Jump to content

Horizontal CounterFlow Testing And Direction Comparison


Recommended Posts

Tldr: Avoid modifying the counterflow such as adding sculpture blocks, drywall, conveyor bridges, tempshift plates and so on. Make your counterflow 17 or less to avoid logarithmic loss of effeciency. (see xxxx)

 

  • General Methodology

Create identical setups with only one variable changing from setup to set, focusing exclusively on the counterflow. Flowing 100C crude oil through the pipe, with petroleum flowing the other. Using gold whenever possibly due to Coppers identical performance and Irons similar results. With a ( ) to clarify material choice if needed.

image.thumb.png.afd066e5aefee5597dcb536d57f0991a.png

This being 3 examples of the setup. Temperature measurements are taken from diamond tiles, with temperature from the infinite liquid sources being constant. RTL flow tests using identical setup only mirrored.

  • Counterflow length

While the curve isn't truely logarithmic does stagnate in growth, so anything beyond 20 tiles wont provide nearly then benefit as anything before that.

If anyone who's better at math than me, wants to figure out a function for the graph am I  happy to include the Data.

(Data for graph based of Left to Right counterflow)

image.png.55351ef9959d8f1b61b064ea434cabd9.png

 

  • Comparing RTL Counterflows with LTF Counterflows

I'm not sure if the results are gonna be "controversial" but it seemed like the bug had a very minor impact on the counterflow (assuming 10kg/s of flow) with it averaging out to RTL being 0,53% more efficient than LTR. With it's effect varrying based on length, albeit neglible amounts. Whether this variety comes down to bad measurements, game mechanics or bugs or a combination of all three is anyones guess.

 image.png.e0d74a1f013ad591cb8b6be8af9d9b8b.png

 

  • Improving Counterflow Efficiency

Without increasing length*

(All test for this was done using LTR flow and at 10, 20 and 50 pipe segment long counterflows)

It seems there isn't any real way to do this, with the following having all equal or worse performance than a lack of modifications.

Drywall every tile, Drywall+Ladder everytile, Tempshift every tile, Tempshift(Gold) every 2 tiles, Tempshift(Gold) every 3 tiles, Tempshift(Gold) every 4 tiles, Conveyor Bridge(Horizontal, Steel), Conveyor Bridge(Vertical, Steel), Sculpting Block(Granite).

If anyone have any more ideas would I happily test them out, do include an image if it isn't simple (am dumb).

 

That's it thanks for reading! And thanks to @Ishamoridin from the ONI discord for help with Math and formating.

Save File

Link to comment
Share on other sites

Thank you!

thermal mass don't increase working counter-flow efficiency -- it just helps if flow is not constant, and we need to keep state

And as we can now see from data, any thermal exchange in reverse direction is bad, so even very sparse thermal plates makes things worse, especially if temperature difference already small at the far end of long exchanger.

 

9 hours ago, DangerTeddyBear said:

If anyone have any more ideas would I happily test them out

Can you, please, test on more realistic numbers, with petro ~400C ? Will efficiency differ at those numbers?

Can you, please, test "improvement" with tiles under petroleum made from diamond (window tiles in place of bottom airflow), and oil pipe going through tiles, not through petroleum?

Can you please, test "improvement" with liquids reversed (oil flowing, petro in pipe) ?

Link to comment
Share on other sites

10 hours ago, DangerTeddyBear said:

If anyone who's better at math than me, wants to figure out a function for the graph am I  happy to include the Data.

When using the polynomial curve fitting, the function will be approximately y = 0.1617119 + 0.15836314 * ln (x), but it is only a prediction, and on it, efficiency in ~100 (and more:-D) percent is reached for length in 200 tiles.

Link to comment
Share on other sites

15 hours ago, DangerTeddyBear said:

I'm not sure if the results are gonna be "controversial" but it seemed like the bug had a very minor impact on the counterflow (assuming 10kg/s of flow) with it averaging out to RTL being 0,53% more efficient than LTR. With it's effect varrying based on length, albeit neglible amounts. Whether this variety comes down to bad measurements, game mechanics or bugs or a combination of all three is anyones guess.

Iirc, that bug affects staircases, not horizontal flow. The slight difference is probably just statistical variance. 

Link to comment
Share on other sites

On 7/26/2021 at 6:28 PM, DangerTeddyBear said:

it seemed like the bug had a very minor impact on the counterflow (assuming 10kg/s of flow)

Are you seeing any heat deletion/creation? I.e. is the heat gained by one fluid identical to the heat lost by the other?

Link to comment
Share on other sites

On 7/27/2021 at 3:52 PM, Hjoyn said:

Iirc, that bug affects staircases, not horizontal flow. The slight difference is probably just statistical variance. 

No it effects horizontal LTR flow as well, the effect is just much more pronounced with staircase LTR flow. 

 

On 7/28/2021 at 4:46 AM, ghkbrew said:

Are you seeing any heat deletion/creation? I.e. is the heat gained by one fluid identical to the heat lost by the other?

There is a amount of heat deletion, but it decreases as counterflow length increases.

image.png.e1e4e465a73de7394d022a149d738ad3.png

Link to comment
Share on other sites

On 7/27/2021 at 10:25 AM, Prince Mandor said:

Can you, please, test on more realistic numbers, with petro ~400C ? Will efficiency differ at those numbers?

Can you, please, test "improvement" with tiles under petroleum made from diamond (window tiles in place of bottom airflow), and oil pipe going through tiles, not through petroleum?

Can you please, test "improvement" with liquids reversed (oil flowing, petro in pipe) ?

Wouldn't a temperature difference of 100C be more realistic than somwhere around 315C since that difference is spread out over several heat exchangers.

Test is underway I'm gonna be letting it go over night.

I did any while their marginally better over all RTL flow is it still worse than no variable.

Link to comment
Share on other sites

12 hours ago, DangerTeddyBear said:

Wouldn't a temperature difference of 100C be more realistic than somwhere around 315C since that difference is spread out over several heat exchangers.

That's about most often situation. We have an oil at 95C from oil well going to petroleum boiler, and exchanging heat with petroleum, just boiled at 402C.

If game model is perfect, than this 100C difference is representative. But in reality it is not perfect. We have different result even on changing flow direction. Game have bugs, simplifications, clapping and other strange things to speedup calculations. So, it is interesting, is it gives same results on most often real usage scenario

Link to comment
Share on other sites

On 7/30/2021 at 9:21 AM, Prince Mandor said:

That's about most often situation. We have an oil at 95C from oil well going to petroleum boiler, and exchanging heat with petroleum, just boiled at 402C.

If game model is perfect, than this 100C difference is representative. But in reality it is not perfect. We have different result even on changing flow direction. Game have bugs, simplifications, clapping and other strange things to speedup calculations. So, it is interesting, is it gives same results on most often real usage scenario

What I meant was, it's not as if you only use a single counterflow, meaning the temperature difference would be divided over multiple counterflow. 
IE it wouldn't bee 317C difference in temperature. 
With most Anywhere between 2 or 4 typically to get a sufficiently efficient exchange of heat, my logic at least being 100C difference being roughly in the middle of the temperature difference you'd see in each counterflow. 

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...