Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[GhostDwemer]

Sometimes, water in my forts will completely lose pressure. I've seen it happen in u-bends before, but now I'm seeing it in a system with no u-bends. I've got water going down a flight of stairs, through two engraved fortifications and a wall grate, over a raising bridge as the upper shut-off, then down twelve z-levels. The water travels horizontally north 13 tiles, west 18 tiles, and then north 6 tiles. It goes diagonally through a floodgate, to reduce pressure, and into the top of my 2 z-level tall cistern.

At the beginning, it all works like it is supposed to. After a few seconds, the water seems to forget that it is under pressure. A 6 water tile has to make it all the way back to the stairwell before more water falls. It should be 7s all the way to the depressurizer. Water should teleport to the next available 6 tile before the depressurizer. It doesn't. However, using the dfhack liquids command to destroy the tile of water under the stairwell instantly restores things to working order, and the system remembers it is under pressure. For another few seconds. Then it forgets again.

I've looked on the bug tracker, and the only reports I can find are from people who don't understand how pressure works in the game, and think water should come back up to the same level like it would in the real world. I'm not expecting that, I know how the pressure system is supposed to work in the game, and it does, but it fails intermittently.

[Mostali]

I haven't done much elaborate with water in this version yet, but what I have done has worked as expected.  .02 and .03 seemed fine also, though I don't display water units - just periodically check to make sure everything is running smoothly.  I may have missed the behavior you're describing if it's too subtle.

Just to clarify, this is .43.05?  And any mods?

[taptap]

The thing is: there is no pressure just a different mode of water movement when there is 1) a path 2) tiles in between filled to 7/7. The hydrodynamics eduction thread is really better in that regard than anything in the wiki (http://www.bay12forums.com/smf/index.php?topic=32453.msg466003#msg466003). Treating "pressure" as a property of the water will lead to all kind of nasty surprises, it does however give a workable worst case scenario - what will fill up if unlimited water is provided here - and that is likely all most people care about.

(I also place loads of linked doors in my plumbing, so when I open one the situation is basically like you emptying a tile with dfhack, everything with no diagonal path fills rapidly.)

@fleeting frames: Imo op refers to water propagating by spread instead of teleportation for no obvious reason - not a complete freeze. With the cavern, lake, tree bug things correct itself as soon as any "spread" happens nearby, so: different situation?

[Fleeting Frames]

Are you doing this across the sub-embark tile 16x16 tile? I recall reports of water-based óverseer-triggered levers failing to work when water wasn't kept in an updating state in the 16x16 square, with water failing to move to adjacent empty squares - and similar effects happening to trees cut in cavern lakes, with the trunks leaving behind empty tiles free of water.

Still, odd.

[PatrikLundell]

Apart from holes in water after trees removed with ballista arrows, I've also encountered pressurized water that sometimes refuse to fill up emptied tiles at the level above. In addition to that, I've also read about reclaimed fortresses with the parts under oceans having water walls on the ocean boundaries.
My interpretation of this is that DF doesn't calculate whether water should move unless either prompted actively (such as the opening of a door presumably also marking the tiles on either side to check for water movement), or by the fact that moving water results in calculations the tiles water moves through (regardless of whether the water actually moves, or it's just a propagating calculation resulting in water teleporting at the end). Given that interpretation, the disintegration of a tree doesn't trigger recalculations to fill the hole, and sloshing at the level above a completely filled level of water may not trigger a recalculation on the level below to push water upwards.
In my case I triggered recalculations (and the flow) by closing and opening a drawbridge in the lower part of the plumbing system. This atom smashed one tile of water, and flow calculations were re-triggered (which I believe is the same thing taptap does with the linked doors).

The above doesn't quite match what the post linked by taptap said, although I read that page of posts only, not the rest of the thread.

[Infinityforce]

The water is just waiting.
Biding its time.
Then it will sneak out when you're not looking and drown your fortress.

[GhostDwemer]

The thing is: there is no pressure just a different mode of water movement when there is 1) a path 2) tiles in between filled to 7/7. The hydrodynamics eduction thread is really better in that regard than anything in the wiki (http://www.bay12forums.com/smf/index.php?topic=32453.msg466003#msg466003). Treating "pressure" as a property of the water will lead to all kind of nasty surprises, it does however give a workable worst case scenario - what will fill up if unlimited water is provided here - and that is likely all most people care about.

(I also place loads of linked doors in my plumbing, so when I open one the situation is basically like you emptying a tile with dfhack, everything with no diagonal path fills rapidly.)

@fleeting frames: Imo op refers to water propagating by spread instead of teleportation for no obvious reason - not a complete freeze. With the cavern, lake, tree bug things correct itself as soon as any "spread" happens nearby, so: different situation?

It should have been clear from my use of the phrase "teleport" that I understand the way water works in the game, but thanks anyway. And no, the situation of you opening a door is not like the one I describe, but again, thanks for your input. It only works if I remove the water at the bottom of the water column, removing water from any old random tile does nothing, the water still only propagates by spread, not teleportation ("pressure").

Fleeting Frames: Yes. This is over two adjacent subembark tiles. This may be the issue.

[wierd]

Water pressure is known to stop when it rounds a corner.

EG, if you run it through a checkerboard pattern, like this:

Code: [Select]

     O===O
 O===OO+OO===O
(IN)+++O+++(OUT)
 O===OO+OO===O
     O===O

pressure will drop to zero on the (OUT) side.

Or, at least it used to in prior versions.  IIRC, this was discovered in the .30X versions.

[taptap]

I didn't mean to offend you or your knowledge of the game. These threads are read by more people than the original poster. Fwiw, i tried it in a fort at hand in overly long pipe (20 tiles north, 20 east, 5 up 5 down 50 south, down, up, north...) and can't reproduce losing pressure. Definitely crosses several 16x16 areas. Water happily teleports as long as there is some in the top reservoir and keeps doing so whenever aquifer fills it up again.

[PatrikLundell]

In my lost pressure pipe i used the water for magma sea obsidianization for mining purposes (from the aquifer many levels above). That means that I only used the water intermittently, and in the mean time the pipe filled up completely. When I opened the drawbridge to let another batch of water out it often, but not always, just emptied the top part, as if nothing informed the top in the aquifer (or anywhere else in the pipe) that it was supposed to push up into the emptying segment. As mentioned above a drawbridge close/open midway through the system triggered the water movement again. Due to lack of forward planning, my pipe went down one or possibly two levels below the troublesome outlet (to serve outlets at earlier places along the way).

@wierd: Yes, the diagonal movement issue has been discussed further up in the thread and has also been ruled out as the cause of the OP issue (and mine, for that matter).

[GhostDwemer]

I didn't mean to offend you or your knowledge of the game. These threads are read by more people than the original poster. Fwiw, i tried it in a fort at hand in overly long pipe (20 tiles north, 20 east, 5 up 5 down 50 south, down, up, north...) and can't reproduce losing pressure. Definitely crosses several 16x16 areas. Water happily teleports as long as there is some in the top reservoir and keeps doing so whenever aquifer fills it up again.

Yeah, it is a random thing. Sometimes it happens, most times it doesn't. I've noticed that it has been an intermittent bug since at least version .28. It's annoying when it happens, as it requires a redesign of the water intake, but providing another intake, or even modifying the initial intake, always seems to fix the issue. Now that I think about it, it seems like it always occurs when the water flows through grates or carved fortifications, but that could be a memory lapse on my part.

[GhostDwemer]

In my lost pressure pipe i used the water for magma sea obsidianization for mining purposes (from the aquifer many levels above). That means that I only used the water intermittently, and in the mean time the pipe filled up completely. When I opened the drawbridge to let another batch of water out it often, but not always, just emptied the top part, as if nothing informed the top in the aquifer (or anywhere else in the pipe) that it was supposed to push up into the emptying segment. As mentioned above a drawbridge close/open midway through the system triggered the water movement again. Due to lack of forward planning, my pipe went down one or possibly two levels below the troublesome outlet (to serve outlets at earlier places along the way).

@wierd: Yes, the diagonal movement issue has been discussed further up in the thread and has also been ruled out as the cause of the OP issue (and mine, for that matter).

Yeah, this sounds like the exact issue I'm facing. Intermittent pressure failure. Due to the fact that I've run into this bug a number of times, I always put a bridge valve right at the beginning of any water system I design now. Which is really not a bad idea in general, it has saved more than one fort when I accidentally mined out the wrong tile and started a flood.