pumps seem useless for long distance pipes?

[llVIU]

Help me understand this. https://wiki.factorio.com/Fluid_system

Let's say I have 1000 sections of pipes, and at the end I put a pump... that pump will be worthless! As the pipe gets longer, the fluid speed decreases... but the pump cannot raise the speed, because it receives a low amount of fluid. You can put a tank but then that tank gets filled at a slow speed so the pump will still pump out slowly. The speed would only suddenly increase if there was a gap of time where there was no requirement for fluid and the tank got full.

This seems completely contradictory to how pumps in real life act. They raise the speed of the fluids, but the ones in factorio don't. So why would I bother using a pump then??

Currently, I only use them with green/red wires to stop or start the flow of fluids with certain conditions, or stop fluids from going into one direction.

[quyxkh]

Any kind of fluid sim running on a single 50-watt core (let alone sharing that core) is going to be a bit crude when compared with reality. Real fluid sims eat top-500 petascale computers burning megawatts and even they can't get it right because it's literally chaotic.

Add pumps between ug pairs at whatever interval keeps your desired flow rate. Every three UG pairs is often enough.

[llVIU]

I have literally no idea what you're talking about

[eradicator]

This seems completely contradictory to how pumps in real life act. They raise the speed of the fluids, but the ones in factorio don't. So why would I bother using a pump then??

The pumps work fine. It's the pipes that are "unrealistic". In reality a perfectly straight pipe would eventually end up with the same fluid mass on both ends, no matter how long the pipe. In factorio this is not the case, adjacent pipes can magically contain different amounts of fluid mass. And pumps can't "fix" that loss in mass, same as you can't use a pump in real life to turn a bucket of water into an ocean.

[llVIU]

so how would you fix the problem of long pipes? Because eventually, you end up with a very low fluid speed on the pipe. And pumps don't help.

[Koub]

The best thing to do is wait for the fluid physics update, which is expected after the first 0.17 stable is out (which in turn should be out quite soon).

[llVIU]

I didn't hear about any of this, is there an article or such that I can read about?

[eradicator]

so how would you fix the problem of long pipes? Because eventually, you end up with a very low fluid speed on the pipe. And pumps don't help.

Use shorter pipes. Underground pipes are calculated as only two pipe segments, and thus suffer far less on long distances than straight over-ground pipes. Then put a pump after every ~10 pairs of underground pipes and you'll get pretty decent flow. If that's still not enough use a mod with longer underground pipes.

The best thing to do is wait for the fluid physics update, which is expected after the first 0.17 stable is out (which in turn should be out quite soon).

That could be anytime from within a few weeks to next year though :p. Enough time to read the thread about fluid mechanics.

[llVIU]

''use shorter pipes'' is the same solution as ''I need more green cards to make blue cards, what is a good blueprint to make more green cards?'' and you tell me ''make less blue cards''.

Okay, so if I use less pipes, that means I have far less oil rafineries, which means I have far less of every single thing that can possibly use oil.

But seriously, weren't pumps meant to fix the issue of having long stretches of pipes?...

p.s. thanks for the link to the post, I'll read it

From the link that you gave me I understand that
-putting a boiler, then pipe, then boiler is bad. It cuts off a lot of the water flow speed
-putting pipe then pump is useless, unless you put a chain of pumps with a max of 1 pipe between each, that way they don't flow speed but they don't gain any either
-pump going in tank directly is good, it fills up the tank directly, and pump going out of the tank is good, it pumps out the liquid very fast with 200 pressure speed.. but pipes have 100 max pressure speed so as soon as I use one single piece of pipe, that pump is useless. Unless I somehow have 2x pipes that connect to 1 pump, then the pump adds up the low pressures from both pipes and can fit it into one single pipe. Like 2 pipes with 50 pressure (because of long distance), put 1 pump, then you have 1 single pipe with 100 pressure.

[disentius]

read better:)
There is a table on the wiki page you mentioned. It gives the max troughput in u/s for pipe-x(pump)-pipe configurations.
Example: if you use underground pipe-pump-underground pipe, you can maintain a flow of 3000 u/s indefinitely.

[llVIU]

indefintely? no, it does say in the article that it counts as a minimum of 2 pipes, the one that goes inside the ground and then the part that goes outside. Only the parts inside the dirt don't count because the fluid teleport from one pipe to another.

so if you have an oil field and put a whole bunch of underground pipes, you still lose fluid speed over distance.

[astroshak]

“One” Pipe to Ground consists of the pipe segment that goes into the ground, and another that comes up. Consider them matched pairs.

To enhance fluid flow over long distances, use pumps every few pairs of Pipe to Grounds.

Pump->Pipe to Ground down-space gap-Pipe to Ground up->Pipe to Ground down - gap - Pipe to Ground up -> Pipe to Ground down - gap - Pipe to Ground up -> Pump

And repeat. Its not a matter of using shorter pipes or less pipes. Its USE MORE PUMPS.

Currently, I believe that the max sustainable flowrate through a pipe is 1400ish. Pumps are capable of a lot more, provided they have sufficient input. So, the highest attainable flowrate over long distances comes from Pump->Storage Tank->Pump->Storage Tank->Pump->Storage Tank->Pump, repeated as long and as far as necessary. Note that this turns into a rather fragile wall that you are not getting past without certain mods, but it is useful in certain circumstances.

[jcranmer]

Let's say I have 1000 sections of pipes, and at the end I put a pump... that pump will be worthless! As the pipe gets longer, the fluid speed decreases... but the pump cannot raise the speed, because it receives a low amount of fluid. You can put a tank but then that tank gets filled at a slow speed so the pump will still pump out slowly. The speed would only suddenly increase if there was a gap of time where there was no requirement for fluid and the tank got full.

Put a second pump in the middle of the pipe. Now your maximum flow doubles. Put more pumps in the middle, and you can increase the flow.

This seems completely contradictory to how pumps in real life act. They raise the speed of the fluids, but the ones in factorio don't.

It's actually decently close to how pumps in real life work, especially if you treat the pipe's current fluid value not as a percentage full/empty but as an indicator of the current pressure in the pipe. In real-life pipeline systems, the pipe is pressurized, and it is always full of fluid, with the volumetric flow rate determined primarily by the pressure differential at the input of the pipe. Pressure loss (or head loss, as it's usually referred to) is proportional to the length of a pipe, as the pipe walls provide resistance to the flow of motion, and so you need pumps every now and then to keep the fluid moving. Furthermore, the amount of pressure a pump can add to a pipe is dependent on what the pressure of the output is already at, so that two pumps that could each pump 60 MGD by themselves might push around 100 MGD together. (Naturally, all the values involved are horribly non-linear and difficult to compute from first principles because fluid dynamics hates us, so values are best determined by experiment).

Single pumps moving fluids long distances don't happen like that in real-life. The Trans-Alaska Pipeline doesn't use a single massive pump at Valdez to move the entire pipeline, it has pumps every 70 miles or so to keep the crude flowing.

[5thHorseman]

Is there a reason no one is suggesting that instead of trying to create the Trans-Alaska Pipeline you instead load that oil onto a train?

[Oktokolo]

Is there a reason no one is suggesting that instead of trying to create the Trans-Alaska Pipeline you instead load that oil onto a train?

That is too easy and would be way too flexible when expanding to other oil fields further away.

[Pi-C]

Is there a reason no one is suggesting that instead of trying to create the Trans-Alaska Pipeline you instead load that oil onto a train?

That is too easy and would be way too flexible when expanding to other oil fields further away.

You could always use the trunk of your car to fit in a load of barrels …

[quyxkh]

Is there a reason no one is suggesting that instead of trying to create the Trans-Alaska Pipeline you instead load that oil onto a train?

That is too easy and would be way too flexible when expanding to other oil fields further away.

You could always use the trunk of your car to fit in a load of barrels …

It's how I launched my first rocket, setting up auto-load/unload stations I could just drive to, stop at, drive off was fun. That was like 12.33 days though, before the barrel capacity nerf.

[Pi-C]

You could always use the trunk of your car to fit in a load of barrels …

It's how I launched my first rocket, setting up auto-load/unload stations I could just drive to, stop at, drive off was fun. That was like 12.33 days though, before the barrel capacity nerf.

Must have been a bit more than just 2 weeks ago, though …

Anyway, I'm doing something similar right now: driving a car with a huge trunk to pick up and deliver raw materials to my base. As soon as I'll have bots, this will be automated with the Autodrive mod -- kind of a poor man's version of AAI Programmable Vehicles where you can program vehicles to path to defined points and wait there until they have been loaded/unloaded by bots. However, I realize that this will turn cars and other vehicles into something very similar to trains -- but it seems to be a good idea for covering shorter distances where setting up a rail network would be overkill.

[Oktokolo]

-- but it seems to be a good idea for covering shorter distances where setting up a rail network would be overkill.

It is never overkill to set up a rail network.

[Pi-C]

-- but it seems to be a good idea for covering shorter distances where setting up a rail network would be overkill.

It is never overkill to set up a rail network.

Right, I didn't mean to use cars instead of, but in addition to a rail network. Say, you have two ore patches far away from your base and every other outpost, but comparatively close to each other. In this case, it could make sense to set up just one train station at the one patch and use a car to bring the ore from the second patch there. You could extend the rails to the second ore patch, of course, but this would mean to provide a bypass at the first station so you could use different trains simultaneously, or make sure that the train cars have reserved slots so the train can be loaded at both stations. Also, variety in transport makes for more interesting playing.