Massive heat deletion/creation bug from liquid boiling -- Part 2: 20 kW from one aquatuner.

[mathmanican]

2 minutes ago, Zarquan said:

I just had a thought.  What if you put all of the liquid vents on the far side of the steam rooms?  Then we could have a minor counterflow effect.  It might bump up the delta a few degrees for pretty much free.

That could work. We could also used staged cooling chambers with escher waterfalls to control precisely the rate of cooling.  Mostly, the design above was a "for science" exploration of what is possible.  For science, we might as well see if we can power 30 turbines, with a >150 degree delta on each one run.  @TripleM999, This sounds like something you might enjoy.  Squeeze all the energy out of flaking phosphorous as possible.   I would summon @Saturnus into this discussion, but he's not really playing ONI right now. 

[mathmanican]

I headed over to reddit to see what people are doing, and found a crude->NG boiler build (very much OSHA like, but prespace materials only so they used thermo regulators).  I want to link it to this thread, as you can see in their initial picture a very clear indication that flaking has hit their build. 

Spoiler

You can spot the places where a single bead split in two, and places where falling petroleum has left bead form.  These are clear indicators that flaking has hit a build. In a build like this, the low temp petro that fell to the bottom of the heating tower will suck out the heat from the magma much faster than you want.  

[Zarquan]

A major limitation on the existing boiler designs is that they are limited to 25 kg/s within a single resource loop.  That is 5 kg boiled per tick and 5 ticks per second.

To this end, I designed two types of parallel boilers that are capable of boiling more than 25 kg/s in a single loop. 

In this one, the heat comes in through the natural diamond tiles.  It heats the refined carbon which boils the sulfur.  The sulfur then moves up to the upper chamber where if can be used in steam turbines and fed back in to the boiler.   I think I could make it one shorter per shelf, but I only wanted to worry about one variable at a time.  This boiler boils 100 kg/s

Pros:

1. Pure vertical stack.  There are no steps, so this is easy to place and scale.
2. 100 kg/s
3. Immune to pressure damage as long as there is enough heat

Cons:

1. A lot of heat is wasted on heating the gaseous sulfur.  It essentially reached the temperature of the diamond tiles.  Ideally, the sulfur should leave at  exactly 340 C, but instead it leaves at 420 C.
2. It is a bit bulky vertically.  But I don't want to deal with 2 qualified donor cells right now.
3. There is appears to be a lot of heat bleed.  The sulfur liquid settles at about 160-170 C, whereas we would prefer it to stick to 150 C.

The second design attempts to solve the heat bleed problems, as my theory was that the sulfur hanging around was the cause.  To this end, I created a chained bypass boiler.

This boiler is again scaled to 100 kg/s. 

Pros:

1. It uses cool ONI physics.
2. It has minimal heat bleed
3. It can take infinite pressures
4. 100 kg/s.

Cons: 

1. I'm not sure how the heat would reach the boiler plates from the outside, as there is a solid wall of petroleum in the way.
2. Not vertically stacked.  Though this can be scaled, it is quite annoying to do so, as it grows at an angle.

On a side note, I think my ice maker magma boiler is a no-go.  There is too much heat loss in the magma boiler for my ice makers to even keep it preheated.

[TripleM999]

3 hours ago, Zarquan said:

In this one, the heat comes in through the natural diamond tiles.  It heats the refined carbon which boils the sulfur.  The sulfur then moves up to the upper chamber where if can be used in steam turbines and fed back in to the boiler.   I think I could make it one shorter per shelf, but I only wanted to worry about one variable at a time.  This boiler boils 100 kg/s

What about reducing the size with the help of some bridges?

 

It seems to me, I only reach 70kg/s at the moment though. Advantage is, sulfur leaves at 385°C.

Now it works at 100kg/s, and astonishing 352°C sulfur exit temperatures.

[Zarquan]

58 minutes ago, TripleM999 said:

What about reducing the size with the help of some bridges?

 

It seems to me, I only reach 70kg/s at the moment though. Advantage is, sulfur leaves at 385°C.

Now it works at 100kg/s, and astonishing 352°C sulfur exit temperatures.

I should have thought of using bridges for heat transfer.  Though I am a little concerned about what happens with the 2 donor cells.  I should un an experiment.

[TripleM999]

7 hours ago, Zarquan said:

I should have thought of using bridges for heat transfer.  Though I am a little concerned about what happens with the 2 donor cells.  I should un an experiment.

In the picture i use both, temp shift plates and bridges, wire and conveyor. It settles for one donor cell... in the beginning i had a 5th cell at the top... didn't worked pretty well... i think, the algorithm had to recalculate each time, which donor to use.

Wow... crazy... got an epiphany. 

If you take this formula:

Donor Cell's New Temperature = DonorOldTemp + ((5 * FlashedSHC * FlashedDelta) / (DonorSHC * DonorMass)) - 10

then set Donor Cell's New Temperature = DonorOldTemp for no temp change at all

you get 0 = ((5 * FlashedSHC * FlashedDelta) / (DonorSHC * DonorMass)) - 10

FlashedSHC, DonorSHC are fix, FlashedDelta is fix too... at equilibrium.

So you get DonorMass = (5 * FlashedSHC * FlashedDelta) / (10 * DonorSHC )

or DonorMass = (FlashedSHC * FlashedDelta) / (2 * DonorSHC )

In this Pic, phosporus base temp is 146°C, flashed temp is 283.5°C (283.45° to be precise), so flashed delta is 137.5... so mass of naphta has to be 24.16kg:

Naptha temperature isn't changing at all, it only follows steam temperature. One has to plan and test for final equilibrium, (or just calculate), but when it is reached, you get infinite energy from nothing. Maybe this is the REAL game goal, to rebuild/recreate the "Temporal Bow".

Edit: This works to some extend too for the parallel sulfur boiler, but there is more heat exchange involved, so correct mass is harder to predict.

This is not yet final equilibrium.

[mathmanican]

4 hours ago, TripleM999 said:

It settles for one donor cell..

FYI. You can flake solids with multiple donors. Haven't tried with liquids but I'll report soon.

4 hours ago, TripleM999 said:

then set Donor Cell's New Temperature = DonorOldTemp for no temp change at all

Bingo. This is key. 

[TripleM999]

I present you... THE INFINITY ENGINE (Modification of @mathmanicans original idea)

Aquatuner is only running around 2 times a cycle to cool steam turbines, not for providing any surplus heat... start temperature was 400°C.

Phosphorus on far left side is 133°C, on flaking point it is 135°C. The heat, it got, is from child. child flakes to 283.45°C, leaves to the top at 283.3°C. The machine implements @Zarquans idea of steam counterflow. Heat exchange is done via thermium metal tiles and bridges. The door on the right turbine was implemented for steam over 200°C, but it never goes that hot. The 27kg naphta could be delivered there as plastic.

Edit: now with only prespace materials too:

Ceramic insulation, diamond windows and tempshifts, wire bridges and radiant pipes are copper. Conveyor bridge there is steel, there are not many options, as the other materials beside steel, thermium and niobium are metal ores with quite low TC. Steel aquatuner is a little bit tricky, as you have not much overhead temperatur left, steambox should be at min 300°C, so only 25K to play with. And to start, it could use a little help, either by preheating the phosphorus, or by other means to heat the naphta, again at 27kg. It is a little bit much, but 26kg is too little again, phosphoros temps would have to be 135.5°C, but will stabilize at around 133.6 to 134.9°C. 26kg could work too, when flake parent temp rises above 135.5°C.

[mathmanican]

Parallel boilers, and liquid boilers, don't require the heat source to be below.  This works just as well. This has one issue, namely that the gas on the lower level can only shoot about 1% upwards each tick. So it bunches up to several tons of gas.  However, there is no heat bleed from the heat source to the incoming liquid (just as in the infinity boiler above) aside from the conduction that 5kg of newly created gas can send to the incoming liquid, and that is essentially constant and unavoidable. Conduction is at an absolute minimum.

9 minutes ago, TripleM999 said:

THE INFINITY ENGINE

I suspect that it's not truly "infinite", but rather you have obtained a balance that results in the aquatuner turning on to cool the turbines more often than is needed to continue flaking the incoming liquid.  If you disconnect the aquatuner cooling from the flaking, you'll be able to measure exactly how much energy is used in flaking. Of course, the closer we get to DonorOldTemp=DonorNewTemp, the less heat required. 

Not sure when I'll have time, but I bet the Naptha donor cell can actually flake from both sides. An aqua tuner below injecting heat could probably enable flaking from both sides. 

[TripleM999]

36 minutes ago, mathmanican said:

I suspect that it's not truly "infinite", but rather you have obtained a balance that results in the aquatuner turning on to cool the turbines more often than is needed to continue flaking the incoming liquid.  If you disconnect the aquatuner cooling from the flaking, you'll be able to measure exactly how much energy is used in flaking. Of course, the closer we get to DonorOldTemp=DonorNewTemp, the less heat required. 

This is probably correct, in debug it is easy to paint exact mass, but i think, 27kg are good enough for everyday usage. 26kg was slightly to low to handle the delta, so the boiler got hiccups from time to time. Another possibility is to optimize the length for a certain mass. But i'm quite happy with the result. 

36 minutes ago, mathmanican said:

Not sure when I'll have time, but I bet the Naptha donor cell can actually flake from both sides. An aqua tuner below injecting heat could probably enable flaking from both sides. 

With a diamond or metal tile and some bridges this is for sure no prob. The steam tunnel has the advantage of more buffer, which can be used for a staggered startup for as many modules, as one wants, and centralizing the cooling.

[mathmanican]

@TripleM999, The temp in your Aquatuner room kept rising (did it level off?)  Every time the aquatuner room temp rises, you end up shooting a bit more heat into the newly created gas (the 24kg of Naptha temp rises, and so you get conduction from donor Naptha to 5kg of child gas). Increase the aquatuner room temp, and this heat bleed increases.  As such, you will eventually reach an equilibrium where the heat bleed into the new gas accounts for the excess heat produced to cool the aquatuner (might not happen till after your aquatuner overheats).  The entire process converges to equilibrium quite nicely, rather than diverging, providing the tuner doesn't overheat itself. Will it stay under steel aquantuner temp ranges (not sure).  There are lots of fun converge/diverge/steady state/equilibrium problems in these setups. The door in the room is one thing I added to help control this. You could add a safety measure into your tuner room to bleed extra heat into the new gas (when needed), to prevent the tuner from overheating, and then it's self contained.  Then the infinity turbine becomes an exercise in trying to get as much power as you can, or in gettting as much power per tile, or some other measure of efficiency. If you had a parallel boiler, you could probably keep 4 turbines at 100% wattage, thus maximing the power per tile issue).  You can always bleed out more heat, down to 100C (split turbines do work), but it costs lots of space.

I've been thinking of using staging pools connected by escher waterfalls with chlorine and carbon dioxide (basically vacuum separation), for even more precise control, especially with much higher temps. However, before doing all this, I really have to figure out a way to completely mitigate (or at least predict) the steam turbine heat deletion bug (and mass deletion), otherwise these problems are left to apparently random/chaotic behavoir.  I want to wisk away several hundred degrees from a heat source (maybe over 1000C), and not have to deal with loosing a large or unpredicatable amount of heat to some as-yet-unpredictable mechanic. If I can figure out to increase the amount of heat, even better.  

I would love to build a contraption that takes 100kg of 125C steam, and instantly with almost no power (so scrap aquatuners, which do this already) turns that steam into 75kg of 100C steam, and 25kg of 200C steam. So basically, want a zero power instant AT. Soon.... Maybe.  

[wachunga]

2 hours ago, mathmanican said:

I really have to figure out a way to completely mitigate (or at least predict) the steam turbine heat deletion bug

Not sure how relevant this would be for you, but one idea is to superheat the returning water via the small packet mechanic so that steam temp differences are as small as possible. If all the steam below the turbine is the same temp than no heat deletion/creation occurs AFAICT.

The pipes aren't run through the steam chamber because that would create a temp difference in the steam. This is running at steady state, but I suppose a large thermal mass at the radiant pipe section would perhaps slow heating/cooling enough to minimize the deletion for your purposes.

 

This is as bang on theoretical power production (and thus minimal heat shenanigans) as I've seen. Steam leaving the giant thermium block heat source is 156.7C, returning it is 123.6C. That's 276.6498 kW consumed by the turbine system. 12.537 (1.5C worth) of that is deleted when the circulating water boils, leaving 264.1128 kW. The turbine is self cooling so we add a fixed 4kW of heat and 10% of the heat consumed by the turbine (and 10% of that and another 10% of that and ...), (264.1128+4) * 10/9 = 297.9 kW. Which theoretically gets converted to ~288.5 W of power (power = heat / 4.179 / 2 / 105 * .85). In the save, the turbine is making ~288 kW of power.

TurbineMinimalDeletion.sav

[mathmanican]

11 minutes ago, wachunga said:

all the steam below the turbine is the same temp than no heat deletion/creation occurs AFAICT.

This is my plan. Deliver heat in a 1 column wide verical spot with no heat anywhere else. All steam will be the same temp. I want to compare if the column location (left right middle) makes a difference. Then try one tile tall horizontal columns at various heights. I also plan to build tick timers to see how random the gas mechanics are. Might be a while or I might get lucky and stumble on the answer.

[Zarquan]

I made a new Borg Cube style cooler. 

It will cool the outer metal tiles towards about 2-3 C, but it can also take heat input directly in to the aquatuner room and delete it.  This heat source is simulated by 1000000000 kg diamond at 523 C. 

In this case, since we are attempting to delete heat rather than create it, we want to use a large donor tile.  To this end, I created a 20,000 kg refined carbon tile for the boiler plate.  This can be done (slowly over time) by making a storage compactor, storing coal in it, then heating the coal to around 280 C. 

No space materials were required to make this build.

It can cool the metal tiles about as fast as a normal aquatuner could, but it also deletes the heat from the aquatuner.  Additionally, it can delete the heat from the diamond tiles almost for free.

[mathmanican]

1 hour ago, Zarquan said:

It can cool the metal tiles about as fast as a normal aquatuner could

Fun!  I thought you were using a solid parent, but nope.  The water is the parent, which goes to steam and then drops back to water because you keep the water cooled with your PW loop.  You then disconnect the water from the donor (huge refined carbon tile), and then it instantly drops back to water.  Just open/close doors, and you drop the temp on the refined carbon tile by 10 degrees each time..   Ingenious. I love it.  I think we have just scratched the surface on flaking abuse. 

@Zarquan, have you tried using other liquids than water as your parent?  I'm guessing you could get the temp much lower if you used something like crude/petro, or maybe even liquid oxygen (provided you had supercoolant for your loop).  I'm guessing the temp things settle at will always be about 3K above freezing point. I'll give it a try at some point. This looks fun. 

[Zarquan]

1 hour ago, mathmanican said:

Fun!  I thought you were using a solid parent, but nope.  The water is the parent, which goes to steam and then drops back to water because you keep the water cooled with your PW loop.  You then disconnect the water from the donor (huge refined carbon tile), and then it instantly drops back to water.  Just open/close doors, and you drop the temp on the refined carbon tile by 10 degrees each time..   Ingenious. I love it.  I think we have just scratched the surface on flaking abuse. 

@Zarquan, have you tried using other liquids than water as your parent?  I'm guessing you could get the temp much lower if you used something like crude/petro, or maybe even liquid oxygen (provided you had supercoolant for your loop).  I'm guessing the temp things settle at will always be about 3K above freezing point. I'll give it a try at some point. This looks fun. 

The issue is that sour gas doesn't condense in to petroleum. 

My main concern with colder liquids is that I would need to have a coolant which could keep them cold.  I'm using polluted water.  This would be more efficient if you were to use LOX, but then I would worry about the huge temperature differentials.  Hanging around above 273 is nice because almost everything is stable.  I think I did choose incorrectly when I chose my parent material (water at about 20 C).  I was still thinking in terms of creating heat, but I should have been thinking in terms of destroying it.  With a few modifications, I should be able to improve the design.

I increased the temperature of the water to 90 C and insulated it.  If it gets too hot, I added a conveyor bridge attached to the cold area connected to a door to keep it below 95 C.

Disregard the damage, I made a mistake at one point and things exploded, but its all fixed now.

[mathmanican]

1 hour ago, Zarquan said:

I'm thinking this could be used to create a super efficient liquid hydrogen machine.

So it deletes the heat of whatever you want to throw at it (either in an aquatuner, or through the door.  All you have to do is make sure the water doesn't get too hot (so a blip of cooling every once in a while).  I like it.  

[Zarquan]

5 hours ago, mathmanican said:

So it deletes the heat of whatever you want to throw at it (either in an aquatuner, or through the door.  All you have to do is make sure the water doesn't get too hot (so a blip of cooling every once in a while).  I like it. 

I went and made it.  It really wasn't that difficult.  The main feature of this build is that flaking is used to reduce the temperature of 2 kg/s hydrogen from 26.9 C to between -243.0 C and -245.2 C.  Since the main purpose of this build is heat deletion, I moved the hydrogen stream extremely close to the donor cell and moved the aquatuner in to the "warm" side of the hydrogen path.  Since the donor plate drove the temperatures down, I am able to precool the hydrogen essentially for free.

The aquatuner is running 20% of the time, so on average 240 W is being used on the aquatuner, which is cooling the final condensation room and the cold side of the flaking room.  So it is a 240 W liquid hydrogen machine that produces 2 kg/s liquid hydrogen.

Spoiler

The big automation piece is used as a giant heat transfer plate.

One odd thing is that the door pump near the boiler plate sometimes fails to work.  Even when it is powered and given 2 seconds per step.

EDIT:  I made a post about this build.  I also found a flaw and improved the efficiency to 120 W.

 

[mathmanican]

12 minutes ago, Zarquan said:

I went and made it. 

Broken... so broken. I love it. Keep em coming.

A minor modification will convert oxygen to solid oxygen for almost no power...

[Zarquan]

52 minutes ago, mathmanican said:

Broken... so broken. I love it. Keep em coming.

A minor modification will convert oxygen to solid oxygen for almost no power...

Do we want solid oxygen?  Because I was thinking I could adapt this design for liquid oxygen pretty easily with little to no modification. 

[mathmanican]

32 minutes ago, Zarquan said:

Do we want solid oxygen? 

Why not.  Just run it behind walls in the base for cooling and oxygen delivery. Though do it sparingly.... We can use a liquid tepidizer to get enough for rockets when needed.  

[Zarquan]

14 minutes ago, mathmanican said:

Why not.  Just run it behind walls in the base for cooling and oxygen delivery. Though do it sparingly.... We can use a liquid tepidizer to get enough for rockets when needed.  

FYI, I made the LH2 machine its own post.

 

[Manarz]

  Even though the discussion already drifted towards cooling i thought its time to share my simple design for heat creation:

This build is designed to solve 3 reoccuring issues:
1. No heat transfer between child/parent and donor
2. Easy Setup and Maintainability for survival (space materials are needed though)
3. The input amount of parent liquid to flake does not matter at all -> no mass deletion, no heatloss through normal conduction

 

The heat transfer concept:

Heat is only transfered into the isolation tile by the conveyor bridge but not towards the supercoolant sitting on top.
As usual the mechanical door controls when to inject the heat.

 

Barebones Build

 

One of the biggest Problems in all builds suggested here was always the conduction of heat between the Donor and the Parent/Child aside from flaking. Now the basic idea in this build is to use insulated tiles for flaking in order to avoid the nasty fluid dynamics setups in survival games and still have no heat transfer at all.

Some maths:

Spoiler

 

So in order for the flaking to be efficient, we need to calculate the temperature at which Supercoolant can flake on an Igneous Isolation Tile. You get
the max temperature delta possible if you set DonorOldTemp and DonorNewTemp to be the same and then solve the formula for DeltaFlakingTemp. Then u need to subtract this maxFlakingDelta from the Boiling Point of the parent.

Doing this for Supercoolant and Igneous Rock:

maxDeltaFlakingTemp = (DonorSHC * 10 * DonorMass) / (5 * ParentSHC) = (1*10*400) / (5*8,44) = 94,8C°
minFlakingTemp = ParentBoilingPoint - maxDeltaFlakingTemp = 440-94,8 = 345,2C°
(Note i am not sure if you should use the resulting temperature of the Supercoolant after the flaking (440C°)
or the real BoilingPoint so i chose the more defensive option)

 

So we know we need to preheat the Supercoolant to 345C° to kickstart the chain reaction (which is exactly what my tests show). 

Flaking only once results roughly in a 4mDTU gain.

Spoiler

Temperature diffrence gained : 94,8C°
resulting DTU for temp gain: 94,8C° * 8,44SHC * 5000g = 4mDTU
DTU lost in heating Donor : depends on exact temperature but its usually around 0.1-0.3C° => Average:  0,2C° * 1SHC * 400000g = 80kDTU

How much Steamturbines can we run with one AT?
According to the wiki an AT produces 1,181,600 DTU/s. So we can in theory flake a little short of 15 times per second overall creating:
60mDTU/s for steam turbines to consume.
The turbines can run either on 3 or 2 Inlets depending on whether u want to handle more cooling for a few extra watts per turbine but
based on the heat production of the turbines (90kDTU-110kDTU) i guesstimate they delete around 1mDTU/s each when used with 3 inlets.
So in theory this setup should be able to run around 50-60 Turbines at full 850W efficency.
Of course distribution of all the heat from one single AT to 50 Turbines becomes a bottleneck.
Also you guys are already on it but steam turbine heat deletion might occur here (i haven't seen it in my tests but it's hard to tell)

Sadly this desgin is not viable with Phosphorus or Sulfur (which are the only other 2 materials in the temperature operating range of a Turbine). The calculations show that no matter which insulation you choose it wont net much DTUs and you are better off with the more traditional infinite energy designs out there (e.g. abusing pile stacking mechanics) or do the fancy liquid shenanigans and abuse chorine as a Donor which i dont deem really viable for survival (yet :D).

Usually i don't do too much ONI maths, so please feel free to doublecheck my calculations.

And here is the upscaled version:

30 Turbines, heat still rising slowly, flaking has not reached full efficiency yet
(-> flaking still happens too fast so it consumes a bit more energy than necassary. more turbines should help)

 

Note that i haven't used any heattransfer methods at all yet. its all just Hydrogen and Copper conducting the heat here.
All the onveyor bridges transfering heat into the Igneous Rock Isolation tiles are out of Aluminium Ore. Everything else is Copper basically.
So the hard part to come by is the amount of Supercoolant for each chamber (i used 400kg per tile for better temp buffering) and the Thermium AT

At the time of posting this contraption ran stable for about 50 cycles. There is still alot to optimise though.

[mathmanican]

9 hours ago, Manarz said:

Heat is only transfered into the isolation tile by the conveyor bridge but not towards the supercoolant sitting on top.

I love it. I kept thinking insulated tiles would be useful but hadn't come up with a design. This is awesome.

[mathmanican]

12 hours ago, Manarz said:

(-> flaking still happens too fast so it consumes a bit more energy than necassary. more turbines should help)

ONI engineering 101.  

• If you can't quite generate enough heat energy to get something to happen, then build a few more heat deletion machines. Deleting the heat will get you more heat energy...

I have been laughing inside at that line for the last 2 hours.  Thanks for making my day @Manarz. Have you tried swapping to another type of insulation, like sedimentary (I think you can drop the delta and then use phosphorous or sulfer). 

Edit. You did check. And you are absolutely right. We'd need a 207.87 delta.