Let's finally figure out the Steam Turbine Heat Deletion Issue

[mathmanican]

2 minutes ago, TripleM999 said:

I'm not that good at designing infernal exploitation devices,

I'll see if I can drop it to 5 cycles.... 

I'm pretty sure you've nailed gas mechanics.

[mathmanican]

Here's an improvement to my infernal design, with an aquatuner on the right working alone, so I can compare. 

 

The cooling rates are similar, the infernal contraption is better though. I'm down from 100C steam to 48C water in 10 cycles. I think it would be faster if the cooling plate were on the left side, not right, so that hot stuff rises instantly (when swapping happens vertically) rather than balancing out on the right side, so this could be faster.

I'm also only using a bypass that activates one time per second, rather than once every tick (0.2s). Changing this would reduce the conduction that can happen between the three tile gas swapping region.  Note that the cold plate is isolated from the tuner with liquid radiant pipes behind it. The hot plates touch the tuner room.  I added one more conveyor bridge going out left, to increase heat bleed into the region right before heating.  That may have been a mistake (or maybe should be moved up one). 

The greater the temp difference we can make, the stronger this contraption will be.  The doomsday freezer (partial melting) is still in the game, so I could keep the cool plate near -150C easily, and this contraption would decimate anything (combining two cooling bugs into one), and that point increasing the hydrogen mass will greatly improve it's power.  

I think some other improvements might be the addition of a few more bypasses to keep the regions before and after the three tile swapping region near vacuum, so we don't get backflow heat we have to deal with. Perhaps just one after the cooling section is enough. 

Here is a save, anyone wants to play along. 

SuperNovaSeveralST.sav

[mathmanican]

On 6/20/2020 at 6:36 AM, TripleM999 said:

a showcase.

So, I've been working on this quite a bit, and had lots of dead ends. Here, however, is one that I think has great promise.

It's dropped the steam temp (3500kg of steam) by about 10C per cycle. The key little device is this

If you use a bypass pump every other second, then you never get to benefit from temp swapping, as the gas and liquid swap spots too rapidly (you basically alternate between supercoolant and vacuum right above the pool of supercoolant. This is why I have a valve (right now it's set to 15g).  The valve causes the bypass to trigger every 3 or 4 ticks, so that you get one or two ticks with hydrogen gas coming in at 10% of the mass next to it.  Then periodically you get the temp swap (dropping 10 times as much gas's temperature down, 10 because the tile was replenished by the neighbor). I have two of these cycling the gas through, and dropping the steam temp by 10C a cycle. 

To improve it, I think playing with the valve settings to guarantee that above the supercoolant pools by the minipump I get an alternating pattern ever three ticks of hydrogen,supercoolant, vacuum, will cause the least conduction heat bleed from the supercoolant pools (which are both cooled by the aquatuner). 

Basically, this device greatly amplifies whatever cooling mechanism you have, and I think it can still be ramped up.  What we need to make it stronger is something that can be kept cold at almost zero mass, next to something that has lots of mass. The never-overpressuring bypass pump from our sourgas adventures gets this (I have tried several other contraptions, but this might be as good as we can get).

I do have one other idea - namely doors. We can pop a door between two gasses, and use a bypass pump to almost vacate a room of gas, and get it super cold. Then open the door and let a few mcg of gas come  in contact with 10000kg of gas. If they temp swap, then BOOM.  Cycling gas, just like above, but blocking flow with three doors (one in the middle to make vacuum, and then cool the low pressure side while the other side heats up).  This kind of device would be banking on random chance, and some times would fail. However, it would just take one interaction to drop thousands of kg of gas in temp down to -100C (as you can quickly cool off a few mcg). 

I'll play with this door idea and share my results. I think it could be way faster than what I've got above (which might be near optimal with bead/bypass pumps). I did go back and check it and temps are still dropping about 10C a cycle on the steam. 

[mathmanican]

The door idea is looking promising.  It needs some fine tuning, but I think we can make a heat destruction device. 

Spoiler

The chamber to the left of the doors needs to be a bit smaller, and I think I want to deliver cold to the chamber via a chilled pool that only connects when the doors are closed and vacuum separated. Drop the temp from 100 to -100 almost instantly, and then reopen things. The faster I can drop the temp on a few miligrams of gas, the more destructive this device will be. i haven't added automation yet, and I think powering the doors will be a good idea. 

Gotta sleep.  If you find something else fun to take advantage of the temp swapping feature, please share.  

The fact that this device is working adds lots of confirmation to me that what you found for gas mechanics is absolutely correct.  They made a simplification in the code under a condition that would not allow equalizing temps, and that simplification (albeit tiny) can be magnified.  

 

[mathmanican]

Found an even simpler approach. It's not a "steam turbine heat deletion bug".  It is a "gaseous matter heat deletion feature."  Just drop liquid form of a gas into the same room (ideally in bead form), and BOOM.  Watch 900kg of the gas drop 20 degrees almost instantly (that was at 10kg/s). The drop is still there with lower quantities, just not as pronounced.

I chose oxygen because I can make LOX prespace with ease. 

[TripleM999]

10 hours ago, mathmanican said:

The chamber to the left of the doors needs to be a bit smaller, and I think I want to deliver cold to the chamber via a chilled pool that only connects when the doors are closed and vacuum separated. Drop the temp from 100 to -100 almost instantly, and then reopen things. The faster I can drop the temp on a few miligrams of gas, the more destructive this device will be. i haven't added automation yet, and I think powering the doors will be a good idea. 

Gotta sleep.  If you find something else fun to take advantage of the temp swapping feature, please share.  

The fact that this device is working adds lots of confirmation to me that what you found for gas mechanics is absolutely correct.  They made a simplification in the code under a condition that would not allow equalizing temps, and that simplification (albeit tiny) can be magnified.  

You gave me a great idea... although it has yet enormous optimization potential, it drops temps quite fast.

The gas valve is at 10g/s. Milli or micro grams are to easily influenced by environment, like the 10g of supercoolant. The liquid vent had to go up 1 floor, the thermal mass was too heavy, and conduction to the gramwise hydrogen too much. Hydrogen is at 500kg per tile, higher will overpressure the liquid vent, and disturb the whole swapping. Aquatuner and pumps are most of the time not running.

[mathmanican]

51 minutes ago, TripleM999 said:

Hydrogen is at 500kg per tile, higher will overpressure the liquid vent, and disturb the whole swapping.

Move the vent left one so liquid has to fall over a ledge and you can crank up the pressure as high as you want. Though you will have to clean up a 10g liquid mess then. Super coolant on top of crude works down to almost negative 40c.

[TripleM999]

40 minutes ago, mathmanican said:

Move the vent left one so liquid has to fall over a ledge and you can crank up the pressure as high as you want. Though you will have to clean up a 10g liquid mess then. Super coolant on top of crude works down to almost negative 40c.

We are approaching... -40°C is not enough, i fear. Coolant loop is at -200°C, temp swap at -100 to -150°C. After 8 cycles, 100°C steam, 10 x 12 x 1000kg,   got changed to this:

Aquatuner was running once to get the loop to -200°C, but never again within the 8 cycles.

One cycle later, without the aquatuner doing anything:

I would really like to know devs mind... is it more like "Wow, there we made a mistake, this is not good" or more like "Wow, they f... up the simulation again... what is wrong with these guys?" 

[mathmanican]

28 minutes ago, TripleM999 said:

One cycle later, without the aquatuner doing anything:

Now we just need a tiny box version that is super easy to make. Then compare it to partial melting and the doomsday freezer.

I think we've got steam heat deletion mysteries solved. Do you want to write the summary post? Or should we wait till the above design is optimized and shrunk and given a snazzy name?

We still don't have my heat creation mystery solved.... I wonder if it is a liquid gas interaction when temps equalize without being swapped.

36 minutes ago, TripleM999 said:

"Wow, there we made a mistake, this is not good" or more like "Wow, they f... up the simulation again... what is wrong with these guys?" 

It evaded discovery for several years (or people discovered it and refused to post here about it). Nice work on the eye bleeding data collection.

[TripleM999]

52 minutes ago, mathmanican said:

Now we just need a tiny box version that is super easy to make. Then compare it to partial melting and the doomsday freezer.

You can reduce it to this, as this is the important part:

You could even scale it by using multiples of this, with a common hydrogen source (like the vent). And you need some auxiliary parts too, like cooling the hydrogen grams. Without the vent overpressurizing you can scale it to any heat deletion, you want. And you need super coolant... there is no other material capable of these low temperatures. Although to some degree you can reach some cooling with other materials.

I think, @Zarquan would make a viable "doomsday asteroid freezing device" out of this... without breaking a sweat. He is much better at utilizing game mechanics, especially the broken ones. 

[mathmanican]

3 minutes ago, TripleM999 said:

But you need some auxiliary parts, like cooling the hydrogen grams.

You can cool 1kg per sec of water without issues. It's not as power efficient but should do just fine. It rarely has to run.

By cycle 20 we could have it ready

[TripleM999]

25 minutes ago, mathmanican said:

I think we've got steam heat deletion mysteries solved. Do you want to write the summary post? Or should we wait till the above design is optimized and shrunk and given a snazzy name?

I am good at persistently trying to reach a goal, but not quite that good at documenting the way towards, or describing it firmly. Although the math is quite simple, when you take out other gases or vacuum. One issue is plaguing me though... even the heat equalization is not good in my opinion. It should at max equalize the heat up to equal temperatures, but as of now it fully equalize the heat, if it has a chance to, which leads to heavy pressure with quite low temperatures, as shown in this post 

The heavy tiles shouldn't be at 23.3°C, when they start at 100°C, but they should average at around 60°C or so. There is probably a not so hard formula for this, but as of now, my beer level denies further insight. 

[mathmanican]

46 minutes ago, TripleM999 said:

but not quite that good at documenting the way towards, or describing it firmly

I'll put out a post then in a day or so.  I would like to optimize this, make some clear demonstrations to show the issue, and then find a few other ways to abuse it.  If you get more ideas (while coming out of that hangover), please share.  The more crazy contraptions, the better.   

[mathturtle]

I guess now we know why the AETN takes 10g/s hydrogen   Too bad the designers didn't realize the effect gets bigger with smaller gas amounts.

You guys are nuts: you start out trying to figure out the steam turbine heat deletion and end up figuring out how to make an AETN.  Keep up the good work, I enjoy reading about your exploits.

[mathmanican]

25 minutes ago, mathturtle said:

trying to figure out the steam turbine heat deletion and end up figuring out how to make an AETN.

Generally that's what I aim for.  If there is an issue that's causing a problem, then a simple way to make sure you have identified exactly the issue is to build a monstrosity that exploits the issue.  Here's a GIF of one of these infernal contraptions in action.  Just watch the hydrogen temp move all over the place, dropping like crazy even while cooling the steam. The -120C temps are at the end of the thing. 

This is small enough for the top region to cool. 

The bottom region can be smaller. Stacking several of these side by side underneath steam should be able to freeze the 3500 tons in less than a cycle.  I'll work on making a small design. 

BONUS:  The bypass pump works great with 0.1g even when dropping from any height.  The liquid only disappears (or gets frozen) upon landing.  That is useful information that I had not known before. 

On a side note, I tried getting this to work with only a 0.1g/s bleed of hydrogen into the bottom chamber. It did not work.  There appears to be some minimum needed. It may be the interaction of the hydrogen with the supercoolant, but not sure. I'll see if I can find the break even threshold (as that could help isolate other issues). 

[mathmanican]

Less than 2 cycles for 100C steam at 1000kg to freeze... Ummm................  Kinda broken.

It costs 0.4g/s of water (neglible), and 30g/s hydrogen (never lost). I need to find a way to remove the 1000kg hydrogen limit (give me a day). Not all the parts you see above are needed.  Still working on min/maxing, and soon I'll add other versions for different purposes (like taming a geyser). 

@TripleM999, this was largely your design, so you get to name it.   I await a snazzy, derpy, title. 

[beowulf2010]

23 minutes ago, mathmanican said:

 

@TripleM999, this was largely your design, so you get to name it.   I await a snazzy, derpy, title. 

AETMMM? 

[mathmanican]

Just now, beowulf2010 said:

AETMMM? 

May need a few more M's.... 

[TripleM999]

3 hours ago, beowulf2010 said:

AETMMM? 

I like... though maybe only AETM, Anti Entropy Thermo-Maimer

[mathmanican]

I'm currently working on a piss cooler. Dump your dupes excrement into a contraption to freeze your asteroid.....

[mathmanican]

I figured out a way to eliminate the pressure requirement.  It does require a bead pump off the right side, as bead pumps off the left side maintain the liquid bead on top.  I might be able to get it off the left, but this works for now.

Both airflow tiles are needed for max efficiency, and that is where all the gas stays.  You can use just the upper one, but efficiency greatly drops. You want the crude to stay cool, hence the metal tile. The bottom area is not optimized at all. 

And here is the piss powered cooler: Note that there is no need for a gas vent in this, though you want to keep your piss level low, otherwise you get too much gas below and then the gas leaks your heat. cool.  

Update: You need one extra tile higher in this one, so that the PW can offgas upwards.  All other gasses work better than PW, as then you don't have to deal with the top level of gas which  takes much longer to chill. 

[mathmanican]

A configuration like this one will let us use a single liquid off the left, and never overpressurize.  It will need some tweaking to get it started (similar to the sour gas boilers), and probably another row of gas tiles across the top, but it will work. 

The key is the insulation tile between the two airflow tiles.  We now are using air lift pumps (thanks @socooo). The extra insulation tile is  needed to stop the waterfall and start the beading.  It does a very  nice job of dropping the temp on 9999kg of gas from 100C down to 0C in a few seconds.... 

The level of abuse is taking on new levels. I hope the devs laugh tomorrow when they come in to work. We love the game guys.  I hope you take as much joy in reading these posts as I take in making them. Now please fix that "temp swamp" issue, and everyone will be extremely happy they don't lose a bit of heat from their turbines (and never realize the extreme abuse that mechanic led to).

[mathmanican]

So, we call it the AETM. I think I can put together a post (hopefully tomorrow, or the next day) that showcases the temp swapping issue. Along with it, we put up several devices for maiming thermal energy. Then we hope for a fix sooner rather than later.  

What shall we do next?    @Manarz, you've been quiet.  Should we tackle the extra heat generation I keep getting from the turbines, and see if we can melt magma with it.  I doubt we can, but you never know.  I didn't think finding the steam turbine heat deletion bug would lead to the AETM...  It's been fun.   Another option might be mass deletion in steam turbine setups (document when it occurs, find out why, and see how to eliminate it - no real benefit from abusing it). Any other ideas? @wachunga, do you have any lingering things you would love to work on? I guess we could try to crack why right flowing stair heat exchangers are bugged... @nakomaru, we'd love to include you in the upcoming Nature paper.  

There are some other bugs sitting around that are known, that at some point need some additional showcasing.  @Blazing Falken left right around when he cracked the debri bug where small piles of debris maintain the mass of the original stack (I think this is still around). This was the key to fridge cooler builds, and the crap he got from it may have led to him leaving (or live got busy). Another bug is the temp reset of newly created debris that affects lava builds (SFT if I remember right).

[Manarz]

5 hours ago, mathmanican said:

There are some other bugs sitting around that are known, that at some point need some additional showcasing.  @Blazing Falken left right around when he cracked the debri bug where small piles of debris maintain the mass of the original stack (I think this is still around). This was the key to fridge cooler builds, and the crap he got from it may have led to him leaving (or live got busy). Another bug is the temp reset of newly created debris that affects lava builds (SFT if I remember right).

I got a cooling tower that works on that principle which cools literal tons of 70°C Hydrogen down to LH without barely any energy and also a small contraption that produces infinite energy. I'd love to share my findings if u want to investigate it further.
Right now i try to use this stacking bug together with flaking and gas swapping in order to be able to call it the "TripplePhaseCooler".
Im not quite there yet. Didn't have too much time but it looks promising. I had some other ideas to force a high gas gradient. Hope i can show it off soon!

[mathmanican]

5 hours ago, Manarz said:

I'd love to share my findings if u want to investigate it further.

Absolutely