Better central cooling

[avc15]

Here's my newest central cooling plant. You could do something quite similar with aquatuners and pwater but I'm using an AETN and all my base's wheezeworts.

Why did you build this monstrosity?

I built my first central liquid cooler because I realized how much work I was putting into hvac chains for my closed-chamber electrolysis systems. And then I realized just how tricky it was to make them efficient. I switched to open-air electrolysis like the layout shown below to save power, and just overbuilt all my electrolysis systems to keep up with oxygen consumption.

But then you have to cool the oxygen in place. So, piped liquid cooling. Much more power-efficient than piped gas cooling because liquid pipes deliver so much more mass than gas pipes.

Setting wheezeworts around the air paths works  at first but seems really wasteful. I wanted a way to harness ALL the cooling and automate how much gets piped in at any time. So, zone controls come next.

(fig 1: my open air electrolysis block, each base contains at least 2. Nothing fancy here)

Why do you say this is better?

Well, someone else's system may be better for you. I don't exactly make my systems simple, just very efficient, modular, and extremely robust. If you want a  central cooling system that's easy to set up, start out with something like Lifegrow posted here. It works great and is much less work to set up.

But, here's why I chose to make this system instead:

- This cooling plant is extremely thermally reactive. The resistivity per unit length is as low as I could figure out how to make it; that means heat fluxes can be high with very small temp differences, and it'll ramp up its cooling supply almost immediately when demand changes.

- The gas, liquid, and plates buffer a TREMENDOUS amount of cooling potential, so that my plant is very resilient to massive load surges. (temperatures hardly move even if you spill a whole lot of heat into it all at once, well above its continuous load capability)

- This particular arrangement BARELY impacts game performance, unlike my last show+tell. (That's one of two interesting parts for this post)

But avc do you actually need something so robust? In a past map I did, when I made mistakes and spilled heat everywhere. But now it's probably overkill unless I want to direct condense steam for a steam turbine or something.

The liquid cooling plant

This is just two chambers:

(1) an insulated chamber built around an AETN, plastered with diamond tempshift plates, every wheezewort in my map crammed into it, and pumped full of hydrogen to 20kg. (that 20kg is important, back to it later)

The gas cooling chamber is built around a liquid cooling chamber. Built around for maximum surface area.

(2) the liquid chamber with controls for turning on and off heat transfer from the gas chamber - regulates crude oil temp to -37C. It's been pumped to a vacuum. This chamber provides surge volume for the doors to close, and surge volume for the cooling loops. Liquid chamber is also criss-crossed with diamond temp shift plates.

Temp control in the liquid chamber

This takes a small bit of explaining because my setup looks different from what we've seen from the community until now.

My liquid cooler uses a different arrangement: from left to right, hydrogen, plastic tile, then two layers of doors (material is irrelevant), then the oil.

 

You're probably used to seeing this arrangement instead (hi brothgar - btw brothgar didn't make this up, it's been in use for a long time):

This more common setup goes tungsten tile -> wolframite door -> tungsten tile. Brothgar here is talking about a heating system, where you can't use plastic anyway, but even in cooling setups most people are using wolframite tiles. (It makes intuitive sense)

So why are you doing it differently?

Well, first, plastic has a thermal conductivity much higher than tungsten's. Plastic comes in at 913 W/m/K (kleeeeiiiiii!) while tungsten comes in at a measly 60. The real world problem with this is that plastic would brittle fracture if you put it under so much thermal strain, not necessarily the conductivity values.

But then, arrangement. The common arrangement is orders of magnitude more resistive. When my doors are open, I have oil and hydrogen separated by one layer of plastic, which amounts to heat transfer almost as good as just having them in direct contact. When brothgar's doors are closed, it's a little bit worse than having a single tile of wolframite between them. My setup is orders of magnitude more conductive per unit length, making for a more robust cooler.

Again, you can only use plastic in systems that will always stay considerably colder than 70C.

Show us how it regulates. 

Fine! The temp thermostat is set to -37C. If the thermostat goes above -37, all the doors open. If the thermostat goes below -37, after five seconds the outer doors shut, at 10 seconds the inner doors shut, and at 15 seconds the outer doors open again.

Probably not best illustrated with an animated gif, but refer to the closeup logic diagram above. I just don't want to make a post on youtube right now.

The system is mechanically mono-stable. That means it shouldn't ever just become unstable and flip back and forth between open and closed quickly, so you don't have to add debounce features to the temp control logic. The doors may immediately enter their closing sequence after opening when the system is barely loaded (opening the door mixes out any temperature gradient), but it shouldn't ever close doors and then immediately re-open them.

Difficulties

Every once in a long while, the lower row of plastic tile takes bit of pressure damage on the order of 10% for one or two tiles. I'm not 100% sure how to eliminate that problem, since I've sequenced the door operation to be very gentle. But, someone could make a suggestion.

Zone coolers

There are a few different ways to set up zone cooling, and I have several different builds in my base.

- Open air zone cooling: I use this for air systems, no surprise. Use some airflow tiles where you want the air to pass through, and this is your cooling boundary. Run some radiant pipes made from tungsten through the airflow tiles. Set a thermostat in the air nearby. Pictured: one of my electrolysis blocks and its zone cooling. Each zone connects a liquid shutoff to the supply header, that valve empties into a liquid flow valve. Zones with more radiant pipe are set to 5000 g/s, and zones with less are set to 3000 g/s. The goal when you choose a flow rate is for your coolant to heat up most of the way to room temp as it passes through radiant pipes. Shorter runs of radiant pipe, lower flow rate.

Each of the 4 zones shown here has its own temp sensor set to activate above 27C.

Control logic:

These cooling supply valves are clocked. Anytime the clock is active (3 sec on 60 sec off) all zones with an active temp sensor turn ON. The end result is that in most cases exactly 20 kg of oil gets delivered through the cooler every minute for each zone hotter than its set point.

- Radiated cooling: This is the most efficient method but it's not always possible to incorporate it into some rooms. I tend to use radiators for cooling geyser output, farms, ranches ...

In cool rooms, the radiator should be made out of plastic; in hot rooms, tungsten. With radiant pipe inside plastic, you don't need a very large radiator at ALL to extract almost all of the cooling from your oil. Tungsten is alright too, but thanks to the log-averaging of conductivities, plastic performs incredibly well. plastic will very quickly just drag all the available delta T out of embedded radiant pipes.

See these two cooling zones in my berry farm. Only two sections of the pipe are radiant, I've got just four blocks of plastic for the radiator on each floor. The flow valves are set to 5000 g/s but oil exiting into the return header leaves at ~0C, having entered at -37C. About 60% utilization, perfect!

 Now, here's one of my steam geysers. I throttle the flow low enough that oil exits back to the return header as near to 100C as possible. My goal is to condense a pool of 99C water, that I'll use for my steam turbine / boiler later. Since this geyser that I'm showing is a pretty slow geyser I only need 100g/s of coolant flow.

I've set liquid shutoff valves on a clock and gated with a pressure sensor, not a temp sensor. Cooling only turns on when pressure is above 1000g at either corner of the room above.

To condense water at a higher temp, I had to slow down condensation. I put the coolers far away from the steam vent, made very small coolers, made the room large, and barely supply any cooling flow. I chose a flow rate such that oil flows out at around 99C.

Return header

We still have to discuss the coolant return header. How you set this up might absolutely kill your game performance if you're careless. It'll be  a little more challenging to explain why.

[PhailRaptor]

For the Plastic vs Metal tile discussion, aren't you still bottlenecked by the conductivity of the gas or liquid that is being contacted?

[0xFADE]

Hmm. I had the bulk of my base cooled to 20c by only radiating through 2 AETN with granite pipes filled with petroleum. Wheezeworts are better for spot cooling

[avc15]

15 hours ago, PhailRaptor said:

For the Plastic vs Metal tile discussion, aren't you still bottlenecked by the conductivity of the gas or liquid that is being contacted?

yes, hydrogen is the limiting factor. But it makes a difference. We switched from a 3-wide thermal contact with total thermal conductivity about 6, to a 1-wide thermal contact with total thermal conductivity about 900.

In this system the resistance of the heat exchanger is basically zero. In the system we see all over ONI content right now (tungsten-door-tungsten), thermal resistance is significant.

Basically this barrier lets you get to the same rate of heat transfer at about a third of the delta T.

15 hours ago, 0xFADE said:

Hmm. I had the bulk of my base cooled to 20c by only radiating through 2 AETN with granite pipes filled with petroleum. Wheezeworts are better for spot cooling

Let's say instead that wheezeworts are designed to make spot cooling very easy.

However, you can't regulate the output of a wheezewort. If you put it in hydrogen and regulate outflow from the hydrogen, you're going to be wasting most of your wheezewort's cooling capacity.

Back to the main advantage of central cooling: you don't waste anything. You pipe just exactly however much cooling to the precise location where it is needed. You'll still stifle your wheezeworts, but you can get that lost cooling back just by building more radiators somewhere else.

Yeah, if you want a simple system just put wheezeworts in your base. I've done it before, but now I'm going to shift my focus into some kind of industry that needs tons of cooling. Haven't decided what yet.

[Carnis]

Its very cool and lovely complexity! Also some new ideas. Would a 2tile-wide plastic at the bottom avoid pressure dmg on the plastic tiles? Like 4 tiles of abyssalite never break under pressure?

Ive never needed this much cooling, I guess insulated electrolyzers with pre-cooling save a lot of trouble.

Still lovely and fresh new ideas. I also use WWs as spot coolers (next to kitchen for instance). But this way is more efficient, if a WW is not enough.

[avc15]

5 hours ago, Carnis said:

Ive never needed this much cooling

Yeah the most reasonable question is "why do you need this much cooling"

I don't know, that's next  Maybe I'll boil some natural gas.

Maybe 2 layers of plastic prevents the damage, yeah. Would be worth a try.

[Oozinator]

so many wheezies lol

[BlueLance]

Thats a lot o wheezeworts, not everyone is that lucky XD

[avc15]

if you don't have many wheezeworts you can still make a really powerful central cooler, just have to supplement aetn/wheezeworts by also aquatuning your coolant.

You want to dump the heat into a pool of Pwater. From there, you can either boil or sieve the pwater to delete heat, and the rest of the system is the same.