Moody chart v/s FEM analysis for u/g tank

[SHBH]

I am designing an underground tank with dim 10x15x5m deep.
With normal soil loading (using triangular distribution) i used Moody chart for short wall (5x10) with three edges fixed. At the same time i modelled this tank in STAAD with FEM mesh size of 0.5x0.5x0.4m thk. My problem is, the moments at the slab and vertical wall junction doesn;t match at all.
One thing which comes to my mind is that in Moody chart the edge conditions are fixed where as in actual STAAD model it is in continuation with base slab.
In staad analysis i have used soil springs based on the modulus of subgrade reaction of 15000KN/m3.
These springs have been desfined in vertical dir and in horizontal dir i have released the moments (i guess this is the closest approx of a subgrade).
Can any body tell me why these moments are so different?

 

[JedClampett]

You're on the right track. The Moody charts give fixed end moments, while STAAD is distributing them based on stiffness.
If you were to model just one wall with fixed edges, you should get the Moody results.

 

[SHBH]

that's true JedClampett, but the thing is 126knm is the moment(base slab & short wall junction) i get from Moody chart where as STAAD gives me just 25knm......!!
can there be a mismatch of this level......moreover this moment varies if i change my spring stiffness...
how exactly spring stiffness controls this moment...?

 

[miecz]

My tank models behave the same as yours. A 3D model will have the most effect of the moment at the base of the short walls, for this reason. Imagine a 2D rectangular frame with long and short sides as in your tank. Under uniform pressure, the corners will rotate, and this rotation will be dominated by the longer walls as they have larger fixed end moments. Your tank wall corners will rotate in a similar fashion, and this rotation reduces the moment at the bottom of your short walls. By the way, 15000 kN/m^3 seems pretty stiff.

 

[SHBH]

Good reasioning jmiec. So as per your logic, if i have a square tank then this moment difference shouldn't be that much. Anyway i will try with difference model.
But still how does spring stiffness control these moments.
If i change spring stiffness, it affects all momemts.

 

[miecz]

When you load your walls, the long walls deflect inward and the joint at the bottom of the walls rotate with the deflection. With soft springs, the mat offers less resistance to rotation, so the joint rotates more. The joint becomes less fixed and carries less of the lateral soil load. More of that load is carried in horizontal bending, so you'll see your wall corner moments and shears increase. Those corner moments cause the corners to rotate more, which tends to bow the short walls outward. This in turn effects the moments at the bottom of the short walls.

Model a square tank with very stiff springs, and you should be able to mimic the tables. (That won't work with a rectangular tank because of corner rotations.) Then soften the springs to see the effects.

 

[271828]

Good job verifying your results and not just trusting the program!

Make some simple models in STAAD to verify your results.

Start with a simple slab that should give you w*L^2/8, etc.

One good intermediate model would be to take your current model and actually fix the bottom and edges. See if that matches your theoretically exact results--it should be close.

Will your deflected shape have the bottom of the slab going upward? If so, will the soil springs be correct or should they be compression only? I think it depends on your situation.

 

[271828]

Sorry jmiec, I didn't see your last paragraph before typing my next-to-last paragraph!

 

[miecz]

271828-

Yeah, I do that a lot myself. Nothing really bothers me here (no headaches or stomach aches), unless I post something I've put some thought into and it disappears, presumably red flagged. I wish I could challenge the red flags.

 

[271828]

LOL, a couple of interesting threads just vaporized about a month ago. I found that to be irritating, but that's the way it goes.

 

[SHBH]

great man....
model is behaving in the same manner as you have said (or the other way round )...
this has really improved my knowledge on underground tank design.....so now i am not worried about the difference in the moments.........

 

[miecz]

Disregard my remark about the soil being stiff. I thought I read 150,000 kN/m3.

Seeing that your moments are a function of your spring constant, you might consider performing two solutions, with a maximum and a minimum spring constant.