I been swamped at a new test lab assignment. I’m doing
product level structural testing related to this discussion for a new product I
cannot disclose. The way design development should go (like in the case of foam
products) is there is a FEA (Analysis) /Finite Element Model (FEM) that is
initially produced from Design CAD Models, these days paperless 3D called Model
Based Definition/BIMS. For deflection on a plate the article discussed, we set
up a compression plate load on the Unit Under Test (UUT) and for deflection we
use LVDTs (Linear Variable Differential Transformers). If interested google,
simply put they are little actuators with a piston rod in a magnetic field
cylinder that measures liner displacements from E-6 – 30”. So if we set up
LDTVs at the ends of a 4 x 8 sheet we could produce mfg data and conform,
certify to a test standard like ASTM. When the test is over, we take those
results back to the FEM and hope it is within +- 10% of safety margins, or, we
still have positive vs negative margins…that data is populated or back
calibrates the FEM as empirical. That is no different than BEOPT, WUFI, code,
design guides, or the better models that have the funding since what I do is
not cheap. I’m also doing temperature, humidity, and moisture (soak) testing
with the LDVTs & “strain gages” attached another linear displacement type instrument.
I’m also doing plumbing operation, proof, burst, and rupture
testing at various PSIG.
If the test did not fail, from there we release it to the
client that witnessed all the test and it goes out for field testing where more
issues may cause a redesign or revisions to code, standards, etc, or better
technology in part or whole causes us to go back to the design models/lab.
Three main modulus are validated, young’s, shear, bulk,
along with other mechanical properties. Creep would take a different test in a
hot box like my temperature and humidity chambers with a “load cell” that puts
compression cycles on the foam that is accelerated over 30 years, or some long life
in a matter of weeks.
That’s what I was referring to above that is missing for
under slab foam. This article lost my attention in the first paragraph when it
stated “EPS has been successfully installed for more than 40 years” when it did
not point to a Final Test Report (FTR) of such findings. I guess readers are
supposed to believe that since some magazine article said so. Reminds me of
post, threads.
The other path the article is based in is called “conformity by
analytical methods” where math or other science is used, more general, not
specific in the article to a manufactured part. It can be specific, for
example, later I’ll be doing fungi testing by analytical methods on the UUT I’m
testing now, like WUFI-BIO modeling, since that is acceptable to this client.
Often, lab testing is not invoiced until it is too late and there are issues,
since it is expensive and lots of analysis.
In this case, it stated we can get a “Modulus of Subgrade
Reaction” (MSR) the test I am doing would provide to a client. I’ve never seen
that in the foam specs, nor creep, as I said unknowns. The article appears to
be static dry analysis since there is no tie to sub soil atteburg limits.
In code we can find soil bearing strengths for different
soils. This is another method vs MSR for “rigid foundations”. In rigid
foundations more foam thickness equals more stiffness, more distributed
pressure if the plates are assumed to be planer. It has to since concentrated or point loads
crack it. MSR is a flex method or well know flexural modulus with a nick name.
More foam thickness as Jon accurately stated causes Ks (MSR) to go down, less
pressure distribution, more concentrated load in flex slabs “plate” is ideal to
keep it from max deflection limit load.
So, the two methods tested rigid foundation using code
bearing & analytical MSR using flex modulus can be combined. At the end of
the day, pressure/settlement is the determine factor.
Crude assumptions are made and the article is anything but “conservative”.
MSR works well for soils, not rigid foam and rigid foam is not a simply
supported beam like a slab that spans between solid footings. Foam can also be
in bending under high subgrade concentrated pressure loads, freeze-thaw,
requiring a geotech report.
Further, most designers and especially DIYs like this thread
will guess wrong, rarely look for mfg MSR soil bearing or PI in residential design until the day code
requires it or a PE. In my jurisdiction PI test is mandatory. The problem is
finding a lab since they only want to deal with pro’s in commercial since they often get blamed when ppl don’t understand test reports. Sound familiar?
Concrete foundations have experienced costly mistakes by
pro’s, how DIY foam is supposed to solve those issues, and articles like this making
it appear as an “over- engineered” industry wide design flaw, is a magazine selling
bogus articles, just like most GBA blogs and Q&A.
So to answer Jon question, “how to achieve compete
non-compressible contact between the foam and underlying base”
By suspending the slab so the compression load is within
allowable design load limits tested by the manufacture as I denoted above. That
takes at least two Engineers, a geotech report to a PE to design to low cost.
That may require stabilizing soil so foam can work.
A “conservative” approach would be to never use foam. Only
use products with high compression limit load like Roxul IS board @ 700 psi @
10% deflection or Foamglass 1200 psi IIRC, or any way you can get the highest
compression for the buck. Less than 100 PSI is not enough margin.
There are some insulative cretes using lime, magnesium,
pozzolans, aggregates, AAC, that are better than foam hygrothermalyl & 100+ PSI compression/better
than concrete deflection properties w/large max moisture contents & liquid
transfer coefficients. The materials cost little but the knowledge and skill
set labor can be hard to find and costly.Not easy to get a ready mix company to do a custom design, many have their own labs BTW but they want bulk sales, not one-offs.
Anyway, I’ll take another look at Chiltrix COP vs CCHPMS and
that BEOPT model asap.