FEA How to model this flat plate

[karthur1]

For simplicity sake, here is what I have.  Say I have a 4'x4' x 1in thick plate resting on the ground (soil).  There is a pipe in the center of the plate that has an load on the pipe (load directed toward the plate).  I am trying to model this to determine the load distribution on the bottom of the plate.  I know the AVERAGE pressure that the plate would exert on the ground would be force/area, but this is only valid if the plate is sufficently thick.  I want to keep the load that the plate exerts on the ground below a certian value.

Question is, how do I model this since the bottom of the plate is not really "Fixed". Its sort of like there is a bunch of soft springs under the plate supporting it. Common sense says that the plate will exert more force on the ground around the center of the plate and less toward the outer edges.

[karthur1]

Been messing with this some more, but still not sure about the results.  I added a very thick polyurethane part to the bottom of the plate. I added a mate constraint between the top of the poly and the steel plate.  In Stress Analysis, I applied a fixed constraint to the very bottom of the poly.  I thought that this would behave similar to a section of soil.

I was thinking that I could see how the stress was distributed around the load when I varied the thickness of the plate. I thonght that as the plate thickness increased, the stress in the supporting medium (poly in this case) would go down. But I'm not really seeing that.

[rmerlob]

I think that modeling the "ground" like your current approach is the only way since the stress distribution would depend on both the stiffness of the plate and the ground, question is, how to get properties for soil, seems to me that this is a problem normally handled with average pressures since a modulus of elasticity would be kind of hard to determine

[whunter]

Basically, you can't do it accurately with Inventor's FEA (which is very limited, unfortunately), since it's a non-linear problem, and Inventor does linear only. You can handle some non-linear problems to a certain extent with a linear solver, but it usually requires a lot of work combined with some analytical solutions.

[cwhetten]

Stress is independent of material properties--it is a measure of force distributed over an area.  So, if you really just need the stress, just model the ground as a flat plate.  If you need to consider deformations of the ground (strain), then this is where the material properties come into play (the modulus of elasticity).