Thermal boundaries

Poly does nothing to reduce heat loss. It is merely a vapor barrier. But, if the crawlspace floor is below the frost line, then the heat loss into the earth will be minimal, since heat transfer is directly proportional to the delta-T, or temperature difference.

The deep ground temperature in any climate zone is the same as the average annual air temperature.

A slab on grade, however, loses most heat at the edges, to the air rather than to the ground. For that reason, it's imperative to insulate the edges and - depending on average heating season temperature - at least the outer perimeter below the slab.

Thanks Robert. I imagine the insulation is much more of an issue if the slab is heated, no?

Yes, if you raise the temperature of the slab you're increasing the Delta-T and increasing the heat loss proportionally. Thus here in the cold Northeast, radiant slabs often have as much as 4 to 6 inches of foam underneath with equal amounts at the edges.

Another approach, which I have used, is to build on a shallow, frost-protected foundation which not only minimizes the amount of excavation and concrete required but relies on warming the earth beneath the home to maintain a frost-free environment and low Delta-T. This requires a reduction in the sub-slab R-value (and less petrochemical foam) to allow some heat loss downward. Once the ground temperature is raised, heat loss diminishes since the large thermal mass of the sub-slab earth stores the heat.

The folks down at Advanced Energy have just released a report that partially investigates this question (www.crawlspaces.org). They did testing on closed crawlspaces with insulation in the walls or the floors above the crawl. In a cold climate (Flagstaff), the wall-insulated crawl used over 50% more energy than the control house that had insulated floors over a vented crawl. The authors suggest the penalty was due to the uninsulated floor chilling the crawl & floors above. I find the size of the penalty hard to believe considering the other care they went to in air-sealing the crawl, but it does suggest some strong caution be used for wall-insulated crawls, if only to boost the wall and slab insulation beyond the levels they used.

A careful reading of the Advanced Energy study reveals that, while the vented-crawl control houses and the floor-insulated closed-crawl houses used R-30 fiberglass floor insulation, the closed-crawl wall-insulated houses used only R-13 foam board and left a 3" gap at both the bottom and top of each wall (for termite protection). It's no wonder, then, that the R-30 floors performed better than the incompletely-insulated R-13 walls.

What's perhaps more interesting is that the Flagstaff homes had to have their crawl-spaces re-vented when high radon levels were detected. This, perhaps, is the greatest danger of unvented crawl spaces that are communicative with the living space, and one rarely discussed by closed-crawl advocates.

Good point regarding radon. The best way I know to evacuate radon gas is to run a pipe down through the vapor barrier on the floor, into a layer of crushed stone, and up through the roof. If everything is properly sealed there should be no communication between the layer of stone and the living space.

IMO heat loss through the ground is widely misunderstood, neglected and underestimated in our current insulation requirements and building practices. Good reference to this topic is the ISO 13370 standard http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?c... which was established in 1998 and revised and updated in 2007-or the Passiv Haus publication #27 which lays out the scientific bases of heat loss to the ground and thermal bridge free foundation construction and means to properly calculate this particular heat loss. Delta-T isn't the only equation in this...and bleeding heat into the ground under the house results in heat loss and not a heat store. The listed publications show very good isothermal graphs which shows this visually rather well -one could go through the exercise to calculate this in Therm to come to the same conclusion. PHPP07 is a good means to calculate this particular heat loss and establish a good insulation ratio for the overall building performance. I have seen very high real world gains in energy efficiency in our homes by increasing the under slab insulation in accordance with the calculated heat loss thought the ground - and adjusting the Insulation ratios between the overall building envelope: wall insulation, roof insulation and under slab or floor insulation to a balanced level. Our standard slabs are now R-50 - more if the client can afford to put more money into insulation.
Good information in regards to shallow frost protected foundations can be found here: http://www.cchrc.org/frost_protected+shallow+foundation+study+.aspx. Radon problems have hardly anything to do with what type of foundation system was used - but with the fact that a installation of a proper radon system was omitted. Installing a radon system is cheap to do if done in the right stage. A starting point: http://epa.gov/radon/pdfs/buildradonout.pdf.

Thorsten,

I'm afraid that you're adding to the misunderstanding. Heat loss through any building element is directly proportional to delta-T between inside and outside surfaces. And, particularly with a shallow, frost-protected foundation (SFPF), the ground isotherms under and around a foundation change due to heat loss downward - reducing the delta-T and hence reducing the heat loss.

The actual ground isotherms, at any particular site, will depend on annual average ground temperature, conductivity of the soil, dampness of the soil, presence of moving water (water table), and snow cover.

But with proper building practices (i.e. good grading, a well-drained site, course granular backfill, good footing drains, good overhangs with gutters and diverted downspouts, etc.), the ground under and around a foundation should be non-conductive enough to develop a radial isotherm that not only protects the foundation from frost heaving but reduces downward and perimeter heat loss.

The entire purpose of a SFPF is to raise the ground temperature sufficiently to both move the freezing isotherm away from the footings and change its angle to prevent vertical heaving (frost heaving occurs perpendicular to the freezing isotherm).

With either a deep foundation or a shallow frost-protected foundation, the average heating season delta-T downward should be no more than half the average delta-T to the outside air. That means that, for the same efficiency, only half the R-value of the walls need be placed under a slab.

And the radon problems encounted in the Flagstaff study homes would also be exacerabated in any home with a conditioned basement or crawl space whenever negative pressures from stack effect or mechanical ventilation exceeded the negative pressure from a properly-installed radon vent.

Certainly not intending to add to any misunderstanding. Based on my research and experience I just simply disagree with some of the statements you made. ISO 13370 is a internationally recognized standard - so is the Passiv Haus standard and both publication I referenced provide methods of calculation of heat transfer coefficients and heat flow rates for building elements in thermal contact with the ground. Both were established over many years and tremendous amounts of research money and funding. We monitored several buildings with data loggers over several seasons and I have seen very good gains in energy efficiency on our buildings by following their guidelines. A Passiv Home build in Germany has on average 12 inches of foam insulation under the foundation for a good reason - or do you think they do so because they simply don't know what they are doing?
Have you read either publications, modeled this in Therm, or done a research project for a minimum of 12 month with data loggers placed into the ground to measure and log the actual isotherms?
A properly installed radon vent includes a inline vent installed in the attic space.

Robert makes a good observation on the Flagstaff study about the R-13 on the walls, with gaps top and bottom. I noticed that when I read the study. Still, I was surprised by the magnitude of the energy penalty. I didn't see the level of wall insulation in the rest of the house - did the investigators just use R13 on the crawl walls because that's what they used in the house walls? Of course the gaps would reduce the effective R-value below R13.
I have not done any modeling of the heat loss to the ground using the tools Thorsten mentions, but I will probably do so when I've got some spare time. I suspect that in the Passiv House standards, the rest of the house is so well insulated that heat loss to the ground becomes very significant. Clearly, ground temperatures are lower than room temperatures, and some level of floor insulation makes sense. But considering that the delta-T from the house to the slab is 1/2 to 1/3 of the delta_T between the house and exterior, I would think that the appropriate insulation level would be reduced similarly. For most houses, this would be substantially less than the 12" that Thorsten suggests, but certainly far more than the zero in the Flagstaff house.
This subject is interesting to me, because subslab insulation is unheard-of in my area, but we live in a reasonably cold climate (coastal NJ). Considering the Flagstaff results, some level of subslab insulation is probably warranted, or at least running the crawl wall insulation continuously down to the footings.
Alternately, considering that the air in a crawl is very quiet and stratifies easily, I wonder whether a low-emissivity surface (radiant barrier)installed on the floor joists would be effective at controlling the heat flow downward at lower cost than subslab insulation?

You have 2 out of 3. Insulate the floor above the crawl space with either Nu-wool Premium Engineered Cellulose by dense pack method or the Nu-wool CMS Panels.

Peter,
Concerning your post July 15.... quoting here:
"But considering that the delta-T from the house to the slab is 1/2 to 1/3 of the delta_T between the house and exterior, I would think that the appropriate insulation level would be reduced similarly. For most houses, this would be substantially less than the 12" that Thorsten suggests, but certainly far more than the zero in the Flagstaff house."

Is it really fair to compare the DeltaT of House to Ground with the DeltaT of House to Air?
Is not the Ground much more conductive?

I think Thorsten's observation about the German Way makes sense:
"do you think they do so because they simply don't know what they are doing?"
I think that there is a good chance that they do know best.

I do not understand our North American concepts about R-value
Is the Answer 60/40/20/10 ?
Is that really based on Physics?
Modern European Standards seem to be almost equal amounts Ceiling /Wall not 60/40
They also seem to have much more respect for THE GROUND.

I think that the North American Mentality is more concerned with the initial cost of changing away from the way we have always done things.
The Euros seem to analyze and test the physics and then apply.

The Euros DO seem to favor Riversong's (more vapor open) approach over Thorsten's (less vapor open) Construction.
The Euros also seem to Avoid Foam Products and use natural materials.....Same as Robert Riversong.
hmmmm

Martin,
I did not understand your comment from July 15
I understand that dust would collect above a horizontal radiant barrier...
Are you saying that dust will also collect below a horizontal barrier?

Martin,
I think that Peter's suggestion was with the radiant surface facing down....
That is why your answer confused me.
It seems to me that Peter's concept would reduce heat transfer ...
There would still be a big difference between temperature at the bottom of the joists and the slab .. so convection would still be an issue?
So ... maybe best solution would still be ample insulation