Nuclear Reactors in 2.0

[kremlin]

Currently, to make a large scale or tiling reactor, you have to use a lot of pumps to keep your water pressure up to get the water where it needs to go. Many of those designs are meant to be plopped onto a landfilled lake to allow the offshore pumps to feed directly. The fluid update in 2.0 would seem to significantly reduce the need for those patterns. You can get a couple of offshore pumps worth of water fed in through a pressurized line with pumps every few pipes, but you can't get the peak throughput unless you do unbroken lines of pumps which isn't really feasible. I'm thinking the new system fixes that problem, and lets us saturate a pipeline by attaching 5 or more offshore pumps to it, without needing the pressure pump lines.

Will have to wait until it drops and experiment with it to be sure, but I think this is a really good thing for Big Nuclear.

[Fletch]

That's what I understood as well from Fluids 2.0 FFF -- pipes will have infinite throughput (same concept as electricity in factorio) which will simplify all the fluid designs.

[PunCrathod]

If there are no throughput limiting mechanisms that were not talked about in the FFF then the only throughput limit for these new pipes is the total capacity of the pipe. If the total capacity of the pipe is larger than the total amount all the connected consumers can consume then the pipe will never be the bottleneck.

[Fel]

Not infinite, but no longer defined by the pipe's length, which means that mods will be able to make pipes with higher capacity with an effect on throughput.

[Fletch]

Originally posted by PunCrathod:

If there are no throughput limiting mechanisms that were not talked about in the FFF then the only throughput limit for these new pipes is the total capacity of the pipe. If the total capacity of the pipe is larger than the total amount all the connected consumers can consume then the pipe will never be the bottleneck.

yes. I should've qualified: as long as the pipe is full (and there is a massive buffer to fill in them), then the throughput goes to infinity. This should mean that as long as "production > consumption" then the pipe will eventually be full. It is only the partially full pipes will be throughput limited (from what I gathered from the FFF) -- you should be able to plug in as many producers and consumers into the same pipe as you want (as long as production stays greater than consumption).

This is similar to having your entire megabase nuclear reactor farm sending all its power out on a single wooden pole (what you can do in the game now). It's totally unrealistic, but it is simpler for the player and computationally for the PC.

[Hurkyl]

I don't think pipe capacity is relevant at all -- except for contributing to the total volume of a pipe network.

The notion of throughput ceases to be a concept for the logistics of fluid transport in 2.0. As long as producers are connected to consumers, it gets there.

(okay, technically the concept would apply to the use of pumps to transfer fluid between two adjacent networks)

[Fletch]

Originally posted by Hurkyl:

(okay, technically the concept would apply to the use of pumps to transfer fluid between two adjacent networks)

But why would you need to do that if throughput is infinite? You will be able to run your entire base off a single water pipe on the "main bus" -- as long as you enough producers creating more supply than there is demand into that pipe segment.

I have a love/hate feeling with the new FFF. I like that it is simpler (and computationally cheaper), but I prefer a little more "realism" where you'd need to use more than 1 pipe -- needs something to constrain overall throughput...

[kremlin]

Inline pumps operate at 12,000 units per second, that's the figure I'm using as the theoretical maximum throughput of a pipeline. I think the practical maximum is closer to 3,000 per second, fed by three offshore pumps with very few pipes connecting them to the first pumps in the pressure line. That gives you enough pipe segments between pumps to add turns to the pressure line while keeping 3,000 per second.

If the maximum throughput of the pipeline in 2.0 is the same as this theoretical maximum of 12,000 units per second and we are able to reach the maximum flow, then it will take a quarter of lines for the same size of reactor.

[knighttemplar1960]

Originally posted by Fletch:

Originally posted by Hurkyl:

(okay, technically the concept would apply to the use of pumps to transfer fluid between two adjacent networks)

But why would you need to do that if throughput is infinite? You will be able to run your entire base off a single water pipe on the "main bus" -- as long as you enough producers creating more supply than there is demand into that pipe segment.

I have a love/hate feeling with the new FFF. I like that it is simpler (and computationally cheaper), but I prefer a little more "realism" where you'd need to use more than 1 pipe -- needs something to constrain overall throughput...

IRL you would need larger diameter pipes. If you wanted to simulate that in Factorio you would need tiers of pipes like with belts maybe tier 2 would require steel. Tier 3 would require steel with copper jackets, etc. but that would just add complexity that they are trying to eliminate and you would have to go back to looking up fluid flow rates on a a wiki, something they wanted to avoid.

[knighttemplar1960]

Originally posted by kremlin:

Inline pumps operate at 12,000 units per second, that's the figure I'm using as the theoretical maximum throughput of a pipeline. I think the practical maximum is closer to 3,000 per second, fed by three offshore pumps with very few pipes connecting them to the first pumps in the pressure line. That gives you enough pipe segments between pumps to add turns to the pressure line while keeping 3,000 per second.

If the maximum throughput of the pipeline in 2.0 is the same as this theoretical maximum of 12,000 units per second and we are able to reach the maximum flow, then it will take a quarter of lines for the same size of reactor.

That would be easy to game. Connect 10 off shore pumps to a set of interconnected tanks and then 1 electric pump to the tank for output.

[Fletch]

Originally posted by knighttemplar1960:

IRL you would need larger diameter pipes. If you wanted to simulate that in Factorio you would need tiers of pipes like with belts maybe tier 2 would require steel. Tier 3 would require steel with copper jackets, etc. but that would just add complexity that they are trying to eliminate and you would have to go back to looking up fluid flow rates on a a wiki, something they wanted to avoid.

Yes, pipe tiers would be nice. Fits inline with belt tiers and electric pole tiers. They could "fake" the throughput with a cheap easy to implement gimmick: only allow XX consumers (or producers -- either one) to be connected to the same pipe segment. It could just show a Red "can't build here, too many pipe consumers" if player tries to violate the constraint.
They do this today when you try to connect an existing oil pipe to an existing water pipe (the game prevents you from accidentally "mixing" fluids).

Then go with your higher tiers idea to allow more consumers/producers per pipe tier. Alternative to tiers is the player just needs to run an additional pipe segment.

[Hurkyl]

Originally posted by Fletch:

Originally posted by Hurkyl:

(okay, technically the concept would apply to the use of pumps to transfer fluid between two adjacent networks)

But why would you need to do that if throughput is infinite?

Because you want/have separate fluid networks and to transfer fluid between them. E.g. to isolate heavy/light oil cracking so they don't consume too much fluid, just like you do in the base game.

Or maybe the numbers work out so that this is relevant

Originally posted by FFF #416:

Machines can push fluid into a segment at an unlimited rate, and can pull from a segment at a rate proportional to how full the segment is.

and you want to separate your fluid buffers from your production machines, so that production is always working with full pipes.

Originally posted by kremlin:

If the maximum throughput of the pipeline in 2.0 is the same as this theoretical maximum of 12,000 units per second and we are able to reach the maximum flow, then it will take a quarter of lines for the same size of reactor.

No maximum throughput. Except maybe something you can only insert/remove "the entire volume of the entire fluid network" per frame.

[PunCrathod]

Originally posted by Hurkyl:

No maximum throughput. Except maybe something you can only insert/remove "the entire volume of the entire fluid network" per frame.

As far as the FFF goes this is how it works. One tile of pipe can hold 100 units of fluid. Which means you can have 5 offshore pumps per one tile of pipe. Or 300 non moduled non beaconed oil refineries per one tile of pipe. It is literally impossible to have enough offshore pumps to max out a pipe. For refineries you would need to make them over 300 times faster to max out a pipe. People who are concerned that a longer pipe will increase the max throughput of that pipe need not bother. Even a one tile long pipe is faster than anything you could throw at it. Unless they seriously nerf the capacity of pipes that is. Which I hope they do not even attempt because the solution would always be to add a long appendix somewhere and voila bottleneck solved. Better to just leave it high enough that you don't need to think about it.