Improved thermo regulator efficiency

[Mijae]

Based on an idea by Saturnus in this thread, I have been playing around with some different configurations to improve the efficiency of a thermo regulator.  The idea is to use feedback loops which mathematically should be able to loop some of the output from a thermo regulator back into itself.

I have had success that the average output from this is almost double the cooling of the regulator by itself.  If volumes were to split evenly at junction points rather than every other packet going though, it should stabilize to exactly double efficiency.

I found I needed to use pipe bridges to force splits to be one-directional.  Otherwise you see packets getting confused trying to change direction within the pipe.  You can have one input with multiple outputs or multiple inputs with one output.  However, if you ever have multiple of both the system gets confused.

With this new configuration, I can put 8 in series to get roughly the same cooling as 16.  Putting one in the front without this gets back to ~17.

The main drawback to this is that (again) since volumes don't split evenly at junction points, you could end up with some packets that don't liquefy and some that solidify.  The solid O2 should be less of an issue once the new update is available.  Since the remaining contaminated O2 will be super-cooled, it might best be to make a separate chamber that loops on itself to complete going to liquid.  At that point it might make more sense to always loop with fewer of these or use 9 to over-cool and just let more solidify.

Theoretically, this concept could be taken to an extreme by chaining together the feedback loops recursively.  This way you should be able to achieve 16x cooling using just 4 thermo regulators.  However, my attempt was a pretty big failure.  Once again, my guess is due to how volumes split/merge.  If anyone can see a problem with this let me know

The main concern about this setup once again goes back to how volumes split at junction points.  I could see a case where the same packet ends up getting stuck in the loop and never makes it out.  This would result in the temperature going negative and crashing the game.

Another option would be splitting the output additional times back into the feedback loop to increase the cooling.  However, this also increases the pressure/volume within the loop.  Again, due to volumes not splitting evenly my tests for this eventually resulted in 10kg packets sitting in the pipes and probably getting destroyed.  If this were to work, I think it could be possible to achieve liquid oxygen with a just single thermo regulator.

EDIT: Corrected a reference from "CO2" to "contaminated O2" since that is obviously a different thing.

[Masterpintsman]

I think we should get (in the sense of: developers, please implement them) thermal filters that filter gas packets based on temperature.

[Mijae]

Additional note: As mentioned in the last paragraph, feedback loops increase the pressure on the system equal to the efficiency gain.  This means they only support (max volume / gain) initial input volume.  If a regulator only supports ~1 kg/s, then you can only use 500 g/s input.  So, if your current system supports 2 input pumps then this will not help.

[heckubis]

I have my compressors running in series with there own power system runoff the hydrogen generator . if a compressor is at low power it waits till it has energy to cool. As it is waiting its packet becomes fuller before it cools and passes it on to the next cooling unit by the time the gasses are leaving the system they are 4-8kg packets instead of trying to use a full unit of energy unit on150g.

splitting it and looping it back is risky  and has to be manually controlled or you get over frozen or uncooled passing through. if you set you system up properly in the first place even at low power you know exactly the temps the gasses are coming out at and can easily set up filters along the chain to syphon off gasses at temp   I have this one running contaminated and steam in this shot but I'm cycling liquid oxegen out of it now from a room to the right

[Saturnus]

@Mijae I agree that the problem is likely down to the game not simulating fluid and gas like they should behave. I think the closest we can come to a stable system with the current modelling of in game physics is to try and anticipate the randomness and run 4 regulators, where the 1st loops forward to the 3rd, the 2nd loops forward to the 4th, the 3rd loops backwards to the 1st, and the 4th loops backwards to 2nd. Now that is annoyingly complicated to set up but doing so should quickly disperse denser and less dense "packets" through the system.

[Octyabr]

3 hours ago, Saturnus said:

 I agree that the problem is likely down to the game not simulating fluid and gas like they should behave.

For now it's ok, but somewhere in time it must be addressed properly. like the complete absence of pressure.
The game could be able to discern enclosed rooms and treat them as a whole instead of watching and propagating gases around for every single square. (Why do I have a mass of ~60g of CO2 below a huge mass of O2 at ~1.2kg?)
That'd fix the pumps and also improve performance.

4 hours ago, Masterpintsman said:

I think we should get (in the sense of: developers, please implement them) thermal filters that filter gas packets based on temperature.

That's a Maxwell's demon, an interesting opportunity for a crossover...

While I congratulate the OP for his effort with the elements we currently have in the game, I'd like to address that all this setup and your proposed tool (the demon) will be unnecessary if the devs implement an air cooler that we can set the target temperature. (consuming energy based on the difference between input and output, of course)

[MoMuffins]

Gas filter with temperature sensing will be needed anyway.  The game is set to have pipe burst when their contents become a solid.  While this will surely prevent the current issue of crashing from going below 0 Kelvin, it will be quite annoying having to fix your pipes all the time.  Temperature sensing of the environment is already in plan for switching on and off circuits so it would make sense to have one for the pipe's content.

@heckubis Gas can queue up at the input, but it will still leave the final output at around 1000g +/- 300g packets because the thermo-regulator is limited to such packet size.

[heckubis]

your right it is about 1k give or take 300 g just checked again but running the cooler at capacity is far more efficient then any setup that only randomly lets the right temperature through wile still having to be cooled the same number of times. what takes more energy for cold air to warm or warm air to cool looping back cold air warms the cold as much as it cools the warm  so if your looping system ends up balancing out mass going through your still ending up cooling the air the same number of times but one system is not letting random quantities of random temperatures through. those random chunks of not liquid are evidence in the flaws of the system. that its either wasting energy freezing too much or not cooling it enough is because its not actually doing an efficient job.

creating a loop is only a means for saving building resources and space at the direct cost in having a stable system. your wasting energy to not have to build a larger system that will do the job with the same number of times through the coolers no matter the volume of gas

[Mijae]

It's not about saving space and building resources. As noted, the difference is a trade off of volume for power.  Yes, a 17 regulator system is going to be more efficient if it running at full capacity.  You are correct that it does not change the number of times an individual packet needs to go through a regulator.  That said, If your regulators are not processing the full volume they can support you are losing efficiency.  Using feedback loops is a way to make use of that lost efficiency.  This is not the same as an infinite loop that you have to start and stop manually.

The theory is sound and if the game improves its fluid dynamics implementation it should be viable.  It not a flawless system, but has utility.

[Octyabr]

4 hours ago, Fatmice said:

Gas filter with temperature sensing will be needed anyway.  The game is set to have pipe burst when their contents become a solid.

What you want is pressurised pipes, not a Maxwell's demon. I'm more inclined in adding more real mechanics in the game than adding a apparatus that violates the Second Law of Thermodynamics. Just a bias of mine.

[MoMuffins]

It's a game, don't need to be real mechanic.  Just good enough.  Pipes are just rail roads with cargo cars on it.  It doesn't behave like a continuous fluid, nor is it needed.  The amount of computation needed doesn't add game play.  So demon or not, the current mechanic is fine.  What's wrong with adding a temperature sensor for the content of the pipe?

[Masterpintsman]

1 hour ago, Octyabr said:

... Maxwell's demon. I'm more inclined in adding more real mechanics in the game than adding a apparatus that violates the Second Law of Thermodynamics.

The real-world implementation of the suggested thermal filter is a split in the pipe with a bimetal agitated flap, when temperature is below the threshold the flap blocks one output (thus the other is open), when temperature is above threshold the bimetal bends in the other direction and the flap closes the other output. Thus the suggestion to let it work on gas packets (same as the filter by type works) in the sense of packet comes in, tempers bimetal (=> sets flap), packet goes out.

It wasn't intended as a maxwell demon to overcome entropy by splitting a gas packets into two (one cold, the other hot).

[jigggy2000]

13 hours ago, Mijae said:

 

 

Putting a valve (with max volume) at the end of this system makes the gas loop quite a considerable amount. I ended up putting  an emergency vent to let out all the backed up almost frozen gas. And that was just with 2 thermos.

Probably not an ideal method but just wanted to see the results. 

[Automatonnn]

I think Mijea his idea to create a single thermo regulator for x16 cooling is possible, but a buffer would have to be added to average the temperature in the system. This should also provide a more consistent output.

Edit: I figured out how to do this.

I managed to create a setup that cools from 30 to around -190 using a single thermo regulator. The output is very stable. However its largest draw back is that it needs to stabelize its internal tempurature for +- 10 cycles. It will jump below and above its final temperature until then. As you can see from the picture it has an output of 130 g/s. What is also nice is that you can change the output temperature by changing the output rate. With a smaller output rate providing a cooler output temperature. 

 

This is the setup. The numbers show the settings of the gas valves.

 

Spoiler