-down transfer is desired, and largely done through all the bridges. Obsidian is the best constructed choice available here (with a lower SHC than Igneous Rock, it will reach steady state in 1/5th the time and have 5 times the minimum ΔT for thermal clamping). You could also try to use regolith itself, or even refined carbon by cooking coal (both TC 1).

11 hours ago, occamrazor said:

Secondly, the convoluted route the regolith takes is because there is a little bit of heat transfer between a rail and the tile beneath it. Is that still a thing?

This is still a thing. Although the path tries to optimize this (10 tiles are transferring heat in this way), half of the time it isn't allowed to happen because the debris never stays in the bridge entrance tiles. The bridge entrance tiles are placed to the outside of the block, which maximizes the time debris spends inside the block.

11 hours ago, occamrazor said:

The thread also mentions off-screen vs on-screen performance. Is that still a concern?

On 3/4/2020 at 7:04 AM, Ipsquiggle said:

Fix solid conduit contents not exchanging heat when the camera was in certain configurations

This has apparently been fixed! The large numbers of conveyer bridges in this design creates a large number of individual conduits. A while back solid conduits (rails) were the only type to not be multithreaded. If that is still the case, this design likely harms game performance more than a design without conveyer bridges, which are now unnecessary due to the fix. You could try a variant like this to preserve the one-tile-below transfer mode.

image.thumb.png.74f4cd6a45a69a944195072acff3d53f.png

Link to comment
Share on other sites

18 minutes ago, nakomaru said:

You could try a variant like this to preserve the one-tile-below transfer mode.

Thanks. Would it be further beneficial to break up the heat exchanger into discrete parts?

Also, what heat transfer equation governs rails in liquid?

20201109220103_1.jpg

 

 

Link to comment
Share on other sites

Short answer: no

In the linked thread you can see some prototyping with some discrete exchangers. If space is no concern, you can probably push the limits with such a design. But in my experience this was unnecessary and lead to worse results within the same space. Still, I would be happy to be proven wrong.

On 11/10/2020 at 1:01 PM, occamrazor said:

Also, what heat transfer equation governs rails in liquid?

This should be the same equation you linked earlier.

For debris-to-tile-below, it's either 1/4th that speed (per 2017 data) or 1/16th that speed (per wiki).

Link to comment
Share on other sites

8 minutes ago, nakomaru said:

It depends on what your input rock is, but if it is at 1409C, that means your ΔT is 13K.

The two smaller designs you originally linked (one imaged, and one described) were performing at ΔT of 1K and 0.3K.

Yeah, I was testing a discrete bridged version and I got the delta down to ~5K. That version did not have the pipe and automation bridges. My timer wasn't as good on that one as it was with the non-bridge version, but it looks like your version engaged the doors ~0.1% of the time.

With a lower ΔT I can reduce the number of steps in the stair heat exchanger.

More testing later.

Link to comment
Share on other sites

13 hours ago, nakomaru said:

The two smaller designs you originally linked (one imaged, and one described) were performing at ΔT of 1.3K and 0.3K.

Is that at a full 20kg/s and do they still work that well in the current game version?

The best I've been able to achieve is a ΔT of 3.9C with this layout:

1846036889_RegolithMelterHeatexchanger.thumb.png.b5b3c7eef3c47c2c2945dac3bc2d6367.png

Igneous 1408.35C -> ~1182.7C and regolith 275C -> ~1404.5C

Link to comment
Share on other sites

Linking the solid tiles with bridges brings the ΔT down a bit more. If one of the lines was always full, conveyor bridges could be used without messing up the flow. Otherwise stuck with power and automation. In theory, not/buffer/filter gates would be better than automation ribbon bridges (clamping in multi-cell buildings based on size) but the difference, if any, is probably tiny in practice.

image.thumb.png.fdf1448c84ecafd08ec8159c95398685.png

 

Edit:

Monolithic block style still seems the way to go.

image.thumb.png.f26132116eff480df81a795a0d306fbc.png

ΔT of 0.4C. If one was confident the igneous line would always be full, you could zigzag the regolith line above the solid blocks for some extra heat transfer time and bring the ΔT down to 0.3C.

Link to comment
Share on other sites

Did some testing tonight on different heat exchanger layouts. The four I tested were effectively identical.

20201111012758_1.thumb.jpg.754d5456181c9a4ba7e5f6f6f85adba5.jpg

The monolithic block (2nd from top) was garbage as expected. Regolith never got much above ~1360° C. So I didn't include it in the benchmark run.

Once everything came up to temperature I ran 100 tons of regolith through the machines and recorded how much of the time the heat transfer doors engaged. I also observed the temperature delta at the HEx regolith exit from time to time as the machine was running full tilt.

The results from top to bottom:

Configuration Delta T Door Usage
56 1x2 blocks ~4K 0.15%
28 1x4 blocks ~2K 0.12%
8 3x4 + 2 2x4 blocks ~3K 0.18%
2 14x3 blocks ~3K 0.07%

 

So the HEx basically performed equally. Good enough for the stair-step HEx to finish the job. I shouldn't be surprised that this is basically free energy, but wow. Using the heat injectors less than two tenths of a percent of the time is crazy.

Materials:

Blocks are obsidian. Rails are iron ore. Window is diamond. There are two diamond tempshift plates between the steel doors. Steel is used in the fork-shaped melter, but otherwise this build is basically pre-space. About 2 tons each of steel & diamond will do it. I did use one insulated tile made of insulation where the melty-fork interfaces with the stairs (but ceramic would work). The build is pretty simple and all the components are pretty "standard".

I don't think the magma freezer ever reached equilibrium. I have it set up to inject ~20 kg of water any time the steam chamber gets above 1405° C. Outgoing igneous rock heats it up as the machine gets humming. Pressure is about 135 kg steam per tile.  The magma has not built up enough on the freeze plate to form tiles of igneous rock (yet).

One thing to note: Until this bug is fixed the stairs should descend right-to-left as pictured above. Descending left-to-right will delete heat from the magma.

Link to comment
Share on other sites

It looks like you might not be using thermal bridges (liquid, automation, etc) in these designs, correct? They significantly improve performance, especially in 4 tile tall exchangers. (Buildings are usually HC clamped, so use high SHC materials like igneous rock and steel)

For discrete exchangers, they are already insulated in the horizontal direction, so use high TC like diamond.

Link to comment
Share on other sites

5 hours ago, nakomaru said:

It looks like you might not be using thermal bridges (liquid, automation, etc) in these designs, correct?

Incorrect. The bottom three all make use of wire, gas and liquid bridges.

 I suspect once the freeze plate stabilizes closer to 1405° C the true performance of the heat exchangers will come out, but they are all good enough for me.

Back to what I'm really working on - a lumber & ethanol heat exchanger....

Link to comment
Share on other sites

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.

×
×
  • Create New...