Factorio Forums

tl;dr
I've found information about 30 units per second, per pump, and only increasing total pipe throughput beyond 30/s when using pumps in parallel, but I'm still thoroughly confused about the mechanics of small pumps, especially in series, because rate of transfer increases when I use pumps in series, in both long and short pipe-distances. I have zero mods.

For instance, when connecting an empty storage tank to a full storage tank without any pumps and at a distance of one pipe, it gradually equalizes. If I place a small pump where the pipe is, it speeds up. If I start from scratch with a full tank and place an empty tank at a few pipes' distance, connect it, and gradually swap in small pumps in series, it further increases the rate of transfer. The rate of 30/s is either never met with one pump or it goes beyond that. Because I'm not familiar with the complete mechanics of the game or how to properly test and measure, I don't understand what is going on or how to apply it to the rest of the game, so the usage of parallel pumps doesn't make any sense to me, because the easiest (accidental) test for me shows that a noticeable increase exists with pumps in series well before reaching equilibrium between the tanks, so pressure relative to height of the fluid vs pump doesn't seem to play a role. To further complicate things (if pressure changes as a result of height & gravity don't play a role), what mechanics (if different from the full tank being on the input side) still allow a nearly-full storage tank being filled by a nearly-empty storage tank and one pump, to be accelerated by additional pumps? Transferring those last few units can take some time, and it speeds up with pumps in series.

I've read about the throughput for various distances of pipes which makes the above situation very confusing for short distances. I also read someone mentioning that "pressure" is visually shown to us, the users, as a mathematical result of various, singular points of mathematical input, rather than being calculated in real-time for distances, turns, heights, pumps, and other dynamics affecting throughput. That makes sense, because that would be a lot of physics to account for in real-time. But, if it's been simplified, what is allowing me to seemingly notice improvements of both long and short pipelines' throughput by using pumps in series in relatively simple systems? It certainly seems like long pipelines fill a storage tank faster with more pumps in series, but for many reasons of game mechanics I've read, I don't understand why this is so.

I'm not an engineer, but I'd still like to know which elements of fluid dynamics (real or fabricated) are in play, because the simple explanations I've read are contrary to my experiences, and I'm building "wrong" systems as a result of pumps in series, but it works for some reason.

Jon8RFC wrote:tl;dr
I've found information about 30 units per second, per pump, and only increasing total pipe throughput beyond 30/s when using pumps in parallel, but I'm still thoroughly confused about the mechanics of small pumps, especially in series, because rate of transfer increases when I use pumps in series, in both long and short pipe-distances. I have zero mods.

Parallel not Series


Simple side, Pumps force whats behind them in front of them...

In series you are just playing leapfrog and there is no performance gain really. Think of it like replacing yellow belt with red belt....Sure everything catches up for a minute because you SHOVED it down the line but it levels off to what you can produce eventually.

In parallel, pumps share a pipe in and a pipe out...The combined power is additive more or less.


I think you are confusing the unassisted rate of transfer. It might actually be slower than one pump depending on the pressure. In such a case, putting a pump in will move things forward more quickly. For your test though I think you are limited by the throughput of your storage tanks. I Think they max out at about 10/s

Small pump throughput is usually tested using steam engines, because steam engines consume a predictable 6 fluid/s (when under full load, or consuming cold water). You can thus measure actual fluid throughput by seeing how many steam engines will run or how much power is generated:
flow = Power_in_MW * 6 / 0.51

There are a few other ways to measure flow, including scripting, but steam engines is definitely the most straightforward.

One thing about small pumps is they are effective at pulling liquid out of a storage tank. Storage tanks only let go of their contents begrudgingly but a small pump will forcefully remove 30 liquid/s until the tank is empty. This also happens to a lesser degree if there is a pipe between the tank and the small pump, a tank will release more of it's contents into an empty pipe than a half full pipe, the small pump keeps the pipe empty and promotes more flow from the tank. Really the fluid flow is all about differentials (liquid moves from a pipe which is more full to an adjacent pipe one which is less full with the rate depending on the difference in how percentage full each is), but there are also some wacky performance optimizations thrown in to make it faster, which also makes it more confusing.

Ah, yes. Unassisted rate is drastically lower even at full conditions, and I didn't anticipate that to be the case. I played late at night over the weekend and I must've been more sleepy than I remembered when I was tinkering, and I spent a few days searching for info (which makes sense why I didn't find it) before posting. Fired up my computer today and even at a quick glance, the rate of one vs more pumps in series is static, and I don't know how I imagined it otherwise. In my test area, I had 8 pumps in series between two tanks, and it's very obvious that two or more in series don't flow more than one, and I have no idea how I decided that it was. Dumb.

Heh, sorry for wasting your time, but thanks for the info on the flow rate of storage tanks and steam engine testing!