Heat Transfer in 266730

[wachunga]

So with all the changes I thought I would do a quick test to see what we're working with in the release. Here are the 4 different heat transfer regimes in ONI. I didn't exhaustively test every combination of materials, there may be edge cases or bugs. Just an overview.

 

1) Tile/Tile: Same as before previews, lowest conductivity wins. Simple test with granite touching wolframite and granite touching tungsten. Each behaved exactly the same, ergo no averaging going on. Single width abyssalite should be just as effective as before, no need for insulated.

https://imgur.com/a/I16itFB

 

2) Tile/Buildings(pipes, wires, plates, etc). Even before the previews, this depended on conductivities of both materials.  Perhaps the specific calculations changed, but the overall behavior is the same. Insulated vs regular vs radiant seems to just be a conductivity multiplier. Radiant pipe is twice as effective as regular which is 32x as effective as insulated. As the deltaTs diverge, rates will change. So in practice, radiant isn't exactly twice regular. Seems like the conductivity displayed in properties is all you need to know. For example, insulated tungsten is similar to regular sedimentary. Insulated invoking a lowest conductivity rule does not appear to be true, it just has lower conductivity.

 

3) Pipe/Contents. This is what changed. Before it behaved like Tile/Tile, now it's like Tile/Building. Like Tile/Buildings, insulated vs regular vs radiant appears to just be conductivity multipliers.

 

4) Tile/Loose(material from digging/deconstruction, including stuff on conveyors). Same as before previews, lowest conductivity wins. The change to pipes may mean that conveyors are no longer the best "radiator". Needs testing. Conveyors used to be better because pipes were so lousy. How does a radiant tungsten pipe filled with petroleum now compare to a conveyor running tungsten through tungsten?

 

For those doing their own tests, do note that buildings change temperature 5x as much as you would think judging from specific heat calculations (compared to tiles or contents).

 

[superhappyfun]

This is exactly what I wanted to know thank you!

[stretch611]

I noticed this in the release notes:

Quote

Do a special pass on the world when loading old saves or starting new games to make Abyssalite temperatures match their surroundings as they conduct temperature very slightly now.

From your testing it appears as that is not so, or it is so insignificant that Abyssalite tiles are still so good that you do not need to bother making them insulated (and using more material.)

Did your testing include pipes and vents as well?

[avc15]

log average of abyssalite against a tile with conductivity of 2 (igneous rock) is very low.

Igneous rock: thermal conductivity of 2,

Abyssalite: 1E-05

log-average of igneous and abyssalite: 4E-03 (a QUITE low thermal conductivity, probably would be hard in practice to differentiate from the old behavior of a single thickness tile of abyssalite)

I have some doubts and I default to believing the patch notes. But, the OP's data definitely bears further investigation.

[wachunga]

I specifically examined the conductivity between granite/tungsten and granite/wolframite tiles. The conductivity is exactly the same. There can be no averaging, whether it be arithmetic, geometric, or logarithmic if the conductivity is exactly the same, which it is. It may be that averaging is the intended behavior but there is a bug. Perhaps the bug only effects some systems, of which mine happens to be one. Perhaps the notes are erroneous. Perhaps the devs were fixing a problem that was later reverted.

I can definitively say that every test I did showed zero averaging in tile/tile interactions. If other people have different experiences I would be very interested. It's a pretty easy test to do. In debug, place two isolated 200kg 25K granite tiles in vacuum. Next to one place a 200kg 525K tungsten tile, next to the other place a 200kg 525K wolframite tile. Observe the rate at which the temperatures change. For me it is exactly the same.

[crypticorb]

Another important bit to note is that while abyssalite tiles are premium for building perfectly insulated sectors, insulated tiles are nearly as good for almost every purpose. Insulated tiles take twice as long to deliver/build, but they can be made from FAR easier to obtain materials for nearly the same effectiveness.

Other than sleet wheat farms, volcanoes, and coolant tanks, I use whatever material other than abyssalite for insulation. The conductivity of abyssalite is nearly zero, but the insulated tiles are, in practice, just as good.