Print

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

[CaptainArchmage]

New lurker here. I've been doing some flow science and engineering with pumps and rivers:

If you pump water into a reservoir using a pump, the water will remain flowing after the reservoir reaches 7/7. This flow may remain after the pump has been switched off (verify?). The water flows in all directions. Since the water is at 7/7 everywhere, the reactor does not require refilling every now and then. This looks like the same thing that causes water from aquifers and rivers to be permanently flowing. I also know water moats connected to a river will have flow. The same flow rules should apply to magma (pumps only).

So far I built reactors pumping from a brook and a murky pool into reactors, with enough water feeding the pump to fill the reactor completely to 7/7.

I've Spoiler'd some reactor designs that worked for me.

Spoiler (click to show/hide)

Code: [Select]

Z = 0
   =====                                       ==
~  =77777                                      7=
~%%777777   -> Space for reservoir ->          7=
~  =77777                                      7=
   =====                                       ==

Z = 1

     FFFFFFFFFF                          FFF
     FW  W  W                             WF
     GW  W  W   -> Space for reactor ->   WG
     FW  W  W                             WF
     FFFFFFFFFF                          FFF

Key:
~ Water (beneath tile)
7 Water at 7/7
= Wall
% Pump
F Floor
G Gear Assembly
W Water wheel

Here I put the wheels perpendicular to the pump, so you only need the gear assembly above the pump to transmit power down and support the wheels. A gear assembly or other machine on the other end of the line of wheels will be needed if you want to switch the reactor off.

You expand the reactor in any direction and to any size, and the water will still remain flowing. You could probably add extra levels.

[Urist Da Vinci]

use [code ] [/code ] around ascii design diagrams to get the letters to line up

Code: [Select]

123456789 (sample)
1x
2 x
3  x
4   x
5    x
6     x
7      x
8       x
9        x


[Hans Lemurson]

I believe that all surface-level water on a map with a brook, stream or river in it has "flow" in it.

[CaptainArchmage]

I believe that all surface-level water on a map with a brook, stream or river in it has "flow" in it.

Streams and rivers have flow from the moment you embark unless there is no way for the water to flow off the edge of the map, which happens if your river goes into the ocean or a lake that hits the edge of the map without any other way to drain. In these cases, the forums suggest draining the upstream end of the river so those tiles producing water, which causes the river (and lake or ocean) to start flowing.

[morgoththegreat]

Yep, it's 100% doable. In fact, you don't even need pumps, you can sometimes do it with a 1-tile hole designated as a pond and some buckets (it doesn't always seem to work).

Regarding the "brook makes all tiles flow," that may be true, but you definitely don't need a brook/river/etc to make the exploit work. Try it and see!

Couple months ago I made a thread on basically the same thing, there is some interesting stuff in the replies (apparently it was on the forums a few times before then, although my search-fu is too weak to find examples) (sorry about the shameless self-promotion).
http://www.bay12forums.com/smf/index.php?topic=64078.msg1492231

[CaptainArchmage]

Yep, it's 100% doable. In fact, you don't even need pumps, you can sometimes do it with a 1-tile hole designated as a pond and some buckets (it doesn't always seem to work).

I haven't tried filling reactors with buckets yet, but it would help to have a reactor pool at 7/7 producing power. Of course, when you have a reactor producing 10,000 units of power the 10 units of power for the pump are not going to matter unless you're using up nearly all of it. All of this testing has been done "Above Ground" so far, so I am going to do some testing underground.

Buckets may work because they "create" a 1/7 tile of water in the tile being filled a bit like pumps, aquifers, and rivers. I have actually tried turning off a pump system before and the flow seemed to remain.

Regarding the "brook makes all tiles flow," that may be true, but you definitely don't need a brook/river/etc to make the exploit work. Try it and see!

When I mentioned brooks and flow I meant they don't get gen'd through flow-stopping bodies of water (as far as I know). They don't seem to get gen'd in any world tile with a river or lake either.

If people dwarves want power generated without windmills on embarks without flowing water, they seem go for the dwarven water reactor. If the 7/7flow reactor works, it will be more effective for powering 8-wide magma pumpstacks, since 12 or so waterwheels should be sufficient to power one over 100 z-levels, including power conduits, and that will fit in a 5x14 space over 2 z-levels.

[morgoththegreat]

All of this testing has been done "Above Ground" so far, so I am going to do some testing underground.

It works exactly the same below ground as above.

Of course, when you have a reactor producing 10,000 units of power the 10 units of power for the pump are not going to matter

Good luck getting flow in 300 tiles with 1 pump: the limiting factor to how big you can build the reactor seems to be how quickly you can fill it: it only seems to work if all the tiles get to 7/7 (or another equilibrium level above 4/7) at about the same time. If you try to fill a really big hole, not all the tiles reliably have flow, especially (in my experience) those that are further from the pump output. You can use multiple pumps to fill it up faster and get reliable flow in all tiles.
Also, if you don't want to power the pump when it's all said and done, just disconnect it from the power train (or don't ever connect it and use dwarf power instead) or put a floodgate (or bridge, door, etc) between the output and the reservoir and deconstruct it.

Other things of interest, which are relevant to this topic:
I did a test embark on the head of a brook.
1. I put a waterwheel in a 5-tile murky pool on the same level as the brook, and there was no flow.
    a. I channeled the 5 tiles into 7 tiles, which split up the water into seven 5/7s and made the waterwheel perpetually spin.

2. The little pools at the head of the brook have flow.

3. When I channeled a 5x4 pool next to the brook and overflowed it from the level above, the wheel in it did not spin (which was my goal, I had pumped it in starts and stops in an attempt to prevent the perpetual flow in the pool).
     a. When I channeled out a diagonal between it and the stream, the wheel did not spin.
     b. When I gave it a full orthogonal connection, the wheel did not spin.
     c. Only when I took dug out pretty much the entire side of the pool did the wheel pick up the flow from the brook.


I guess this is probably just another example of the state of water getting "locked" (presumably in a pathing-reduction measure by Toady). You can also see it in u-bend pathing, if you make a u-bend and put a hatch above the lowest level of one side, fill up the other side with buckets (such that one side has several z-levels of water, while the other has only one, locked below a hatch), and open the hatch, the water will stay in place. You have to disturb the water by giving it an "obvious" exit (one that doesn't involve pathing upwards) in order to make it start moving.

[CaptainArchmage]

When I did the above ground testing, I managed to get enough reactor for about 10,000 power, with flow in every tile. This did pump from a brook though, so there wasn't any water shortage. I think what happens is when the reactor is built as something like a block of 10 3x10 reservoirs, so the water reaches 7/7 close to the pump first and then reaches 7/7 at the end last, because water is very viscous in Dwarf Fortress (which is also why no intelligent creature will drink water when there is booze available). The sections that reach 7/7 still have flow in them.

I think the programming is related to making rivers flow.

You can tell when water is flowing because the graphic on the surface tiles changes between ~ and a double ~. This isn't the case with magma, but I suspect when pumped in the right way, magma will have flow in the same way as water.

I noticed the flow is maintained through diagonals, unlike pressure.

[Schmlok]

I was finding what I thought was permanent flow while testing, even when elevated above rivers.

But they seemed to subside after a while, and or possibly save/reload.