Mike's passive solar greenhouse design/build

There is heat coming from the centre of the Earth all the time.  Sometimes the sun is warming the surface of the Earth and sometimes it isn't.  But over the course of a year, there is a net flow of energy out of the Earth.

If we insulate the perimeter of a foundation with the insulation nearly horizontal, what is seen from the point of view of the Earth?  There is a "ring" in insulation, through which little or no heat can flow.  The heat that would normally go through that area, has to divert to either flow further away or to flow through the foundation of the greenhouse.  Any heat that diverts to the foundation, helps keep our greenhouse warm.  If the insulation is vertical, there doesn't seem to be a diversion needed.

The larger a diversion of heat a given plan produces, the more heat we have reporting to the foundation of the greenhouse.  Or footprint.

It may be that this presentation breaks down when the greenhouse has significant heating within itself.

Concrete is a good conductor of heat, you do want to insulate the outside surface of the foundation, the top of the foundation outside of where the pony wall sits, and the outside of the pony wall.  That is to "save" the heat coming up from the Earth, that flows into the concrete perimeter (and can be easily lost).

But, when viewed this way, it may be that there are locations where a little extra insulation does well.  Do we want to have a little extra near the corners, or away from the corners?

Thanks Ebo, that's a book I haven't read yet.  Luckily our library system has it so it will be in my hot little hands soon.  I heard of the horizontal insulation a few years ago as a "shallow frost protected foundation".  I like it because I don't have to dig as deep and it makes the design more flexible for other folks with rockier soils.  I don't know if it's better at capturing or holding the heat bubble under the greenhouse as compared to a vertical insulation system.

Good point on building the template with two sides.  I could also just remake the template after doing one pattern (screw on the blocks instead of gluing them).  So I guess that isn't much of an issue after all

Hi Gordon, I like the idea of viewing the heat as it leaves the core of the earth.  Given that perspective, the horizontal sloped insulation should divert some of that heat from outside the foundation into the footprint.  Basically a 3' skirt could send 1.5' of heat into the greenhouse.  Yay.  

Now I'm not sure what happens when the greenhouse is also adding heat to the soil.  Is that heat held in better by a vertical or horizontal insulation?  Is that heat overwhelmed or underwhelmed by the heat coming up from the core?  Beats me

I think for frost protected foundations they do add more insulation in the corners.  But that may be for unheated structures so I'll have to look into it a bit.  Unlike "normal" buildings I won't be clearing snow around this one (sidewalks, driveways around a house).  So there should be snow also protecting the surrounding soil from the cold.

I am going to insulate the outside of the pony wall and footing.  If I lay it out well, the distance from the pony wall to the edge of the footing will be 2" so the insulation can just shingle over the adjacent piece.  I'm also hoping that I go with styrofoam on the surface of the North trusses so that it can also be contiguous with the pony wall insulation.  

@Mike, there is an online PDF scan available.  If you look back in the link I think I found it here.  Maybe you are referring to another book and I'm ust confused.

Anyway, I agree that when you have a situation where you need to compromise, that this is a good approach, but it would be interesting to either find someone that can run some simulations or instrument around the greenhouse so that you can actually measure it and figure it out.  Not sure what the protocol is without installing some simple sensors in place (like the DS1820 1-wire sensor and an arduino -- something like this https://www.aliexpress.com/item/Free-Shipping-1PCS-Digital-Temperature-Temp-Sensor-Probe-DS18B20-For-Thermometer-1m-Waterproof-100CM/32501970086.html which you can get locally, but are expensive enough that it is worth buying 50 of them and burring them around the place...  Just a thought...

@Gordon, I agree.  As I said I do not have an intuition.  I would be very interested in seeing if some student of such things did a paper on the cost/benefits of angle.  In fact I bet that if I/we dug around in the scientific literature a bit we would find it.  I could see it as working very well, or basically negating the effect of going straight down -- because there is a direct connection with a lower temperature soil, but it has to travel further...  

Maybe these *might* help:
 https://aceee.org/files/proceedings/2004/data/papers/SS04_Panel1_Paper15.pdf
 
hmmm... need to go to be as it is a "school night".  Will look into this a little bit later if I have time.  For my applications I am physically constrained and down is the best way to go.

This looks at the problem, but does not give numbers, just suggestions:

   https://www.ceresgs.com/solar-greenhouse-basics-insulating-your-foundation/

really quickly these might be of some interest:

   http://www.engext.ksu.edu/files/engext/publications/envelope/builderguide.pdf
   http://www.greenbuildingadvisor.com/blogs/dept/qa-spotlight/best-way-insulate-foundation
   https://web.ornl.gov/sci/buildings/conf-archive/2001%20B8%20papers/188_M_Swinton.pdf
   https://energy.gov/energysaver/weatherize/insulation/where-insulate-home

now it is time for bed...

Ebo, I just realized that scan I mentioned above IS the book you were talking about.   When I pulled it up on Amazon I didn't recognize the cover.  So the confusion is all mine.

Burying those sensors would be pretty cool.  But unless I have someone who's going to analyze them, I probably won't bother  I should check with the university to see if they're interested but past attempts did not reveal much interest.

I was dreaming about heat flows most of last night.

There are places which are boring from a geological and geographical basis, for me SE Saskatchewan would be one.  Heat flows from down deep would parallel the gravitational vector (or rather, be anti-parallel).  But if the location involves a slope, a lake is nearby, there are geothermal features in the vicinity, or maybe just fault lines, they could alter how these heat flow vectors would move towards the Earth/air surface.  I think what happens when the heat vector hits the Earth/air interface, it orients itself to be perpendicular to the surface.  So at the interface, there could be something analogous to a refractive index where there is a discontinuity in the heat flow vector.

For me, living on a long north facing slope or someone living near a lake, I think the heat approaching our foundation from deep will see that the heat approach is not orthogonal.  This then may bias where a person puts extra insulation to maximize how much flows through the footprint of the building.

I think where these converging heat flows meet (such as the heat flow from deep and the heat flow down from the surface when warmed by the sun) is just manifested in a raised temeprature.  Which would be where the vertical insulation would act to prevent this heat from exiting out a shallow bottom.

I had looked for source code to run thermal simulations, but drew a blank.  I've been building up "horsepower" to tackle some weather problems.  I have 34 amd64 cores on my LAN now, and a few Polaris GPUs.  For the most part, all they are doing is BOINC.  But at some point, I want to either write my own BOINC stuff, or use MPI to get these computers working on problems.  So, if you happened to know of code, I might be able to run a simulation or two.  (My server has not had its hardware upgrades yet, it is to get 6TB of RAID-10, a SSD for just a dbase to use, and swap a HD6450 for a RX-550.)

Albuquerque.  See I can do it too.  Snorkel.
      -Riley Poole in National Treasure

Hopefully someone gets the joke.  

I do have a shallow lake about 200 yards away from the greenhouse site.  And the water table is probably 8-10' beneath the greenhouse.  I have no idea if that helps or hurts.  My shallow well puts out 48 degree water in January so that water table is probably holding, limiting or distributing the deep earth temps on my site.

Your greenhouse is 8 feet higher in elevation than the lake surface?  My guess is that you just extend the lake surface under the ground.

I think it was last spring, but I looked at the edges of my dugout in the spring time.  There were tunnels from near the high water mark going back into the bank all over the place.  Im guessing mice tunneled in, looking for a warm winter home.

If the water table at your greenhouse, is still substantially at the lake level, I would say it is going to substantially affect temperatures.  It will be much harder for frost to get into the ground near your greenhouse, than someplace where the ground is dry.


There was some story I think out of Las Vegas, about a curved building and reflecting the suns rays onto the pool area.  But in any event, with your north wall being curved, there is the possibility of you focusing light at some distance inside the structure.  Having a diffuse paint on the north wall would tend to minimize that, but it is something to look at.

Yup, I think the greenhouse is in the realm of 8-14' above the lake surface.  There's a rise between the two so I can't eyeball it.  There's also a wetlands about 8' below the greenhouse and to the West about 100'.  So the lake is 600' east and the wetland is 100" west.  It's basically sitting on a large island of ground that is above the water table and dry.  The wetland and the lake are connected several hundred yards to the South.  There is minimal water flow in the wetland so I believe it is only slightly higher than the lake surface.

I did see a few write-ups about a trial greenhouse somewhere in the SW USA where they framed up the South wall deliberately to focus sunlight on a big pool of water.  My impressions were that they couldn't get enough heat into the water to make it worthwhile.

I sketched the reflected light lines on the winter solstice and they may generate a band of bright sunlight near the North wall.  As the season progresses the band will change and move.  I do plan to paint that wall with some diffuse white paint-like substance to spread that light around.  So hopefully I won't burn any ants or start a fire in there on a sunny day

You could rig up a spirit level, and measure how much your greenhouse is above the lake surface and/or wetland.  Plastic pail, long length of tubing, some long enough sticks to pound into the ground to make levels on.

If you had 100 feet of clear tubing, you would 6 sticks to transfer the level of your greenhouse to a stick in pounded into the lake bottom.

WRT your joke, I missed it entirely.  Mind you, I've never been to Albuquerque.  When doing my M.Eng. at C-MU many years ago, there was one student from New Mexico (I am not sure where) and one from El Paso.

Yup, I'm sure I could figure out the elevation if I was ambitious enough.  So far I haven't been that ambitious

And the joke is at a point in the movie when the experts are bouncing jargon back and forth using their weird words.  Riley has no idea what they're talking about so he gets into the conversation by just making stuff up.

@Mike,  there are a LOT of messages that I have to read through to catch up, BUT for starters -- I analyze environmental data as a NASA contractor day in and day out.  If you are willing to instrument it up I will process the data for you.  I might even be able to automate it (so you can check things out online, but I have to look at the long term data availability, etc.).  The big question is establishing the experimental methodology and protocols.  It is likely that is already defined in some peer reviewed paper some place.  Let's look around and see if we can find something appropriate...

As for mistaking the scan of the book, etc., think nothing of it.  

@Gordon on simulations..., at one time or another I have programmed in 40 different langues (well if you could the different scripting, shells, and assembly langues separately).  I mainly now do Python, C/C++, Fortran with a liberal splash of bash.  Before you jump headlong into MPI, and if you do any Python, you might want to look Dask frames (which are basically pandas statistics on a fully distributed/parallel delayed computing model -- really efficient).  Ping me off list and I may be able to point you to some interesting tools if they are related to anything I do at work.  BTW, if you have any experience with thermal propagation modeling through soil we should collaborate to get this going.  Hmmm.. there is other stuff we should talk about as well...  BUT those are discussion for another day.

@Maike, Gordon has it right with establishing a level with a clear tube.  We can look around and see if anyone has done a video explaining  how to do it, but it is quite simple and should give  you a good idea of the water table depth to within a good inch depth within in 100'.  Actually, if you are Type-A about it, you should be able to get it down to about 3mm over the 600'.  Also, this is a cheap and very good technique to establish the level of you greenhouse foundation.  Other than that, you could drive a slotted point test well.  Heck, that is so close you could drill a well and do hydrothermal (water cool geothermal).  Also might be a good source of water for your greenhosue.  Check your local regulations...

BTW, I may be in and out of reading the site for the next few days.  To much on my plate, but I would love to set up some code to analyze the heatflows (I have not don that before), and I may be able to help Gordon on weather/climate analysis (but I need to know what he hopes to accomplish).

Hi Ebo.

I prefer FORTRAN, but I too probably have worked in 60 languages.  I've done a lot in Perl.  My M.Eng. thesis was a dynamical system of differential equations on a VAX 11/785 in FORTRAN.  I had doubly linked lists and garbage collection to try and cut down on thrashing the disk with page swaps.

The Peace Country is the watershed of the Peace River in Alberta and British Columbia.  It is about the size of Germany, with maybe 300,000 people.  I think we have 30 official weather stations.  I want to find a way to use global circulation models to predict future weather.  Which means learning how to downscale.  I also have a dugout I want to protect from evaporative losses (I live 5 miles downwind of a 130MW wind farm).  I have WindNinja and one other source package to try and tackle that one.  There was a VM that NASA out of JPL produced for doing weather stuff, that I was going to try and start with for downscaling.

It wasn't obvious how pinging would work on this site.

Since most of my work is currently in Python, is Python one of your 60 languages?  I have been on the list less than a week, so I am still learning the cultural lingo.  They call it something like "moose messages" or something.  I am not sure how to initiate, but I have replied to one 8-/

Something I looked into, and I will send you some links off-list, is NASA's 9 month climate forecasting model.  Basically it starts every month with the current assimilated state of the world, and models the gross precip, temp, and other variables out for the next 9 months.  The cool thin is that it incorporates the effects of El Nino, PDO, and other global scale phenomena.  Then next month it does the same, so you end up with 9x global maps of current temp, percip, and I forget what all, followed by the next 9 month forcast...  When I looked into this as part of a drought forecasting intern project with deliverables to native communities in the Rio Grande valley (I would have been the science advisor for the project BTW), I discovered that the archived products are not available on-line, but they have hind-casted these models to the mid 80's to overlap various NCAR and historic satellite records.  We can get access to those, but it is one of those things that you have to know who to ask or go through an entire FOIA request -- people do not care about releasing them, it is a matter of paying for the disk space on the archive and maintaining it...  I might be able to find downscaled products for you -- what resolution do you realistically expect?  BTW, besides working on identifying all the trees and shrubs across the entire sub-Saheran Sahel, my MS in Plant Biology was in multiscale integrative modeling.

Anyway, we should move this off-list and post back a synopsis what we end up with (because it might just be useful to a number of folks).  We are getting SERIOUSLY off topic, but joyously so ;-)  @Mike, I have not forgotten about you 8-)

WindNinja?  Just looked it up.  Fought forest fires southwest of Albuquerque in the Magdalena Rangers District (remembering the joke about Albuquerque... Snorkel... never  quite got the joke in the movie, much less that the intended reference, but I grew up just south of there and lived a decade or more in ABQ, and the nearest snorkeling available is about 100 miles south in Elephant Bute)...

Hello Ebo.

I started a thread, which may be a place to move some of this.

https://permies.com/t/76820/Computational-permaculture

I will look into the purple moosages.  I once got one for a problem, and I thought that was ll they were for.

I like the revision Mike!
Couple of thoughts:
Horizontal insulation is used in some situations to keep the cold water away from the interior. It adds what is called wing insulation (at least at this site: Wing Insulation ) Water in the soil adds to the chilling effect of the earth, so keeping it dryer helps keep it warmer. Not sure if that is the most useful effect for you, I'm not up on your climate at all, worth thinking on it from that perspective though.

Trusses: A triangle is stronger than a single supported V for the roof, so count on horizontal bars too, and a ridge beam will be needed to keep them spaced, as well as to make your cover make the angle (I qualify that with the comment that I don't understand the whole double poly/blower bit, I may be off base when that's involved.) Have you seen Travis Johnson's Sheep Barn? Interesting construction technique there, might be worth looking at if you missed it. Sheep Barn The part I like is the trusses can be linked together at the top to hold it all still, and stood up and, the weight disperses nicely. Might make asymmetrical trusses easier to work with. The plywood gussets could hold up the ridge beam, perhaps? Not sure.

And yeah, I have to admit I got lost with whole programming stuff, not sure what that's all about, I didn't read it closely, as I don't program and wouldn't understand it. A synopsis would be nice for us lost ones....

Hello Pearl.

I'm sorry for putting too much computer programming stuff in this thread.

I don't think that a person can assume that anything of this nerd nature raised, is going to help much in the near term.  It's entirely possible that it could help down the road.  My inclination is to do things in a freely available nature, but that is not what all people knowledgeable in programming want.

Certainly, when I get around to building my own greenhouse, I want to situate it to the best of my ability.  By and large, because I have to do so much of the work by hand.

At one point, I taught weightlifting (because I kept getting injured playing soccer).  So, I can do a fair amount of manual labour.

Thanks for the link to the wing insulation Pearl, that's exactly what I'm talking about doing.  It looks like they suggest straw as the insulating material for the wing.  I'm thinking I'll go with styrofoam just to make sure it still works 5 years from now.  Unless I'm totally uninformed  I'm not too worried about moisture due to my sandy soil but it will be good to direct water away from the footing.

I'm hoping to avoid horizontal bars if I can.  The footing and the two trusses make a triangle.  But it's a big triangle so I'll have to be ready to add cross braces if I notice it deflecting under snow load.  The cover will stop at the peak so I will need a board that goes E/W along the ridge to attach the poly to.  But I believe it doesn't have to be a full sized ridge beam.  I will have diagonal boards going across the face of the South wall to brace it plus the North wall roofing should also provide good lateral bracing.

Thanks Ebo for the offer!  That's awesome!  I'm fairly dumb at electronics though.  I've heard of arduinos but I have no idea how they work or how to make them work.  But if the data collection part is easy enough for a mechanical guy like me to do, I'll happily take you up on the offer!  

I wish I had a university nearby with some excited horticulture grad students looking for a project...

Arduino is a wonderful search topic, all by itself.  Instructables has lots of Arduino projects, but lots of other exist as well.  Having the hardware be Open, tends to make the software Open as well.  One high school age person (or maybe early 20's?), built a plane simulator for his living room.  It had pneumatic actuators to tilt and slope the "cockpit".  I think it was all Arduino.

There is a tendency for people doing electronics to go surface mount, which as you get older can be harder to focus on.  There is a company (Schmidt ....) which makes "daughter" boards, which are easier to solder a surface mount chip to, so that you can take "pins" out to other boards.

Mike Jay wrote:Hi Kenneth, welcome to the conversation!  

I am a bit worried about the compost IN the greenhouse for the ammonia reason.  Some of my thinking/hoping:

With a primarily "brown" wood chip compost it may not produce as much ammonia as a "properly" C:N balanced hot/fast compost pile.  I believe the NAI/Fulford system and the Bioshelter Market Farm have compost inside but both are manure intensive.

By putting it inside the GH I get to harvest all the heat it generates, not just the bit I can circulate away with water.

If the gasses are overpowering I could seal the compost chamber to keep the ammonia in.  Aeration air would still route heat through the South grow bed but then exhaust outside.  Downside is no CO2 benefit from the compost.

I don't have any room on any side of this greenhouse for an external pile.  So if I want it outside, that reduces the size of the greenhouse.  If I had a big field to put it in I would have more options.

But my hopes/thoughts are based just on reading and dreaming.  Do you think your compost would have offgassed as much if the C:N ratio was tilted more towards "browns"?

My main compost heat harvesting methods will be:

Radiation/convection off the surface of the hopper

4" pvc pipes that enter the chamber near the bottom, rise up through the pile and then exit the top to create chimneys of ambient air heating

Aeration air being sucked through the pile and then blown through the South planting bed

Option to add coil of pipe on the underside of the roof of the chamber to circulate water through.

I'm hoping to not coil pipes through the middle of the compost so that loading and unloading won't be a massive pain in the ass.

I do like the simplicity of running low grade heat through pipes that are shallow in the ground.  Perhaps on the surface and covered with mulch so that when I'm digging I don't have to worry about puncturing pipes.

I did see the Builditsolar plan for the check valve at one time but the link isn't working any more.  Do you have a copy of it by any chance?

Mike, I guessed that you might not have any space to move the compost outside, based on the clipped corner on the other end. ;-)
Sadly, no copy of the BIS plan.
Making the compost more towards "brown" would definitely have solved the ammonia problem.

So the pipes in the ground could be for water or air to store heat in the soil, and can be placed deep enough to not be dug into when planting.
The warm air could be from the peak or from your compost... and... the compost exhaust can be bio-filtered through finished compost under the seedbeds (pgs. 96-97 of GB's CPWH shows one example) which handles the ammonia and odors, but also keeps the CO2 inside.
You've got three different methods of extracting heat from your compost planned, keep in mind that as ambient temps drop the pile activity slows way down and can stall.

Why not make your North wall/roof more conventionally framed? (like the Univ. of Missouri solar GH) It could have a vertical wall that might make shelving or hanging of your SHW pipes easier, and make that clipped NE corner easier to frame.
It would save the time of making the (now a different shape) trusses for the North side, and maybe shorter time to complete. (My projects seem to take three times longer than I hope...especially if there's "custom" parts I'm making)
Lots of things get simpler, flat wall panels for insulation and sheathing without cutting strips, joist hangers and brackets that fit the lumber...

@Pearl, sorry for going all geeky.  Some of these things are SO esoteric that I am overjoyed in finding someone that has thought about the problems.  I'll bet that Gordon and I can come up with something useful -- as a note I have started dreaming about using some of the AI machine learning techniques I have been using to optimize various greenhouse design parameters based on historic (and possible projected IPCC scenarios over the expected lifetime of the greenhouse -- say 30 to 50 years).  Imagine looking at a rough design and having a program telling you what the optimal size/spacing of the vents, depth of foundation/insulation, glaze angle (and type of glazing)...  A lot of this stuff can be optimized based on local conditions.   To be clear, writing that would be a labor of nitpicky love, and my time would be better spent elsewhere, but I might implement it while I am studying the equations -- BTW, I did the same thing when studying geodesy.  My instructor was floored when I walked in with a hand crafted C program which reduced the observations to the surface without least squares and I closed my measurements 1  in 65,000 over a 1km transect.  After I wrote that program and took those measurements, no one questioned whether I understood the material or not!  So Gordon and I need to fet our curiosity and see what fruits we might have to share afterwards...

@Miake, we should be able to feed you with hardware/software designs, etc. to make it happen.  The sensing is little more than:

https://create.arduino.cc/projecthub/everth-villamil-ruiz/temperature-sensor-ds18b20-3decfc
https://chrisramsay.co.uk/posts/2014/04/soil-temperature-monitoring-part-one/
https://chrisramsay.co.uk/posts/2014/04/soil-temperature-monitoring-part-two/
https://chrisramsay.co.uk/posts/2014/11/soil-temperature-monitoring-part-three/

multiplied several times over (and the scaling is trivial).

You might also look around at the high school 4-H students for a research project.  The heavy lifting was mostly done decades ago, and it is mostly a little literature review and building out standard designs.

 More later,

 EBo --

Kenneth Elwell wrote:Making the compost more towards "brown" would definitely have solved the ammonia problem.

Awesome, that's reassuring!

Kenneth Elwell wrote: the compost exhaust can be bio-filtered through finished compost under the seedbeds (pgs. 96-97 of GB's CPWH shows one example) which handles the ammonia and odors, but also keeps the CO2 inside.

Yes, I've seen those "biofilter" options, they do complicate things but they are a possibility.

Kenneth Elwell wrote:Why not make your North wall/roof more conventionally framed? (like the Univ. of Missouri solar GH) It could have a vertical wall that might make shelving or hanging of your SHW pipes easier, and make that clipped NE corner easier to frame.
It would save the time of making the (now a different shape) trusses for the North side, and maybe shorter time to complete. (My projects seem to take three times longer than I hope...especially if there's "custom" parts I'm making)
Lots of things get simpler, flat wall panels for insulation and sheathing without cutting strips, joist hangers and brackets that fit the lumber...

I've tried to come up with layouts with a vertical North wall and a sloped North roof but the issue usually boils down to then needing a ridge beam and posts to support the peak.  Or making the north wall and roof into triangle trusses of some sort.  While I don't need a clear span structure, I kind of want one.  Maybe I'm overthinking it but I can't figure out how to easily make flat surfaces work on the North side.  I am definitely open to ideas here though.

If I go with styrofoam insulation on the North side, I can use thin enough sheets to bend to the radius.  Then if I run horizontal purlins I can bend corrugated metal roofing to the radius.  Or I could try the billboard tarp idea.  

Ebo David wrote:@Mike, we should be able to feed you with hardware/software designs, etc. to make it happen.  The sensing is little more than:

https://create.arduino.cc/projecthub/everth-villamil-ruiz/temperature-sensor-ds18b20-3decfc
https://chrisramsay.co.uk/posts/2014/04/soil-temperature-monitoring-part-one/
https://chrisramsay.co.uk/posts/2014/04/soil-temperature-monitoring-part-two/
https://chrisramsay.co.uk/posts/2014/11/soil-temperature-monitoring-part-three/

Thanks Ebo!!!

@Mike,  you are welcome ;-)

"I've tried to come up with layouts with a vertical North wall and a sloped North roof but the issue usually boils down to then needing a ridge beam and posts to support the peak.  Or making the north wall and roof into triangle trusses of some sort...."

If you look at truss ag buildings, you make the cross section of south to north wall a truss, and you can make the walls flat or curved as you like.  It is a bit more complicated, but not overly much so.

probably @Mike, but this might be generally useful:

Bisoniya, et al 2014."Study on Calculation Models of Earth-Air Heat Exchanger Systems". Journal of Energy 2014(859286): https://www.hindawi.com/journals/jen/2014/859286/
The nice thing about this one is that it discusses a situation where they use a U-shaped heat pump installed in a water filled borehole.  You probably can set this up in your place.

Some information on energy loss:  http://www.sensiblehouse.org/nrg_heatloss.htm

more later...

I'm wading through a 2015 NREL report on foundations.  Or is it 2017?  One of each?  Doesn't matter.

NREL tends to use EnergyPlus (from Lawrence-Berkeley Labs) for modeling building thermal performance.  This is just the building.  Supposedly this software is "free", but it would seem only in so far that you can get it without cost (even on Linux).  But, to run the software, you need a password.  And I don't see a software source code archive mentioned.

If we look closely at heat transfer across an interface, it is the microscopic roughness of the interface that determines the actual contact area between the two volumes heat is being transferred between.  In the case of a building and the ground, there is an "atmospheric" volume which can be air with a varying water content, or it can be (impure) water.  There may be radiative transport of energy across the gap as well.

A particular program meant to handle that foundation/soil interface was BUFET (from UMinnesota).  As I read it, this was intended to be a freely available _database_ of building data (not a program, just data).

As of 2015, the problem of modeling a building with a basement in the ground could be handled by many runs of EnergyPlus and BUFET.  I think the estimate was 30 CPU hours.  The reason is that the two models don't couple well.  If the temperatures for the basement are higher for BUFET than they are for EnergyPlus, the model is stable and inaccurate.  If the BUFET values are less than the EnergyPlus values, the coupling is unstable and the physics breaks down.  So, you have to run the two many times to creep up on the answer.

It may be because the greenhouse doesn't have a basement, that one doesn't need  BUFET, and can avoid the problem.

More recently (this is the 2017 part), the professor(s) at UMinnesota that are involved with the engineering company in quesiton, rewrote BUFET and it can now couple of EnergyPlus on this problem much more easily.  But, BUFET-B is now proprietary.

----

It's entirely possible that CSIRO (Australia) has a comparable thermal modeling system.  From what I've observed in the past, if CSIRO has a solution, there will be no freely available stuff.

So, maybe some place in Europe has software?  I'll look.  I think I need to read more about EnergyPlus and BUFET so that I can more of the jargon clear in my head.

----

I do wonder if perhaps some kind of physical model might not help.  I'm think two different kinds, one for use in the summer and one for use in the winter.

We prepare a level piece of land (civil engineering wants to build on undisturbed soils) that is the kind of soil the foundation is on.  From working in agriculture (GIS stuff), there are often 1 or 2 layers of soil at the top, that are different than the deeper soil.  Maybe there are places where there are 3 layers?  Anyway, you would need to get below that, and then make a level spot of sufficient size.

Our model consists of an aluminum plate which we will coat on the bottom with a thermal grease (to couple it to the Earth), and some other parts.

In the summer, what we want is for our model to melt ice and have the melt flow away, we need some "drainage" ditches cut into the surface of the aluminum, I will guess at a grade of 1%.  Which means our level spot has to be level to better than 1%.  On top of the aluminum plate, we set a 3D printed mold.  The inside skin is smooth, and has holes to allow the melted water to drain out.  The outside is either "wrapped" in insulation of some scale thickness, or is of some fractional density to represent being insulated.  We fill the mold with a piece of ice (probably a piece cut out of a flat sheet?), and then set that onto the aluminum.  We set a piece of foam insulation on top (so that all the heat for melting comes from the ground).  What we are essentially looking for, is how long it takes to melt the ice.

For winter, the mold doesn't have a flat bottom, it has creases in X and Y, so that when the water freezes, it deforms upwards and doesn't break our mold.  And I will guess we put cameras even with the top of the mold, and we are looking at the deflection of the ice as a function of time (telling us how much water has frozen).

Having a thermal grease that is nontoxic and doesn't freeze to the ground is probably needed.

----

I haven't spent a lot of time thinking about this physical model, it was what came to mind in trying to invert the problem.  There are probably holes in it.

Not enough coffee.  I see that you can get old source (tarball or zip).  Maybe that is the place to start.

Tarball unpacked fine.  Cmake did its thing, and nothing seems to be missing.  Make is about 60% through.  Some interesting warnings in the compiling.

Compile finished.  Just warnings, no errors.  Hopefully none bad.

I need to look around a bit, but it would be nice if there are test cases it can run to see how things are working.

I ran across a Python facility called eppy, which shows how to run EnergyPlus.  I think I will need to "define a greenhouse" in terms of walls, ceiling, airconditioners, furnaces and the zillion other things that EnergyPlus has in it, as well as come up with a weather file for somewhere in Wisconsin.

Hey Mike, where's the nearest airport to you?  Or, the one you think has the weather closest to you?

I'm not finding other software which solve the same problem more simply so far.  But, the example using eppy didn't show that it takes long.  It looks like the time consuming part is defining a "house" and getting the weather data.

What I think would be nice for a test case, is the steel shed you can buy at one of the box stores.  Maybe 10'x10' in size, and probably 8 foot tall.  Made out of steel, no insulation.  Set it on a concrete foundation (or slab) with a gasket between the steel and the concrete.

Thanks for the links Ebo!  I think the sensible house one on "Calculating heat loss" was interesting.  If I understand it, the chart that most closely represents my greenhouse would be the Unheated Slab one.  If that's the case, vertical insulation would be better than wing insulation.  I'm not sure if either model is accounting for the heat rising from the earth.  The rainbow color picture on the right may but I'm guessing the picture on the left doesn't.

It also suggests that for horizontal insulation, R7.5 is almost as good as R15.  For vertical the R matters a bit more.  

So if I'm still using wing insulation, I can skimp down to R10 and save some $$$.  Also, my layout will be 1' vertical and then 3' horizontal.  That isn't in their model but I'm guessing I should do R20 on the vertical part and then R10 for the horizontal.

One other reason I'm liking the wing insulation is that I don't have to dig as deep for the footing.  With the wing I can just pour a footing at grade and put a pony wall on it.  If I was excavating for a 4' deep footing, then I'd definitely go for the vertical insulation.

Using the earth temps to heat has always been a struggle for my brain.  I know I could dig way down and get to 47F soil.  Circulating air through that would give some heat to the greenhouse.  But my intuition tells me that if the greenhouse is getting cold (say 40 degrees), pumping 40 degree air through 47 degree soil will raise the temp to 45.  Blowing 45 degree air into a cold greenhouse doesn't seem like it will do too much good.

Now using the temperature of the earth to warm incoming exterior air when I'm venting from the roof on a sunny day in winter does make sense to me.  40 degrees is much better than -10.

Gordon, the closest airport for weather reports is probably RHI.

If you think of a refrigerator, it might help.  We pump low pressure working fluid (could be air) into the ground.  It approximately equilibrates with the ground.  We do work on the working fluid to increase its pressure.  A side effect of this, is that it also gets warmer.  We extract the heat from the high pressure working fluid (to equilibrate it to the warm side conditions).  And then we drop the pressure (which cools it).  The cycle repeats.  But to me, that is what a refrigerator or heat pump does.  I am not a mechanical engineer, who lives with that kind of thermodynamics.

---

Do you know Sketch Up?

I ran across a paper from a couple of guys out of Montreal, who looked at a greenhouse on the top of an existing building.  But, basically a slab on grade situation, possibly a little colder than you might see.

The paper suggests there is a way to look at the SketchUp file, to come up with a data file to use with EnergyPlus.

Comparison of Different Mechanical Systems Models for a Passive Solar Greenhouse With Two Thermal Zones
Leveille-Guillemette and Monfet
Department of Construction Engineering
Ecole de Technologie Superior
Montreal
IBPSA proceedings of some kind 2016

And I am missing diacritical marks, and I spelled some stuff wrong.

That paper references a paper by Piscia and Munoz, which looks interesting to me.

Again, I apologise if I mangled names.

I could put in a heat pump to harvest deep soil temps but I'm hoping to keep this whole design simpler and cheaper...

I got SketchUp last winter when I first started on this design.  I quickly learned that by the time I got a design decently represented in SketchUp, I was overly invested in it and it was hard to do design changes.  I'm sure if I was more capable at SketchUp the changes would only take a few minutes.  So I've moved on by using (gasp) graph paper and a pencil.  The current layout is my 40th geometry.  Hopefully we're done with the major shape changes.

So now I could put it in SketchUp.  If I get enough time/ambition to make it worthwhile.  I can see how a temperature model could be a useful output.  The only other reason to do it is to have a pretty picture to share.  Maybe there are others?

There are some weather files included within the EnergyPlus tarball.  I think the closest one to you is Probably Chicago Ohare.

I'd guess that Minneapolis would be closer to my conditions than Chicago if it's an option...

Next closest places I think were in Colorado (Golden and Denver I believe).  Not at all comprehensive.

@Gordon, Woo Hoo... thanks 8-)  It will probably be a few days before I can try to catch up, but this is exactly the direction I was hoping to go.  Depending on were we go with this I can look at throwing some machine learning techniques to optimize the parameters of a given design -- like depth and thickness of foundation insulation, angle of glazing (given long-term weather/climate data), etc.  BTW, I have to crank out a few deliverables for work before I can expend that kind of mental energy even on my private time, BUT...

@Mike, depending on what the data looks like I may be able to pull the weather/climate data for you (or give you instructions on how to download and maintain your own).  I had 3 emergencies come up today that I must focus on before focusing on this.  Also, in a few spare moments today I found this: http://www.homeintheearth.com/tech_notes/basics-of-earthsheltering/earth/soil-properties/soil-temperature-experiment/  I have not had time to review the full protocol, but it is along the lines of what I was looking at...

Do the two of you speak git or mercurial?  Maybe we can set up a github or bitbucket repository and share specific versions of code, data, etc.  Let me know and I will set it up.  I will want to cordinate with Gordon to make sure we are using exactly the same code, etc...

 EBo --

The most exposure I have to git is from OpenWRT, but I have git and mercurial installed.  And I believe they can work with each other (sort of).  I am not really comfortable with any version control system, but I sort of know what they are trying to do.  I have one box here using etckeeper to track changes to /etc with git.  But so far, there hasn't been any need to actually use git to undo anything.

(I use ImageBuilder with OpenWRT to build images for my router.  I don't do this often enough to be comfortable with things, but some things rub off.)