Bloody pipes

[Indifibulator]

OK, I know pipes have been an issue for a lot of people, but up until now I have been fine with them. Before coming here I have read the wiki and the pipes manual, but nothing is working.

In short: I pump water up quite a bit to a factory (already a bad idea with the way pipes work) but it starts up and runs fine. The pipes all fill and supply the factories. Then... it stops. After a few minutes I notice some backflow happening right at the entrance to the factory manifold and the water stops flowing entirely. The pipes at the manifold run dry. The vertical pipes are full up to the top one. The three mk 2 pumps each report a head lift of between 45 and 48 meters, so just as they should be, but there isn't flow.

I flush the system and the whole thing happens again. I have heard that if there is sloshing that goes back to the water extractors, they can get bugged and stop working, so stuff I have tried that doesn't work:

-Upgrade all pipes in the system to mk2, even ones that don't carry that much water.
-Put valves right in front of each of the two water extractors set to max flow to prevent backflow.
-Install a pipe loop from the beginning of the factory manifold to the end.
-Install a valve set to max flow right before the water gets to that loop.
-After the long vertical run of pipe, keep it running up another 10 meters, then back down to the manifold, so the last section before the manifold is gravity fed.
-Delete an rebuild all pipes from the extractors to the manifold.

Nada, zero, zilch. Same problem. I'm at a loss as to what to try next. I've stuck some fluid buffers in different places, but they seem to make no difference. Since the system runs fine for a while before stalling, it seems to be some sort of issue with the game itself, but damned if I can figure out how to bypass the issue.

[Bobucles]

Connected pipes all share the same head lift, so one of the tricks is to build a "water tower". Take a pipe that does nothing, pump it up to a high altitude, and drop it back into your network. It will set the headlift for everything else, and the actual flowing pipes won't have glitching pumps to worry about.

[Indifibulator]

Originally posted by Bobucles:

Connected pipes all share the same head lift, so one of the tricks is to build a "water tower". Take a pipe that does nothing, pump it up to a high altitude, and drop it back into your network. It will set the headlift for everything else, and the actual flowing pipes won't have glitching pumps to worry about.

I tried incorporating this by continuing up uphill run another 10 meters or so up past my big vertical lift, then back down to the factory. The water flows for a while, then everything past the top of that tower dries up. Water still makes it to the top and the closest pipe still has extra available head lift (reading around 46 meters out of 50), but no flow past the top of the tower.

I guess I could try making a separate tower, before the vertical climb, but I'd need the same number of pumps and a lot of extra pipe.

Hell, nothing else works, I may as well try it!

[turvarya]

An additional valve also often solves such problem.
Is their maybe a section somewhere, where there could be backflow, because a pipe goes a little bit down?
Something like
Pump1 ____
Pump2 _/
So you should put a valve were Pump2 flows into Pump 1 even if the elevation difference is small

[Zak]

What works for me when pumping water up an incline is putting a tank at the top. So, the pump goes at the bottom, right before the incline, then long pipe all the way up and a tank at the top. There is no need for a valve on the long pipe going up as the pump should prevent and overcome any backflow, but you may insert one there, though I never do.

Then I wait for the tank to fill up before making further connections.

I also don' think it's necessary to put valves on extractors, I never do and I never have any problems. I generally go easy on valves as I rarely find them useful.

Also, I avoid putting pumps too close to one another, within the other pump's headlift, as I think this may be causing weirdness. Too many pumps does not mean a better flow.

Something like this:

https://steamcommunity.com/sharedfiles/filedetails/?id=3348086993

And then, beyond the tanks on top, there are coal generators on long, flat, level pipe manifolds, no more pumps since those large tanks provide 12 m of headlift. It's been running without a hiccup for weeks.

[Indifibulator]

Originally posted by turvarya:

An additional valve also often solves such problem.
Is their maybe a section somewhere, where there could be backflow, because a pipe goes a little bit down?
Something like
Pump1 ____
Pump2 _/
So you should put a valve were Pump2 flows into Pump 1 even if the elevation difference is small

I don't think so. It looks like this:

___
|........|..................................................._Manifold loop (same height)_
|........|..................................................|.............................................................|
|........|__________Valve___|_Manifold to feed factories__|
|
pump
|
|
|
pump
|
|
|
pump_______junction from two extractors____valve___extractor

Edit: The periods are just for formatting this diagram right

[Zak]

Originally posted by Indifibulator:

Originally posted by turvarya:

An additional valve also often solves such problem.
Is their maybe a section somewhere, where there could be backflow, because a pipe goes a little bit down?
Something like
Pump1 ____
Pump2 _/
So you should put a valve were Pump2 flows into Pump 1 even if the elevation difference is small

I don't think so. It looks like this:

___
| | _Manifold loop (same height)_
| | | |
| |__________Valve___|_Manifold to feed factories__|
|
pump
|
|
|
pump
|
|
|
pump_______junction from two extractors____valve___extractor

Too many pumps, too close to one another IMHO, but if you must use several pumps in a row then try putting small tanks between them.

[Indifibulator]

Originally posted by Zak:

What works for me when pumping water up an incline is putting a tank at the top. So, the pump goes at the bottom, right before the incline, then long pipe all the way up and a tank at the top. There is no need for a valve on the long pipe going up as the pump should prevent and overcome any backflow, but you may insert one there, though I never do.

I originally didn't use any valves. I have been throwing them in to try to solve the issue, but they don't seem to make much of a difference either way.

Then I wait for the tank to fill up before making further connections.

Also, I avoid putting pumps too close to one another, within the other pump's headlift, as I think this may be causing weirdness. Too many pumps does not mean a better flow.

Each pump has between 45 and 48/50 head lift. That seems about ideal to me, right?

Something like this:

https://steamcommunity.com/sharedfiles/filedetails/?id=3348086993

And then, beyond the tanks on top, there are coal generators on long, flat, level pipe manifolds, no more pumps since those large tanks provide 12 m of headlift. It's been running without a hiccup for weeks.

Very pretty!

[Indifibulator]

Originally posted by Zak:

Too many pumps, too close to one another IMHO, but if you must use several pumps in a row then try putting small tanks between them.

Why too many pumps? Each one is operating between 45 and 48/50 meters of head lift, so I could adjust the positioning a little, but I would still need that many to get up that incline.

[Zak]

Originally posted by Indifibulator:

I originally didn't use any valves. I have been throwing them in to try to solve the issue, but they don't seem to make much of a difference either way.

Yes, I rarely find any use for valves as means to prevent backflow since I don't seem to have any backflow problems. My pipe networks never give me any headaches. I try to keep my pipe networks flat and level and if there is any incline, I try to overcome it one vertical hop, thus avoiding multiple pipe ups and downs. Like on the screenshot above. I also use lots of tanks to stabilize the flow. I have no idea why, but maybe that's why I don't have any issues with pipes?

Originally posted by Indifibulator:

Each pump has between 45 and 48/50 head lift. That seems about ideal to me, right?

50 is fine for Mk2, that's the recommended value but it can even go a bit higher, 55 maybe even, but I try to stay below 50, yes.

If you have to move liquids higher than a single pump can do, more than 50 m, then first rethink the design. Maybe there is another way to do this? I never needed two Mk2 pumps in a row. But if you have no choice, then move them further away and put a buffer tank between them. Let the tank fill up and observe it for a bit. Use tanks to trace and test the flow, as they can both send and receive liquids, before hooking up actual machines or generators.

Originally posted by Indifibulator:

Very pretty!

Thank you :) It's the blue color and the curves, our brains like those :)

Originally posted by Indifibulator:

Why too many pumps? Each one is operating between 45 and 48/50 meters of head lift, so I could adjust the positioning a little, but I would still need that many to get up that incline.

It's just my "feeling", I'd never do that. Then try to rethink the situation. Why do you have to pump that high up? Or try the trick with tanks between the pumps. I think that buffer tanks are underrated and underused in this game while valves are overrated and overused.

[Indifibulator]

Originally posted by Zak:

Yes, I rarely find any use for valves as means to prevent backflow since I don't seem to have any backflow problems. My pipe networks never give me any headaches. I try to keep my pipe networks flat and level and if there is any incline, I try to overcome it one vertical hop, thus avoiding multiple pipe ups and downs. Like on the screenshot above. I also use lots of tanks to stabilize the flow. I have no idea why, but maybe that's why I don't have any issues with pipes?

This is the first that has caused an issue for me. It must be the big incline. The very hard part about diagnosing this problem is that everything works for while before it just stops.

If you have to move liquids higher than a single pump can do, more than 50 m, then first rethink the design. Maybe there is another way to do this? I never needed two Mk2 pumps in a row. But if you have no choice, then move them further away and put a buffer tank between them. Let the tank fill up and observe it for a bit. Use tanks to trace and test the flow, as they can both send and receive liquids, before hooking up actual machines or generators.

I absolutely will in the future. I am just loathe to completely destroy my current setup in order to accommodate this issue. Also, this should be an opportunity to gain insight into how the game's fluids function, but I just can't figure it out.

It's just my "feeling", I'd never do that. Then try to rethink the situation. Why do you have to pump that high up? Or try the trick with tanks between the pumps. I think that buffer tanks are underrated and underused in this game while valves are overrated and overused.

Up until this point I have no need for either of them and they seemed to me like a crutch to fix issues that could be handled by better balancing, but I'm grasping at any straw here.

[Indifibulator]

Originally posted by Zak:

What works for me when pumping water up an incline is putting a tank at the top. So, the pump goes at the bottom, right before the incline, then long pipe all the way up and a tank at the top. There is no need for a valve on the long pipe going up as the pump should prevent and overcome any backflow, but you may insert one there, though I never do.

Tried this. Deleted all pipes from the extractors to the manifold. and rebuilt with no valves and a tank at the top. Same issue. Everything flows. The tank begins to fill slowly though most of the fluid continues past to the refineries. Everything functions. Then the system dies. The vertical pipes stay filled up to the top when they turn horizontal to go into the buffer. That horizontal pipe, the buffer, and everything after go empty. Its like the system just gives up. Oddly, the last pump before the before the buffer is still reporting a normal head lift (the last one has around 40 meters) and a flow rate around 200m3/minute. I'm guessing that is back and forth sloshing as the pipe it is on has no visual indication of flow.

I guess I'll try your buffer between each pump idea.

[Zak]

Originally posted by Indifibulator:

Up until this point I have no need for either of them and they seemed to me like a crutch to fix issues that could be handled by better balancing, but I'm grasping at any straw here.

Buffers are great. They are crutches in a way but they work and they are often necessary. Think of a disk cache or network buffers - same idea. Crutches? Maybe, but they help you overcome throughput limitations and fluctuations.

[Zak]

Originally posted by Indifibulator:

Originally posted by Zak:

What works for me when pumping water up an incline is putting a tank at the top. So, the pump goes at the bottom, right before the incline, then long pipe all the way up and a tank at the top. There is no need for a valve on the long pipe going up as the pump should prevent and overcome any backflow, but you may insert one there, though I never do.

Tried this. Deleted all pipes from the extractors to the manifold. and rebuilt with no valves and a tank at the top. Same issue. Everything flows. The tank begins to fill slowly though most of the fluid continues past to the refineries. Everything functions. Then the system dies. The vertical pipes stay filled up to the top when they turn horizontal to go into the buffer. That horizontal pipe, the buffer, and everything after go empty. Its like the system just gives up. Oddly, the last pump before the before the buffer is still reporting a normal head lift (the last one has around 40 meters) and a flow rate around 200m3/minute. I'm guessing that is back and forth sloshing as the pipe it is on has no visual indication of flow.

I guess I'll try your buffer between each pump idea.

Hmmm.... so even the tank at the top goes empty? I can only think of distancing the pumps further away from one another and using buffers between them. But maybe you're just trying to do the impossible? Maybe there is some other limit in the game I'm not aware of. Like I said, I never did more than one pump, I never needed to pump that high up. Perhaps that is not doable after all?

Try to overclock the extractors too, so combined, they pump a little more than 600 m³, as the extractors output fluctuates too maybe? So the first pump would be getting better, more reliable supply of water from the extractors.

Just throwing out ideas on the table at this point.

[Tirandys]

I don't know if it matters, but when I build a fluid tower I always drop the return line all the way back to the extractor level. It works for me, but I don't know if it is necessary.

I have seen mention on these forums a couple of times about dropping the fluid back in on top of the fluid network, I could never get that to work for me. Something I did not understand more than likely.

[Indifibulator]

Originally posted by Zak:

Hmmm.... so even the tank at the top goes empty? I can only think of distancing the pumps further away from one another and using buffers between them. But maybe you're just trying to do the impossible? Maybe there is some other limit in the game I'm not aware of. Like I said, I never did more than one pump, I never needed to pump that high up. Perhaps that is not doable after all?

Yup. That is what happens. Maybe I am.

Try to overclock the extractors too, so combined, they pump a little more than 600 m³, as the extractors output fluctuates too maybe? So the first pump would be getting better, more reliable supply of water from the extractors.

That is how I started with it, but someone suggested that it can result it backflow pressure waves that stall out the pump, so I reduced the extractors to the point they produce just slightly less than my total consumption. I might try going back.

Anyway, update: buffers between each pump result in exactly the same problem. But I think I am narrowing in on the issue. I watched one of the extractors the entire time from a flush to a stall. Here is what happens: The pipe can handle a bit more than the extractor produces, so the reservoir in the extractor slowly depletes. Fine. That should happen. But, when it is empty the extractor does not the just pump out at the rate it is extracting, the flow rate goes right to zero! The reservoir in the extractor then fills back up. When it is full, the extractor shuts off entirely. What the hell!?