Question - Existing slab on grade analysis

[PE_JRM]

I am renovating an old cabin (DIY for myself) and having some trouble determining if the existing slab is adequate to support what I want to do. The existing slab was not well implimented. The existing slab is 4" thick, old enough (1965) to be only 3000 psi, probably unreinforced, no thickened perimiter, sort of level and flat and no sub-base. 34x20 feet dimensions. Poured directly on stiff clay with about 1500 to 2000 psf bearing). The only thing the slab has going for it is that it is uncracked after 60 years although it is lightly loaded.

For various reasons I want raise the floor a few feet by building a wood framed structure above the slab, essentially making the slab the floor of a crawl space.

Framing this conventionally with long beams and joists with edge supports exceeds the strength of the slab. This could be dealt with by placing a square footing under each beam support but this is expensive and a lot of work if I do it myself. This approach is probably the simplest and best solution but I'm looking for a way to transfer the load to the slab in a distributed fashion.

I've been using excel spreadsheet GRDSLAB and BOEF as well as LISA FEM (free limited node license) to compare various approaches. Due to the limited nodes and uncertainty in the soil parameters I don't really trust the FEM analysis for much more that a gross understanding of where the load is going. I've been using it as a gauge of the practicality of various ideas.

The idea I'm currently exploring is a grid...beams supported every few feet (maybe 4'?) and perpendicular joists 16" OC. This would give me a 4'x4' group of point loads on the interior region of the slab. The slab edge load can be controlled by the edge spacing around the grid. I've been referring to 1) Slab thickness design for industrial floors on grade and 2) ACI 360R-06. The issue I'm having with analysis is that I'm off the nomographs for these references, probably because industrial slabs are typically thicker than 4". I've also referred to the OLD Army manual TM 5-809-1/AFM 88-3, Chap. 15.

I'm a retired structural engineer (PE), although concrete was never my specialty. Any ideas you might have that I could use to show my existing slab good with a bunch of point loads on a grid would be appreciated. I'd like to keep the solution simple and something I can do with a calculator, pad and pencil since my buget for the project is small. This solution is kind of fiddly but its worth considering.

Another idea I've considered is to pour some load distributing reinforced pads (2'x2') on top of the existing slab and just let the existing slab crack (could saw cut around pads) if it wants to. These pads would be for the beam supports. I don't really like this solution but it has its merits as being the easiest to DIY impliment and probably the cheapest.

 

[XR250]

The slab is only going to be as good as the subgrade. If it is stiff clay, you should be fine.

 

[PE_JRM]

XR250: Yes. I agree, BUT, I can't explain that to the oounty when/if they come to inspect. None of the details of the slab meet code.

 

[jhnblgr]

You’d be surprised how many buildings are still around that were built his way. Does the slab exhibit structural cracks? I have see 3” of washout concrete on clay in NY stand up to years with excavators and tractors going up it. If it does settle you can always jack it up and do a footing on top of the slab or run a steel plate to distribute the load. I would frame up a wall with 2x12 to get the most load distribution.

 

[PE_JRM]

jhnblgr: One third of the slab was a porch and most of that is spalled and powdery. The other 2/3 of the slab is uncracked. However, I'm completely changing the floor plan and the roof. The roof on the building currently is corrugated steel sheets supported only by Zee shaped purlins (no deck) 2 ft OC that run the long direction spanning 15 feet - no insulation. My estimate of the roof weight is less than 2000 lbs. I'm putting a code compliant roof on the place which will increase the loads dramatically due to the minimum design load of 20 psf. The walls are uninsulated 8' stud wall with aluminum siding with light masonite paneling... no exterior sheathing, no drywall. I intend to remove the paneling, insulate and drywall the interior at a minimum. I may also remove the aluminum siding, sheath the exterior, and put up hardiplank siding. This will dramatically increase the loads on the edge of the slab no matter how I manage the 50 psf floor loading required by code.

So, the fact that the slab has served 60 years without cracking is a testament to its extremely light loading more than its toughness (my view).

 

[PE_JRM]

Since no one commented on a good free tool for analizing my slab I've downloaded and tried about a half dozen different things. Most of them are too restricted to be helpful. HOWEVER, spMats from the PCA is both easy to learn quickly and the 15 day trial is enough time to get a good feel for what I'm dealing with. I've spent an afternoon trying to figure out how to specify a slab property that has no reinforcement and still haven't figured it out. I'm assuming the only thing you can do is just choose any of the reinforcement options (ignore them) and then do the calculations by hand using the results. Is there a better way to do this?

 

[XR250]

The software is only going to be as good as the inputs. i imagine the most important ones are subgrade modulus and concrete strength. You have guessed at both of these. Garbage in = Garbage out.

 

[jhnblgr]

The whole slab analysis is a waste of time. The slab is too thin and under reinforced for it to be applicable for what you are doing. You would need to check the shear capacity of the concrete and the appropriate width it could be assumed the loads would be distributed by the footing.

If you want to bring it up to any code standard you are going to want to likely just bit the bullet and install concrete footings

Checkout wood foundations which are still in the code.

 

[PE_JRM]

XR250: Thanks. Not really helpful. garbage in = garbage out is a good admonition for an ignorant audience. Seat of the pants engineering and enveloping can accomplish what I want to determine. What I'm trying to determine is if the slab will crack or not based on some assumed loading conditions (building details) so I can choose the best approach to distributing the loads. If the slab cracks its not the end of the world. Nothing terrible is going to happen - maybe some settlement. I own the place and I'm willing to take the risk.

My screen name says I'm civil/environmental which is incorrect. I've asked the moderator/admin to change it to structural and it never got done. So, in spite of how it may seem I'm fully capable of having an intellegent discussion with collegues even though I changed from civil/structural to aerospace/structural 30 years ago.

jhnblgr: Thanks for your comment. I agree that the slab is marginal. I don't agree that the analysis effort is a waste of time. What I'm trying to do is determine the best way to distribute the loads to either eliminate (maybe not possible) or minimize potential cracking. spMats actually is a great tool for making quick comparisons. Unfortunately, spMats insists on specifying a reinforcement detail in order to run without errors. The spMats manual shows an unreinforced example but says nothing about specifying dummy reinforcement information in that case.

 

[TLHS]

jhnblgr is saying the analysis effort is a waste of time because the method you're using isn't really relevant. A thin unreinforced or underreinforced slab is only an elastic mat for a very small range of loads before cracking and if you treat it like one then everything is going to show it failing. The amount of cracking that's acceptable is a hard call and has very little to do with the slab criteria. It's so very much governed by the subgrade that without actual investigation of the subgrade below the slab you're output doesn't mean anything.

I can have a four inch slab on a nice twelve inch prepared gravel pad that will hold up to heavy trucks every day. I can also have a ten inch slab on random soft and wet subgrade that cracks to hell just sitting there minding its own business through a couple of winters

My real question is, why are you going through this effort? If it cracks do you care? If you don't, and you don't want to spend money on footings, then I'd just be sizing the bearing points to some conservative bearing pressure and be calling it a day. Basically a footing on top of the slab. If you really want, you can sawcut through the slab locally around these footing locations too. That's assuming there isn't an uplift that needs to be dealt with and you don't have frost considerations to worry about. I'd still probably smack a couple of holes in the slab to get eyes on it though.

 

[XR250]

Thanks. Not really helpful. garbage in = garbage out is a good admonition for an ignorant audience.

No comment

 

[PE_JRM]

TLHS: yes I agree the pads on top of the slab with saw cuts is a good, cheap and easy to impliment solution.

If you read back through the thread you'll see that I posted this a few days ago (oct 6):

In writing this post I've collected my thoughts and think that my best option is pouring reinforced pads right on top the existing slab. These pads would be coupled with CMU piers. Around the perimeter I can fill between the piers with an apron of some sort. This solution would be cheap, easy to do myself and if I saw kerf the existing slab around the pads I'm not relying on it to do anything structural. The only downside I can think of is that I might get some differential settlement from concentrating the loads on piers. That said, although I'm designing the floors for 50 psf and the roof for 20 psf its unlikely the building will ever get loaded that heavily.

As a structural engineer with 40 years of experience I think the pad on top solution is fine for our climate (Atlanta, GA). I will probably just go ahead with this if i can't show the distributed loads for the pony walls good. The prescriptive allowable bearing stress here is 1500 psf which is easy enough to design for. In truth if I had the tests done the bearing stress is probably more like 2500 psf. Currently the structure is zoned agricultural and I don't have to follow the county code but there is a very good chance that it will get rezoned while I'm working on it and the county will then have jurisdiction. I don't want to get stuck halfway through my renovation having to satisfy a county building inspector who doesn't like the foundation details.

I've had a couple of people suggest a series of perimeter and interior pony walls (noted to be the obvious solution that would be fine without even checking) only to flip on the idea a few days later. I have no idea why they flipped. My crude hand cals show that the pony walls stress the slab up to the "allowable". This is all based on very old school and very very conservative design approaches used in the 20's and through WW2 and after by the Army. My intuition tells me the slab is OK with my loads but I would like to check it with a more rigorous technique.

I already know and understand all of your points about the subgrade. Like most civil/structural engineers I took a few courses in soil mechanics and foundation design. I would still like to proceed with some assumed k values to see if its worth the money to have the soil tested by a geotech. Normally you would just do the tests to cover all the bases, protect your reputation and your license. My approach has been to do the analysis with an assumed subgrade modulus (K value) of about 60 pci and see if the slab checks OK. 60 pci is recommended by Terzaghi for a stiff clay if you dont have test values. The GADOT design manual recommends much higher k values for preliminary road designs in this area. If I can get the slab to check out for an assumbed k value I'll have the soil tested and proceed (I believe its closer to k =100 pci). If I can't show the slab good with conservative assumed values not I'll just do the pad on top of the slab solution.

 

[Tomfh]

My crude hand cals show that the pony walls stress the slab up to the "allowable"

Provided the walls can smear the load out over a sufficient area then it doesn’t matter.

Do it your way if you like, but a grid of walls will work. It works on soil alone, and will work with a slab on top of soil.

 

[PE_JRM]

Tomfh: Thanks. I appreciate your comments. Sufficiently "smearing" out the load has always been my objective. As always,the devil is in the details. Knowing that my plan, whatever it might be, adequately accomplishes that objective has been the detail I can't really quantify. I'd strongly prefer that the slab doesn't crack even though it doesn't matter and I know that. I've been using cracking and a slab on grade analysis as the only means I could think of to quantify if the "smearing" is smeared enough. I've found the issue to be frustrating due to limited analysis resources and data I have available to me.

So, maybe I need to change my approach. I've made an effort to think of some simple analysis solutions and I've come up empty on that. Just making sure that I get a distributed pressure under the pony walls under 1500 psf (soil allowable) and also checking punching shear is as simple as it gets. But, it almost assures me that the slab will crack somewhere, probably parallel to the edges. There is no gravel sub-base beneath it to spread the load or channel water away and there are probably some soft spots under the slab. That said, a rigorous analytical approach really doesn't preclude problematic soft spots either.

How would you be satisfied that you've spread the load sufficiently? I can always build more pony walls or post supports to lower the applied loads but how would you determine what is enough?

 

[Tomfh]

If the stress in your walls and piers is below the allowable soil bearing stress then it will be fine. Widen the bottom courses to suit. If the slab cracks it cracks. Slabs crack.