OK, starting to get back to this design. Have been kicking around a variation of an Alaskan slab/slab on grade/frost protected shallow foundation/whatever you want to call it... Picture attached. In the pic, the black cross hatched area represents a single ICF block such as
Introduction to Insulated Concrete Forms | Reward Walls. ICF blocks add $1000 to the project, but also supply about $400 worth of foam insulation, so their convenience for forming and rerod placement really only costs $600. I can see lots of forming if not using the ICF, and almost none if using them, so it seems like a great addition for labor reduction. Also, some concrete saving will be possible with the ICF blocks, since the quantity will be so closely predictable using them, as opposed to normal forming technique (mine anyway).
Rerod not shown for clarity, but you can maybe imagine horizontal and vertical sticks in the "wall" part of the drawing. Outside, inside and under the black ICF area, the deep blue color is to represent extruded polystyrene (XPS). Total depth into the ground of the outer XPS is about 24", so this is not a frost wall, although it is somewhat protected. Plenty of insulation on the outside of the foundation and floor, about 4.5 inches thick total, for R value of about 18. R-10 under the slab, which is enough since this is a shop building afterall, and so will not be heated to 72 degrees air temp.
The idea would be to pour the entire ICF building perimeter square with rerod, nearly full of concrete, then backfill the inside and outside of the ICF "form" with aggregate, then pack and level, then the XPS under the floor area on the aggregate, next lay all the floor 1/2 inch rerod on 2 foot grid, wire tie intersections, all set up 1 inch above the surface of the XPS. Then the RFH tubing zip tied to the rebar grid. Then a second pour to the top of the finished floor height. Then metal brackets such as
http://www.permacolumn.com/drill_set_models.aspget
which are bolted (hammer drill the holes after the pour) to the surface of the floor to bolt the laminated 3 times 2x6 pole assemblies into. Then the bookshelf wall gets assembled. Pink line is Tyvek outside structure and inside steel siding. Gradeplank shown is not set into poles, but added conventionally.
A big advantage of this is the way the floor and wall float (I bet very little) together. So doors always fit properly all winter long. Floor pour should be way easier than pouring after the building walls are up. The perimeter wall is quite tall and with rerod dependably placed in the trays in the ICF, the wall should be really stiff. Also, I think I have designed (with your help) a pretty thermally well isolated mass for the RFH method. One minor thermal nosebleed area is the overhead door thresholds, which will radiate heat outside. A possible band-aid would be to keep the tubing further from those areas, to use the concrete slab's own low R-value (resistance to heat flow) to minimize the total BTUs lost this direction. In fact, I'd keep all the tubing a couple feet from the outside edge of the floor, since I am not worried about an uncomfortable cold spot. Most of the wall space gets cabinets or tools or benches anyway. Maybe even three feet inside the outer edge would work for the first run of tubing.
Your feedback welcome. Again, this is a heated shop in MN, frost penetrates 48 inches.