Compact 4.8 GW Nuclear Reactor

[Lisp Wizard]

You all know the problem: You absolutely, positively need to power your 26.000 electric furnaces or 10.000 beacons, yet most nuclear setups are ugly 2xN tubes, powered by thousands of robots, so you fall back to solar energy and spend the next months paving the whole map with solar panels. Fret not and have a look at this 32 reactor setup!

Sales pitch:
- adaptive fuel consumption
- zero two robots required
- beautiful rotational symmetry
- almost perfect neighbour bonus: Only 3% less efficiency than a 2xN setup

The reactor requires 12 water pipes from each side, so it is recommended to place it on a small island. You really need water on each side, trust me . Apart from that, the only prerequisites are an incoming belt with uranium fuel cells and an outgoing belt with nuclear waste. The blueprint contains enough roboports such that most of the reactor can be built autonomously. The only exception is the center of the reactor, as it is a logistic network of its own. Once the construction is complete, don't forget to place at least two logistics robots in the reactor core's roboports.

Reactor Overview

Reactor Overview

reactor_overview.jpg (1.35 MiB) Viewed 32743 times

A simple combinator setup prevents the system from consuming too many fuel cells. Each reactor is numbered from 0 to 31. Only reactor 0 is unconditionally active. The current level of steam is used to determine how many reactors should be active to meet demand. Every 375 game ticks (200 seconds * 60 ticks/second / 32 reactors), one reactor is inspected to see whether its number is less than the number of reactors that should be active. If so, a single fuel cell is delivered to this reactor.

Reactor Circuitry

Reactor Circuitry

reactor_circuitry.jpg (1.27 MiB) Viewed 32743 times

Suggestions for improvements are most welcome.

Blueprint URL:

- Version 1.1 https://pastebin.com/t2TWkmm4
- Version 1.4 https://pastebin.com/qrmCd5QF (Now with roboports, better aesthetics and more reliable circuitry!)

[naknak]

I don't know why this has no replies. I've tried a lot of nuclear blueprints and yours is the first that works as advertised. What's more, it's clear that a lot of thought went into the design, and a lot of effort into refining and debugging. The result speaks for itself. Thank you very much for making and sharing this.

My only criticism is that it won't self-build from a surrounding logistic network, because it's bigger than a roboport's green square. I added ports to some empty space near the four outer corners, just close enough for the unbuilt center to be reachable by construction bots, while far enough to avoid infringing on the standalone logistic network at the center of the finished product.

[Jap2.0]

Looks like it would probably work with water barrels as well, so putting it on an island is optional (although by the point where you need 32 reactors, you should be able to make a few landfills).

[vanatteveldt]

Pretty!

I think scaling the # of reactors is not the most efficient way to step down the plant, since fewer active reactors also means less neighbour bonus. It is more efficient to activate/deactivate the whole setup at once and buffer the energy as steam.

[Lisp Wizard]

Hi guys, thank you for the feedback.

@naknak: Thank you for the reply. It is good to hear that someone finds use for this beast. I personally considered self-building not so much an issue --- I built some temporary roboports to lay out the concrete and placed the rest from my personal roboport. But maybe I could add an outer layer of roboports, if only for some occasional repairs of the walls.

@Jap2.0: Sure, barrels would be fine. But do not underestimate the water intake of this reactor. As you can see on the many pumps in the design, moving the water/steam for 32 reactors is quite a challenge.

@vanatteveldt: Always activating all reactors at once is not much more efficient. In the current setup, I activate the reactors in a clockwise order, so that most reactors actually have full neighbour bonus. This is only 2-8% less efficient than your proposal, but permits more fine-grained control over the fuel consumption. I estimate one would have to place 400 additional steam tanks to buffer the energy of an all-or-nothing setup. At the scale where you need such a reactor, the CPU utilization caused by the fluid mechanics in 400 steam tanks is probably a bigger concern than these few percent reactor efficiency. Bonus: With the clockwise activation scheme, your reactor visualizes the current utilization of the electric grid.

[darkfrei]

Can you make same reactor for games without any drones?

[impetus maximus]

i'm not even going to attempt to understand the circuits used here. over my head.

i didn't manually feed any fuel. i just loaded the belt, and let it do it's thing.
i noticed 3 reactors weren't getting fed fuel. after a while all but 1 were humming along.
this is under 100% load. now 4 to 6 reactors are idle, and the power fluctuates between 3.5 to 3.8GW.

i do realize the steam tanks are for peak power draw times. just wanted to report on my findings.
i'm an overclocker, can't not test under full load.

[BlakeMW]

This is rather interesting. One thing to note is that reactors can buffer an awful lot of heat just in the reactors and heat pipes/exchangers - for example a 160MW reactor doesn't require steam storage at all to buffer an entire fuel cycle (you need some way to detect low stored heat, most easily that a steam tank on the periphery of the heat network is getting low), even larger reactors can buffer entire fuel cycles assuming they are being run at a significant load (i.e. 30-50%). A potential optimization would be to activate reactors in blocks of 8 and exploit the free heat buffering, this would get almost full neighbor bonus and simplify circuitry.

[Hamsterman]

Hi guys, thank you for the feedback.

@naknak: Thank you for the reply. It is good to hear that someone finds use for this beast. I personally considered self-building not so much an issue --- I built some temporary roboports to lay out the concrete and placed the rest from my personal roboport. But maybe I could add an outer layer of roboports, if only for some occasional repairs of the walls.

@Jap2.0: Sure, barrels would be fine. But do not underestimate the water intake of this reactor. As you can see on the many pumps in the design, moving the water/steam for 32 reactors is quite a challenge.

@vanatteveldt: Always activating all reactors at once is not much more efficient. In the current setup, I activate the reactors in a clockwise order, so that most reactors actually have full neighbour bonus. This is only 2-8% less efficient than your proposal, but permits more fine-grained control over the fuel consumption. I estimate one would have to place 400 additional steam tanks to buffer the energy of an all-or-nothing setup. At the scale where you need such a reactor, the CPU utilization caused by the fluid mechanics in 400 steam tanks is probably a bigger concern than these few percent reactor efficiency. Bonus: With the clockwise activation scheme, your reactor visualizes the current utilization of the electric grid.

I very much agree with naknak - I would prefer it to be self sustaining and building.

[Aeternus]

There is room in this design to add a roboport grid. The heatpipes are in a double row configuration. They don't need to be, each heatpipe has a 1GW top transfer capacity, which in this design isn't even half utilized. Make the heatpipes a single row and the open space width in the center rows increases to 5. Along the diagonal too there is room for roboports. Combined they should have a construction area wide enough to self-build the entire plant.

This design does not have enough room for belts inside the plant, so it won't work without robos.

[CyberBill]

I was looking for a large nuclear reactor and came across this one, and now have it up and running. I have some notes for people who are trying to build & run this...

* WATER. Holy hell this takes a lot of water. My first attempt was to use barreling... Failure. In order to get this to run you'd need around 48 Level3 assemblers with speed modules to de-barrel water, and an equal number to barrel them, and like 18 fully saturated blue belts of barreled water. I tried to just get a SINGLE row of heat exchangers going and I failed. I ripped up the reactor and placed it in an area where 3 small lakes were well positioned for water pumps, and then added extras that wrap around to the last side.

* You should add Roboports! On each side, where the water comes in, you can put 2 roboports. In each diagonal that runs from the outside to the core, you can put two roboports. Then you can remove a few heatpipes and put one more roboport in each straight area as close to the core as you can get without linking the roboport networks. This gives you a total of 20 roboport (5 per quadrant) and lets it self build. It is very important to not allow the central roboport network to connect to your main roboport network!!

* Remove the concrete from the blueprint. It uses 36k concrete to fully cover an enormous area. Once you import the string and make the blueprint, go in and edit it and right-click the concrete to remove it. Do the same for the walls, gates, or lasers if you need. The concrete adds significantly to the build time as there are 3x as many items as *all other items combined*.

* You need to add a few Logistics Robots to the central roboport network to get it to run. I added 10.

* You also need to supply lines on the South side of the base for nuclear fuel and spent nuclear fuel. Instead, I pulled up the underground belts and used an active provider chest to get rid of the spent fuel and a requester to bring in fuel.. Again, be sure you put these OUTSIDE the core, so that it uses your main logistics network.

It really works quite well, so thank you very much.

[ddschmitz]

I was wondering if anyone who has built this could offer me some advice on getting this thing up and running. I have it fully built minus the turrets and defense stuff, the main stuff is in place. I had to start it myself by putting some fuel in the reactors but after that it doesn't seem to refill itself. I have fuel cells in the steel chest so it's not that it's empty. I can't tell if the individual reactors are checking against how much steam is available or what? I can provide screen shots of whatever is needed.

Edit:
I've detached the power plant from the rest of my base and have it running on solar power. I thought I could isolate it and wait and see if it would read that there is 0 steam in the system and start fueling itself but it's still not doing anything. It won't even put fuel in the default 0 one that OP stated.

Edit 2:
So I got the blueprint to work and the error was on my part. I had tried to replace the concrete in the design with refined concrete and some how along the way I broke it. I removed everything and used the blueprint provided by OP and it worked just fine (0.16.51). Now that it is in place and working, now I can start fiddling and adding my own touches .

[Lisp Wizard]

@ddschmitz The system should work fully autonomously. The reactors are filled whenever the steam levels drop, which happens whenever you draw more power. But if it does not even power the reactor zero, it must be another issue. Either some broken circuit, or you forgot to add the two robots.

The good news for everyone: Now that 0.17 is out, I will spend some more time refining the design, adding roboports, etc. So expect a new version of the reactor in the near future.

[spraynard]

Thanks a lot for this design. I've been using it in a couple of bases and it works just as advertised.

I too do like CyberBill's suggestions and omit the concrete and lasers before placing the blueprint, but I think it's useful to have the option in case it's needed or wanted.

Looking forward to new version!

[janneclaes]

for each of the eight HX/Turbine sections i count 104 Turbine and 60 HX.

you mentioned that at max load, power output drops from design 4.8GW to 4.5GW. my guess is that the turbines could consume a smidge more steam if a smidge more was available.

(104 turbine) * (60 steam per second / 1 turbine) * (1 HX / 103 steam per second) = 60.58 HX
citing https://wiki.factorio.com/Heat_exchanger for HX water demand / steam supply number. so you need about another half HX per section or four more across the design to be able to meet steam demand under sustained full load.

personally i like to run ratios with (100 steam per second per HX) to err conservatively and to simplify design. that would point you at 63 HX per section or 1 more HX in each line. then assuming sufficient water supply i think the reactor will sustain that 4,826 MW.

[Tekillaa]

Hello,

Awesome work, i like it a lot, do you got a version with direct feeding, with no circuitry? a version burning fuel all time, because like you said, my factory works all time, so it need all time energy : i got other little central to make the difference when needed, so i just need a big energy provider 24/7. It would be very nice, thanks a lot for the bp already existing.

[phi Seeker]

First of all thanks for a great and interesting blueprint, I personally was inspired to try very similar nuclear setup with the goal to improve to the desired theoretical 4.8GW but I only managed to get 1 GW per side without prebuffering in tanks and without breaking the shape too much. To manage to get the steam in all the turbines seems to be a bit tricky. Also I was testing this setup in sandbox and creative mode(among other setups) where you can have unlimited power consumption and after some afk time I noticed the power production to not go down to 4.5GW but rather to oscilate up and down(in 0.17). My opinion is that what is flawed is actually the circuitry that tries to save the fuel. After I tried to remove all of it, the power went straight up, but this tested only for a while.

@janneclaes

The ratios which were used are actually correct. You dont count the number of heat exchangers from turbines but vice versa. You have 8 reactors per side which count as 30 with neighbour bonuses so it is 30*4=120 exactly heat exchangers and then 120*1.718=206.16 turbines which is just rounded up to 208.

@Tekillaa

All you need to work for 24/7 is to remove the circuitry which is quite easy change I would say.

[Feltcherab]

The good news for everyone: Now that 0.17 is out, I will spend some more time refining the design, adding roboports, etc. So expect a new version of the reactor in the near future.

That would be really great! I tried loading the blueprint, but got an error, maybe due to 0.17?

My current base is big enough to require more than 24 reactors and again I find myself constantly adding an 2*4 reactor blueprint. But I want something more beefy, and yours looks just what I need... I'll keep a close eye on this thread and hope that you'll update your bp!

[phi Seeker]

The good news for everyone: Now that 0.17 is out, I will spend some more time refining the design, adding roboports, etc. So expect a new version of the reactor in the near future.

That would be really great! I tried loading the blueprint, but got an error, maybe due to 0.17?

My current base is big enough to require more than 24 reactors and again I find myself constantly adding an 2*4 reactor blueprint. But I want something more beefy, and yours looks just what I need... I'll keep a close eye on this thread and hope that you'll update your bp!

I can confirm that copying the blueprint string from that site can be tricky(I got an error as well first time I tried that, due to incorrect selection of the string on the website), but the blueprint actually works very well even in 0.17 as I loaded it and used it successfully already.

[Lisp Wizard]

The good news for everyone: Now that 0.17 is out, I will spend some more time refining the design, adding roboports, etc. So expect a new version of the reactor in the near future.

That would be really great! I tried loading the blueprint, but got an error, maybe due to 0.17?

My current base is big enough to require more than 24 reactors and again I find myself constantly adding an 2*4 reactor blueprint. But I want something more beefy, and yours looks just what I need... I'll keep a close eye on this thread and hope that you'll update your bp!

I finally found enough time to create a new version of the reactor. As usual, feedback is most welcome. Have fun!