Jump to content

Counterflow Heat Exchange


Recommended Posts

1 hour ago, biopon said:

Ok so you're suggesting this, but with diamond instead of metal.

image.png.131a2c4a2b341a7c75fb3d867bab9184.png

This has tested significantly worse than tempshifts and a gas as a medium, except when using thermium for the metal tiles. Diamond is marginally better than metal. Or am I missing something?

Terribly sorry I forgot you cant put tiles on tempshift plates (lol). Here is the correction which will yield the expected results: 

Make the tiles 1 apart from each other with tempshift in the middle. So from top to bottom repeat this: diamond window tile - diamond tempshift plate - diamond window tile.

Link to comment
Share on other sites

1 hour ago, Saturnus said:

Here's a picture from early September 2017 of a polluted water distiller I built that could process up 5500g/s. This was before geysers was introduced and when sand was only available in very limited amount in the starting biome. I later made a version that could process 10kg/s but this one is more interesting in the sense that it uses both a counter current and a co-current heat exchanger in the situation where each type is the best option.
5c572a52e9c80_2017-09-07(1).thumb.png.11c2e5dcb1f0cd353200925d2cc30ab1.png
 

I remember that! I posted a petrol boiler around the same time but I can't find any pics of my build. 

Link to comment
Share on other sites

51 minutes ago, MorsDux said:

Terribly sorry I forgot you cant put tiles on tempshift plates (lol). Here is the correction which will yield the expected results: 

Make the tiles 1 apart from each other with tempshift in the middle. So from top to bottom repeat this: diamond window tile - diamond tempshift plate - diamond window tile.

You definitely do not want temp shift plates anywhere in a counter current heat exchanger (unless as used in the OPs set up) as they defeat the entire purpose of equalizing temperatures in discrete segments. Temp shift plates will just make the whole thing the same temperature and thereby equal to a co-current heat exchanger.

Here's an example of a very effective counter current heat exchanger that is temperature controlled as you can set the doors to open and thereby cancel heat transfer in that section when the desired temperature is reached.

image.thumb.png.7a8f755f549a226454ddf7efe095dd6f.png

Link to comment
Share on other sites

2 minutes ago, Saturnus said:

You definitely do not want temp shift plates anywhere in a counter current heat exchanger 

He meant it like in mine, vertically. I'll test it, i think it may be very good, better than doors in your post. :p Cancellation option aside however.

Link to comment
Share on other sites

1 minute ago, Saturnus said:

I think you missed the point of the doors then...

No sir, you missed my edit. :p

6 minutes ago, Saturnus said:

I think you missed the point of the doors then...

Actually, you may be onto something with the doors, other than the ability to open them I mean. Are the top and bottom halves tied together as far as its temperature is concerned? I'm looking at my game and it seems all doors have uniform temperature across their 2 tiles. If that is true, it's perfect heat conductivity, and using doors may actually be the best possible option if you need 2 tiles because you're doing liquid/liquid or gas/gas.

Link to comment
Share on other sites

51 minutes ago, biopon said:

Actually, you may be onto something with the doors, other than the ability to open them I mean. Are the top and bottom halves tied together as far as its temperature is concerned? I'm looking at my game and it seems all doors have uniform temperature across their 2 tiles. If that is true, it's perfect heat conductivity, and using doors may actually be the best possible option if you need 2 tiles because you're doing liquid/liquid or gas/gas.

That is correct. Buildings that occupy more than one tile, in this case two tiles, have the exact same temperature in all tiles so it shows 100% efficient heat transfer between it's tiles.

In my distiller example you'll notice the statues. They are used as they are 3 tiles tall heat exchangers. The top is in the steam, the middle is in the outgoing clean condensed water, and the bottom is in the input polluted water.

Btw, the example you have shown here indicate a oil cracker build. In such counter current flow is typically achieved by submerged pipes. Here's an example of one of my earlier builds that cracks 4000g/s crude oil into petroleum.

Crude oil input is 90C. Petroleum output is 116C, so cold enough that the pump can be made of gold. No space age materials is used, not even ceramics or steel is used anywhere. Heating is done with obsidian pipes in magma filled with superheated steam. It uses about 30g/s of superheated steam to cook the crude oil from the 390C the heat exchanger is limited to with the shut off valves and temp sensors to 405C.

image.png

Link to comment
Share on other sites

Ok so I put together a fairly comprehensive test, here are the results:

image.thumb.png.b0fd99b234a2b95c9798f6a0db0a0001.png

Building and pipe material is gold and ceramics, unless otherwise noted. Testing methodology was:

- Counterflow 402 Celsius petroleum and 90 Celsius crude at 10kg/sec.

- Reach 80% efficiency. (152.4 Celsius petroleum and 339.6 degree crude.) When criteria met, remove a segment, until criteria are no longer met. Then re-add a segment to reach or exceed 80% efficiency again.

- Score is the number of pipe segments in the exchanger.

Save file (QoL Mk2 dev branch):

The Antigravity Sewer.sav

From worst to best:

image.png.3e7561842aecc371db5e6b500067c54a.png

J 143+: Gold amalgam doors. I was very surprised to get such a low ranking for this. It came close but couldn't reach 80% in the space I had. I guess on the one hand, doors perform instant and perfect thermal transfer within themselves, but on the other hand, they receive an 80% malus for being a building.

image.png.2e91b79696efb7e9fc5e0f5824494dfc.png

Q 143+: Long flat flow. Liquids are gooey. This is garbage. :p

image.png.956b993d4fbd4017dc65de3235281522.png

G 119: Surprisingly low rank for what I thought was sure to be one of the top ones.

image.png.57181d1495cf13c33a0465e829487b4c.png

A 109: A bad idea I wanted to test. I guess the tempshifts affecting the ceramics is causing a fair bit of inefficiency.

image.png.ebbde44c0ea7c389b6d360667220aee2.png

I 107: Suggested by @MorsDux, I expected more from this. Diamond tiles with diamond tempshifts.

image.png.0eb10d4c4e111dc3ab20f89dca4720ae.png

H 103: Thermium tiles and gold pipes. Very poor result given the space material used.

image.png.1db6594c1ef8c1287edfc1494446038a.png

K 101: Thermium doors with gold pipes. Also disappointing.

image.png.5a14d85acbaa5fdcdf84e2de6bae7dad.png

E 67: 750kg petroleum per tile, diamond tempshifts in the center. It's not a very good score, and takes ages to get up to operating temperature, but it has unmatched buffering capabilities and should handle uneven/intermittent flows very well.

image.png.2608bbaed6274b61844169e640cf5663.png

F 64: Dumb as a rock, gold tiles and gold pipes within. I am shocked this did so well; much better than gold tiles with 1 space apart. I can't explain.

image.png.c326243680a8966558a807ee7b766941.png

D 60: Iron pipe for top (hot), and gold pipe for cold (below). Hydrogen gas medium, diamond tempshifts. This should have placed better according to @wachunga's findings.

image.png.d4c41ad610061b2f639ad7656f0e6606.png

C&B 57: Gold pipes, otherwise same as above. B is cold above, C is hot above. According to @wachunga the "cold above" setup should be worse, but the difference between the two are negligible.

image.png.c461ba729d9ea793bd8f301371e57672.png

P 40: Amazingly good. Only beaten with space stuff.

image.png.0b926900a1d635f2172b0ebf902b863e.png

M 33: Dumb as a rock, thermium tiles. Very good for the effort I guess.

image.png.efbf48c142c18e8c92b8bd999527f532.png

N 24: Same as C but with thermium pipes. I was very surprised to see it place so well.

image.png.bdb16c98828a82ae7222e5910e7b1365.png

L 23: Same as E, but with thermium pipes. 

  image.png.261f87034454128df0ec23446f4230da.png

Q 20: Same as P but with thermium pipes.

EDIT: Revisiting the doors, made from thermium and with thermium pipes, they scored 29 for 4th place, but they're also by far the most expensive option.

Link to comment
Share on other sites

image.thumb.png.e74120604fb749e0811a529d244cc1e0.png

This is what I mean by vertical exchangers with hot above cold. The left is copper pipes with hot petroleum entering at the top and cold crude entering at the bottom. The middle is the same but with gold. The right is copper pipes but with cold crude on top and hot petroleum on bottom. This particular test doesn't show the difference between copper and gold. It does show that hot above cold is better than cold above hot and that vertical is superior to horizontal (offscreen above). Adding shift plates improves things, but the order of performance is the same.

Pipe Test 2.sav

The problem with doors is that the unified temperature thing Saturnus mentioned only kinda works. When a door closes, 2 tiles phase into existence. It seems those tiles take their temperature from that which is on the "airlock" info box. However from that point on they act as 2 separate adjacent tiles. You can click more than once to get the info box for the tiles underneath the door. They will have different temperatures in a heat transfer situation. The "airlock" info box takes it's temperature from the tile with the automation port. I don't recall if it has been this way forever or changed, it was a bit broken in the QoL1 preview IIRC.

Free liquid in direct contact with a pipe works well because there are only 2 steps. Free liquid to the pipe, then pipe to the piped liquid. Everything else has extra steps which slows down the transfer. Thermium works well even with the extra steps because it's OP.

Personally I use free liquid to pipe for a crude to petroleum cooker that is the same in principle as what Saturnus posted. It's a bit different in configuration as my goals are a bit different than his. The 30 g/s thing he does to not overcook his petroleum is quite interesting and not something I had thought of. I also use hot steam, but from a copper volcano and with a different way to prevent overcooking.

 

Link to comment
Share on other sites

This is very useful, thank you!

Any thoughts on why the consecutive metal tiles are so good, while the spaced out metal tiles are not? 64 wide versus 119. The latter is so wide because of the spacing, of course, and it'd have to be 128 wide for the horizontal transfer in the 64-wide tiles to not matter at all. (I hope this logic makes sense.) It's not quite 128 but not far from it. 

Am I correct that for the radiant-to-tile equation this is the one to use:

q = k1 * k2 * dT * HC / 10

For tile-to-tile, this one: 

q = klow * dT * 200

In the first case with gold/gold we have:

q = 120 * 60 *dT * 50 * 0.129 / 10 = dT * 4644

or, if the tile is hotter,

q = 120 * 60 *dT * 100 * 0.129 / 10 = dT * 9288

In the 2nd:

q = 60 * dT * 200 = dT * 12000

So tile-to-tile should be significantly stronger. Except... You talk about tile-to-tile transfers being capped at dT/4. If ~313 degrees of difference is distributed along 64 tiles, that's 4.89 degree difference per tile so dT/4 = 1.22. Which explains why radiant gets to owerpower adjacent thermium tiles.

Do I have the right read on this?

 

Link to comment
Share on other sites

Looks mostly good. For the radiant-tile, the masses should be in grams, so bump those numbers by 1,000. Also a pipe is a building and it's HC gets reduced by 1/5 (because Klei). I briefly retested the tile-tile, pipe-tile, and pipe-contents equations and they are the same except for something with the tile-tile that is wonky. Doing some more tests.

Looks like there a x2 multiplier for tile-tile interactions involving thermium, if it's thermium-thermium then it's a x4 multiplier. This includes liquids and gases. Tile-tile seems to be limited to dT/8 which would explain the adjacent metal tile thing even more.

Link to comment
Share on other sites

the problem with free liquid design is you have to pump twice.   I guess I see efficiency different.  once in a pipe, I leave it there and design the best exchanger for that.  glad to see "F", which I posted and was critiqued on, did good enough.  nice and compact and easy to build.

Link to comment
Share on other sites

12 hours ago, biopon said:

image.png.ebbde44c0ea7c389b6d360667220aee2.png

I 107: Suggested by @MorsDux, I expected more from this. Diamond tiles with diamond tempshifts.

I think you got this one the wrong way around.

I think (and it is only a guess) that the intent was for it to be:

T V V T V V T V V ...

D V V D V V D V V ...

T V V T V V T V V ...

 

where T is a diamond shiftplate, D is diamond window, and V is a vacuum.

Link to comment
Share on other sites

21 minutes ago, Yunru said:

T V V T V V T V V ...

D V V D V V D V V ...

T V V T V V T V V ...

 

where T is a diamond shiftplate, D is diamond window, and V is a vacuum.

That would imply a tempshift plate behind the pipe segments and considering the adjacent tiles are in a vaccum, it would imply a vacuum behind the pipes.

=> This would exchange no heat at all ;)

 

But I would contribute an other design I like to use:

V D D D V ...

V D T D V ...

V D D D V ...

V = Vaccum; D = Diamond window tile; T = Diamond tempshift plate covered by 1t of water

It´s a block of 9 tiles with the middle one beeing a tempshift plate covered by water (later sometimes steam).

The liquid inside provides a good thermal stability.

(I am using just gold tiles and a diamond tempshift plate in the middle, but replacing the metal tiles with diamond should improve it.)

Link to comment
Share on other sites

Maybe larger sequential groups of metal blocks?

The best luck I had in trying to optimize a crude-> petro boiler was multiple sequential blocks of metal with radiant pipes in them. 4x4 ish

I noticed if they were too small not enough heat transferred, and if they were too few it would normalize the temp when flow was slower.

 

 

 

 

Link to comment
Share on other sites

TO Q 20 P40:  add mesh tiles and you can make it one tile smaller

To K101: Uee manual doors instead of automatic ones. Will reduce it to 91 or 93.

H103: Split the pipe into 2  (6200 to 3800) so that the transfer is over 4 tiles wide. Also use thermium for the pipes. Reduces the size to 31

After testing a bit you want the high conductive material to be the pipe and the low specific heat material to be the tile with that i could decrease the size to 21-23 tiles

M33: use gold instead of termium for the tiles and you can further decrease it by 4 tiles -> 29 tiles

20190204210917_1.thumb.jpg.fb0f1dbaecf87f370a2c4bd45b202b11.jpg

 

Link to comment
Share on other sites

14 hours ago, biopon said:

image.png.3e7561842aecc371db5e6b500067c54a.png

J 143+: Gold amalgam doors. I was very surprised to get such a low ranking for this. It came close but couldn't reach 80% in the space I had. I guess on the one hand, doors perform instant and perfect thermal transfer within themselves, but on the other hand, they receive an 80% malus for being a building.

You have to put it into context here. Last I tested these things extensively there was no radiant pipes, no window tiles, no refined metals, etc etc

The best we had for tiles was doors, iron ore doors (not gold amalgam which is much worse), and we typically used the manual doors due to them being twice the amount of metal per door.

The best we had for pipes was wolframite but that's always been scarce so we typically used granite pipes.

That makes for a horrible heat exchanger which is why we never did it that way. Instead we typically sucked up the added power consumption by running the liquids directly on top of eachother with tombstone heat exchangers between the two liquids.

Link to comment
Share on other sites

1 hour ago, Saturnus said:

You have to put it into context here. Last I tested these things extensively there was no radiant pipes, no window tiles, no refined metals, etc etc

The best we had for tiles was doors, iron ore doors (not gold amalgam which is much worse), and we typically used the manual doors due to them being twice the amount of metal per door.

The best we had for pipes was wolframite but that's always been scarce so we typically used granite pipes.

That makes for a horrible heat exchanger which is why we never did it that way. Instead we typically sucked up the added power consumption by running the liquids directly on top of eachother with tombstone heat exchangers between the two liquids.

kids today.....uphill both ways......

Link to comment
Share on other sites

2 hours ago, Saturnus said:

iron ore doors (not gold amalgam which is much worse), and we typically used the manual doors due to them being twice the amount of metal per door

The issue isn't so much the material, it's that the doors aren't really one thermal entity. There are two distinct ore tiles behind the building; the lower one matches the temp of the building while the upper one can and will have different temperatures.

image.png.bc61e4d74be27d795571e52e5841d1bb.png

By the way, I retested it just now, both for manual and auto airlocks, made from ore. They both scored 137; 18 worse (so rather close) than refined gold tiles. That's what the materials appear to account for - a significant difference but one that's dwarfed by the overall design.

 

 

 

Link to comment
Share on other sites

2 hours ago, Iluinrandir said:

To K101: Uee manual doors instead of automatic ones. Will reduce it to 91 or 93.

H103: Split the pipe into 2  (6200 to 3800) so that the transfer is over 4 tiles wide. Also use thermium for the pipes. Reduces the size to 31

M33: use gold instead of termium for the tiles and you can further decrease it by 4 tiles -> 29 tiles

K101: That should not make a difference if same material.

H103: If we're using multiple flows it's not an apples-to-apples comparison. And yeah, thermium is OP. :p

M33: That's the same as F64, which is way worse.

3 hours ago, Hellshound38 said:

Maybe larger sequential groups of metal blocks?

Cutting, for example, F (64) into the smallest possible chunks gives G (119) at the extreme. G is "more efficient" in that it uses 60 tile pairs to achieve what F does with 64, but in almost 2x the space, and with almost 2x the pipe. 

I'm not sure if using intermediate chunk sizes would make any improvement over a single big block.

Link to comment
Share on other sites

18 hours ago, biopon said:

image.png.2608bbaed6274b61844169e640cf5663.png

F 64: Dumb as a rock, gold tiles and gold pipes within. I am shocked this did so well; much better than gold tiles with 1 space apart. I can't explain.

Not as good as 2x2 block with space in between

The best I've found with normal materials is this where stream is split in two 5kg/s lines going either direction interleaved with 2x4 blocks.

image.thumb.png.0b4da57985e99a3116bf572c1e143781.png

Link to comment
Share on other sites

1 minute ago, Saturnus said:

Not as good as 2x2 block with space in between

I will test this. (I so wish you could run multiple copies of ONI.)

2 minutes ago, Saturnus said:

stream is split in two 5kg/s

You can split even further if you want to, but that would also apply to most other designs as well. I was trying to keep this an apples-to-apples comparison. :)

But I guess if we wanted to find the most efficient way to do this (aside from the naked option) I probably should test multiple streams.

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...