So here's my conundrum: I want to make a large, tileable, basically infinately expandable nuclear plant. That plant needs... Water. And lots of it. I'm trying to figure out the best ways to get that water where it needs to go, but am running into trouble.
The stack I am using has 18 heat exchangers per row (180MW, 1800 water/s needed). Since I'm still (possibly futilely) trying to incorporate both chemical and nuclear power into one plant, these rows are wide. REALLY wide. Over 3 screenlengths wide. So anything designed has to become a part of the row itself. The rows are 9 deep (max 10 deep, 2 reactor widths)
Here's what I've tried thus far:
- Pipes / close to the shore. Was my old setup, but the requirement to keep heat pipes short makes this impossible to maintain. Putting the turbines across from the reactors also doesn't work - underground pipes just dont have the reach - and I want to build on both sides of the reactor row anyway.
- Pipes / long distance. Quickly dismissed the idea of piping it in since even with pumps, the parallel pipe mess would spiral out of control very quickly. Also, long distance pipes have flow issues.
- Belt/Water in barrels. Blue belt has a throughput of 40 units/second. At 250 liquid/barrel, that means a max throughput of 10000 liquid, constant, across any distance. That's enough for 5 rows (or 900 MW). 4 Assembler 3 plants with one speed module each can unload water at 1875/s, this can be put in the row as the water generating source. Decent, but still not expandable, it limits the design to 5 rows (1.8GW total)
- Barreled water / Logistics bots. Has potential but the system has an annoying tendency to jam under load. Using active provider chests to push away the empty barrels works, but that forces bots to bring them to storage, far away. Isolating the logistics grid of the power plant is an option but makes the whole thing not as easy to use, and also requires special handling for fuel rods and spent fuel.
Other ideas, not tried yet:
- Every 5 rows a train station to unload/load barrels. Then belt it from there. A good idea in theory, but it'd mean a dual homed, short range train, squeezing in a narrow train station... also that train would need to be very frequent or at least 4 wagons long. At 40 barrels per second, it'd only take 12 seconds for a chest (10 seconds for a wagon) to be completely emptied. Fluid wagons are even worse, the content of those would be consumed in ~8 seconds at full load.
- Every 5 rows, split off a new belt all the way to the end (1 in, 1 out). This would cause some belt madness at the edges, but far less then the piping madness.
- Pure logistics grid, but isolate the powerplant logistics grid from the rest of the factory. This is different from my usual style where I stick to a single logistics grid, but the need to have separate storage kinda forces the grid isolation. Would also mean that the water loading sections need to be on that separate grid.
Anyone else got any brilliant ideas on how to get a scaleable amount of water (1800 per row) across any length? I'm open to suggestions, as this is probably the last piece of the puzzle to my new nuclear plant design.
[Edit] I know there's mods that have water generating structures, like WaterWell. But I want to stay vanilla for this one.
Watery woes: Looking for help on water throughput
Re: Watery woes: Looking for help on water throughput
These are the methods I would use. Utilize pipes near the shore as much as possible (use landfill to make a nice long, straight shoreline). This should get you through a bunch of reactors if done right. But for a more infinite solution you'd want barrels + bots.Aeternus wrote: - Pipes / close to the shore. Was my old setup, but the requirement to keep heat pipes short makes this impossible to maintain. Putting the turbines across from the reactors also doesn't work - underground pipes just dont have the reach - and I want to build on both sides of the reactor row anyway.
- Barreled water / Logistics bots. Has potential but the system has an annoying tendency to jam under load. Using active provider chests to push away the empty barrels works, but that forces bots to bring them to storage, far away. Isolating the logistics grid of the power plant is an option but makes the whole thing not as easy to use, and also requires special handling for fuel rods and spent fuel.
In my setup I have everything related to nuclear power all in one place one an isolated logistics grid. I bring in iron plates and uranium ore by train and also use a train to export excess U-235 and U-238 for ammo/bombs. I have a bunch of centrifuges to process ore, run kovarex enrichment, and also an assembler to make fuel cells. It should be a simple matter to add water barrels to the setup.
Re: Watery woes: Looking for help on water throughput
I'm looking for an universal solution. IE no "pipe these few things, then use that for the rest". It might seem the best way, but it gets in the way of a uniform, repeatable pattern.
But I guess, if you want to keep it expandable into infinity, the logistics network is the only system that will make that happen. Sort of, if you expand too deep you run into queues at the roboports. But 40/s at load means 10 bots/s making deliveries per 5 rows, which shouldn't overload the logistics grid at all. Even 3 or 4 rows deep (and with 900MW per such row, and the ability to mirror it to the other side of the reactors, I very much doubt you'd need to expand any further then that).
But I guess, if you want to keep it expandable into infinity, the logistics network is the only system that will make that happen. Sort of, if you expand too deep you run into queues at the roboports. But 40/s at load means 10 bots/s making deliveries per 5 rows, which shouldn't overload the logistics grid at all. Even 3 or 4 rows deep (and with 900MW per such row, and the ability to mirror it to the other side of the reactors, I very much doubt you'd need to expand any further then that).
Re: Watery woes: Looking for help on water throughput
I guess my thought was if there is a shoreline close enough you may as well use pipes for whatever you can reach and bots for the rest. Of course I am the opposite of OCD, so I don't mind mixing things upAeternus wrote:I'm looking for an universal solution. IE no "pipe these few things, then use that for the rest". It might seem the best way, but it gets in the way of a uniform, repeatable pattern.
But I guess, if you want to keep it expandable into infinity, the logistics network is the only system that will make that happen. Sort of, if you expand too deep you run into queues at the roboports. But 40/s at load means 10 bots/s making deliveries per 5 rows, which shouldn't overload the logistics grid at all. Even 3 or 4 rows deep (and with 900MW per such row, and the ability to mirror it to the other side of the reactors, I very much doubt you'd need to expand any further then that).
Re: Watery woes: Looking for help on water throughput
Any particular reason you're using stacks of 18 exchangers? In 2 by X arrangement, corner reactors will produce 120MW, while the middle reactors will produce 160MW, which can be equaled by 12 and 16 heat exchangers respectively. I've made just such a design, and it's definitely possible with a little help from some pumps.
Re: Watery woes: Looking for help on water throughput
The rows are 9 wide, so you're going to have a surplus of power generation no matter what. I chose 18 heat exchangers because that lined up nicely with the 1800/s water transfer rates of 4 assembler 3's with a speed module and 3x efficiency (max power savings there too). I would have done more, but then you're getting flow capacity issues again.
Re: Watery woes: Looking for help on water throughput
I suppose. Although given that pumps are 30kW/each, I'm not sure how hard you'd want to focus on efficiency at 1800/s water transfer If you haven't seen this, it might be worth a look. In a nutshell, at 1200/s fluid flow, you can go about 16 pipes between pumps, at 1500/s, you're down to 6 pipes/pump, and at 1800/s, you're good for only 3 pipes/pump. Feels like you might be hitting a pretty hard wall, that's all.
That said, you could probably extend this kind of design a couple extra exchangers (16 and 12 exchanger versions shown) if you added some extra pumps on the steam output side? Not sure if even that would be able to give you a useable 1800/s, though.
If you haven't seen this, it might be worth a look. In a nutshell, at 1200/s fluid flow, you can go about 16 pipes between pumps, at 1500/s, you're down to 6 pipes/pump, and at 1800/s, you're good for only 3 pipes/pump. Feels like you might be hitting a pretty hard wall, that's all.That said, you could probably extend this kind of design a couple extra exchangers (16 and 12 exchanger versions shown) if you added some extra pumps on the steam output side? Not sure if even that would be able to give you a useable 1800/s, though.
Re: Watery woes: Looking for help on water throughput
If vanilla had a way to artificially create canals like we can do landfill it'd be a nonissue, you could just plan pumps somewhere as in your screenshot. That'd be my preferred solution but it's just... not possible at present. Also, heat propagation from the reactors becomes an issue. At 32 distance from the reactor I'm already seeing a 300 degree difference between reactor temperature and the last heat exchanger... Difficult to build out wider if you intend to use the reactors as heat capacitors to a degree.
Re: Watery woes: Looking for help on water throughput
I guess I'd need a better idea of what your exact goal is for the form factor. It was an interesting problem, regardless, so I hashed out this little contraption which I *think* might be useful to your requirements. The design is too long to post a screen of, but see what you think?
That said, if you're going to run off both sides of the 2xN reactor strip, I'd remove a couple exchangers/turbines since this will overdraw on heat otherwise. I used 18 exchangers as proof of concept, but it seems reliable.
LongReactor is Looooong
180MW output. Maximum 5 spaces wide (1 reactor width) and no interference when stacked next to each other. It takes in 1800/s water but splits it across two pipes, so lots of flexibility for water input (shore pumps included in blueprint, but could easily be adapted to other sources) Even after the buffered steam in the turbines runs out, it holds it's sustained output (I let it run for 20 minutes with no issues). Just give all the heat pipes a chance to get up to roughly 600+ before putting it under full load.That said, if you're going to run off both sides of the 2xN reactor strip, I'd remove a couple exchangers/turbines since this will overdraw on heat otherwise. I used 18 exchangers as proof of concept, but it seems reliable.
-
iceman_1212
- Filter Inserter
- Posts: 256
- Joined: Wed Aug 17, 2016 9:49 am
- Contact:
Re: Watery woes: Looking for help on water throughput
Might be able to achieve this by taking the inverse approach: landfilling over large landmasses and leaving a strip of water for pumps.Aeternus wrote:If vanilla had a way to artificially create canals like we can do landfill it'd be a nonissue, you could just plan pumps somewhere as in your screenshot. That'd be my preferred solution but it's just... not possible at present. Also, heat propagation from the reactors becomes an issue. At 32 distance from the reactor I'm already seeing a 300 degree difference between reactor temperature and the last heat exchanger... Difficult to build out wider if you intend to use the reactors as heat capacitors to a degree.
(Personally, fluid flow rates are the one aspect of this game where I don't feel comfortable min-maxing. I typically budget only ~800 l / s along a pipeline up to a maximum of 1000 l / s.)
-
SpeedDaemon
- Fast Inserter
- Posts: 124
- Joined: Fri May 22, 2015 3:31 pm
- Contact:
Re: Watery woes: Looking for help on water throughput
My instinct on how to solve this would be to design the largest plant practical without having a lot of water supply problems, then "tile" it by smacking those down next to the nearest convenient body of water when I need more power. Use a train to deliver fuel/remove spent fuel.
Depends on whether the small efficiency loss from having several 2xN arrays rather than one huge 2xN array is acceptable or not. Every time you need to "break" the line and make a new plant, you basically only lose 160MW, which really isn't much if you're going this big. (1.2GW for a 12 reactor site instead of 1.4).
Depends on whether the small efficiency loss from having several 2xN arrays rather than one huge 2xN array is acceptable or not. Every time you need to "break" the line and make a new plant, you basically only lose 160MW, which really isn't much if you're going this big. (1.2GW for a 12 reactor site instead of 1.4).