Testing nuclear reactors ratios and layouts in 0.16

[wvlad]

I tested how many exchangers I can place for which layouts and what behavior happens using void pipes connected to exchangers output. I want to share and discuss the results here.

1. The heat is shared between neighbor reactors. You have to turn on or off all of your reactors, not by one, because the heat is shared between them and all connected heat pipes. You can connect 2 lines of exchangers to one reactor and zero to its neighbor and they will balance.
2. Neighbor bonus really makes more heat (not changing cell usage time) so you need more exchangers.
3. Single steam pipe can take input from up to 12 heat exchangers.

1 single reactor can handle 4 exchangers as stated in wiki.

1x2 reactors are very inefficient so I gave up at tweaking them.

For 2x2 reactors the maximum I could reach was 52 exchangers working (7-8 in a line) which is ~13exchangers per reactor.


For 2x3 reactors the maximum I could reach was 83 exchangers working (8-9 in a line) which is ~14 exchangers per reactor. Interesting that extra neighbor bonus didn't add any significant effect.


For 2xN heat pipes allow to connect up to 18 working exchangers in line per reactor though I'm not sure how many of them will be really used (should be tested with different N, theoretically it's 16 for 2x100 and 15 for 2x10). According to #1 you can take an average number and use this length for each reactor, even at corners.

So 2xN beats them all, right? What do you think?

Also the values are a bit bigger than in the wiki https://wiki.factorio.com/Tutorial:Nuclear_power

Update
The current winner is 11-12 in a double-row + some "corners":

[Aeternus]

The pipe braid you are using is known to cause throughput issues. I'm working on an extendable row based design. It's tricky, for an 2n+2 kind of design in reactor rows, you need to bleed off 540MW per 4 reactors, and that gives you 10 tiles width to play with. That's enough for 2 rows of heat exchangers feeding off a single heat pipe, and you can fit those on each side, requiring 13 length (for 520MW bleedoff), with separate water input and steam output. This requires 26 turbines.

I'm not worried about the heatpipe throughput, which is capped at 1GW. Adjacent reactors have a 5GW transfer rate so those should be practically transparant to eachother - a reactor row for practical purposes can be seen as a single heatsource.

Maximum pipe flow rate for liquids is, to my knowledge, around 1500 across short distances. Assuming that daisy chaining the heat exchangers won't run you into flow issues, it should be possible to fit 26 heat exchangers on each side. A double row of 13 (peak flow requirement 1300), with the turbines not far behind in a 10-6-10 pattern with room for some power poles and substations. Some buffer tanks inbetween the exchangers and turbines is also not a bad idea.
Note that this requires 3 offshore pumps per heat exchanger row.

[wvlad]

I don't understand what do you mean by throughput issues? As long as total number of working heat exchangers is equal or even bigger than theoretical maximum I see it as a perfect solution.

[mrvn]

I don't understand what do you mean by throughput issues? As long as total number of working heat exchangers is equal or even bigger than theoretical maximum I see it as a perfect solution.

The heat pipes are the problem. The heat gets bottled up in your setup and only a fraction reaches your heat exchangers.

[wvlad]

I like your idea of using two sides of a heat pipe. Since the row is less or equal to 12 it makes possible to put turbines just at the same row instead of limiting reactors count to 24 (2x12) and putting turbines at sides. This design works exactly as it was theoretically calculated.

[wvlad]

The heat pipes are the problem. The heat gets bottled up in your setup and only a fraction reaches your heat exchangers.

Then they would not heat to 500 C. My tests show that they actually would at that distance.

[Aeternus]

You can't really chain more then 14 heat exchangers... 15 at most. More then that and you run into a limit of how fast liquid can flow to a pipe. 1500 has been the demonstrated maximum over short distances, and each heat exchanger at maximum capacity will require just over 100 water per second to produce steam. I see you've basically cheated this by putting infinite water producers on both ends of the chain. Typically, it'll not work like that.

Try putting 16 of em in a chain, then put one or two water generators only on one side of them. The last heat exchanger in the chain should not be receiving water if all of them are under full load.

[Edit] Your setup above looks somewhat similar to my old 0.15.11 design: viewtopic.php?f=208&t=48150
I'm working on something similar to this with the intention to be able to tile it outward. It's the only way to make an extendable nuclear plant.

[wvlad]

*Updated

I'm curious now what reactor temperature is considered good. With 272 exchangers it's on 980 C (previous screenshot) but if I connect more (at the top and bottom) it cools to 700 and some distant exchangers stop working. The maximum count of working exchangers is then 280.

So is it better to go with 272 at 980 C or 280 with 700 C? I think 280 would just produce more energy.


So the best design I see is to put 9 in a row and add 4 additional rows with fewer exchangers at 4 corners.

For the water pipe limit I said that you can chain up to 12 of them and put turbines at the right or left side, NOT at the same row, but this would limit total reactors to 24.

[Aeternus]

Heat exchangers will produce steam as long as they're above 500 degrees. Any lower and they're off. Any higher and they're on and produce 10MW of steam. They produce this by reducing their own heat, which then has to flow from the heat exchanger. As long as you have more power production then heat dissipation in the exchangers, the temperature in the reactors will continue to rise.

[wvlad]

*Updated
Ok, so the theoretical calculations are really bellow what you can do. I tested the same design in bigger scale. There you have to use 11-12 exchangers in a row and less corners but still necessary:

The temperature is 870 С and stable.

Also it seems that you have to turn on or off all of your reactors, not by one, because the heat is shared between them and all connected heat pipes.

[Kelderek]

My setup has rows 16 exchangers on each side with a water pump on either end of the exchangers. I'm not sure how good or bad that is for the fluid flow, but it seems to be working fine for me so far. Some creative use of landfill allows me to place the water pumps right where I need them. I then have a line of 28 turbines in a straight line. The first row of exchangers is only 12 per side, and ideally the last row would be the same, but since I plan to expand on one end of the row of reactors, I just keep adding exchangers in batches of 16 per side.

The trick was finding the right body of water so that I would have enough room to expand (using landfill), but I should be able to extend this setup for quite a long number of reactors before I run out of water for the pumps. I will eventually run into an island and have to end the setup.

Birds eye view

20180208110255_1.jpg (340.06 KiB) Viewed 6224 times

The outermost exchangers are at about 630 degrees and as best that I can tell, the steam is able to make it all the way to the end of the turbines okay. I also have room to fit the robo ports on each side for feeding the reactors. It takes about 1000-1500 landfill for each new pair of reactors I add when extending it. Perhaps someone else can test this in a bigger setup with max load on the creative mod.

Closer in

20180208110409_1.jpg (943.94 KiB) Viewed 6224 times

[mrvn]

*Updated
Ok, so the theoretical calculations are really bellow what you can do. I tested the same design in bigger scale. There you have to use 11-12 exchangers in a row and less corners but still necessary:
The temperature is 870 С and stable.

Also it seems that you have to turn on or off all of your reactors, not by one, because the heat is shared between them and all connected heat pipes.

Only reactors that are on give you a neighbour bonus. So use wires to refuel all reactors at the same time every time. Regulate the amount of power you generate by how often you refuel, not by how many reactors you fuel.

[Aeternus]

Yep, and store surplus energy in storage tanks for steam. They make awesome energy buffers, and the treshold of the steamtanks nearest to the reactor is a good gauge for when the reactors need to switch on and produce energy. Reactors should be on an "all on or all off" setting, so tie all of them into a single control circuit. You'll need to be able to convert a lot of heat into steam fast during the 200 second heat burst in the reactors - if you don't want to waste fuel, you'll need to be able to store roughly 24GJ per reactor - or be able to consume that much energy over a 200 second burst. The reactors themselves store 5GJ each in heat. The heatexchangers and heatpipes have some capacity too but relying on those isn't wise as the reactors will reach max heat before their capacity is fully utilized. That leaves 19GJ per reactor to store. A single storage tank of steam can store 2.4GJ roughly - so to fully buffer your reactors you'll need about 8 storage tanks per reactor. In my setup above I used a 5x3 cluster fed by 2 rows of heat exchangers, with turbines hooked up to the other side of the tanks.