Inconsistent power output with steam engines

[cogwheel]

Discovered here: https://www.reddit.com/r/factorio/comme ... er/d6xulx4

Set up: 8 water pumps. Each pump feeds into 14 boilers. The outputs of two boiler lines are combined directly into pipes. An underground pipe connects that to a line of 20 steam engines. There are 4 equivalent setups side by side for 80 engines total. A massive array of laser turrets are placed to provide constant power drain.

Filling every boiler with rocket fuel, the power output is the expected 40.0 MW

Replace the last boiler with a pipe in every other line, turning it into a 2-27-20 and the power production increases to 40.6 MW.

Replace the other final boiler so it's 2-26-20, and it's 40.5 MW. It is an expected decrease from the previous one, but still noticeably higher than with all the boilers.

Remove one more boiler, and the production finally dips below 40

[cogwheel]

Actually, you don't need to alternate the removed boilers. If I make two of the lines 2-26-20 and the other two lines 2-28-20, I get the same 40.6 MW as having all 2-27-20s

Here's a screenshot with the setup and the power display:

[Loewchen]

This is not enough for it to be a bug, chances are the limiting factor is water throughput and not heating power so removing boiler would increase the power output.

[cogwheel]

This is not enough for it to be a bug, chances are the limiting factor is water throughput and not heating power so removing boiler would increase the power output.

You're right. I was under the mistaken impression that boilers output 500 KW. Seems they're 510 instead.

[cogwheel]

Actually, that still doesn't quite make sense.

I have 4 lines of 2-28-20 producing pretty much exactly 40 MW (with occasional blips to 40.8). These are completely independent lines, with no water crossing paths. If I turn any two of them into 2-26-20 lines, the total power output increases to steady 40.6. If I turn all four of them into 2-26-20 lines, the total power output reduces to 39-ish.

There is definitely something not making sense here.

[orzelek]

Actually, that still doesn't quite make sense.

I have 4 lines of 2-28-20 producing pretty much exactly 40 MW (with occasional blips to 40.7). These are completely independent lines, with no water crossing paths. If I turn any two of them into 2-26-20 lines, the total power output increases to steady 40.6. If I turn all four of them into 2-26-20 lines, the total power output reduces to 39-ish.

There is definitely something not making sense here.

I'd still bet on water flow.
Line of 20 steam engines is pretty long and fluid flow can be quirky. And you are merging output of 2 pumps into one common line for steam engines.

Try adding 4 parallel water pumps between boilers and steam engine line and see if it changes.

[cogwheel]

I will try that. For what it's worth, though, the end boilers are all have full production bars and > 1 water

And again, I only get the full increase in power if I leave boilers in some of the lines. If it was a universal flow issue, then turning them all into 2-26-20 should not result in a decrease.

[cogwheel]

I agree it's a flow issue, but I still think there is something wrong with the behavior.

If I go back to the 2-28-20 setup and put 4 pumps on one of the boiler outputs, the power output does tend to stay around 40.7. However, doing the same with any of the other lines does not result in a similar increase.

In fact, the amount of water reported in all the engines is much less than it was when the power output was 40.0. Most of them are less than one, many say 0.0 or 0.1, but still show performance at 100 % and temperature at 100 degrees. O.o With the original setup, the engines at the end all had around 5-7 water in them at 100 degrees suggesting it is *not* a flow issue.

Adding a 5th pump brings the water levels back up, but does nothing to the power output.

[cogwheel]

Please consider moving this back out of "not a bug"

Yes, it does seem to be a flow issue, but the issue is not due to anything I constructed. Attempts to increase the flow of a single line seems to change the power output for a whole set of lines, even when that change decreases the overall flow. Everything I've looked at suggests the flow is just fine. All of the engines have plenty of water at 100 degrees and yet, with any of a number of minor modifications to a single line of a 4 line system, the whole system changes in an unexpected way.

[cogwheel]

My best guess at this point is that there are periodic blips where power production seems to be less than max, but those blips line up with the power UI refreshing. So the production may be 40.8 most of the time (the theoretical maximum) but the UI is displaying 40 with an occasional blip up to 40.8. By making a minor change to the system, the power actually goes down from 40.8 to 40.7, but the UI displays it as 40.7 with the occasional blip down to 40.0.

This is the simplest kind of bug that would explain every behavior I'm seeing.

[XKnight]

The answer on the intial question is weird fluid mechanics.
Each pipe replace operation change order in which entities will be proccessed, as a result this has influence on the fluid flow.
The same behaviour was observed some time ago viewtopic.php?p=126066#p126066
As for me, this bug is "by design".

[Loewchen]

The topic was moved by kovarex.
If we manage to pin point the cause and can prove that it is not just fluid mechanics (like a pressure oscillation) or rounding issues I will reopen it. But even if so, it will most likely just end up in minor or wont fix after all so don't get too exited

[cogwheel]

Fair enough