Print

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

[Lummox JR]

I have an architectural/engineering question about a structure which is working only partially in my world. In my fortress I decided to build a waterfall room for my dwarves, which required building an aqueduct since the location of the proposed room was too far away from the river. Here is a rough schematic of the system:

Top level; | represents the wall grate

OOOOOOO
O#+#+#O
O+++++OOOOOOOOOOOOOOOOOO
O#+#+#|   <-         ÷÷_
O+++++OOOOOOOOOOOOOOOOOO
O#+#+#O    _÷÷ ->
OOOOOOOOOOOOOOOOOOOOOOOO

Ground level; waterfall room is inside fortress

+++++++O                  W~~
+#+#+#+O                 *W~~
+++++++O              OOO|W~~
+#+#+#+O              O ÷÷~~~
+++++++OOOOOO         OOOO~~~
+#+#+#+O_÷÷ O             ~~~
+++++++OOOOOO

Basement level, drainage:

+++++++O
+++++++O
+++++++O
+++++++O
+++++++OOO
+++++++++O
+++++++OOO

These are crude schematics; the structure is bigger. The main water wheel is also connected via a couple of gears that can be disengaged by lever, as a safety shutoff. Also not shown is the drive mechanism for the pumps on the left, which run off a water wheel in the incoming channel (top). As you can see, the plan is to have the river (flowing south) drive a water wheel which pumps river water up two levels via connected pumps, into the main aqueduct channel, into the top room and through the grates. The top room should then rain down into the indoor waterfall room (ground level), and as water is coming into the inflow channel it's driving a second wheel that drives the pair of outflow pumps, so the drainage room can be emptied and send its water back to the river through the aqueduct's outflow channel.

My problem is not that this structure does not pump water; it does. I have successfully pumped water into the top room and verified that it does power the second wheel and the outflow pumps, which in turn send water to the river. The problems I did encounter, however, were twofold:

1) Water hitting the grates in the interior waterfall room (ground floor) also tended to spill out of the room a little, a problem that was magnified when:

2) The drainage room filled to capacity relatively quickly, despite being pumped out.

I have no explanation for #2. In Dwarf Fortress physics, the second water wheel should provide just as much power as the first, so the outflow pumps should be pulling just as much water out as the inflow pumps. In practice this is not happening. Water is flowing out, but apparently not at a rate that matches the inflow.

What I'm wondering is, what's the fatal flaw in this design? Is it that the drainage room can't fill up to 7/7 at the pump's source quickly enough so the pump is maybe pulling only 4/7, 5/7 at a time? If so, would a second nearby pump, emptying into a shared trough, alleviate the problem? E.g.,

+++++++O                  W~~
+#+#+#+O                 *W~~
+++++++OOOOOO         OOO|W~~
+#+#+#+O_÷÷ O         O ÷÷~~~
+++++++OO|O O         OOOO~~~
+#+#+#+O_÷÷ O             ~~~
+++++++OOOOOO

+++++++O
+++++++O
+++++++OOO
+++++++++O
+++++++OOO
+++++++++O
+++++++OOO

Even if two pumps could fill up the small reservoir fast enough though, the second-level outflow pump would probably have the same issue. Should I perhaps have designed my outflow system for double capacity to ensure overfilling was impossible?

[Albedo]

Should you take "extra precautions"?

I think you know the answer to that one. 

[puke]

What I'm wondering is, what's the fatal flaw in this design? Is it that the drainage room can't fill up to 7/7 at the pump's source quickly enough so the pump is maybe pulling only 4/7, 5/7 at a time? If so, would a second nearby pump, emptying into a shared trough, alleviate the problem? E.g.,

this could be, if youre filling froma 7/7 source and emptying something less than that due to the dimensions of the area letting water spread out more. 

also try watching the power in your system and make sure its not sporadic or intermittant for some reason, possibly causing one of the pumps to pause.

my last idea is that the place you are emptying TO may its self not be emptying fast enough.  you cant pump fluid into an area that is full or already 7/7, so your pump will be ineffective until the area it is pumping to has a chance to drain.

[Lummox JR]

also try watching the power in your system and make sure its not sporadic or intermittant for some reason, possibly causing one of the pumps to pause.

my last idea is that the place you are emptying TO may its self not be emptying fast enough.  you cant pump fluid into an area that is full or already 7/7, so your pump will be ineffective until the area it is pumping to has a chance to drain.

Yeek! Either of those could be the problem. The incoming stream might not always be high enough to drive the second wheel (although I don't think that actually is a problem), but I think the bigger issue might be that the outflow channel is only one tile wide. In terms of pressure that wouldn't be a problem but I forgot about the bottleneck keeping a pump from working altogether. Of course, expanding the aqueduct is gonna be a pain in the butt. I wonder if I can just build a second level of outflow and link it to the other two outflow pumps.

[Lymojo]

I was trying to do the exact same thing in my current fotress and faced the same problem. Are you trying to pump INTO a 7/7? That could also present problems.

I eventually gave up and just had a closed water circuit that could be pumped out gradually if I want to empty for maintenance.

[Lummox JR]

Good to know I'm not the only one with this problem. I constructed my waterfall based on a very limited concept description in the wiki (under "Swimming"), where the impetus behind it was to increase happy thoughts and teach swimming. It was implied that pumping out the drainage area would be easy; maybe it would if the drainage area was huge. Of course the description there also has an intermittent inflow based on water dropping in massive quantities from above when triggered by a pressure plate, which I deemed impractical and kind of pointless. My waterfall room is on the same level as the river anyway, not deep below it.

I am currently having my dwarves expand the outflow channel of the aqueduct to 3 tiles wide, which should let the water disperse quickly if pumping to a 7/7 is indeed the root of the problem. If the problem turns out to be more a case of intermittent pumping power though, then that won't work. What I may end up doing in that case is constructing a second and maybe a third wheel on the river (if I can kill the two new carp that have shown up there--I had to have my marksdwarves take care of the last one), and building a long axle. Possibly I could tie the first wheel in as well. Then these would continuously power the outflow pumps.

With as useful as water power is anyway, I probably should have a generic power house built on one or both sides of the river at some point, which I can also use as a wide bridge. My miller has been doing everything by hand at a quern.

[Lummox JR]

Just to give an update on this, I expanded my aqueduct to use a 3-tile-wide outflow channel, and that did the trick. As soon as I turned the system back on, the drainage room--which was still mostly flooded--started clearing out right away faster than new water could flow in. I'm not sure this will avoid the spillover problem entirely, but for now the system works.

[Lummox JR]

It seems I spoke too soon. The drainage room still fills up, I think because liquid doesn't get to the pump's square quickly enough. My next plan is to wait until I can get back into the drainage room, then build a pressure plate to turn off the inflow until the level subsides.

[Eater of Vermin]

I would do two things.

Firstly, I'd add a second - but seperate - drainage pump parallel to the existing one.  (It's always a good idea to have the capacity to remove the water quicker than you can pump it in.) Of course, this also means either adding another outflow tunnel or widening the existing one as two pumps into a one tile wide channel will only handle 1 pumps worth of flow.

Secondly, I wouldn't have the pumps sucking direct from the drainage room... instead I'd run a single tile wide channel across the floor of the drainage room, to a pair of stacked pumps.  To act as a sump.  This helps reduce the "not enough water getting to the pump" problem.

Well... actually...  with twinned pumps, I'd have two such channels, one for each pumping system.   

[Kanddak]

One thing will likely fix all your problems: Extend your drainage system down a z-level.
The new bottom level will still fill up and stay full. Water moves very slowly across the same z-level.
Water on the old-bottom/new-second-to-bottom level, though, will teleport instantaneously to the pump intake due to the water pressure rules. Unless there is nowhere for the pump to put the water, it will be pretty much impossible to fill that level and make water back up into your fort.

[Lummox JR]

Water on the old-bottom/new-second-to-bottom level, though, will teleport instantaneously to the pump intake due to the water pressure rules. Unless there is nowhere for the pump to put the water, it will be pretty much impossible to fill that level and make water back up into your fort.

That's a pretty good idea; I'll have to see if that's doable with my current setup.

I haven't really run across any good info on pressure yet except what's on the wiki, and I find that confusing on a couple of points. It mentions that a pump moving pressurized water has to be powered by a machine, but says the gear/axle can't be next to the screw pump and there can't be a channel above. But if I don't dig a channel above the pump, can it still be connected to an axle or gear from above?

I'm not sure that info is relevant to this particular issue, but I have started a longer-term underground reservoir project that I'll use to feed a drowning chamber and anything else I care to via floodgates, so undersanding pressure will be important.

[Konis]

I've been having some adventures with this one, trying to get a feed system 2 z-levels down from the brook to pump into my moat.

I'm not sure precisely what the deal is with the phrasing on the wiki, but here's what I've found: the ONLY spot to power a pressurized pump is from directly above the 2nd block of the pump (i.e., if it pumps from the south, the north block; if it pumps from the west, the east block, etc).  Lemme take that back - I've done it, and it works.  Haven't tried from below.  Because the source water is under pressure, it will raise up through a channel dug to power the 1st block.  Same thing with going around to either side if trying to power from the side.  Basically, the 2nd block is the only one that will successfully plug a gap, and a channel from above (or maybe below) is the only way that doesn't give the water a diagonal to sneak through.

And don't forget to dig a z-level down for the intake, too; makes it much quicker.

[Kanddak]

I haven't really run across any good info on pressure yet except what's on the wiki, and I find that confusing on a couple of points. It mentions that a pump moving pressurized water has to be powered by a machine, but says the gear/axle can't be next to the screw pump and there can't be a channel above. But if I don't dig a channel above the pump, can it still be connected to an axle or gear from above?

Ok, first, you should read this post I made a while ago for just such purposes: http://www.bay12games.com/forum/index.php?topic=32453

The concern with pumps in water systems under pressure is that if the pump ever stops or doesn't have anywhere to put the water, the water can come up out of the channel being pumped from and cause flooding. There's no magic rule wherein getting wet hurts your mechanical system if pressure is involved.

You can definitely put a channel either above or below the output block of the pump, and put power in that way. That's the principle that makes pump towers work.

[Lummox JR]

Very useful info, Kanddak, thanks!

I wouldn't be surprised if the non-moving-items caveat of the pressure rule is changed in a future version though, since it should be possible to generate flow along the found path.

[zchris13]

(Awkward phrasing)

Summary:  Pumps can act like gear assemblies.  They can transmit and receive power from all directions.  The solid block of the pump is the one receiving the power.