Factorio Forums

Allright. New design! And this one is worthy of a megacomplex... Tileable, expandable, somewhat modular even. And capable of taking both Chemical and Nuclear fuel. Reactor is throttled, so no fuel wasted. Thought it wasn't possible anymore? Neither did I. Wether or not this design is of practical use is up to you guys to decide.

First screenshot should give you an idea of the scale. Can't do an overview of the entire powerplant in one screenshot... It's just too wide, by a -large- margin.

Yea. This one won't squeeze in between a couple of assemblers...

The Reactors + controls:

The reactor controls are fairly standard. On fuel insertion a latch is set, preventing further fuel from being inserted into the reactors until spent fuel is ejected. The logistics network is used to move fuel and spent fuel cells. There's a steam level latch that checks if the steam buffer is low enough to warrant more fuel. One more that checks if spent fuel is piling up for some reason.
There are 2 Constant Combinators which can be used to control the plant as a whole: Red disables nuclear fuel entirely, Yellow disables chemical and switches the plant to pure nuclear. Setting them both disables power generation and shuts the plant down.
Note that the chemical boilers are optional, if you want this tileable design but use only the nuclear part, just set the yellow combinator here and omit the chemical boiler section lateron - works just fine that way.
Control signals are sent to the other reactors via the Red circuit. The Green circuit is used in the rows to control switching between chemical and nuclear. Both signals should be connected through to all reactor substations.
The first two reactors carry the controls. The remaining reactors are just add-ons in blocks of 4. Note that this design has less heat generation then heat exchanger conversion capacity until you have 10 reactors in the row.

The Heat Exchangers + Buffer:

Getting to the meat of the design, and this part took me a while to get right. The issue is water... in combination with the requirement for the heat exchangers to be close to the reactor cluster, water in a (for practical purposes) near-infinite pattern is troublesome to get where it was. After asking around and some headscratching, I've concluded that the only viable way to maintain scalability, is to use the logistics network. Which means bots have to carry barreled-up water. That feels strange but the design works. Of course, you'll lose a small portion of the energy produced to the logistics network and the assemblers that load and unload the barrels, but the design is sound, and given that this design is built to be used in very large plants (that don't have enough at even the static 5GW designs).
The screenshot shows 2 rows. The heatpipe extends straight off of the reactors into a cluster of 18 heat exchangers. Since these rows are identical, -all- parts of the plant will have identical steam production. Therefor, the steam latch can be safely set on the first arm you make (red circuit going to the tank in the bottom row).
The water unloading stations each have, with a single speed module and 3 efficiency ones, a speed of 1.875. Combined, they unload 1875/s water at full speed. Each heat exchanger will take 100/s, for a total of 1800/s, so this lines up neatly. A water tank is used to catch fluctuations in water pressure and have a slight buffer.

I've chosen a straight 12 tank buffer per arm. More could be added, but are likely not neccesary. In combination with the heat capacity of the reactors and the partial heat capacity still usable in the heatpipe and exchangers, should be enough to buffer the energy produced in a single burst of the reactors. Pumps are used to forcefeed the turbines this steam, and to be able to act as valves between the chemical and nuclear power generation portions.
In the screenshot, the plant is on chemical power so the red signal for the pumps feeding nuclear steam is lit, and the pumps are disabled. Note that even if the plant is on chemical power, it will in parallel still fill the buffer for nuclear steam if that one is drained.

The Turbines:

Not much to say about these. 30 turbines over 3 rows, with Substations to carry the power and control signals. Pumps on both ends act as valves and feed steam into the turbines at maximum pressure. Maximum power generation is 30*5.8 = 174 MW, this is the top generating capacity per row. If fed with steam from the chemical boilers, the rows have a top capacity of 54MW. This is slightly less then the generating capacity of steam, so the buffers should always refill.
If you don't plan on using the chemical boiler at the end, you can add one more turbine per row to bring the power generation per row to 179.8MW max.

The Chemical Boilers:

An optional part of this design - but dang it, I wanted to prove that hybrid can be done so here, it IS done. Another water unloading station, configured slightly different but with the same function: To deliver 1800/s water to the boilers. Since the turbines use either chemical -or- nuclear steam, you only need 4 water loading stations per arm. I'm sending in rocket fuel by blue belt, because of it's high energy density. At 225 MJ with 40/s going across a blue belt, you could produce 9 GW of boiler steam from one blue belt of rocket fuel, which is enough for 166 boiler rows in my configuration.
The chemical boilers have some control logics of their own. They deactivate in favor of nuclear steam if:
- No chemical fuel is at the entrance of the boilers (belt is empty).
- Yellow constant combinator at the reactor controls is set
- No water in the water tank (since the nuclear part buffers steam and the boilers do not, water shortages are more problematic here)
- Power production shortages detected. This is done by measuring the level on an accumulator located in the row, if it falls below 75% the plant is considered overloaded and will switch to nuclear fuel until the accumulator is back up to 95%. Naturally, a single accumulator is not going to catch load shortages for a power plant this size, so a large battery park at a random location is recommended.

Water loading station:

To fill the water barrels. A simple design that makes use of the speed module again. Assembler 3 units can load roughly 400/s with one speed module, so you can use 3 per pump. Alternatively, you can use other means to barrel up and transport in water. Trains plus an off-site loading station is an option, but those trains would have to be either very long or very frequent. I simply put these assemblers near the shoreline and let the bots figure out the rest. Note that you need 4 water loading stations per row, and that a single offshore pump can supply at most 3.

Failure modes:
- Underload: While under low load, chemical power is used exclusively and the nuclear power parts are inactive, unless the reactor control to disable chemical power is set.
- Overload: While under more then maximum load, nuclear power is used exclusively. Brownouts may occur. If the overload is severe and prolonged, the water loading/unloading stations may end up not working fast enough causing the plant to starve itself of water, eventually blacking out. Short term overloads are not a problem due to the water and steam buffers.
- Blackout recovery: This plant does not recover well from a blackout, mainly because the steam-feeding pumps won't feed steam into the turbines initially. A small solar plant with battery backup, plus a latch that disconnects the plant if battery power is at 0%, could solve this.
- Fuel shortage, chemical: Plant switches to nuclear power.
- Fuel shortage, nuclear: Plant will use whatever steam is in the buffer to keep power generation going, then black out. Current design does not fail over to chemical power, but it wouldn't be too difficult to add this.
- Water shortage: Plant cannot generate steam. The nuclear portion assumes this to be a low buffer level and will continue to feed fuel to the reactors. This wastes fuel. Could be handled with a water level detection in the first row, and interconnecting the steam buffers to prevent partial power generation failures.

Notes:
- This plant works best if on an isolated logistics grid. It's not needed, but the damn thing has a tendency to push the empty barrels to storage - and preferably a storage far, far away. As a result, this plant probably also works best if it is built a good distance away from your main factory. If this is done, locating the uranium refinery and fuel processing locally is a no-brainer. The plant needs only some metal for fuel cell production (and barrel production), and either the chemical fuel you plant to burn if any, or oil to produce into said fuel.
- I don't consider this design as elegant as my previous 2. Using the logistics net... it feels a bit cheat-y to me, but there's simply no other way to keep it scaleable into infinity.

Bravo! I was playing with infinitely scalable designs but gave up on the water requirements. Never occurred to me to use bots, well done.

Using a train would be interesting . At 1800/s per row, assuming you would have a wagon per row, you need a train every 40 seconds, which sounds doable enough. That same train can then carry away the steam as well so you can build the turbines somewhere else. Who needs power poles when you have steam trains!?

If on a separate network, perhaps. Using trains for steam transfer might work - even for remote outposts, plop a steam buffer or two here or there and have local turbines. No power poles needed. But a problem at the rail grid would quickly escalate into a blackout. Still, it's an interesting idea... detaching the turbines from the heat exchangers alltogether, and seeing the steam as yet another resource to be distributed wherever it is needed. Water in, steam out. Problem - again - is the high volume required... about 10 water/steam per MW. I'm not sure about the power draw of mining outposts precisely... Could a few outposts be supplied by one or two twowagon steam trains?

That said, even with bots, the scaling is not infinite. At about 15 or so rows, the amount of bots that streams up and down grows so numerous that the roboports get backlogs of charging bots that just grows and grows... it slows down the bot train. Also, the amount of energy the bots demand make supplying standard boilers with water via bots... less then efficient. It works, but isn't practical. A more practical design is probably to toss the chemical boilers alltogether (and put them in a separate powerplant), move the turbines out a little further, and use 5 or so beltlines. Dunno. If just the underground pipe could reach a little further you could do the coastline - pumps - heatexchangers - reactors - steambuffer - turbines setup. It'd still be messy and probably difficult to fully utilize reactors - you'd need to convert 160*4 = 640 MW per reactor quad to steam to fully utilize the produced power. To avoid flow issues you'd need to keep each row in a separate steam pipe. Did a quick test with a setup of that kind, you can fit 5 rows per 4 reactors... so you'd need 12.8 heat exchangers per row. Given a single water pump supplies 1200l/s, rounding that down is sensible. This would support rows of 20 turbines.
It'd be a very neat design, but the underground pipe needs to be extended by just 1 tile to make it possible.

I've tried to get a train based setup working, but I'm not happy with the results yet.

Have you considered belts+barrels? A blue belt is 40 items/sec or 10k water per sec, which should supply 166 turbines. Can you side load a belt through your turbine setup? It seems like there might be space together with the substations? Alternatively, you could belt the barrels to provider chests just outside of the turbines so the distance the bots have to cover is constant?

Edit: it's a shame that you can't barrel steam, although very understandable from a realism perspective

I have considered belts, it does fit in with this setup. You need 1 belt per 5 rows in this setup (~150 turbines that way). It could work if I sneak in a service row every 5 rows, for roughly 1GW per belt of barreled water. But that's still not expandable into infinity ('though arguably, with ~1GW per arm, dualsided, I can't imagine a factory drawing more then 10ish).

Hi Aerternus,

I like your design! I have a (stupid) question, though. You call these chemical boilers. How is this chemical? Are you not using just standard Boilers in conjunction with heat exchangers to provide a mix of steam temperature to your steam turbines? I love power optimization and would love to know if I'm missing out on something

Thanks!

Fire is a chemical reaction, so the boilers are the chemical part.

COol idea with the water. Might have to steal that, curious what's the output? Got my own design i was going to share soon, curious how it compares.

Cool idea, feels a bit pointless though since providing enough chemical fuel is going to be much more difficult than nuclear

Also, the logistic delivery system is still not scalable to infinity (eventually you'd hit charging problems due to travel distance I suppose).

Nice ideas and design, any chance you could post the blueprints? Thank you!

Scratch this design off as well now that barrels don't hold as much water anymore. That's the third design of elegant power generation that gets axed by updates . Back to the drawing board... Bots are a bust!

Version 4 will likely require some geoshaping to pull off (creation of canals/waterways). The only way to make a tileable nuclear plant that can handle full load performance and be extendable to an arbitrary length, as far as I know, will be to use both sides of the reactors to convert the heat, as each 4x4 reactor block will generate 540MW, requiring 54 heat exchangers to process the heat. That's 27 on each side, but a single pump will only support 12 of them, so 24 will be about the maximum for a total generating capacity of 480MW per row.

I'm tempted to detach the chemical boilers entirely from this design, replace it instead with a railway that'll choo-choo in steam from a suitable location. Chemical boilers have chemical fuel so fueling trains shouldn't ever be an issue. But with the fluid wagon's capacity reduction even that may be an issue.

Hey Aeternus... Long time no see.

Aeternus wrote:Scratch this design off as well now that barrels don't hold as much water anymore. That's the third design of elegant power generation that gets axed by updates . Back to the drawing board... Bots are a bust!

They really should have reduced the stack size instead of what a barrel can hold... because as I remember the reason for the nerf was that a cargo wagon could hold way more fluid than a tank wagon.

Aeternus wrote:Version 4 will likely require some geoshaping to pull off (creation of canals/waterways). The only way to make a tileable nuclear plant that can handle full load performance and be extendable to an arbitrary length, as far as I know, will be to use both sides of the reactors to convert the heat, as each 4x4 reactor block will generate 540MW, requiring 54 heat exchangers to process the heat. That's 27 on each side, but a single pump will only support 12 of them, so 24 will be about the maximum for a total generating capacity of 480MW per row.

I think the lost 12% total energy output are bearable. Except if one suffers from OCD.

That said couldn't there be 3 Offshore pumps on each side instead of 2? Should be enough space there.

Aeternus wrote:I'm tempted to detach the chemical boilers entirely from this design, replace it instead with a railway that'll choo-choo in steam from a suitable location. Chemical boilers have chemical fuel so fueling trains shouldn't ever be an issue. But with the fluid wagon's capacity reduction even that may be an issue.

I would say leave it out entirely. It was a nice thing to have back when one could still use pre-heated steam... but now it is just cumbersome. Haven't tried to do it anymore ever since... but then again I was pretty pissed off about as far as I remember.



But then again I am not really that much into endlessly tileable Nuclear Power Plants anymore because of how there is never enough space anywhere for the ridiculous size such a thing has.
And if one doesn't use bots (their usage will probably getting impossible anyway when the devs finally decide to nerf the bots) one requires a mod to place watertiles, which is also ugly... or a very, very big lake/ocean... but then the landfills get tiresome.

But the ugliest thing is that they are prone to updates messing everything up... and then the layout doesn't work anymore but you are stuck with a particular position of the watertiles. Urgh.

Yea, well... with the devs making Cliffs removable, and landfill being able to convert water to land... It's a nobrainer to make some kind of demo charge to blow a canal and turn land into water. The bot thing is an issue, yes, but not one unsolvable. The design can be stretched slightly near the reactors, axing the bot stations in favor of a double belt row (fuel in, spent out). I still want to use buffer chests since this design relies on timely fuel in/spent out. A backed up belt would be an issue. Will need to put fuel production and reprocessing near the head of the plant in that case... and grow it in a line away from the base.

Yeah it's a shame, I think this was by far the most original use of bots

Since barrels are no longer really useful for water transport, I guess the only way to make a tileable design would be to have pumps on the side and move the water past the turbines to the exchangers, but if that requires two sides to be used you would somehow need an infinite rectangular piece of land with the right width...

(and yes, blasting canals would be fun, especially if they also teach biters to swim, maybe only in shallow waters and make them slower but more resistant while swimming?)