Indoor waterfall features are nothing new. They usually have a pump that moves water to the top where it flows down over a back plate. This creates a pleasant trickling sound and a little bit of a random visual effect due to unpredictable turbulence in the water flow.
To make this a bit more interesting, we can make the back plate of the waterfall more dynamic. To do this, the backing should consist of a grid of metal tiles separated by thin insulation. Behind each tile, a thermoelectric cooler/Peltier device is bonded, complete with an associated heatsink. When a voltage is applied, the individual tile begins to cool. Ultimately this will cause a layer of ice to build up. If activated long enough, this ice will form a protrusion and the waterfall will flow pleasingly around it.
Tiles can be activated individually or in groups to make a constantly changing ice layer over which the water is flowing. Of course, the interaction will have additional complexities. Water will cut channels leading to dead zones and accelerated melting. Occasionally, ice will destabilize and detach, possibly in large sections, maybe high up leading to the unfortunate crushing of the resident pianist. Art always costs.-- bs0u0155, Feb 15 2019 I like this. [+]-- doctorremulac3, Feb 15 2019 In humid environments, you might not even need to add the water!-- bs0u0155, Feb 15 2019 Hmm. Are Peltiers really up to generating ice in a flowing stream of water?-- MaxwellBuchanan, Feb 15 2019 Yes, if you stack them. But other heatpump technologies are available.-- 8th of 7, Feb 15 2019 //Yes, if you stack them.// I don't believe that to be true. Stacked Peltiers can achieve greater temperature differentials (though not as great as you might hope), but I don't think the flux increases much.
I think this would only work if the flowing water were kept at very close to 0°C.-- MaxwellBuchanan, Feb 15 2019 //Are Peltiers really up to generating ice in a flowing stream of water?//
A lot depends on how efficiently you remove heat from the hotside. I recommend ice water as an excellent candidate for this. Don't ask where that's coming from.-- bs0u0155, Feb 15 2019 Where's that coming from?-- MaxwellBuchanan, Feb 15 2019 The Thwaites Glacier, probably.-- RayfordSteele, Feb 16 2019 Haha!-- pertinax, Feb 16 2019 One figure picked out is that with the most powerful 135W peltier, the difference would be only 8.44e-4 C.
What if they were recursively stacked in insulated layers? The cool of one, is dropped by 0.0008.44 by next until the ice happens. Only a horn of 13,000 odd tiles.
It would be a heavy sculpture in body and energy.
On second thoughts, a good ice maker with a pinscreen conductive interface would work, especially if the water was precooled.-- wjt, Feb 17 2019 [+] Ok, if a society has enough spare solar energy to make public space ice sculptures then i'm all for it.-- wjt, Feb 17 2019 // What if they were recursively stacked in insulated layers? The cool of one, is dropped by 0.0008.44 by next until the ice happens. Only a horn of 13,000 odd tiles. //
Peltier coolers are not 100% efficient. (I've heard they're usually around 5% efficient, but one company claims they should be, IIRC, somewhere around 80–90%.) Therefore, each one in the stack must pump not only the heat from the thing being cooled but also the waste heat of every Peltier cooler below it in the stack. This is why I have never heard of anyone making a stack of more than two.-- notexactly, Feb 17 2019 ^ Would not a tiled cone stack be sensible, heat output on the outside and the temperature stepdown starting from a ring of tiles around the largest volume?-- wjt, Feb 17 2019 Lovely idea.
In cold climates it could be a bismuth 'waterfall'. The melting pount of Bismuth is 271.4 degs C. The bismuthfall could therefore double as a local area heater.
It would not be a problem to cool the flowing bismuth so that it forms crystals dynamically. It should be straightforward to prevent the bismuth from crystallising locally by means of using randomly sequenced electrical heaters. Bismuth crystals are also rather attractive.
There may be a problem with longer-term operation arising from formation of oxides. In this event, the bismuthfall would need to be either protected by a blanket of inert gas, or submerged in hot oil to reduce oxidation.-- bhumphrys, Feb 17 2019 Oxidation will be a huge problem. The USSR had endless difficulties with their lead-bismuth cooled submarine reactors and eventually gave up on the technology. You would have to put your device in an enclosure filled with dry Argon.-- 8th of 7, Feb 17 2019 Instead of Peltiers, perhaps a system of valves that can squirt liquid nitrogen or something on demand?-- Wrongfellow, Feb 17 2019 Why not just have a liquid nitrogen fall then ? That would be cool (literally) and would coincidentally permit the inclusion in the display of other fascinating physical phenomena like the Meissner effect.
Have a collecting trough for the LN, below which is a powerful magnet, and on it place a number of "pebbles" of superconducting material. The pebbles are subject to the Meissner effect and therefore float above the liquid until they heat above their critical temperature. Then they splash back in, cool down, and repeat the cycle.-- 8th of 7, Feb 17 2019 //and repeat the cycle// Why would they repeat? When they fall into the LN2, they will cool and become superconducting. But then they will be locked in place by the Meissner effect. What lifts them up again?-- MaxwellBuchanan, Feb 17 2019 Tiny purple robots.-- 8th of 7, Feb 17 2019 random, halfbakery