All the different ways to increase the efficiency of a nuclear power plant

[mrvn]

the correction that make the design unstable you mean the thing in the red box called "not necessary" ? you mean it transform a design that is 0.06% overheating into a system that's 0.0002% unknown direction unbalanced ? and you are mad about it ? pal that's why it's in the red box called not necessary in the first place.

Bold is from me. And that is the point I'm trying to make to you for ages now. One direction is harmless the other deadly. So "not necessary" should be labeled "potentially harmful".

[mmmPI]

It's not even the case that math says it's potentially harmful in 2 millenia.

(unless you have more information on how the 10/0.097 that seems to fits everywhere was computed ?)

maybe something with the 10 000 000 joules/s that a heat exchanger is supposed to transfer from heat to steam in my math i used it as 10 MW and converted to MJ/S , when a cell of 8 GJ takes 200 second that means 12000 ticks to burn there are roundings. (2/3) that would be rationnal critic.

maybe the 10/0.097 is what the heat exchanger produces if it function 12000 ticks, but when you have a nuclear plant, you have heat not a direct conversion, so it must happen that not all 48 heat exchanger function 12000 tick with the energy of 1 fuel cycle. maybe only 47 and 1 function 11999 and there is 0.000000001°C of heat that is lost during rounding and therefore cannot leave the reactor under the form of MW.

that's just theory, it would require to measure heat precisely which is the whole point of the topic haha.

Yet that's funny because you can also go with >= 82 350 steam, and then you'd have a security margin, as explained, in this case the mechanism is not "potentially harmful" just not necessary, because it makes you save so few compared to what you are already giving yourself as margin, but it still does sometimes reduce the amount of waste. which is not necessary

on a more abstract level the amount of imprecision could force taking a margin, if you cannot make sure the impact are utterly neglectible.

This margin to me is the energy lost equivalent of the amount of energy lost during measurments when you have a mechanism that measure overheating and overcooling.

That's necessary to make the difference between a static system because it's perfect, or because it's dead, you need to poke it a little, and doing so it react, that's the measurment energy loss.

Or you make sure when it's stable to have a little signal, like emiting something you can read to know it's fine which is an energy loss too, that's the margin of security you take with perfection and try to minize.

[mrvn]

Please just stop. Now you are mixing what other people said with what I said as if they had disagreed with your design.

[mmmPI]

i'm not, i'm trying to stay on the topic, GeorG mentionned an interesting fact that he encountered during the making of his power plant, 12000 +1, that is something that could be investigated to understand how the roundings behave .

the number from OP the 10/0.97 seem to fit everywhere means that it must be the correct number, but it doesn't look like a number someone would have deliberatly choosen over 103 if not for a particular reason.

The whole point of those math is calculate the amount of energy in the buffer by counting/timing either what's going out or in without measuring inside the blackbox of the thermal buffer since no component allow to connect the signal circuit to heat.

anything that is explaining more in detail how to compute the energy is to be taken in consideration to compute precisely the imprecisions and rounding

the design is not important that's just one thing, but the math on the otherhand could be used by other person to make their own power plant, the informations are interestingly (imo) all collected at the same place ( concerning the nuclear power plant and the efficiency). Which i'm happy to try and contribute to.

agreeing or disagreeing with a design is a personnal opinion, a choice one make, which we don't even need to know, unless there are arguments, which can convince or not, or be valid or not, some are unvalid because not correct from math standpoint, or application in game, that's different. Some argument are just personnal taste and there is not much to argue about who's personnal taste is the best.

[Khagan]

the number from OP the 10/0.97 seem to fit everywhere means that it must be the correct number, but it doesn't look like a number someone would have deliberatly choosen over 103 if not for a particular reason.

The factor of 0.97 turns up naturally (and annoyingly) in Factorio nuclear power plant design because the temperature range from water to high-temperature steam is 485°(i.e. 97% of 500°).

But surely you knew that.

[mmmPI]

no i had not realised it was the case despite it being written in the wiki https://wiki.factorio.com/Heat_exchanger

Heat exchangers produce 103 steam/second.This can be calculated by relying on steam turbine data: A steam turbine consumes 60 steam/second and produces 5.82MW (assuming 500°C steam). This means a single unit of 500°C steam has 5.82MW / (60/s) = 0.097 MJ of energy. A heat exchanger produces 10 MJ a second, therefore it produces 10MJ / 0.097MJ = 103.0927835 steam per second.

The steam production rate can also be calculated using the energy consumption: 1 Heat exchanger consumes 10MW, so it's putting 10,000,000 joule of energy into heating water/steam per second. To heat up 1 unit of water 1 degree, 200 joules are needed, so the heat exchanger is heating up water by 50,000°C in total. But the water only gets heated up from 15°C to 500°C, so by 485°C. So the 50,000°C are enough to heat up 103 units of steam per second, since 50,000 / 485 = 103.09. Since steam is produced from water in a 1:1 ratio, this also means that 103 units of water are consumed per second.

[mrvn]

the number from OP the 10/0.97 seem to fit everywhere means that it must be the correct number, but it doesn't look like a number someone would have deliberatly choosen over 103 if not for a particular reason.

The factor of 0.97 turns up naturally (and annoyingly) in Factorio nuclear power plant design because the temperature range from water to high-temperature steam is 485°(i.e. 97% of 500°).

But surely you knew that.

All that realism with correct units and honoring the start and end temp and then they ignore that a ton of energy in making steam is evaporating it.

[Khagan]

All that realism with correct units and honoring the start and end temp and then they ignore that a ton of energy in making steam is evaporating it.

Yes, indeed most of it. In the real world the main point of high-temperature steam is its higher pressure, not its greater heat content.

[mrvn]

All that realism with correct units and honoring the start and end temp and then they ignore that a ton of energy in making steam is evaporating it.

Yes, indeed most of it. In the real world the main point of high-temperature steam is its higher pressure, not its greater heat content.

They should have made the steam engine/turbine work when the steam level in the pipe has a certain height.

[gGeorg]

All that realism with correct units and honoring the start and end temp and then they ignore that a ton of energy in making steam is evaporating it.

Yes, indeed most of it. In the real world the main point of high-temperature steam is its higher pressure, not its greater heat content.

They should have made the steam engine/turbine work when the steam level in the pipe has a certain height.

More physical issue is, steam is not consumed by turbine/steam engine but transformed into cold steam, then goes into a condenser, then into heating process again. As result, power plant dont consume tons of water, but only a bit.

[mrvn]

All that realism with correct units and honoring the start and end temp and then they ignore that a ton of energy in making steam is evaporating it.

Yes, indeed most of it. In the real world the main point of high-temperature steam is its higher pressure, not its greater heat content.

They should have made the steam engine/turbine work when the steam level in the pipe has a certain height.

More physical issue is, steam is not consumed by turbine/steam engine but transformed into cold steam, then goes into a condenser, then into heating process again. As result, power plant dont consume tons of water, but only a bit.

The steam engine has 3 cylinders in it. That's usually a design where the first runs hot steam, the middle medium steam and the third cold steam. The output of each is used in the next. Then again I think in those designs the cylinders have different sizes. So maybe it's just 3 independent engines running in parallel, like in a car. So you always have one of them pushing.

There are mods with steam in different temps but the devs want to get rid of liquid temperatures. Needless waste of computations for the few mods that can mix fluids with different temps. One can always make separate liquids for the different steams if mixing them isn't a required feature.

Note: you can't read steam temps so any controlled mixing is rather elaborate and I think nobody uses that. They just have engines/assemblers needing different temps to work.

[mmmPI]

I made another one using barrel with some extra utility next to it :


viewtopic.php?f=208&t=100464

large2.png (1.43 MiB) Viewed 5802 times

There's extra informations there : viewtopic.php?f=208&t=100464

The principle itself is not much different, counting trains, this time trains that brings the water barrel in the power plant.

I used no-signal train network, still no combinators, no storage tank, but also no chest.

ofc no storage tank filled with water x)

This time i added many switches to turn off areas automatically, like the fuel production can isolate when the main grid is in brown-out, this to preserve fuel production, and other mechanism to save power.

The assembly used to un-barrel the water are solar powered for maximum resilience. they also switch off between 2 trains, for roughly 1 second of drain saved each train, same for the loading pumps.

Also this time nothing comes from infinity chests, the train fuel productions is shown and the recycling of fuel cell too, all without combinators. oh and no water tank either.

There are more heat exchanger than necessary, but the refuel is also too more frequent than necessary ( 86% efficiency ). This means the cores are always 999° or so. But if you draw more than 280MW, the core will lose temperature for 1H30 while providing 291MW. That's some sort of extra safety in case someone is very very afraid of brown-out.

Every necessary step for fuel production can isolate and be autonomous in case of a major power draw that exceed the capacity of the power plant, just in case 1 second of brown-out used to scare you, fear no more !

Waterdelivery service16.zip

(6 MiB) Downloaded 173 times

=> this version has a copper cable that shouldn't be there
Here is an updated patched version :

Waterdelivery service17.zip

(5.98 MiB) Downloaded 141 times

[mrvn]

Here is another control mechanism using trains that I think is rather easy to understand:

reactor-train-controlled-overview.png (4.23 MiB) Viewed 5746 times

reactor-train-controlled-control-logic.png (4.15 MiB) Viewed 5746 times

The refueling logic is simple:

a) The train stops for empty trains output their train count.
b) one inserter is set to remove spend fuel cell when at least 6 empty trains are waiting for steam (red line).
c) inserters are set to refuel the reactor when a spend fuel cell get removed (green line)

The fuel belt has one side spend fuel cells, other side fresh fuel cells and some circuit logic to park exactly one fresh fuel cell in front of every reactor. You can replace that with buffer chests or logistic chests or just fill the belt. Whatever your favorite mechanism is. Doesn't really matter, just make sure there is fuel. I didn't include a fuel check for this proof-of-concept.

Now some explanation of why this works:

A 2x2 reactor can sustain 48 heat exchangers, enough to fill 2 trains with 8 fluid wagons. My trains have 10 fluid wagons and 60 heat exchangers. So more heat is converted into steam than the reactor can produce if trains are in both stations. There are some pauses when trains switch but overall more heat is consumed than produced. So as long as trains keep coming the reactor is cooled down, it can't overheat.

But what if trains stop coming? That's condition (b) above. 6 Trains are more than enough to convert all the heat from 4 fuel cells into steam and load it. The reactor stops getting fueled when there aren't enough trains to soak up the heat. Since condition (b) is a simple level trigger the control logic can't drift due to imprecise math or measurement and break down.

Note: Given how I made the rail network there is also some extra circuit control in the train stop to keep the top an bottom balanced. The train stops for empty trains get enabled in pairs and only when the last pair has trains coming in. And the full trains are released in pairs too although that doesn't work quite as perfect. Without that the stackers can get unbalanced. Should have made a single stacker each for input and output but this looked nice being symetrical.

[mmmPI]

that's a smart way to avoid detecting overheating or overcooling.

only pitfall i see is that if train filled with steam don't move due to low consumption AND there are 6 empty train, then the power plant will enter a state where it burns fuel non stop for nothing.

This shouldn't happen in lab conditions like in your example where distances and train number is fixed but in real conditions this means you will need to add stacker room on the exit as you add trains, or just have trains that are only designed to move in a closed non-evolving loop a bit like tanks and pipes.

[xaetral]

Yeah using the steam level automation on train-based setup is the go-to because while having less than 14 steam tanks per reactor is annoying on a 2xN setup, that's only a bit more than 3 trains containing 4 fluid wagons, which isn't so much.

[mrvn]

that's a smart way to avoid detecting overheating or overcooling.

only pitfall i see is that if train filled with steam don't move due to low consumption AND there are 6 empty train, then the power plant will enter a state where it burns fuel non stop for nothing.

This shouldn't happen in lab conditions like in your example where distances and train number is fixed but in real conditions this means you will need to add stacker room on the exit as you add trains, or just have trains that are only designed to move in a closed non-evolving loop a bit like tanks and pipes.

Can't happen, that's what the stackers are sized for. Any train in the output stacker frees up a slot in the input stacker. There aren't enough trains for it to become a problem. And yes, you always have to resize the stackers when you add more trains.

You can make it more complex and check that you have empty trains and space in the output stacker. Then you only need num-trains + 6 slots in total. For a setup with a lot more trains that would make sense. But if power consumption then spikes the empty trains might have nowhere to go for a while till the reactor catches up. The current design garantee a train can leave the power station when a filled train is ready to take it's place at the latest.

Note: The power stations buffer a train too. So adding one train per additional power station is OK. The stations always accept a train and hold it there till empty. Don't make them only accept a train when low on steam or you loose that buffer spot. You can enlarge them to buffer 2 trains if the outpost is far away but the tank is probably enough buffer for a long travel time. 2 trains and no tanks can work too.

[angeldeath29]

Im make a nuclear power plant with control of fuel loading into the reactor by power consumption.
The power consumption is calculated by the rate of steam consumption in one turbine unit (8 boilers, 14 turbines). The calculation is carried out using two tanks connected by a pump. The boiler output is connected directly to the first tank. The second tank is connected to the first through a pump, which turns on when there is less than 5k steam in the second tank, and turns off when there is more than 20k steam in the second tank. Counting is active when in the second tank has steam from 8k to 18k. The current value in the tank is stored, and a new value is stored after 30 ticks. The difference is consumption in half a second.
Next, we determine the time between fuel loads by the power consumption. For example, if the power consumed is 60%, then the time spent in the reactor should be 60% of the time between loads. 200 = T*0.6, T = 200/0.6.
Additionally, the reactor has a determinant of the steam production. To do this, an additional boiler is used, connected to the last row of 16 boilers. The boiler outlet is connected to a tank from which steam is pumped out by a pump. When there is less than 1 steam in the tank, it means the boiler temperature is less than 500 degrees. Warm-up indication in the symbol H of the green network. If a lack of heating is detected, then several downloads are made immediately after working out the previous one. The number of such cycles is determined by the power consumption.

blueprint

[mrvn]

If only pumps could send their throughput to the circuit network....

[angeldeath29]

If only pumps could send their throughput to the circuit network....

it cannot
but I measured it, and it matches what the pump tooltip shows

[mmmPI]

Next, we determine the time between fuel loads by the power consumption. For example, if the power consumed is 60%, then the time spent in the reactor should be 60% of the time between loads. 200 = T*0.6, T = 200/0.6.

I like this !