Mike's passive solar greenhouse design/build

@Gordon, good to know.  I can walk you through the simple stuff, and it we break something I can probably fix it.  The idea is to share working content to move forward.  So basically I will not ask you to do anything more than pull, status, add, commit...

@Mike,  have you ever used git or mercurial (hg) before?  Basically you do something like "git clone http://... /some_project.git" and you end up with the current version of all the stuff in the project. From there it is standard stuff...

That Home In The Earth experiment was pretty cool.  So he basically found that rain percolating down through his sandy soil didn't have much of an effect on the temperatures underground.  If he had, putting additional plastic down as skirting may have been worthwhile.  FYI my soil is sandy loam for the first 12-18" and then quite sandy from there on down as far as I know.  I suspect that some clay appears down at the depth of the wetland.

Ebo, don't get in trouble at work over this project.  Get your other deliverables done first

I have absolutely no idea how to speak git or mercurial.  My background is more mechanical and equipment process improvement based.  I worked with electrical engineers and I know a lot of what can be done.  I just have no idea how they make it happen.  I can install sensors, run wires and hook them up to things.  The data logger in the above experiment looks like something I could handle.  I'd be completely fine with a place to store/share data but I've never done anything fancy like that before so I may need some guidance.

So if it's ok with you guys, I'll happily let you do all the modelling/analysis/predictions/documentation/white papers/geeking out/etc with my deepest appreciation.  Just tell me what to install and where to do it and I'll do my best.  Keep in mind that I'm trying to stay within a budget...

Thanks guys!

Mike

I have a friend that lives a few hundred miles north of you, in SE Manitoba.  He has told me that the soil is lots of sand.

One tree I suggested to him, to try and grow, is the tallest/largest pine known, the sugar pine.  Your latitude is not that far removed from northern California.

If you have an interest in learning how any of this works, try to be involved.  Things on the Internet have made it much easier for people to work like this (I am not very far ahead of you in that regard).  But for people to work together on a programming source code project, is exactly the same problem as people working together to make a document.  It is just a chunk of text.  And that is how git, mercurial, svn, vcs and a bunch of other things approach things.  Git and mercurial are among the newest of that list, and (maybe) they have better ways to do this.

If you decide to work on a significant document (how to wire birdhouses) by yourself, there is no problem.  You write, you save data occasionally (in case the power goes out or some other problem).  Eventually you are done, and you publish it (in some way).

If you decide to work on a significant document with 1 other person, you and this other person can make edits (changes) that affect the same sentence or phrase of the document at the same time.  Same being defined in terms of when you both "checked out" what was there "now" compared to what both of you saved at some point in the future.

It is possible that this other person deleted "Fred Flintstone" in a section of text that he later saves, where you  made changes complaining how "Fred Flintstone" was underpaid by "Mr. Slate" at the time.  There is a conflict between the two edits, and tools like git and mercurial are there to help you resolve the conflict.

So that is a conflict between two people and edits.  You can have conflicts between many people and edits.  They are more involved as to how to resolve.

But, what git, mercurial and others are trying to bring to the table, are tools to resolve conflicts in text.

And it doesn't matter if it is source code, or screenplay for a movie.

Oh, I think my budget is smaller than yours Mike.  

I am going to work with Gordon and Mike to set up some collaborative stuff.  While it is intended as a notepad for our play, it should be there for awhile (maybe forever) to help model and understand these structures.

Ok, I'll try to learn the data/software stuff.  If I have to.  I guess.    (Red Green reference)

Maybe I'm missing something but is the goal to measure the performance of the greenhouse so we have some good data to share and guide optimization for the next person?  Or is it to predict performance beforehand to guide the design of this greenhouse?  Or both?  Or neither?

I love Red Green (thanks for the allusion ;-)  

My personal goal would be to predict the behavior of my own modified design, and then measure where appropriate to post validate.  It is likely that I would not go the full monte on validation, but easily 1/4 of the instrumentation can be used in actively manage the greenhouse operations or passive monitoring to tell me that "Houston we have a PROBLEM".  Getting to the point where I can do my thing, I am glad to help others achieve their dreams and goals.  For me to build a passive solar greenhouse has been for decades a fantasy.  With the references that were passed along here I think I can get pretty close to having it fully passive and build it myself.  NOW...  The stuff others on this list are interested in are close enough that helping both of you gets me two steps closer to what I dream of ding, AND you are both willing to help some.  I call that a WIN, WIN, WIN!!!

Keep your stick on the ice Mike.

Gordon Haverland wrote:Keep your stick on the ice Mike.

I love you guys ))

Update on EnergyPlus.  I installed the current source tree a few days ago.  If you only run cmake, it just sets up the base package.  If you run ccmake, it gives you options to "compile" lots of other things.  One of the things you can compile, is the documentation.

The documentation doesn't compile.  This morning I updated texlive, and it still doesn't compile. Every time if fails on renaming a file, which really isn't a TeX error, more like a makefile error (I think).  

There are a bunch of FORTRAN routines for supporting EnergyPlus, and they compile.  There are a bunch of other packages which can also help.  They have always been failing with a GITHUB_TOKEN error.  There is a bug thread about this which I read most of.  I thought this was fixed, but maybe my system is hooped in some way?  In any event, this error seems to be involved with the production of a changelog.  I kind of wonder why a problem in producing a changelog (entry) is a compiler error, and not just a warning.

I have other things to do, so maybe I'll get a chance to go looking into these errors in a while.

Ebo David wrote:I love Red Green (thanks for the allusion ;-)  

My personal goal would be to predict the behavior of my own modified design, and then measure where appropriate to post validate.  It is likely that I would not go the full monte on validation, but easily 1/4 of the instrumentation can be used in actively manage the greenhouse operations or passive monitoring to tell me that "Houston we have a PROBLEM".

Hi Ebo,

Are you saying that you are going to work to write some greenhouse automation/monitoring software? I'm a developer by trade and I'd be interested in collaborating with you on this.

@Kevin, basically yes, although in the post I was referring to applying a technique called "evolutionary engineering" applied to the parameters of a basic greenhouse design.  By that I mean if you roughly know the style of the greenhouse you would like to have, but am not sure the optimal angle for the glazing, ventilation area, or thermal regulating mass, then you can write a program to modify those parameters and rerun an energy balance analysis tool like EnergyPlus.  So in the end if you changed the glazing angle by 5 deg, and increased the insulation by an extra R7, it would make the difference between year round rowing with passive heating and cooling, and having to heat or cool the greenhouse a month or two out of the year.  That, or you learn that the cost of multiple glazing is not worth the extra cost if you install an automated thermal blanket (this is something that people have been learning about their greenhouses in the last 20 years or so).  

More to your comment though, yes I plan to write or adapt available open source software to data log, monitor, and control various aspects of the greenhouse.  Any yes, I would be glad of your participation.  I already have some collaborations sketched out, but nothing is nailed down yet.

@Gordon, I'll try to take a look at this and see if I get any traction.  I have had a couple of must-do-now's come up that will suck up almost all of my time for the next couple of weeks, but I have seen this type of thing before.  I'll email/post back if I find anything.  Best of luck on the tasks at hand.

After too many days of problems, I supposedly have EnergyPlus compiled and internal documentation generated.

My initial look is that if a simple diagram input is available, it is hiding.

EnergyPlus requires a data file to work from, and having seen one or two examples, I don't think that much of this is going to come from a simple drawing.

By and large, greenhouses are simple buildings.  They tend to have a rectangular footprint, the long axis is E-W or close to it.  Conventional greenhouses are something like parabolic in cross section.  Chinese style greenhouses have a vertical north wall, and a slanted front wall/roof.

Apparently a general rule in building thermal analysis, is that all surfaces are declared as heat storage units, unless you know that the surface is a heat transfer surface.  Well, in a conventional greenhouse, there are 3 external surfaces, and they are all heat transfer surface (and have vanishing amount of heat storage capability).  A Chinese style greenhouse has one surface that is all heat transfer (north wall), one that is all heat transfer (glass front), and two end walls which could be all heat storage, all heat transfer, or some of each.

Typically a greenhouse has no basement.

It may be, that either type of greenhouse is simple enough, that some kind of "20 questions" interface could write the data file.

EnergyPlus thinks that whatever building you give it, it has a furnace (to make up heat on cold days).  I think the program would prefer that you have air conditioning.  Most of these structures are not ventilated (don't admit unconditioned outside air).  It may be that to work with EnergyPlus, a person need to install a ficticious furnace and AC unit in a "basement" crawlspace.  The idea then being to adjust conditions such that the size of the furnace and AC unit gets driven to zero (not needed).

Or, that is an early impression of mine.

I guess there are some number of real glazing systems.  Glass is one of them, or actually more than one.  Probably single pane and double pane, and maybe more panes.  And possibly a thermal film midway between glazings.  There are probably at least 3 non-glass glazing systems.

Anyway, I've got more reading to do.  And non-greenhouse work to do as well.

Ugggh, too much reading.  Mind you, it's not all one topic.  I got a reminder that I have "deer" problems a couple of days ago, when a cow and calf moose stopped by to eat branches, including plantings on the lawn of birch, mountain ash, apple, ....

I am finding that people are simulating greenhouses with EnergyPlus.  What I am not finding, are example data files.  Many times people are using EnergyPlus via DeisgnBuilder, which is proprietary.  NREL has another "free" package called SOLAREL (?), which is trademarked and source code is not available.  You can download it subject to a license acceptance, but I would guess it is only available for M$ and Apple.  It is meant to be used on buildings that have a LOT of glass, such as a greenhouse.

EnergyPlus has a page which talks about where one can get weather information (and this is of an international nature).  There are more than one source of weather information in the USA.  Some are free, some you have to buy.  A common part of these weather descriptions, is that the data is "typical", it does not include extremes.  One thing I am seeing, is that a lot of this data is not current.  Some I would call old.  In an era of climate change, I think people are going to have to "fabricate" weather information for their location from a combination of locations close by (distance wise) and close by (climate wise).

For an example, I haven't looked at whether I can get Canadian data for some place in the Peace Country, but I suspect that it is data from pre-1980.  Dawson Creek was probably considered firmly Zone 2 then.  I think it is on the transition to Zone 3 now.

I believe some of this weather data may include soil temperature data, which is needed for how the house interacts with the ground.  Otherwise, a person needs to go find soil temperature data too.

Some of the reports suggest that it is relatively quick to run an EnergyPlus simulation.  If that is true, then a person can "fuzz" the weather data in various ways.  For example, you read in a "template" file, and write a file which has "noise" added to the values.  Then run your simulation and collect results.  That is one trial, so you end up doing many trials.  And you are looking to look at the distribution of energy usage (and other similar outputs) that EnergyPlus produces.  To prepare yourself for that kind of thing, you need to look up weather extremes: first frost, last frost, extreme rain/snow events, snow in summer.  That can guide you as to how much noise to add in making these "trial" weather databases to run against.

====

Mikes greenhouse does not appear to be a Chinese style to me.  What sets the Chinese style apart is the "north" wall is filled with thermal mass, and well insulated outside of that.  To me, Mike's greenhouse looks more like a well-insulated greenhouse that isn't symmetric.  So, this might make it more interesting to come up with a data file for it.    (I may be remembering his drawings wrong, so please don't shoot me.)

I think that a useful idea to pursue once I figure out how to get greenhouses to work, is a limiting case.  So, on my north wall I have variable amounts of water in tanks that are an epoxy/carbon nanotube composite (aluminum, copper and silver are all good thermal conductors of heat for metals - they are blown away by diamond, carbon nanotubes and graphene: copper is about 400 (units don't matter), diamond is over 2000).  So, the tank is naturally black, and it is filled with water (what I will vary is the thickness of the tank).  Polyurethane foams get you abotu R8 per inch when new, air filled aerogels can get you about R10 per inch.  Evacuated aerogels can get you R20-R50 per inch.  Of course, as soon as someone pokes a hole in the bag with an errant nail, or a mouse or squirrel chews on it, the vacuum is gone and you might get R10/inch.

So, we assume we can have R100 in the west, north and east walls (no glazing), and look at yearly performance as a function of how much thermal mass is on that north wall.

====

Anyway, enough of a break, back to reading.  I think next break means fabricating crude protein in white clover values.

Hmm, back again.

Comparison of Energy Consumption: Greenhouses and Plant Factories
Harbick, Albright (Cornell University)

Among other things, this paper is looking at predicting how much photosynthetically available radiation is going to be received "today" based on the first few hours of the day, and has lights and other stuff installed so that the plants get the optimum amount of light each day.

A plant factory is a building devoted to growing plants.  It is not a factory making something else (like steel).

A different approach to growing things in greenhouses from what I've seen before.

@Gordon, people are starting to look at next climate zones in ecology for comparing what ecosystems will likely do in the future (http://onlinelibrary.wiley.com/doi/10.1111/j.1466-822X.2004.00088.x/full), so that has a sound scientific basis.  As for temperature, we can look at the satellite record and downscaling that.  It has been a few years since i have worked with anyone downscaling temperature data, but I can try to track that one down (I know people, and know people that know people).  But for now we should be able to use published records for our cases and scale up globally later (for those interested try looking at https://giovanni.gsfc.nasa.gov/giovanni/ ).

The automation of the "plant factory" is intriguing.  I will have to look into it more.  R-100 seems a bit excessive, but R-49 is now required for roofs of new homes in the northern part of the USA.  Not sure what is standard for our brothers to the north, but I doubt it is 2x that.

I'm dealing with renovation stuff at the moment and will not be able to play much, but if you send me specific inquires I can try to help.

BTW, do any of the people using EnergyPlus on greenhouses on this site?  Would you be willing to share some of the data or information?

BTW, I doubt that we could generate the data from simple line drawings, but may be able o from something like SweetHome3D=>DXF or a parametric software (I will have to ask some of my 3D printing buddies what they use, I only looked at it once years ago...)

Having too little thermal mass in the north wall of a Chinese style greenhouse  means the temperature of that wall becomes too high.

That paper about plant factories gets into a lot of evapotranspiration issues, which any greenhouse needs to take into consideration.

Another paper, seems to talk about using solar panels as if they are shades.  If you are getting too much solar input on a particular day, not only can you have a "blanket" to roll over the greenhouse to try and keep heat in, but you might have solar panels that you can "unroll" to provide some shade (and convert the incoming light into electricity, to use for what?).  Conversion efficiencies for solar are at best order 20%, but if you need to shade some part of the greenhouse, this might be the best you can do?

Why not mention the paper?  They provide a reference 2006 to EnergyPlus being useful but difficult to use, and a 2015 reference to a proprietary program which can write data files for EnergyPlus.

It does present evidence about polycarbonate being a better glazing than polyethylene for greenhouses on a couple of points.

If we are going to use solar cell panels for "shades", and only 20% of the incoming energy is being turned into electricity, what happens to the other 80%?  Maybe we need to put cooling into those "shades", to extract some of that 80% for our thermal wall?

I could have more, but I needed to go cook dinner for visitors.

There is some kind of project at JPL (OpenWeather? - No RCMES Regional Climate Model Evaluation System), which among other things can be used for downscaling (more than 1 of N different kinds).  But what they distribute is a VM, and it doesn't look like it is meant for things like MPI to work across a LAN.  I have a copy of the VM here, but haven't played with it yet.

I have seen a few times mentions of using SketchUp for input to EnergyPlus.  But there seems to be lots of things in the input files which are not involved with a drawing.  Kinds of materials is part of it.

I don't think our building codes require anything like R100 here, maybe R60?  But in conventional construction, R60 is about 8 inches of polyurethane foam.  If a person just wanted to see what "excessive insulation" could do, putting in a couple of inches of evacuated aerogel doesn't influence the building design anywhere near as much as trying to find a place for 8 inches of foam does.

Gordon Haverland wrote:Mikes greenhouse does not appear to be a Chinese style to me.  What sets the Chinese style apart is the "north" wall is filled with thermal mass, and well insulated outside of that.  To me, Mike's greenhouse looks more like a well-insulated greenhouse that isn't symmetric.  So, this might make it more interesting to come up with a data file for it.    (I may be remembering his drawings wrong, so please don't shoot me.)

You got it.  It's a passive solar non-symmetrical curved truss building with just an insulated N roof/wall, E wall and W wall.  I was aiming for about R20 in those walls/roofs.  When I did my thermal loss math, the diminishing gains on more insulation was negligible compared to the loss through the glazing.

When it comes time to install sensors for the modeling, how many are we talking and where do they go?  I'm assuming it's 10-20 and some would be under the foundation (to measure heat loss through the soil) and others would be in and out of the greenhouse.  Is that approximately correct or am I massively underestimating the amount of data collection?  

I'm also thinking we could install more insulation on one section of the foundation to see if that makes a difference.

[Somebody changed something, the reply page looks different.]

I think the place for seeing how extra insulation is useful, is a research building.  Which is typically paid for by a grant of some kind.  Not for most ordinary builders.

But, in terms of using EnergyPlus to model buildings, I have a paper now (out of Portugal) where they used EnergyPlus to model a "Living Roof".  Which is going beyond a greenhouse (or even a greenhouse on a roof).  So, I think that this modeling engine has what it takes.

In terms of placing sensors, I think a person needs to first determine what sensors.  The airflow and water vapour movement inside a greenhouse are supposed to be complex, and I think a lot of enthalpy (heat) flow will show up if we can look for those small mass flows inside.

A person can take advantage of the symmetry of a building.  If one was looking at an ordinary greenhouse (which has 2 mirror planes for symmetry), you would probably place sensors  from the centre to the close wall, and from the centre to the far wall.  Most people with multiple sensors, place them according to an equal spaced grid.  I think it is useful to have a few sensors placed in "replicated positions" just to get some information on whether the mirror symmetry really is there.  I also prefer the idea of placing sensors at spacings derived from the zeros of a Chebyshev polynomial, instead of being equally spaced.  This results in wider spacing near the centre of the building and closer spacing near the walls.

Could you post the citations?  These look interesting.  I agree with the insulation -- more is better, but how much is cost effective...  Also, changing the amount on one side or another is a research project.  I would be surprised if the burred insulation made much of a difference, but then again, the north side will never see daylight and I would expect it to be colder.

Here's a paper out of Brazil, on finding useful and cheap ways to look at evapotranspriation issues.

http://www.scielo.br/scielo.php?script=sci_arttext&pid=S1415-43662004000200004

And our living roof (aka rooftop greenhouse without the glazing).

https://fenix.tecnico.ulisboa.pt/downloadFile/395146457482/Extended%20Abstract_Marcelo_Silva.pdf

Ok team, I have a bit of new info to report and some questions.  I was scrounging at the local Habitat for Humanity Restore and the manager said he could collect up some sliding glass doors for me.  If my calculations are correct, six doors on their sides would be about perfect to act as the lower vents on the 40' South face of the greenhouse.  I could just rig up some hinges and have a post that they attach to when open.  

I figure I'll leave these vents open all summer, closed all winter and maybe one open in the spring/fall.  Snow will pile against them in the winter which I'll shovel away as needed.  I like the beefiness of the glass for that snow load.  

Two concerns that I'm curious about are:
1.  When the vents are open and at about a 20-30 degree angle from horizontal, would they be able to handle quarter sized hail?  Or dime sized hail?  I've only ever seen 5/16" hail here...
2.  I think many sliding glass doors come with low E and other films applied to them.  Will they cut down the solar gain?  Double poly should transmit 82% of the light and energy.  I'm hoping glass will be even better but maybe not with the films?  Is there a way to test them before I buy?

Truss construction
The trusses will be rip cut pieces of LVL beams glued to chunks of 2x4 per Dan Huisjen's post here on Permies.  I got some full sized 2x4s from a buddy so I can use them, I could use normal 1.5x3.5" two by fours or make anything in between.  So the 2x4s can match the width of the LVL strips.  The LVLs are 1.75" thick and a variety of depths and lengths.  

I'm wondering if I rip the LVLs into 1" by 1.75" strips, will they bend enough for my curved trusses?  The radius of the tighter curve is 16'.  I could rip them thinner (3/4"?) but then I may need to use two for each side of the truss for strength.  Does anyone have a good feel for this?  I'll make one as an experiment anyway but I figured I'd try to start with what we think will work.  I'll rip them on the band saw and plane them smooth if needed.

Other design notes:
My source of 4" cement blocks wasn't worth the transport distance so I'm going to go with standard 8x16x16 blocks.  I think that means I won't need to fill too many of the cores with mortar.  I'll still put rebar in the core that is lined up with each curved truss.

I still haven't contacted my insulation guy to see if he has recommendations for insulating the North, East and West walls.  I'm leaning towards a ton of styrofoam but maybe he'll have another option.

I found the cheapest source of metal roofing to be Home Depot.  The wavy roofing may bend to my radius but it's unlikely.  I think I'll segment the North roof into three facets and cut the roofing into 5' lengths.  The radius of the bend works out pretty good with a 2x4 purlin at the top and bottom of each 5' chunk of roofing and the middle of the roofing resting on a purlin sunk into the styrofoam.  Plus I can make an access panel to get at the compost without building out a dormer.  Just pop off a couple pieces of the roofing and I'm in.

I'll likely build a sample truss in the next month while I wait for our 18" of snow to melt.  Then the digging begins

Before spending the money to purchase LVL, cut them in strips, and glue them back together (at least that is what I think you are planning to do here), why not build them up from plywood laminated on our own jig using exterior grade adhesives and make your own LVLs?  Then you can make something similar to these Gothic arches http://www.starwoodrafters.com/gothic-rafters.html

If you use pieces cut out of plywood, then you can also, with a bit more work, laminate in the cross bracing, or make something more akin to a curved version of an I beam truss https://www.apawood.org/i-joist

It is easy to bend your own laminates https://woodgears.ca/bend-wood/index.html

Hope this was interesting and useful

Thanks Ebo, I considered building the shape up from either plywood or wood strips.  The issue with plywood is that many of the plies are running in the wrong direction.  For a 7 ply piece, 4 would be in the right direction for strength and the other three are just providing bulk (in my application).  Plus there are often voids that take the 4 useful plies down to 3.  

With plywood strips or just plain old strips of wood, I was a bit worried about the joints.  I can't get either type of wood in lengths over 16' so it would necessitate joints.  I'd stagger them but then at those points I'd still only have half the strength of the intended design (assuming two strips of wood).

The LVLs are really well put together with the plies running in the direction I want.  I may be wrong here but I think they'll be very structurally strong when I'm done building them.  I think I may switch from 2x4 spacers to 2x6 spacers to further increase the strength.  

I'm also going to orient the strips so that the plies are parallel to the poly.  That way the screws to attach them to the 2x6 spacers won't have a tendency to split the plies apart.

My approach will be to get a narrow kerf circular saw blade and clamp a guide onto my saw to give a fairly precise 1.5" cut.  I'll slice off a 1.5" strip of LVL (1.75" thick).  Hopefully it will bend enough for the truss and I'll be done.  If it doesn't bend easily enough, I'll rotate it 90 degrees and cut it in half, basically removing one of the plies.  The the remaining 1.5" by .75" pieces will certainly be able to bend enough.  When I assemble the truss I'd just have to glue those two pieces together (like with your last link) to form the curve.

FYI getting 22' LVL's home with a Tacoma is fun.  Truck Bed Extenders are well worth the money.

I would have to look at the LVL's that were put in my house, but as I recall every other layer is oriented 90 deg which increases the strength overall.  Also, I have seen joints in LVLs laminates.  I think you could build them up the same.  Regardless, if you find something that works for you then you have rung the bell ;-)

Best of Success!

The individual plies have joints but since they're spread throughout the LVL the strength is still very good.  The ones I bought have the grain all running lengthwise.  I think all LVLs are that way but I could be mistaken.  

I'll start cutting on them and we'll see how it goes

Best of luck!

I have a quonset hut style passive solar greenhouse that I like because the wind slides over it.  I have a mild winter, so I am no expert on your temps, other than to say maintaining light levels is just as important as maintaining temps.

Sorry if I missed full discussions because I couldn't read every post, just in case this helps:

I had to back off the north wall, slightly cloudy ceiling panels away from the center ceiling (12:00 o'clock) to about 1:30 o'clock, if the ground is 3:00 o'clock, and put clear panels.   The high sun in the summer was cut off by the ceiling panels and tomato plants were really small and late.

Since heat rises, I keep the ceiling height at 7 feet, just high enough for me to walk in with a hat on, so the heat will be down around the plants.  

My east and west walls only have 3-foot high cement blocks filled with 1.25" rock for insulation, painted flat black, and clear plastic to the ceiling, because early morning light is crucial for plant stimulation and growth, and it's surprising how much heat comes in through that east wall before noon.  The west wall captures even more heat as the sun slowly heads towards dusk, and contributes a lot of heat going into nighttime.  Not sure if your "east and west walls fully insulated" meant they would be blocking light.

I have rodent-proof screen on all windows and doors.

I do hugelkulture trenches underground, and plant in those.

With an 8-foot water table, not many trees will be happy with that.  Roots of fruit trees will often go twice a mature height, about 16 feet, and will be stressed in that much water.

My doors and path are wide enough for my wheelbarrow to have a straight shot in, no curves to have to maneuver around tomato cages and heavy greenery with long-handled tools sticking out over the front and back of the wheelbarrow.

Not sure about the pine tree idea.  I have had to remove pine trees for 50 feet around my greenhouses for light levels and root intrusion. Pine roots are shallow enough that in a few years they can grow thick and lift walls of greenhouses (and houses and sheds).   They have really annoying pollen for several weeks in the spring that is not good to breathe.  It will stick to the inside plastic or panels because of condensation.  I have to clean solar panels daily when the pollen is around.  Their sap makes a mess on gloves, tools, anything that is underneath them, and roots have sap, too, inhibiting root growth of annual vegetables.  They are drought tolerant trees and a lot of water kills them.

There are fruit trees that actually need the chill hours that Wisconsin has to offer, and if they are on dwarf rootstock they could be a good fit in a greenhouse, as long as they have another pollinating tree nearby.

The only thing that isn't working as well as I'd hoped is the cement block wall on the south side.  The slanted wall of the greenhouse and winter winds puts enough pressure on those blocks to make them lean, no matter how much I prop them back up with boards and metal fence posts.  I've had to lower them to one course in the ground and two courses above, painted flat black, then filled with 1.25" rock.  Making sure the panels or plastic on that south wall is solid to the ground without rubbing against the rough edges of the cement block is another sticking point because of the wind.

Because of short days from November to February, regardless of temps, not many plants will thrive in that, so don't be too hard on yourself in those months.

Thanks Cristo!  The thread has gotten a bit long but that's probably a good thing.  Here's a picture of the latest geometry.  I shifted the ridge to the North a bit so that I get sun on nearly the whole floor on the summer solstice.  The solstice and equinox sun angles are on the right side of the picture (showing the shadow from the ridge).

Heat will rise on me but I want/need the headroom for bigger trees and to allow snow to shed.  I'm planning on some ceiling fans to mix the air and get the leaves moving a bit.

My East and West walls will be insulated and solid.  Light through them in the summer would be nice but in the winter we get very little strong light until the sun is SSE.  So while it would let a bit of light through, the heat loss for the remaining 22 hours would be too costly.  That's mainly due to our latitude here, design considerations would be much different farther south.

I'm not sure my water table is 8' deep.  If I said that I was possibly mistaken.  I dug a 7.5' hole and it wasn't wet.  There is a wetlands downhill from this site and I'm guessing it is 10+ feet of elevation to the water.  The trees around here don't mind but they are cold/temperate trees (pine/spruce/oak/maple/birch/hemlock/larch).  My apple trees don't seem to complain (but I'm still learning their language).

I'm not sure what you mean by the "pine tree idea".  The nearest pine trees are about 200' away.  I will be cutting two of them down to give me 30 extra minutes of afternoon sunlight.

With your leaning cement wall issue, are your blocks mortared together or dry stacked?  Are they on a footing?  My plan is now to use 8" cement blocks 3 courses high on a footer with mortar.  Where the trusses land I'll have rebar in the void of the block and fill that void with cement or mortar.  

I'll try to not be too hard on myself if it doesn't work  I figure if I can't grow tropicals, I can back off to citrus, or to Mediterranean crops, or to USDA zone 7 trees, or to kale...

I have been researching for about 3 years on this.

Here is what looks to me like the best idea for earth banking heat.

self heating green house

Now a possible way to pull the heat down into the ground would be a solar collector used to generate air flow instead of heat.  The window screen box collectors seem to be winning the efficiency here.  The blower fan can still go on the inlet end but the system would then do something even without electricity.  Especially useful for getting a bit of annualized geothermal in the summer season.

Now some other things to add to your thinking.

Why are the trees at floor level?  If you go deeper with them then you have more height to work with inside the building.  So if your floor isn't level and you build some raised walk ways the base of the tree can be deeper still letting you have more height to work with.  You still need proper air flow to the roots  so you can't completely cover it but would pallets holding compost heating 6 inches up off the lower ground let you mostly surround the tree with compost heating?

The walls outside.  Many of the do it cheap green houses uses straw or hay bales around the perimeter outside as ground insulation.  It will add to your snow removal problems and but might be worth considering.

As for being able to do tropical plants are you also looking at doing greenhouses within greenhouses?  One of the commonly given rules of thumb is that you get roughly 1 zone for each greenhouse layer.  You might be able to surround your banana with water barrels on the back side and put it in its own green house.  For my thinking since the goal is earth banking I need to pull that air down somewhere so If I am pulling the hottest air in the greenhouse down inside a separate greenhouse I may be able to create a hotter climate still without losing much during the day and by getting the water storage warmer carry some of that heat through the night as well.  If you are constantly during the day flowing the hottest air in the building down past the barrels stratification wouldn't matter as much.  For night time use a simple waste gate that allows the airflow to bypass the inner green house would let you keep that heat in the inner green house.

What are you doing for summer ventilation if what you have shown isn't enough?  The worst case rule of thumb for natural ventilation says it needs to be 1/3 of the square footage of the green house.  It seems to vary between that and saying you need 20%(10 in 10 out).

Have you looked at maybe building a ridged foam board multifold door to cover the glazing inside?  

How are you using your other paths in the building.  Some of the stuff suggest doing vermiculture under them.

Hi C, welcome to the discussion!  I had watched that Deep Winter Greenhouse presentation before, the gentleman has some great info.  I briefly clicked through it again just now to refresh my memory.  I believe he was suggesting either a gravel bed or CERES style soil heat storage system.  I don't really want to do the kind of excavation required for the soil system and I'm also not horribly enthused about importing the amount of gravel needed for a rock based system.  Or how the tree roots would interact with the gravel.  But I am still interested in the concept because storing heat on sunny winter days is important.

I was thinking that my phase change battery system with active harvesting of heat off the ceiling would help a lot.  But I do wonder about just trenching 1' deep and running perf pipes across the greenhouse to store ceiling heat as a "poor man's" earth battery...

I had the trees at ground level because that's where trees normally grow  I could put them "underground" a bit but I only have 18" of sandy loam to work with before it transitions to more pure sand.  I think I have enough head room in the structure to not have to worry about height.  Alternately, burying the trees could allow me to shrink the building height but that starts to affect the angles of the glazing and snow shedding.

Yes, surrounding the greenhouse with straw/hay would add insulation value.  I'm not sure it would do much (diminishing returns) if the walls are already R20.  Most of the heat loss is through the glazing.  If I had a good source of straw/hay I'd still do it but there aren't any organic grain farmers in my area and I'd want organic straw/hay so that I could use it for mulch later.  I do have fallen deciduous leaves and the nice people in town collect them for me in bags at the side of the street.  If I could pile them up in a sexy and productive way they could add some free R value...

I kind of had a greenhouse in a greenhouse idea, but it wasn't very elegant.  It was to just allow for dividing the greenhouse into a hot half (east) and a cooler half (west).  That would just concentrate the compost heat on the more tropical plants and let the more cold tolerant ones (citrus, veggies) get closer to freezing (or get their cold treatment).  But having a more enclosed second greenhouse is interesting.  Even if there was just a fan to draw hot air into late in the afternoon so it carries its heat through the night a bit better.

I was toying around with barrels of water or phase change material around the plants that need heat more throughout the night.  Can you elaborate on your suggestion of a waste gate for night time use?

My summer ventilation plan is about 15% in and 15% out for a total of 30% of the square footage.  I'm also thinking I'll grow vining plants up against the glazing in summer to help shade the rest of the greenhouse (per Jerome Osentoski).  If that isn't enough I'm not sure what I'll do.  Maybe shade cloth...

I haven't considered a rigid folding insulation for the glazing.  My current schemes involve a rolling insulation (like Reflectix) that rolls down at night on guy wires or tracks connected to the curved trusses.  I've typically thought that folding insulation would take up too much room when collapsed and be hard to mechanize/unfurl.

I was planning on worm farms under the paths with pallets over them to walk on (per Jerome).  I'm also thinking I'll adjust the terrain in the greenhouse so the paths create a downhill channel that leads cold air back to the compost heater.

Thanks for the ideas, keep 'em coming!

Mike Jay wrote:With your leaning cement wall issue, are your blocks mortared together or dry stacked?  Are they on a footing?  My plan is now to use 8" cement blocks 3 courses high on a footer with mortar.  Where the trusses land I'll have rebar in the void of the block and fill that void with cement or mortar...

My cement block wall has one course underground, and 3 courses above with thick rebar pounded down in each opening of the cement blocks.  It's dry stacked because the blocks are filled with 1 1/2" rock that heats up and keeps the blocks warm.  They are also painted flat black, not sure if I mentioned that.

I think it's the wind that puts the most stress on the blocks because of the curved wall shape.  You can see in your drawing the disparity in length between the height of the blocks and the height of the wall.  Windy conditions and storm gusts on that whole wall surface will put huge stresses on the wall and potentially push the blocks inward, especially when the wall base inside the blocks.   I've moved the blocks to the outside of the wall, covered the tops of the blocks and outer rough corner with old carpet, and pulled the greenhouse plastic down the wall, then out over the blocks.  So the wall uprights are inside the greenhouse, not outside the blocks, which is how mine were in the beginning.

I could have moved the blocks inward, but I lost too much space inside that way.

If your end walls have no light coming through them, you might want to think about grow lights that come on early in the morning.  If you're going to have power for fans, you'll have power for grow lights.

Since your experiment is so location specific, you might want to try something half the size you plan, being prepared to discover over the next couple of years what tweaks need to happen.  Then add on the second half of it with all of that experience in mind.

Thanks Cristo, my blocks will be mortared together so I think they'll have more strength.  Plus they'll be sitting on a 12" wide footing that is just below grade.  Each curved truss will be 4' apart and where they land on the block wall I will have rebar in the hole (and footing) and have it filled with cement/mortar.  The trusses will be attached to the block with a "post base" to keep the truss just off of the cement to avoid wicking and to securely hold down the trusses.  I think that will be good enough but if folks think I'm under-designing it, please let me know.  Or if you think I'm on track, please let me know.

I like the idea of building a half size test greenhouse and just adding on later.  The challenge is that the length to width ratio would become square so the solid East and West walls would block light for more of the plants more of the time.  So I think I'm kind of set on the size but that is a very valid suggestion.

I agree on the grow lights.  I'll probably have some anyway for seed starting or supplemental lighting for key plants.  My main plan is to not have too many plants in the NE and NW corners of the greenhouse (utilities and work space) so that I won't miss the sun in those spots as much early or late in the day.

Here's an update on truss building.

I cut a 1.5" strip off of the shorter LVL to test the bendability of the material.  FYI LVL's are 1.75" thick.  That piece of wood (1.5"x1.75") was very stiff and wouldn't bend much at all.  I should have realized that since a 2x2 doesn't bend much either.  

So I took that piece, rotated it 90 degrees and ripped it in half on the table saw.  Due to the plies on the LVL, I basically removed one ply as I ripped it on the saw.

The new pieces (1.5x.75) were flexible enough to bend to the form!  Plus the plies of the LVL material are parallel to the curve of the truss.  Now when I drive screws through them the screws will penetrate the plies.  In the other orientation the screws would tend to split apart the plies.

I laid out the truss curves on several chunks of think plywood in the driveway.  Then I built a form on the inside wall of my barn.  That way I can glue/screw/clamp together the trusses while comfortably standing up and working out of the weather.

I'm tempted to start a new thread on the construction of the greenhouse since this thread has mainly focused on the design.  Good idea or bad?

Sorry for disappearing for awhile.  Lots going on...

While doing some more background research I came across U. Minnesota's Extension program design of their Deep Winter Greenhouses (DWG-v2) <https://www.extension.umn.edu/rsdp/statewide/deep-winter-greenhouse/>; .  They include plans, and other useful information.  While I have not had time to read through all their documentation, it is really informative stuff.   I also have a little more to report on the modeling front, but will wait to give a report when I sort out what I got working and what I did not...