From seawater to drinking water, with the push of a button | Hacker News Reader

From seawater to drinking water, with the push of a button

news.mit.edu

258 pointshappy-go-lucky99 comments

Edit: I wrote out the comment below criticizing the efficiency of this device, when compared to reverse osmosis RO desalination (100x less efficient).

However, they’re not trying to be more efficient than RO devices, they’re trying to be more compact, portable, and avoid the need for filters.

They’re solving for a different problem than I was measuring them against.

In my race to criticize, I overlooked those key details.

I still think my comment is worth reading to learn about the efficiency differences but I guess that measurement is not important for the viability of this technology.

Original criticism below: ———————

I’m skeptical about how practical this actually is.

While it’s refreshing to see a drinking water solution that isn’t based on inefficient dehumidifier technology, this still seems impractical.

From the demonstration video, their portable ~50 watt solar panel took 30 minutes to get a few ounces of drinking water.

It looked like a cold winter day, so let’s be generous and assume the panel only produced 15 watts of power during those 30 minutes.

Let’s also be generous and say they got 5 ounces of water.

That means they produce 1 ounce of water for every 3 watts.

That’d be 333 ounces per kWh.

To put that into perspective, a desalination plant produces more than 100x that much.

Yes, there is lots of additional infrastructure to consider with a traditional desalination plant but It’s not like this solution wouldn’t require the same complexity to scale up.

I don’t remember the cost of dehumidifier drinking water “solutions” but I think this is slightly more efficient than those horrifically inefficient solutions.

From the article it sounds like they are looking for ways to scale down (i.e small portable units are the target form factor, not a mere demo), and this is where it compares favourably against things like reverse osmosis for efficiency.

I was so quick to criticize technology that I completely overlooked those points. Reading the article again, it’s embarrassing to see that I missed it when they were emphasizing it so much.

I have one more lingering criticism that I don’t think they addressed (I reread it twice).

What happens to the salt after it’s removed? If they don’t have a filter, where does it go? Do they pump brine water back?

You can pump brine water back but if I recall you can't just dump it all back at once because it's so highly concentrated. Ideally you would be able to find an alternate use for it.

For these small amounts, I doubt pumping it back out would be an issue but is that what they’re doing?

I thought the metal plates were collecting the salt but the article and video don’t make it clear.

In CEDI at least the ions are trapped in a resin then are driven across a membrane by a charge, on the other side of the membrane you run water to flush them out.

Some comment in the article left me thinking they end up trapping them in the media and have to reverse charge to release them so I assume it would operate pulsed and presumably have a valve to dump the output during regen.

Though I'm a little confused in that normally CEDI is used after RO because the CEDI media is pretty sensitive to fouling and also doesn't work well when the water conductivity is highly variable. Maybe they solve the fouling with charge reversal.

...who knows, because popular coverage of this stuff never hits on the important parts and almost never links the relevant publications. There are many ways to make drinkable water from seawater-- making them some useful mixture of energy efficient, cost effective, portable, waste-water efficient, and reliable is the actual hard part.

given the size of the thing it's likely to be positioned very close to the water source, so I expect it just dumps it outside the device. I mean, if I was running it on a boat, or next to the sea, I'd just basically pour the waste out right next to the device.

And I agree, at ounces per hour there's no need to dilute the brine, just dumping it is fine.

Note that they are not merely aiming for portability, but efficient portability.

This particular passage is suggestive of their device being more efficient than RO at small scales, although no figures are given for comparison:

> Commercially available portable desalination units typically require high-pressure pumps to push water through filters, which are very difficult to miniaturize without compromising the energy-efficiency of the device, explains Yoon.

Good point. I wish articles like these would include figures. I’d love to know what they calculate RO efficiency to be at small scales and how far away they are from beating it.

Sure, that might be boring for some readers but they could throw a couple of sentences about it at the bottom.

FWIW, I think you made a minor math error; you lost track of the 30 minutes. It would be 1 ounce of water for 3 watts of power over 30 minutes, which is 1.5 Wh of energy. That's 666 ounces per kWh.

It wouldn’t be the first time I made a mistake like that by not clearly identifying my units but I think I got it right.

Let me rephrase and use watt/hour units.

I was assuming it was a 50 watt panel that would only produce 30wh from the weak winter sunlight.

Divide 30wh by two (since it was only ru Ning for half an hour) and get 15 watts generated to purify 5 ounces.

Had they let it run for the full hour, they’d get 10 ounces from my assumed 30wh of energy the panel produced.

30wh divided by 10 ounces would be 3wh per ounce (or unit of water produced).

How many units of water can be produced by 1kwh? Let’s divide 1,000wh by 3wh to get the result of 333.3 units (ounces) produced.

Wh is a measurement of energy, but W is a measurement of power (that is, the rate of energy). They are different units of measurement and you're mixing them up here. Here, you've doubled the original rating of 15 watts up to 30 watts; that's the result of the confusion between energy and power. You don't divide the wattage in half when you run it for half the time, nor double it when you run it for twice the time; that's not how power works. That's how energy works.

In your original post, the number was 15 watts, and you figured it produced 5 ounces of water in 30 minutes. So you then figured that 3 watts can produce 1 ounce of water in 30 minutes. So far so good; that's correct. If the panel continuously produces 3 watts of power (remember: watts are a rate of energy) over a period of one hour, that's 3 watt-hours of energy. For 30 minutes, it's 1.5 watt-hours of energy. Thus, it took 1.5 watt-hours of energy to produce 1 ounce of water. (1000 watt-hours) * (1 ounce / 1.5 watt-hours) = 666 ounces.

It may make more sense to translate this into distance, which I think is more intuitive for most people. Watts (power) are like miles-per-hour (velocity), and watt-hours (energy) are like miles (position). The panel runs at 15 MPH. After 30 minutes it has run 7.5 miles and produced 5 ounces of water. Thus, it could produce 1 ounce of water by running just 1.5 miles. It still runs at a speed of 15 MPH no matter what.

You’re getting confused by their typo on units:

They meant 15 watt-hours in both cases, which would be half of 30 watt-hours — or a 50 watt panel for 30 minutes at 60% efficiency. (Which is the stated working premise.)

15 watt-hours -> 5oz would imply that 3 watt-hours -> 1oz, and hence 1 kWh is 333oz.

I generally try to assume best intentions — so if they repeated otherwise the same correct math, but for misnaming a unit, I try to assume they’re correct and typo’d.

OP here. Thanks for understanding what I’m trying to say and putting it in the correct terminology.

However, I think I’d still make the same mistake in the future because I don’t understand how to express what the (assumed) 50 watt rated panel is generating over half an hour.

If it’s running at 60% efficiency then it’s generating at 30 watts per hour.

After an hour of running, we’ve accumulated 30 watts of energy.

If we only ran it for half hour, we’d have 15 watts.

Dividing that out tells me that 3 watts of energy will provide 1 once of water.

Does it matter if those 3 watts are provided over an hour via a 3wh power source or over 10 minutes via a 16wh power source?

This is what the other person was trying to say:

“Watt” is the unit of power which is the rate of change in “joules”, which is the unit of energy. In the way that “velocity” is the rate of change in “position”.

1w = 1J/s

A “watt-hour” is another way to represent joules/energy by integrating watts/power over time.

30 watts * 30 minutes = 15 watt-hours or 54 kJ.

30 miles/hour * 30 minutes = 15 miles or 24km

> Does it matter if those 3 watts are provided over an hour via a 3wh power source or over 10 minutes via a 16wh power source?

You have the units backwards:

You can get 3wh from 1 hr @ 3w or 10 min @ 18w.

Thanks. I’ll need to read that a few times because I struggle with this constantly.

I think part of the confusion is we don’t normally have units for rates of change.

So let’s create one: the “walk” is the speed a person can walk — 3mph.

Then I can say the store is “3 walk-hours away”, even if I normally drive the 9 miles with my car.

3 walk-hours = 3 hours @ 1 walk = 18 minutes @ 10 walk (or 30mph)

Batteries are rated in “watt-hours” for the same reason:

3wH = 3 hours @ 1 watt = 18 minutes @ 10 watt

Then I’d say you walked 1.5 miles in 30 minutes. The equivalent to saying generated 15 watts of power over 30 minutes from a panel that’s providing 30wh of power.

I’m assuming what I’m saying is wrong, I just can’t spot it. Thanks for the help.

Whelp... you just killed this tech by linking it to the Devil's number. hope you're happy. /s

I want to thank you for your edit, and for keeping the original text. Recognizing something you said ignored portions of the argument, changing the argument, and keeping the original portion to show _why_ you were wrong but still adds to the conversation is very valuable. I commend you for your ability to double guess yourself

Thank you. And you’re right, I am amazing ;)

But seriously the main reason I write these comments is to work through my thoughts as I try to articulate them and get feedback on what I’m thinking.

I’m actually happy when someone points out a mistake because I learned something.

When I’m right, I haven’t gained anything. Sure, my ego enjoys the satisfaction of being right but my already inflated ego doesn’t need anymore stroking.

Heh sounds like what Socrates said - it’s better to be corrected than to be correct.

It's even better to be correct after being corrected by yourself.

My biggest issue is when I hit send on something when what it really needed was another editorial revision.

I almost instantly hit the edit button after I post a comment.

The article says they can get 1 liter for every 20 watts. So that'd give them 17,597 ounces per kWh.

It does say that; and you've converted between watts and watt-hours. The article's claim doesn't make sense.

I feel like nobody watched the video (although I do wish the article text had been clearer).

The video specifically shows (at 1:34) that the machine operates at 15Wh/L (for the larger version, the smaller version is 20Wh/L).

This is 15Wh/33 oz (since we seem so set on using customary units), or 2200 oz per kWh.

I didn't; I don't generally follow video links, unless I deliberately went looking for a video. Specifically, I don't follow many video links from HN; I treat HN as a text/plain site.

So my remarks were only about the article text.

The video claims that the machine operates at 15Wh/L (for the larger version, the smaller version is 20Wh/L).

This is 15Wh/33 oz, or 2200 oz per kWh.

That's roughly one order of magnitude greater than your estimate.

> The suitcase-sized device...can also be driven by a small, portable solar panel,

> Yoon and Kang used machine learning to find the ideal combination of ICP and electrodialysis modules.

These bits feel like they were added because ML and solar power will get more clicks.

> The research was funded ... the Experimental AI Postdoc Fellowship Program of Northeastern University, and the Roux AI Institute

That's the cause of the AI part.

The solar panel bit probably came from the Army requirements, who's another sponsor.

True, especially using ML to determine the size/layout of the components. That could be very easy assessed by varying the sizes in a consistent, linear way, and measuring the effectiveness in subsequent tests, rather than throwing it into the black box of ML.

ML maybe, but solar power definitely not. They show an example use-case as a single-family house in a site of flooding and devestation. An emergency desalinator would definitely need to run off of a small generator or solar.

If this unit is shown to be durable, I imagine it's the kind of thing a lifeboat would have as part of its equipment. Low power means low volume of drinkable water (as energy is needed to separate ions from seawater, by the laws of entropy), so that's really unavoidable with this system.

Compare to things like concentrated solar for complete removal of solids from wastewater and seawater, here's an interesting example (relies on condensation of steam):

https://influencing.com/pr/99083/novel-solar-powered-water-t...

This thing is going to cost more and be far less capable than a traditional suitcase or even hand-pump desalinator and you never see those on lifeboats.

>> The researchers also created a smartphone app that can control the unit wirelessly and report real-time data on power consumption and water salinity.

No no no no. I was all onboard with this seemingly wonderful product until I hit that line. The things that provide fresh water, that actually enable life functions, should NEVER involve controller aps. Switches. Dials. Positive on/off switches. Maybe the occasional touchscreen. But please do not put the need for a working/charged/connected cellphone between the user and their source of fresh water.

And they say that this thing uses about as much power as a cellphone charger. Wrong. It requires the power of a cellphone charger to filter water, plus a second charger to power the cellphone running the app.

A cell phone app would be handy for remote monitoring and control. Say you lived a mile away from the ocean where it is deployed. You could turn it on and off depending on the tide and know when the container you are filling is full.

Having just completed a home remodel where all the home automation features have an app, having apps is a handy way to monitor and control your hardware without leaving your couch, or even if you are off-premises.

But does the pipe bringing fresh water into your house use an ap? There are things in this life that are expected to "just work", that are so basic they should not be trusted to aps. This product isn't for a fancy house in a city. It is a portable product meant for use in remote areas, by soldiers no less, far away from working cellphone towers or reliable power supplies. Serious consideration should be given before attaching anything more complicated than an on/off toggle switch.

Right, the more I learned about computers and automation, the more I want dumber stuff. Intelligent doorlock? Alexa lights? Ring? Nooo thank you, I don't want to be left out of my home or without light because of some internet error.

I like my electronics like my dogs: dumb as a mule.

I'm sure there is a manual override if your phone is dead and you have to walk over to activate it.

>> The researchers also created a smartphone app that can control the unit wirelessly and report real-time data on power consumption and water salinity.

> No no no no. I was all onboard with this seemingly wonderful product until I hit that line. The things that provide fresh water, that actually enable life functions, should NEVER involve controller aps. Switches. Dials. Positive on/off switches. Maybe the occasional touchscreen. But please do not put the need for a working/charged/connected cellphone between the user and their source of fresh water.

You are objecting to a strawman of your own creation. Before you hit that line, you read “The technology is packaged into a user-friendly device that runs with the push of one button.” and “The researchers designed the device for nonexperts, with just one button to launch the automatic desalination and purification process.”

And then comes the line that triggered your reaction: “The researchers also created a smartphone app that can control the unit wirelessly and report real-time data on power consumption and water salinity.”

The key word is “can”. “Can control”, “can report”.

So you have the simplest possible device: press a button on the device, extract drinkable water. You want status updates / remote control? Use the app, which is optional; nowhere in the article it is stated that the app is necessary.

So you might reconsider your stance and get back onboard with the product.

> requires only 20 watts of power per liter.

So if you want a liter, you have to supply 20 watts. For how long? Forever? Does the water instantly become turbid when power is removed?

Tsk, tsk, MIT. I would have thought an MIT journo would know what a watt is.

> Their prototype generates drinking water at a rate of 0.3 liters per hour, and requires only 20 watts of power per liter.

Not great phrasing I agree but the information is there.

Nah, still doesn't make any sense either way.

Watts is a measure of power, not energy. The amount of power (given they use a pump and electrodialysis) shouldn't depend on the volume. Energy OTOH does.

So either their device requires 20W of power - which sounds reasonable as the image shows a <100Wp solar panel next to it - and the volume figure is meaningless.

Or the author left out crucial context (e.g. is the power draw correlated with the speed, i.e. 20W @ 0.3l/h).

The amount of energy would be 67Wh/l regardless given those numbers. It's just a confused mix of performance (processed volume per hour) and power requirements (which is independent of volume and should only depend on the performance).

The trailing "per liter" seems the problem: I think maybe what they mean is that 20W gets you 0.3 liters per hour? So their machine produces 0.3 liters per hour, and draws 20W?

I wish journos wouldn't bandy around terminology and statistics they don't understand. It's getting worse.

It isn't, though, because that would be 60Wh per liter, but the video actually specifies it's 20Wh per liter.

They just stupidly used watt instead of Wh in the article text.

When someone uses watts as a unit of energy, they mean watt hours. It's very common.

... and in the video they specify 20Wh/L.

If I had a dollar for every project MIT's press office has said will be revolutionary and end (insert major world health/water/resource/energy/food problem), I'd be a very, very rich person.

I'm not sure why MIT's press office gets so much favor here, particularly given their/MIT's track record.

Post articles by science journals and news outlets that will tell us if this is actually going to work or if it's just yet another go-nowhere project.

At least this instance has a workable demonstration. Looking at you battery tech...

If this works the way it looks like it does, these devices will be a literal life saver. Would love to see a buy-one-donate-one purchase plan.

Southern California is beginning to see more water restrictions - each and every desalination tech advancement is a very hopeful development for the region, and many others in the world that can't get enough freshwater. I hope they can generally work this up to factory scale and put it behind a solar farm and see if that can supply a lot of homes.

The problem is that southern California is trying to grow a shit-ton of crops that aren't suitable for the environment there. The almond industry is particularly guilty in this regard.

Desalination will never scale to a level suitable to supply that industry. If anything they'll take that water and just grow even more crops, while still sucking the aquifer dry.

Thing is, aquifers compact when you draw water from them too much. That compaction can never be undone. They are slowly rendering that land permanently uninhabitable.

> Desalination will never scale to a level suitable to supply that industry.

Are you sure about that? It's simply a question of economics. As long fresh water from other sources is significantly cheaper, no one's going to invest in large scale desalination.

Once this changes it becomes a question of which is more expensive - shutting down the agribusiness or running large scale desalination (which has options, from nuclear to solar to the use of metamaterials).

It is true in California, because you need to dump the concentrated brine somewhere, and doing that at industrial scale will run up against environmental protections.

There's a big bucket right next to California called the Pacific. Perfect place for sourcing water to desalinate and dumping brine back into without doing anything measurably harmful to salt concentrations in the Pacific. It's like a literal drop in the ocean. Most big Californian metropolitan areas are right next to it. There is no water shortage; just a reluctance to pay a fair price for it.

The challenges in California are not technological but political and policy related. Local consumers of water seem to assume that water is like manna from heaven that has to be perpetually subsidized by the government. They feel entitled to it but are not really willing to pay for it; or even invest in common sense ways to reduce their consumption of it because it doesn't really cost them anything.

I recall similar in function devices already existed for military use - except they made water from air, which is even more available than seawater:

https://www.honeywell.com/us/en/press/2021/05/honeywell-led-...

https://edition.cnn.com/2014/04/24/tech/innovation/machine-m...

Fun fact, there's vodka made from San Francisco fog: https://time.com/4317269/fog-vodka-san-francisco-hangar-one/

Am I reading this wrong? It uses less energy at higher production rates?

This is a verbatim copy of the text at 1:32 in their video:

Seawater >> Drinking Water

~ 20Wh/L for 0.3L/h of production rate

~ 15Wh/L for 1.0L/h of production rate

so it uses LESS energy to get MORE drinkable water per hour?

That seems odd.

I assume this means with a larger device...

Okay that makes sense

They show that the 1L/h version uses more elements.

I assume there are some efficiencies as you get larger.

Those specifications are much worse than 30 year old PUR-06. It makes one liter per hour with 20 Watts and fits in your pocket.

I tried solar panel and windshield wiper motor. But direct-drive windmill was much better as wind blows 24/7 on Baja. https://youtu.be/9xYXWISWv5I?t=390

Durability was the issue. I destroyed three units and average was 200 days / 1000 liters, because all-plastic construction. Found no limits on filter durability.

I was planning to make those crucial breaking parts from Carbon Fiber myself, but I was such a poster-boy for PUR that they sended me free PUR-35 ($5000). It makes 3 liters per hour and if you replace the solid-iron pumping shaft it weighs about a kilo.

Except one crucial high-pressure valve cannot be made of plastic. I made better one from fiber glass. Thereafter it been running ok for 20 years.

What about filtering particles which don't have a charge like plastics or biological matter?

I guess you can always just boil the water after it's been processed by the portable machine.

I didn't quite get this part either -

> The membranes repel positively or negatively charged particles — including salt molecules, bacteria, and viruses ...

Are bacteria and viruses indeed electrically charged? This sounds strange.

Yes [0], "Bacterial cell surfaces possess net negative electrostatic charge by virtue of ionized phosphoryl and carboxylate substituents on outer cell envelope macromolecules which are exposed to the extracellular environment."

[0] https://doi.org/10.1016/S0167-7012(00)00224-4

Not sure this is better than reverse osmosis... They mention ~15Wh/L @ 1L/hr... The Katadyn Powersurvivor 40 (pretty compact and 11kg) uses 8.5Wh/L @ 6L/hr...

https://www.spectrawatermakers.com/us/us/89650-8013438-Katad...

Once an RO filter gets clogged, it stops working.

From the article:

The limitations appear to be expensive materials and scale.

20w per liter .3 liters per hour

The article claims 3x more efficiency in a confusing sentence:

”Their prototype generates drinking water at a rate of 0.3 liters per hour, and requires only 20 watts of power per liter.”

So the unit only produces 0.3 liters an hour but if you wait 3 hours and provide 20 watts of power (7watts/hour), you’ll have a full liter.

> So the unit only produces 0.3 liters an hour but if you wait 3 hours and provide 20 watts of power (7watts/hour), you’ll have a full liter.

That doesn't make any sense. 20W of power draw over 3⅓ hours is ≈67Wh (Watthours, not Watts per hour), so 67Wh per litre (energy use; independent of time) and it takes 3⅓ hours to get this litre of potable water with 20W of power.

In other words to get 1 litre in an hour, the device might require ≈67W of power (assuming that kind of perfect scaling is even possible with their design).

The wording in the article is specifically:

> requires only 20 watts of power per liter.

Whether the article is accurate or not I think you missed that.

The point is that 7W/hour is nonsensical. It's a non-unit and physically doesn't make sense in any context. The power is independent from the volume. Total energy use is volume dependent.

20 Watts per litre is mix of units that's nonsensical. Watt is Joules (Energy) per second (time), i.e. power. Either time is irrelevant, that is the device draws 20 Watts of power, in which case you can ditch the litre. Or the power draw is dependent on the performance in that more power equals higher throughput.

Neither case is in any way shape or form correctly described by "20 Watts per litre". Hope that makes it clearer now.

Science reporting is not great but colloquially people use 20W per <job done> as a way of meaning Watt-hours.

If this is correct, they're saying it takes 20 watts (meaning Wh) to deliver a Litre of water, over three and a bit hours. The time is provided just obfuscated by their "we deliver 0.3L/hour" statement, so by saying 20W/L it does sort of make sense within the context.

~6W/hour (if it means we use a total of 6Wh) would be different than claiming the system runs off 0.1 Watts. Given the size of that solar panel, I suspect they're claiming they are using ~6 Watts for a period of one hour to produce 0.3L. Alternately, they're just flat out wrong and they're using 20 watts for one hour to produce 0.3L.

It could be worded better but I think this is what they're trying to say is that they're consuming 6W.

> Science reporting is not great but colloquially people use 20W per <job done> as a way of meaning Watt-hours.

After watching the video I can confirm that it's actually 20Wh/l for 0.3l/h and 15Wh/l at 1l/h (I'm a bit puzzled how that works, but hey - that's what the inventor says).

> It could be worded better but I think this is what they're trying to say is that they're consuming 6W.

I would be happy if they'd manage to simply repeat what was said in the video. This is the official MIT news portal after all and I would expect it to be run by people smart enough to use the correct units, but alas...

> expensive materials and scale

Aka, we'll never hear about this ever again.

20 watthours per liter of drinking water is not bad and it's way more efficient than the dehumidifiers. Looks like a great solution even at that slow production rate.

Anyone else notice in the beach video that the researcher filled the cup using tubing that was previously on the ground, then drank it? Looks cool though :)

He did wipe the sand off the output tubing... good enough?

I guess if I dropped a drinking straw on the ground in an e.g. parking lot, I'd throw it away, but if I dropped it on sand in the beach, I would think "just wipe the sand off and it's clean!"

Back in the day when I went camping with my parents we used to "wash the dishes" using seawater and sand from the beach. Nowadays people are unnecessarily germophobic.

This is a pretty cool development, even if they made this thing the size of a 20ft shipping container it'd be a pretty good development.

With advances, perhaps one day it will be the size of a pack of cigarettes. And thus portable enough to include in an emergency kit!

So how is this distinguished from existing commercially available continuous electrodeionization?

Uses less power and is portable? That's pretty much all I could gather from the article.

That's its claimed advantages against reverse osmosis.

Amazing! Is this a game changer?

Maybe. I think Atmospheric Water Generators are much More likely to change the game since they don’t need a body of water other than the atmosphere.

There are big thermodynamic problems with those, the phase transition requires a lot of energy... Generally the main places where they work with any kind of acceptable efficiently and decent yield (where there's high enough humidity) are also the places where it tends to rain a lot!

“Atmospheric Water Generators”

At first I thought you were just using a marketing term for dehumidifiers but I looked it up before posting.

I see that term is used for tight mesh fabric that’s hung vertically as a passive matrix to collect condensation.

While those are actually game changers, they unfortunately only work in special regions that have the needed weather and topography to make them work.

I think there’s only one place on earth that has the perfect combination for AWG: weather, drought, and humidity. It’s an awesome technology for them but that’s about the extent of it.

I’ll try to find the region and update my comment with more info.

Edit: here’s a video that talks about what I think you’re talking about.

https://youtu.be/YxRONAZoMDk

It’s providing water in the Atacama Desert near Lima Peru.

Edit 2: They’re called fog collectors and apparently Morocco uses them too. https://youtu.be/0F7CQMd6mQ4

I wonder if you could use a similar sort of device with weather balloons to harvest water from clouds.

I could totally see a Project Loon style craft which automatically surfs the crosswinds, going up above the clouds to radiate heat to the open sky at night, and descending to condense water.

Would dry regions that need water have enough cloud coverage for that to be viable?

My guess is that any place with sufficient cloud coverage is already getting enough rain water to satisfy their needs.

I think there are parts of coastal California where there are clouds more of the year than rainfall. I'd also suspect if you just raised them up with balloons to cool them off and then brought them down where it's warmer, you could condense evaporation from the ocean, without spending energy directly on cooling.

You’d have to raise them up pretty high to get them colder than the dew point.

The energy it’d take to raise and lower the material would make it cost prohibitive.

I considered that the balloons would be a balancing force but don’t forget the rope length changes as the balloons go up and down. That changing length shifts the weight around. You can’t be balance against it.

> It’s providing water in the Atacama Desert near Lima Peru.

I think one big issue is that in winter (May-September) not every day is foggy, so sure its helpful but not a perfect solution. It does get really foggy though in June-October. Thus we get natural vegetation in otherwise arid hills around Lima, the most popular called Lomas de Lachay[0].

[0] https://en.wikipedia.org/wiki/Lachay_National_Reserve

Atmospheric water generation (AWG) uses technology to produce potable water from surrounding air.

Here's a cool example of using the sun to do just that...

https://youtu.be/VQRAtwz3Igs

https://en.m.wikipedia.org/wiki/Atmospheric_water_generator

Probably not. It's more expensive than available alternatives (according to the article they use expensive materials) and the performance is poor compared to commercially available alternatives.

Might be useful for certain niche applications, but outside of that it's more of a proof of concept at this point.

Reply on news.ycombinator.com