Indirect Evaporative Cooling an H13 Hexayurt

[Plumitt]

Over the past 9 years, I've been building, rebuilding, and otherwise tinkering with different cooling systems. I've arrived at a "solution" which keeps my H13 Hexayurt at 70-75F at 40% relative humidity using only solar power and 4G water/day.

This year was the first where I was able to run entirely on solar and have the yurt cooled every day I was there Setup time is about 12-14hr solo, less than half that with another person helping. I still brought a backup generator, because I lack faith and hate heat. I just didn't use it.

20180902_191830.jpg

Along the way, I have built a mister-driven evaporation column, 4+ different evaporative coolers , two air-to-air heat-exchangers, three solar power systems, a mylar dome cover + door (16' 2V geodesic ) , a shade structure for 20 people, an H13 Hexayurt (2" RMax + radiant barrier), a 6' stretch Hexayurt with a novel sandwiched material walls (mylar, corrugated plastic, RMax), and even a data logging system to track temp/RH throughout the cooling system. In other words, I'm obsessed.

Disclaimer: I'm not asserting the following is an easy, cheap, small, or minimum power/water solution to cooling a hexayurt. It does work, though. All prices are approximate.

20181009_095120.jpg

OVERVIEW
The system I'm now focussed on couples an evaporative cooler with an air-to-air heat transfer unit, i.e. Indirect Evaporative Cooling as they say in the HVAC-biz. Ambient indoor temperature inside my H13 Hexayurt is about 25F below outside temperature (e.g. 73F on a 98F afternoon) at about 40% relative humidity -- and virtually no dust! This is achieved using about 4Gal of water/day and ~300W of power.

In the current configuration, the heat-transfer happens AFTER The evaporative cooler, and is used ONLY to cool the internal air. This allows me to get the cold, but not the humidity. This is not power efficient, but it leave the inside of the yurt completely comfortable.

20180831_152557.jpg

SOLAR
I'm finding that solar panels on the playa do not deliver more than about 60-65% rated efficiency at the peak of the day. 600W of panels deliver about 2.5-3KWhr/day, or over 8-9 cooling hours, about 300W. Dry blowing or sweeping the panels gives about 5% over doing nothing (the first day, it'd get worse I'm sure) but pouring water, squeegy-ing, and wiping with clean paper towels improves things by 5-10% additional, i.e. it's significant. Panels are 6 100W Renology ($800) on the ground facing south with the appropriate inclination as advised by solar panel calculation sites. They are mounted in pairs onto 1.5"x4"lengths of wood using normal z-brackett flipped over to make the wood mount closer to flush, and a horizontal length of 3/8" square tube steel to provide rigidity. Each pair weights perhaps 40-50lbs, and has a simple setup using nylon cord and three lengths of wood as legs to provide the tilt. They are staked down in case of high wind.

I had run the system with 12V with panels all in parallel using the cheapest MPPT charge controller I could find. The lack of sufficient information about system state, plus the improvements in line-loss, reduction in wire size, etc lead me to the current setup with 24V batteries (2 80AHr AGM batteries in series, $400) using a Victron Energy Charge controller, battery monitor and battery protector ($275 total, has bluetooth monitoring!), a charge balancer ($30) a 12V inverter with a 24-12VDC converter ($200), DC circuit breakers on the solar, batter, and inverter ($60), cables & a couple cheap inline voltmeters ($50).

I removed the standard solar power connectors and went to spades and terminals as I found dust in the connectors plus repeated connection/disconnection was making them unreliable. The battery hardware is mounted in an old tourister hard-side makeup-box carry-on, and the inverter/controller/etc are in an old CDJ+mixer flight case to keep them safe in transport and dry during winter storage.

EVAPORATIVE COOLER:
The evaporative cooler is built into a plastic footlocker and uses modern cardboard-honeycomb evaporative media to achieve 30degF+ temperature deltas. I would not use the footlocker again, as after 5 years the tape began to fail and I had to use a bunch of spray sealant and tape to keep it watertight -- yuck. Building into a rubbermade or a large food cooler (both watertight, the later has a drain) is a a better option

The evaporative media is vaguely like corrugated cardboard, but with bigger holes, wavy layers, and variable density -- http://www.coolingmedia.com/celdek/ . I use an 8" thick section, about 16" x 20", though the usable area is likely closer to 12"x18". Water distribution is via 4 lengths of 1/2" PVC pipe with tiny holes drilled every 1-2",. fed by a single 1/2" flexible hose from a 12V pump submerged in a 5 gallon bucket, with the hose ziptied to a lenght of rebar to hold the pump down. (don't ask about the less clever ways I mounted pumps.) The cooler sits just above the bucket and drains water back into it. The bucket is a perfect place to cool drinks (beer, coconut water, etc) down to about 58F.


HEAT EXCHANGER
To avoid bringing the humidity inherent in evaporative cooling into the living space, the cold air is used to cool in the internal air *without mixing* via an air-to-air heat exchanger unit. The HXU is built out of corrugated plastic (eg. election signs), cut into about 40sheets 1'x3' and sandwiched with channels going in alternating directions. By feeding internal air through one path, and the cold-wet cooled air through the other, the internal air is chilled. The now-slightly-warmer wet air is vented out. (by using two baffles, im effectively getting a counter flow heat exchanger with 3 cross-flow stages out of a single 1'x1'x3') stack of plastic) (Exlcuding salvage plastic, $50 for tape and caulk etc)

Typical steady state mid afternoon temperatures are:
Stage 1, evaporative cooler: Intake air 95F. -> Cooled air: 64F. (Water supply: 58F)
Stage 2, heat exchanger, wet side: Humid cooled air: 64F -> Exhaust Air 73
Stage 2, heat changer, dry side: Inside Air 75F -> Output air (70F)

FANS/DUCTING/AIRFLOW
I've been using 6" AC centrifugal duct fans (about 100W each), but the inverter losses and the rather-painful inductive load at startup is leading me to consider fully switching to 4" marine bilge fans (I've used two midway through the main airhandling loop with good effect, but they are LOUD. (duct fans: $120, bilge fans, $80, ducting $100?)

The air intake/outflow vents are venturi effect vents (roofvents.com) which produce airflow via the wind blowing across them -- reducing power usage. ($100)

I've tried using flexible plastic ducting, but the static pressure loss is unacceptable. Rigid-flexible metal ducting seems to be a good compromise. The ducting is set up (one T-juction and a couple quick-connect hoses) to allow me to easily run just stage 1, or just the exhaust.

A good door, thorough taping, plus "mudding" the bottom edge of the yurt on the outside is necessary to avoid airflow, i.e. convective heat loss



Next steps:
Back to a cooler for the evaporative cooler container.
Add a precool stage using exhaust air
Switch to entirely DC.
Reflect further on the fact a generator and AC would be cheaper, lighter, and easier in every way.

Questions, comments, thoughts, welcome. More pictures to follow someday.

Apologies for typos, out of time for today.

[Plumitt]

Lost two pictures. Odd.

20180831_152557.jpg

20181009_095120.jpg

[Traveller in Time]

Wow, impressive, extensive and overdone to an artistic extend

Welcome to ePlaya

[Canoe]

... couples an evaporative cooler with an air-to-air heat transfer unit, i.e. Indirect Evaporative Cooling as they say in the HVAC-biz. ...
EVAPORATIVE COOLER: The evaporative cooler is built into a plastic footlocker and uses modern cardboard-honeycomb evaporative media to achieve 30degF+ temperature deltas. I would not use the footlocker again, as after 5 years the tape began to fail and I had to use a bunch of spray sealant and tape to keep it watertight -- yuck. Building into a rubbermade or a large food cooler (both watertight, the later has a drain) is a a better option...

wow
a lot of work

How many years use are you getting out of the "cardboard-honeycomb evaporative media"?

I'm working out a swamp-cooler design that goes in a similar rectangular plastic storage bin. Aimed at being easier to fit on a table or on the floor of an RV. Going dual bin, the evap cooler above and whatever sized reservoir below. This seems that it will be the most flexible in builders modifying it to fit their needs.

... To avoid bringing the humidity inherent in evaporative cooling into the living space, the cold air is used to cool in the internal air *without mixing* via an air-to-air heat exchanger unit. The HXU is built out of corrugated plastic (eg. election signs), cut into about 40sheets 1'x3' and sandwiched with channels going in alternating directions. By feeding internal air through one path, and the cold-wet cooled air through the other, the internal air is chilled. The now-slightly-warmer wet air is vented out. (by using two baffles, im effectively getting a counter flow heat exchanger with 3 cross-flow stages out of a single 1'x1'x3') stack of plastic)...

I'd love to see photos of the details of the air-to-air exchanger. This is extremely relevant for me for Heat Recovery Ventilators.
Getting the benefit of counter-flow with the difficulty of feeding the flows vs. ease of cross-flow construction with coroplast, I've never found an easy or elegant DIY for this. (the early R2000 era HRVs in my area seemed to favour a crappy cross-flow coroplast design... and wanted $2,000 to $4,000 for them)

... I've been using 6" AC centrifugal duct fans ... consider fully switching to 4" marine bilge fans...

Yup. Those certainly seem like a noisy and inefficient way to move the air. Take a good look at the VDC Endless Breeze fans. Take the screens off and they're even quieter - and easy to build into a box. With the right sized rectangular box for the evap filter box, the Endless Breeze can screw directly to the side of the box (inside or outside, depending on how you want to do the fan switch) (only has a 1/2" or less water level in that box, as the return water flow is out the bottom of the box into the reservoir box below; or through a side fitting to exit to the reservoir, if you've got a strong enough pump to feed from a remote reservoir).

... Add a precool stage using exhaust air ...

Pre-chilling is certainly worth doing, but perhaps not with exhaust air. But with what you are able to build, you could certainly take some of the chilled water and use that for a liquid-to-air pre-chiller for the incoming evap cooler air. Which would be a classic two-stage evap cooler: output air should be around 50% R.H.. That's only 10% above your target of 40% R.H.. That would mean you can skip the interior-isolating air-to-air stage (with all its ducting, noise & power requirements) and feed the 50% chilled air directly to the interior, which gives you the benefits of:

• chilled air at 50% R.H.
• interior air is replaced with chilled fresh air that has the dust filtered out of it in the evap cooler stage
  (this also pushes out the dust floating in the interior air)
• slightly pressurized interior (all of those little leaks will leak air to the outside, not allow hot dusty air in)
• drop/remove all of the secondary air circulating fans
  • lose all of the push-pull fan in/out balancing requirements
  • lose the noise and power requirement
• only the noise & power for primary fan(s) moving air from outside through the evap-cooler)

Perhaps you could use the air-to-air for the pre-chiller but chill the pre-chilling air with the chilled reservoir water. Less efficient, more noise & power for such a three-stage over the classic water-pre-chiller two-stage, but perhaps easier to build from what you already have?

[Canoe]

You inspired me.
I went googling.
I found a description of dual-core plate cross-flow, and how the actual airflow (not the diagram arrows), not just that it is two stages, has better than cross-flow performance.
.

dual-core cross-flow.gif

.
Which lead to looking at commercial counter-flow cores. Suddenly I had the obvious for an efficient and easy counter-flow DIY out of Coroplast!

• For whatever size, a full length sheet for the airflow in one direction.
• No sheet for the next-layer (opposite direction): instead, cut & glue strips of Coroplast for spacers, sealing the edges and creating channels. (can also add smaller strips at an angle within a channel for creating more turbulent airflow for enhanced air-to-surface heat transfer)
  I've shown the counter-flow in and out from opposite sides as that will be the most equal pressure/flow, but it could be directed for in/out on the same side of the constructed core.
• Also means single surface between the different airflows. But that's not so important as the surfaces are so thin, but it's got to help.

Build it inside a box made by gluing rigid-sheet-foam?
.

Coroplast counterflow.gif

[Canoe]

And I've found the sheet-strip-sheet-strip, etc., Coroplast construction method online, but they were all building cross-flow designs:

• One guy started using silicone (but not clear, so you can see it), but switched to using double-sided tape as it was easier.
• Also examples of heat-welded at the edges using a soldering iron with the heat turned down.
• Instructions saying to use silicone or PL-200.

[Canoe]

With that Coroplast construction I'm pretty sure I can work out putting headers on an extension of the straight-through layers for a water-to-air heat exchanger. But the cross-section would need to be over double that of the airflow cross-section/pipe-size.
It'd be nicer if I could use the straight-through layers for the airflow and have the between layers for the chilling water, but there's weight and pressure containment to deal with.
The commercial liquid-to-air pre-chillers use cross-flow stages stacked in series for counter-flow, so perhaps there's an easier construction method for a liquid-to-air exchanger in Coroplast.

[Canoe]

... Add a precool stage using exhaust air ...

Pre-chilling is certainly worth doing, but perhaps not with exhaust air. But with what you are able to build, you could certainly take some of the chilled water and use that for a liquid-to-air pre-chiller for the incoming evap cooler air. Which would be a classic two-stage evap cooler...

The reservoir's chilled water may not have enough heat capacity to adequately pre-chill the incoming air. But you could run a second evap cooler to provide air-to-air pre-chilling and its chilled reservoir for a second water-to-air pre-chilling stage.

[Plumitt]

Here is an example of a DIY air-to-air counterflow HX. Very Thorough.
https://www.builditsolar.com/Experiment ... escription

Here is a comparison of three different crossflow designs. Interesting that the sheets-and-spacers method performed poorly. I'm guessing insufficient turbulent flow.
https://ecorenovator.org/forum/conserva ... er-37.html


My (somewhat gainly) design is a 1' x 1' x 3' block of cross flow layers, with baffles on the sides to make one air path S shaped, and baffles on the top and bottom to make it easier to hook up ducting. The baffles are also coroplast. there is a simple metal frame made of predrilled 1/16th? angle iron to hold the whole thing up and keep the bottom baffle from being crushed by the core.

Indeed, piecewise? counterflow from 3 crossflow stages.

The layers were glued together with Locktite 401. All edges were taped and sealed on the inside with caulk.

Screen Shot 2018-10-10 at 2.03.33 PM.png

[Canoe]

Thanks for the details, and the links.

Interesting to see the baffles out of Coroplast. I wouldn't have expected that to be strong enough. That will certainly make construction easier.

Good knowing about the Loctite 401.
Are you happy with the 401? As in, would you use 401 again, or would you try PL or heat-welding?
Hmmm. Shows Loctite 401 as a Cyanoacrylate. Have you had any vibration failures with it and the Coroplast?

For a full length counter-flow, using spaced layers, easy to add some short strips to create some turbulence, and you have single surface, but no meaningful amount of rib surface in the spacing 'layer'. But I've no idea how that would perform vs. alternating sheets for rib surface in all layers, but that has back-to-back surfaces (although you can add some thermal paste between them ($).

Series of cross-flow layers is the standard in liquid-to-air pre-chillers, and fits into most units nicely as a 'pre' stage, but something says go counter-flow.

I figure I'm just going to have to build some smaller test cores. And some bonding tests.

[Plumitt]

The adhesive challenge hasn't been affixing the alternating layers in the core. As far as I can tell, that is rock solid, 5ish years after putting it together. (I'll confirm this weekend, but I've observed no problems.)

It's been tricky to seal any corner where the baffles meet HX core. I've used silicone caulk (mostly, I think). However, I do not have perfect sealing between the two paths, and I haven't yet diagnosed where the problem is.

If I were to do this again, I would consider making an inset wooden frame for the HX core to avoid trying to seal sheet coroplast to the HX block along its edges. Perhaps even better would be to use inset angle for the 4 vertical corners, The baffles aren't super strong (gotta carry by the bottom frame, and can't lay it on it's side), and it is an awkward object to move around. I'm also not at all sure how much flow or uneven flow Im getting through those U turns.

ALSO: I gave you the wrong Loctite Glue. I've been using Loctite Go2 - Polyoxysilane. (I think I used a numbered one in the past, but I can't find a record of what it was. Maybe there was a product rename in the last 5 years?)

Finally -- it]'s almost time to go out snag election signs. Free coroplast in quantity.

[Plumitt]

I'd really like to build a regenerative IEC or M-cycle IEC.

This guy has written a good overview of various evap cooling techniques and built a cool prototpe R-IEC.
https://scholar.google.co.uk/citations? ... AAAJ&hl=en

This prototype notably doesn't need a combined evap/heat exchange core like Coolerado (recently acquired by a larger HVAC firm) M-cycle coolers, and instead uses heat pipes, metal fins, and porous ceramic sleeves... Something of this ilk seems easier to construct than a tight wet/dry matrix.

Heres a pic of that prototype. Note that all humid air is exhausted. The path goes:

intake -> finHX-1 -> ... -> finHX-15 -> (remove half of air into room) ) evapStage-15 -> ... -> evapStage1 -> exhaust
where the finHX and evapStagres are each connected with a heat pipe (fluid copper tube thing.)

Screen Shot 2018-10-11 at 11.40.14 AM.png

Having just been at RePC ... I wonder if a series of (cheap cheap) CPU heat exchangers could work coupled with a series of small-ish evap stages... water distribution being the biggest problem that immediately occurs to me.

[Canoe]

LOL!
I was going to build similar as a roof-installable for my car and my trailer. Only copper, not ceramic. With the upper water also being an evap "pond" for chilling, with a lift up to an evap area by capillary sheets.

But that drawing shows a tight fit for the water jacket. Looks like it's relying upon convective flow to keep the top of the HP cooled. Looks like the jacket water temp could rise and reduce the delta-T of the HP ends. I like extending the HP directly into a large & common water chamber that is kept appropriately chilled to maintain delta-T. Perhaps the ceramic has a meaningfully longer life in the water than finned/pinned tops to the HPs. I think I'd prefer a pinned top for greater surface area and turbulent water flow, just give it a surface cover of ceramic paint (brake caliper paint for the DIY).
(Finned HP laid horizontal between an intake and outtake airflow works very well for HRV, but all the HPs are pricey)

If the orientation of the HPs is to remain vertical, then the HP function is solely as a thermal diode. The HPs don't need sintering, as gravity will do the return faster, spilling outside of any sintering or channels and into the vapour flow. A wide enough pipe (which you want for surface area anyway), and you don't need to worry about protecting the returning liquid from being blown up by the upward pressurized vapour flow, so loose the sintering (and its expense and slow return) and let the liquid return adhered to the pipe's smooth surface.

I'll have to do some reading.

p.s. Thanks about the Loctite Go2 - Polyoxysilane
You could try the adhesive silicone (like used for aquarium glass) instead of silicone caulking. There are cheap 'not aquarium approved' tubes that are used by DIYers for aquarium glass and are regarded as the tried & true gotos. Check Reef Central for the current recommendations.

[Canoe]

Coroplast glue test
https://www.christinedemerchant.com/coroplast-glue-test.html

[Token]

Wooly Fucking Mamoth Batman!

I just have Honda generators with an outboard tank and get AC going in 2 - 3 minutes tops ... but damn ... that is some sexy fucking science.

Crazy but awesome!

[Plumitt]

Here are photos of the heat exchanger and the evaporative cooler.

https://photos.app.goo.gl/w79neTgqb1yPTq318

Comments and questions welcome.