Atmos Proposal - Whole Station Air Recirculation (HVAC)

[ArtisticRoomba]

An idea I had that enables the recirculation of air throughout the entire station. This is needed for regulating station temperature (something that Atmos seriously lacks) and ties into the Atmos Roadmap.

This also leads to cool emergent gameplay, and it can also lead to the implementation of future systems like humidity and comfort based on temperature.

[ArtisticRoomba]

completely totally related thing I forgot to mention in the Pros section but:

given enough money, time, and willpower, the clown has the potential to break into atmos to dump like 40 cans of nitrous oxide into the station distro and sleep everyone without internals on

personally this reason alone is enough for me (godo)

[Partmedia]

Thank you for taking time to write this proposal. I think this is a lot of fun, and as you explain this idea ties well with the objectives stated in the Atmos Roadmap. Let me get input from the engineering workgroup and the PM's to make sure we're all generally onboard with this idea.

Some technical investigation on the performance implications of this should be done and added to the roadmap if needed. My guess is that having more flow would increase the number of overall excited groups, but as long as they are excited together and distributed in a room with monstermos I suspect the performance impact will not be significant.

[ArtisticRoomba]

Thank you! I've played CE for a while and it always rubbed me the wrong way when I saw atmos just spacing gas willy nilly and refusing to use the recyclernet, primarily because I knew in the future this wouldn't viable due to the reasons outlined in the design doc.

I just realized another major problem, burn chambers involving oxygen.

As a compromise I think I'm going to keep gas miners, but have them mine at very slow rates to promote gas recycling, without completely stranding atmos with an entirely spaced and unfillable station, as well as making burn chambers possible, as this absolutely murders burn chambers.

[K-Dynamic]

Hell yeah, we'll have space radiators like those giant panels on the ISS or Tiangong stations

[Jezithyr]

I love this! It's very similar to what I had written down for a potential atmos rework but more fleshed out and developed :D Nice!

[gusxyz]

Like everyone else has said this looks amazing and I look forward to seeing this implemented in the future.

[ArtisticRoomba]

Hey, I made a change to the proposal about the gas miners section asserting my previous realization about burn chambers. In essense:

Gas miners should have their output reduced significantly to encourage the use of the station's recyclernet, but there should be a reasonable amount of oxygen flow to enable burn chambers (as I would still like to operate the TEG and a heat burn chamber for station heating).

[Ilya246]

as already mentioned, yes i think this should ideally not reuse scrubbers but add a new device specifically to steal air back, aka some kind of "reverse air vent" which steals gas above atmospheric pressure as opposed to injecting gas below atmospheric pressure

[Ilya246]

also space radiators are terrible at cooling to room temperature, and a burn chamber would require oxygen to be really cheap (i have a PR up for this) to be viable with no/slow gas miners

[ArtisticRoomba]

also space radiators are terrible at cooling to room temperature, and a burn chamber would require oxygen to be really cheap (i have a PR up for this) to be viable with no/slow gas miners

Can you elaborate more for me? I haven't had any problems with cooling gasses down to near room temperature or below.

Ideally this air mixing chamber is important enough to get its own "control room" mapped and have a massive array in space to cool down the working gas as much as possible, since we are dealing with the heat of an entire station. As someone said, think ISS size radiators.

[ArtisticRoomba]

Also on the topic of canning gasses, we could do something like a "bluespace oxygen canister" that has enough oxygen for supporting a burn chamber.

We'll have to see what the Engineering group thinks about it because they have better ideas.

[Ilya246]

Can you elaborate more for me? I haven't had any problems with cooling gasses down to near room temperature or below.

it depends on amount of gas, an entire station's worth of gas will cool extremely slowly

[ArtisticRoomba]

it depends on amount of gas, an entire station's worth of gas will cool extremely slowly

I agree. This encourages savvy Atmos techs to choose gasses with high specific heat capacities like Water Vapor or even Frezon, as well as find ways to optimize the current radiator setup they will be given (increase flow rate and find out the difference between radiators in series or parallel).

"Atmos, get off your ass! Science just made 500 items out of the hyper-convection lathe and they're toasting the entire station!!!"

[ArtisticRoomba]

Oh and another note before I forget: the air rushing sound will have to be tuned to only happen above a certain pressure differential. Or else you'll just be hearing it all shift.

It means it's working but this could be white noise torture for other spesspeople.

[Ilya246]

I agree. This encourages savvy Atmos techs to choose gasses with high specific heat capacities like Water Vapor or even Frezon, as well as find ways to optimize the current radiator setup they will be given (increase flow rate and find out the difference between radiators in series or parallel).

no i am fairly sure this will not help, the only thing that will help with station air cooling is somehow having hotter gas go through the radiators

[Ilya246]

if we had heat exchangers which just steal heat from one gas and transfer onto another, then yes, radiator cooling setups could be made even for lower temperature targets

[ArtisticRoomba]

no i am fairly sure this will not help, the only thing that will help with station air cooling is somehow having hotter gas go through the radiators

I think I understand what you're talking about now. Are you saying that it's hard for air to be cooled to exactly room temp or something else? The idea with HVAC is that it will output just hotter/colder gas then station air like an air conditioner to combat it being hot/cold, not exactly at 20C. Everything would "balance out" given time.

if we had heat exchangers which just steal heat from one gas and transfer onto another, then yes, radiator cooling setups could be made even for lower temperature targets

Heat exchangers would definitely be something cool to see but I would want a physical room that we can see and interact with, not just "this one tile machine can freeze or fry the station"

Implementing exchangers requires coding work that I'm not adept at doing lol.

Better idea: Maybe for stations it could be a 3 by 3 "station heat exchanger" multiblock that atmos would have to feed gas in/out. The machine could precisely sap heat away from gasses to control station temperature precisely. Honestly I think this is a better idea but it's a lot more work and station-wide temperature control can be achieved right now, albeit not beautifully. This is probably the biggest drawback from the cooling setup I'm proposing.

Today I'm going to set up my proposal on a large station and do some tests to see if it's viable.

[Ilya246]

I think I understand what you're talking about now. Are you saying that it's hard for air to be cooled to exactly room temp or something else? The idea with HVAC is that it will output just hotter/colder gas then station air like an air conditioner to combat it being hot/cold, not exactly at 20C. Everything would "balance out" given time.

no, due to how radiators work even using frezon won't really help you cool gas down, frezon as thermal storage gas is only really useful if you have a coolant bottleneck, but cooling down station air will cause a radiator bottleneck which can only be fixed by making more (vastly more, in this case) radiators
so yes, a radiator loop, unless absurdly large, would very likely simply be too slow to adequately cool down station air
however, i will just go and test this tomorrow

Heat exchangers would definitely be something cool to see but I would want a physical room that we can see and interact with, not just "this one tile machine can freeze or fry the station"

what i'm proposing wouldn't work as such an "one tile machine", it would simply let you shunt heat from one pipenet to another, so you could use it to shunt heat from station air into the radiator loop so that it's more viable at cooling down station air, basically a refrigerator but the hot side gets cooled by a radiator loop

Implementing exchangers requires coding work that I'm not adept at doing lol.

if i can be bothered to i could do it, i'm pretty sure tg has something similar too which can be a source for the sprite

so, what is the problem exactly? lattice radiators remove heat from the gas passing through them proportional to T^4, and this means they're absurdly slow at lower temperatures and absurdly fast at higher temperatures, so the hotter the gas passing through the radiators, the more heat they remove

[ArtisticRoomba]

lattice radiators remove heat from the gas passing through them proportional to T^4, and this means they're absurdly slow at lower temperatures and absurdly fast at higher temperatures, so the hotter the gas passing through the radiators, the more heat they remove

Ah I see now, thanks for clearing it up. I'll do some tests to see what happens. I honestly don't see it as a massive issue to have a radiator array as big as the smallest solar array. Station heat is a big issue in real life, look at the ISS.

If it's a real life problem that atmos techs are forced to solve, so be it. It's time for them to explore some creative solutions.

[ArtisticRoomba]

I ran an experiment (it's still running, actually, I'll be monitoring it further over the day and performing more scenarios).

The experiment is (on Meta station):

• A centralized heating/cooling setup was made similar to the diagram in the original design doc
• All air scrubbers were replaced with siphoning vents with an ExternalBound set to 1 ATM.
• All vents were changed to output gas at 120 kPa.
• Temperature atmospheric upsets were introduced in multiple areas, like fire anomalies, ice anomalies, to see the effects it had on the air and how the system reacted to it

The findings were:

• Performance hits didn't seem to occur, although this was only measured by my own perception. There might be a noticeable effect when polling performance in-game.
• Air vent placement is often right next to each other and as such the system largely leads back into itself in these specific problem rooms.
• Current gas manipulation machines (pumps, filters, mixers) don't work fast enough (not enough throughput) to enable constant air flow on a station-wide scale, without major limit increases. Even when allowing flow based on pressure, not pumps, it was still an issue.
• The system did significantly increase the length of time rooms could remain habitable when a major atmospheric upset (fire anomaly or hyperlathe) was introduced.
• The system was able to cool down the entire station after hot gas was purposely recirculated, however, as @Ilya246 predicted, a large radiator array is required for cooling down extremely major upsets across multiple rooms
• The gas chamber air alarms needed to be adjusted to decrease the allowable warning and danger ranges to promote low and high-strength cooling
• And MUCH, much more, there is too much to put down here. I will add my findings in the design doc, however I have enough down to write a modification to my proposal, which may satisfy concerns raised by @Partmedia as well as increase the effectiveness of the system:

The proposed modification is a hybrid system (based on improvements I made during the experiment):

• Instead of siphoning air vents, the humble air scrubber would be modified to have a PressureBound system.
• When an air alarm detects a large temperature upset, it would enable a new Recirculation mode.
• The Recirculation mode changes the behavior of air vents and scrubbers. Air vents will now output at 120+ kPa, and air scrubbers will now scrub all gasses, with an added PressureBound set to standard atmospheric pressure.
• This would allow gas to recirculate on a room-by-room basis.
• If an air alarm senses a dangerous level of heat, it will enter a new High Recirculation mode, where air vents would have their PressureBound increased to 200kPa to force the recirculation of more air.

This basically means that an atmosphere would be maintained as normal, but if high temperatures were sensed, it would start recirculating the air in that room. This air would be temperature-controlled and sent back to the room.

There's a LOT more to the hybrid system and I have a lot more findings in my experiment, so I will write them down in my design doc and modify my proposal. I think this is a good compromise for server performance, breaking changes, and respecting the current limit of atmospheric devices.

[Ilya246]

If an air alarm senses a dangerous level of heat, it will enter a new High Recirculation mode, where air vents would have their PressureBound increased to 200kPa to force the recirculation of more air.

yes i wanted to suggest this
also did some math and if i'm not wrong radiators do around -12 C/s/tile for tiles of room-temperature air

[ArtisticRoomba]

I have updated my proposal to show the findings of my experiment, and a compromise based on the results.

[ArtisticRoomba]

also did some math and if i'm not wrong radiators do around -12 C/s/tile for tiles of room-temperature air

Thanks for the work! This will be really good for helping size space radiator arrays to different sized stations. Bigger stations will have bigger cooling arrays, similar to the "1 ton per 500 sq ft" rule prevalent in real life air conditioning setups.

[ArtisticRoomba]

This has been converted to a draft before I decide the avenue Atmos would like to go in terms of gas recirculation (a compromise room-scale recirculation or whole station recirculation).

I'll rerun some tests and probably ditch the whole air alarm idea as current air alarm code is hell and I do not want to touch it.

[ArtisticRoomba]

I'm marking this PR as open for review, as right now I don't really have the time to test this proposal further, and I'd like to see if the engineering workgroup agrees upon this idea or some changes are required, before I dedicate more time to testing.

You can view my reverted commit to see the results of my previous testing, but know that it was slightly suboptimal and flawed in at least one aspect, but it still reflects the idea quite well.