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Please be aware that the content of this thread may be outdated and no longer applicable.

mathmanican

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

Recommended Posts

TripleM999    235

Just as an addition for the heat deletion bug, my interpretation of a sewage based cooler.

"Fresh Breeze™", based on usual grid, build from only the cheapest, early game materials.

FreshBreeze.png.420be87091826ad5f3e465da41a6a491.png

Performance isn't that overwhelming, but ok. Thermo regulator will overheat, when PO pressure is below 20kg/tile. From 100kg/tile onwards, it will actually cool something. You can augment the PO-production with slime, or even morbs. I used the thermo regulator, cause it easier to control temperatures, cause an aquatuner will easily overfreeze PW, although the aquatuner would provide much higher efficiency.

Share this post


Link to post
Share on other sites
mathmanican    3256
22 minutes ago, DefectiveNu said:

I had a bug report about it.

Thanks for sharing.  Both setups you describe involved multiple gasses, which is definitely known to delete mass, and has done so for ages (hence the SPOM mass deletion paranoia). I would really like to decipher why single gas (only steam) builds delete mass. 

@Manarz, the design you made for flaking liquid could have avoided deletion with a row of tiles separating the supercoolant from the steam, still achieving the same result. But partial evaporation has been hopefully fixed, so no need now. Your design hopefully no longer works...

Share this post


Link to post
Share on other sites
Denisetwin    242

So does the latest update that just hit 30 minutes ago "fix" this?

<<Fix solid flaking temperature deletion. Uses the actual energy required for the melt, as well as the SHC of the transition element.>>

Share this post


Link to post
Share on other sites
mathmanican    3256
1 minute ago, TripleM999 said:

"Fresh Breeze™"

I think I need to wait a day or two before I put out a summary post.  These are getting better and better. :) 

1 minute ago, Denisetwin said:

So does the latest update that just hit 30 minutes ago "fix" this?

<<Fix solid flaking temperature deletion. Uses the actual energy required for the melt, as well as the SHC of the transition element.>>

That will hopefully fix a different bug.  I don't think it is related to this one at all.

Share this post


Link to post
Share on other sites
TripleM999    235
5 minutes ago, mathmanican said:

I think I need to wait a day or two before I put out a summary post.  These are getting better and better. :) 

Nah... I was only testing, how liable an early game cooler of this type is. I think, it isn't, the pressure buildup just takes too long to make it usable, albeit it works as expected with all this cheap material, when pressure is high enough. It could work though as an early cool oxygen provider, with some modifications to clean the oxygen.

The borg cube on the other side now sits stable at below -240°C, anything lower on the settings, and there will be liquid hydrogen within the cooler, which is not desired. I think it would be able to destroy the heat within the magma biome within some cycles.

Share this post


Link to post
Share on other sites
mathmanican    3256
17 minutes ago, TripleM999 said:

Performance isn't that overwhelming, but ok.

My plan was to scrap the regulator completely, and just build a few PW ice tempshift plates. As for getting enough polluted oxygen, a dedicated dupe can mop 140kg/s quite easily of PW, and then that off-gasses quite quickly. Reduce the number of tiles it has to congregate in (the original piss cooler), and you can get high pressure quite quickly.  The challenge is then to keep the flow trotted to less than 140kg/s for mopping. :) Escher waterfalls are quite fun for moving liquid around without power. Or just pump the stuff into liquid storage tanks, deconstruct, a few, and wait.  

I think for early game cooling, using hydrogen and crude might be simpler (though some people don't consider crude early game...) The nice part is we have lots of options. 

Share this post


Link to post
Share on other sites
Manarz    36
3 hours ago, mathmanican said:

Here is my save.  Could it be a Linux/Windows thing? Want to share your save, and I'll check it out as well?

SuperNovaSeveralST.sav 299.23 kB · 1 download

After i posted i actually found the problem. I didnt use insulation between turbines and steamchamber thus the bleed was quite massive.
Right after i changed itall made sense. Trying to fix this oversight let me to some interesting new revelations though.


Here is my current best turbine:
grafik.thumb.png.969ab41654e7996552ff1ac9660c45d2.png

Runs stable at this power and only needs 2 tiles height!

 

@mathmanican

Here is the save.
TurbineIteration.sav
I am on Windows btw.

I'll look into yours now too.

 

Some observations and rules of thumb i found while iterating over diffrent materials and designs:

- usually the higher the temperature of the AT gets the more power output i get (keep in mind i am always producing the same amount of heat)
- slower heat transfer towards the right performed better (also steel AT is better than thermium because you can get higher gradients)
- lower steam mass leads to higher gradients which again perform better


TDLR steeper horizontal temerature gradients perform better.
 

TurbineIteration.sav

Share this post


Link to post
Share on other sites
mathmanican    3256
7 minutes ago, TripleM999 said:

I think it would be able to destroy the heat within the magma biome within some cycles.

A 30 second YouTube video:  Temperature Swa(m)p Borg Doomsday Device takes on Magma Core.  This would be great advertising. :lol:  Not sure it's needed though (I hope the devs see these shenanigans). 

My current project is to use an aquatuner to display the liquid temp issue affecting right flowing heat exhangers.  I'm hoping to build a Thermal Amplifier, that can either accelerate heat creation, or heat deletion, based off the bug. If it succeeds, then we get even more tools in our temperature regulation annihilation tool kit. 

3 minutes ago, Manarz said:

Runs stable at this power and only needs 2 tiles height!

You can get higher. 

  1. The pipes in the steam room are coming in contact with the cooling medium.  As such, you loose heat from the water to the hydrogen. It's really small, but there.
  2. I see in your build that your steam level is not enough to keep 2000g/s in the pipes.  As such, part of the gain you see in watts comes from not loosing 1.5C to phase change. I have a turbine in another save that gets above 568W continuously. I'm pretty sure your will be that high if you just swap from hydrogen to supercoolant, and bridge over the super coolant. 

So now the question is, "Why?"  What's causing this increase.  It isn't temp swapping (at least according to the working theory that only results in heat deletion from temp swapping).  It also requires liquid in the equation. So there must be a "liquid-gas" interaction that is slightly bugged.  If we can figure out exactly how, then it's just a hop skip and jump to infinite heat. 

Share this post


Link to post
Share on other sites
TripleM999    235
9 hours ago, mathmanican said:

A 30 second YouTube video:  Temperature Swa(m)p Borg Doomsday Device takes on Magma Core.  This would be great advertising. :lol:  Not sure it's needed though (I hope the devs see these shenanigans). 

30 sec isn't probably enough, but nonetheless i have thrown "Borgie" against the magma biome. I'm not even sure, this shows the magnitude of problem. Starting conditions were 10000kg hydrogen per tile at 0°C, 950kg supercoolant per tile at 0°C, slight modifications to protect vital parts, and to conduct more heat VS. magma biome at 1840kg, 1640°C per tile.

1. Round:

Spoiler

MagmaShowcase1Cycle.png.20e879a233ee828423097fe60febe97c.png

I was afraid, the plastic pump in the contact chamber would not hold, as it will melt at 159.9°C. But no sweat, hydrogen temps are now at -10 to -15°C, super coolant temp, and with it hydrogen injector temps are at around -35°C.

2. Round:

Spoiler

MagmaShowcase2Cycle.png.e2f79a28397bfe00b3dbca22134dafa8.png

The biggest obstacle to the "death" of the magma biome is obviously its own conductivity, obsidian is ok, ignious rock and magma itself are not so. Hydrogen chamber sits at a comfortable -62°C average, supercoolant is at -82°C, hydrogen injector spikes to -118°C, cause of the additional cooling for the hydrogen.

3.Round:

Spoiler

Nothing worthy to mention happened. Overall atmosphere is frosty. Hydrogen chamber is down to -106°C, aquatuner is still running, near magma is down to 1462°C.

4. Round:

Spoiler

This fight will probably take some time. Magma can't show anything of his former glory... his punches are weak, and without goal. Borgie takes it easy... denying magma of his valuable heat, wherever it can get a hold of it. It tries to stay cool, at now -150°C. You can almost feel the first hit coming.

5. Round:

Spoiler

Lightning strike, Borgie landed some hits, Magma is pushed further into defense.

MagmaShowcase5Cycle.png.bf9e4ec2d77f7310d981556ad128cd64.png

Borgie is chill at -185°C, while Magma has lost quite some of his temperament with only 1420°C in near vicinity.

...

10. Round:

Spoiler

While reaching final operating temperatures at Cycle 8, Borgie assimilated more and more space around it.

MagmaShowcase10Cycle.png.29fce43081df77badcc200b53e7c7397.png

Magma is giving a fight, but finally... resistance is futile.

 

Share this post


Link to post
Share on other sites
mathmanican    3256
5 hours ago, TripleM999 said:

"Borgie"

Magma biome: we will melt you.

Borgie: You... Will... Be... Assimilated... Resistance... Is... Futile...

Spoiler

assimilated.gif.9400dfd8b19160963e6c6e62

 

Share this post


Link to post
Share on other sites
wachunga    374
3 hours ago, mathmanican said:

@wachunga, you mentioned in the waterfall boiler that you saw backflow conduction

You got it in your earlier post, there is conduction but the liquid cells fall so conduction can never travel up the waterfall. A caveat to this is hot liquid below cold liquid which causes a random averaging in addition to regular conduction. So hot can travel up a waterfall but not cold. Of course in a petroleum boiler you typically don't have hot below cold so it isn't particularly relevant there.

image.thumb.png.aafcb447ee501394f3bdc1c8e8ce0ba8.png

Top is starting condition (all are the same 20@400 over 10@300), bottom is after one tick. Conduction is clamped by what would cause a 1/4 of the temperature delta change. So the 10kg liquid raises to 325K with the 20kg dropping to 387.5K. Then fluid flow occurs. Everything is calculated correctly by my math with the understanding that rightwards flow is with the mass that is leftover after a leftwards flow (the rightmost example, 2.5kg goes left and 25% of the remaining 7.5kg goes right and the 20kg falls to combine with the 5.625kg).

Share this post


Link to post
Share on other sites
mathmanican    3256
4 minutes ago, wachunga said:

A caveat to this is hot liquid below cold liquid which causes a random averaging in addition to regular conduction.

This sounds like the old drip cooling bug. They probably used the same "temp swap" algorithm that appears in horizontal gas swapping. If that is the issue, then hopefully a similar fix will prevent Borgie from taking over the asteroid. 

6 minutes ago, wachunga said:

Everything is calculated correctly by my math with the understanding that rightwards flow is with the mass that is leftover after a leftwards flow (the rightmost example, 2.5kg goes left and 25% of the remaining 7.5kg goes right).

Same as what I've seen. Time to build a device that abuses this. 

11 minutes ago, wachunga said:

but the liquid cells fall so conduction can never travel up the waterfall.

If the liquid level exceeds viscosity (water = 125, supercoolant = 150) and there is liquid below to start with, then the backflow should be visible, and we can track if it moves left and.or right.  Any mass above the viscosity level (provided there is liquid below) will follow regular rules for single layer liquid flow (dump viscosity below, and then spread), if I recall correctly. 

This isn't really practical in a petro/crude exchange, but useful in tracking down bugs. 

Share this post


Link to post
Share on other sites
wachunga    374
23 minutes ago, mathmanican said:

If the liquid level exceeds viscosity (water = 125, supercoolant = 150) and there is liquid below to start with, then the backflow should be visible, and we can track if it moves left and.or right

That is indeed what appears to happen.

Share this post


Link to post
Share on other sites
mathmanican    3256
34 minutes ago, wachunga said:

Conduction is clamped by what would cause a 1/4 of the temperature delta change.

This little bit will help greatly with testing. Thanks for the reminder. 

Share this post


Link to post
Share on other sites
wachunga    374
16 minutes ago, mathmanican said:

This little bit will help greatly with testing.

Also note there is other clamping that sometimes shows up as 1/8. The wiki page on conductivity goes into detail on several of these edge cases. Props to whoever the editor was. 

Share this post


Link to post
Share on other sites
TripleM999    235

And heat transfer actually needs to be able to transfer any heat.

I'm not sure, this case was even intended, but when absolute heat difference between two tiles is zero, they are not able to transfer heat, even when at quite different temperatures.

Share this post


Link to post
Share on other sites
DefectiveNu    13
2 hours ago, mathmanican said:

Thanks for sharing.  Both setups you describe involved multiple gasses, which is definitely known to delete mass, and has done so for ages (hence the SPOM mass deletion paranoia). I would really like to decipher why single gas (only steam) builds delete mass.

The reason I linked that particular one was also because, it seemed to work fine in one version but then ate the steam in the next.  I wonder if testing that particular version's changes is warranted

My theory at the time was that I dripped the water back onto a large crude/petrol lake but I was probably wrong.

Share this post


Link to post
Share on other sites
mathmanican    3256

While Borgie keeps on widdling away at the magma biome, I'm cooking and cooling something else. 

image.png.113f4d1422194f230ecb7807ced1407e.png

The only difference is the location of the aquatuner.  And the results...  The left picture generates heat, and the right picture deletes heat. The temp below the fall is what gets magnified. This design confirms what I suspect is the issue (which is just what @wachunga noted already). The more stairs, the faster the magnification. The pump adds 2 kDTU/s, so the cooling build is a tad slower at magnifying things than the heating build 

All materials started at 100C (painted 9999kg of hydrogen to set everything, including machines and buildings). Petro pools have same mass. 

  • Left side above is the heating one (aquatuner below). Temp in cold petro room is around 102C. In hot room it is132.8. Everything is getting hotter.
  • Right side above is the cooling one (aquatuner above). Temp in hot petro room is around 113C. In cold room it is 60C. Everything is getting colder. 

This effect is not really fast yet, but it's real.  And it all has to do with the mass being pushed right is not being correctly computed (it uses the old mass of the cell, not the new mass).  A simple fix in the code (if we're right), could fix this. @Ipsquiggle, this example may be enough to track it down, and the few comments from @wachunga might point you guys directly to the issue.  I think it really is just grabbing the right number from the wrong array. I'd love to have left and right heat flow both preserve heat computations (currently left does, but right doesn't). 

Here is a save: FlowTestExample.sav

I briefly searched the bug reports, but didn't find a good one.  So I made a bug report. 

Spoiler

 

@Zarquan, I think you would have fun with this one, as escher waterfalls can greatly increase the flow we get past such a contraption. We could crank up the heating/cooling potential quite a bit at 150kg/s. 

Share this post


Link to post
Share on other sites
TripleM999    235

I have found a use for Borgie... as cooling element for my 240kg-Bolier. :twisted:

240kgReactorFreezer.thumb.png.700f11652828f1b5e234aeb5e29a953b.png

Sour gas temp is at 156.6°C at the boiler exit, so 321.1K temperature difference for condensation. 240kg/s sour gas is a whopping 146267.472 kDTU/s heat removed. That's equivalent to 124 aquatuners with supercoolant, running all but one at full time, together with 186 steam turbines. If my math served me well.

Share this post


Link to post
Share on other sites
mathmanican    3256
1 hour ago, TripleM999 said:

I have found a use for Borgie

Hahahahahaha!!!!!

Heating that back up is now gonna be the main prolem, or maybe you've given our friend the liquid tepidizer another reason to stick around. I wonder how many liquid tepidizers it would take to beat Borgie!  Can he take on an army of 50 liquid tepidizers running full time in supercoolant?  Mwhahahaha!

Spoiler

I bet Borgie will still win.  He becomes more powerful the harder you try to resist.   Resistance is worse than futile. 

On another note, you said the following a few days ago.   

On 6/20/2020 at 2:20 PM, TripleM999 said:

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

I believe you have proven yourself very wrong. Very wrong, indeed.

Share this post


Link to post
Share on other sites
TripleM999    235
9 hours ago, mathmanican said:

Heating that back up is now gonna be the main prolem, or maybe you've given our friend the liquid tepidizer another reason to stick around.

That is the reason for the metal tiles between the sour gas and methane, to get methane above boiling point.

9 hours ago, mathmanican said:

I wonder how many liquid tepidizers it would take to beat Borgie!  Can he take on an army of 50 liquid tepidizers running full time in supercoolant?  Mwhahahaha!

I think, Borgie could. The biggest problem now is conductivity. For the NG-Reactor design i estimate from observation, the limit is around 10000kDTU/s per tile of height.

The higher the temperature delta between the coolant sink and the hydrogen chamber, and the higher the pressure in the chamber, the more heat is destroyed. So the more heat is thrown at Borgie, the more heat is destroyed. And when one replaces the minipump with a bigger model made of steel or thermium, the hydrogen chamber can tolerate far higher temperatures, for a much bigger delta.

Crude approximation is, with -240°C coolant sink temperature, 60°C limit for the chamber with the minipump, and 1/10 chance for interaction happening, you can destroy up to 300K x 2.4(kDTU/kg)/K x 1/10 = 72 kDTU/kg , this is per kg of pressure, although the lower the pressure, the higher the temperature fluctuation.

Share this post


Link to post
Share on other sites
TripleM999    235

The "Tepidizer Freezer™", with 50 liquid tepidizers, that is over 200000kDTU... per second:

TepidizerFreezer.thumb.png.c83c20a4f3ca264753dc9205f1238d35.png

With 20000kg hydrogen per tile at start, it settles at 21°C tepidizer/coolant temperature and -67°C hydrogen chamber temperature. Aquatuner is running around 50% of time, cause super coolant sink would not drop below -225°C.

Now trying with 50000kg hydrogen per tile... temperatures in hydrogen chamber are already dropping fast.

After some cycles, temperatures are now at -40°C for the tepidizers, -140°C for the hydrogen chamber, i think it will settle at around the same temperature delta, as before, but were final temperaturs will be, i do not know.

Share this post


Link to post
Share on other sites
TripleM999    235

I have no idea, how to drive the heat density higher. The TF is as high, as we probably can reach. Magma biome defeated... 240kg NG reactor optimized... liquid tepidizers frozen... what deeds else need to be done?

Share this post


Link to post
Share on other sites
mathmanican    3256
1 hour ago, TripleM999 said:

what deeds else need to be done?

An update from the devs that removes this mechanic from the game?

I'd love to have someone come join me in exploring the new flaking mechanic.  I've just about got a spreadsheet built for solid flaking (partial melting) that accounts for conduction as well flaking. The only thing left is to identify if +3K is actually the right phase change temp.  I actually thing it might be closer to 2.97 or so. Gonna test this, then post the sheet. 

Share this post


Link to post
Share on other sites
TheMule    219
13 hours ago, TripleM999 said:

I have no idea, how to drive the heat density higher.

Sorry I'm not following 100% here but is the problem located at the temp shiftplates and the metal tiles to the left of the hydrogen? What's the temp there?

Share this post


Link to post
Share on other sites