FEM results at corners

[dutchie]

I have a pretty basic question I was hoping I could get some enlightment on. I recently began doing some FEM analysis on Robot ver 16. My current model is the pump house roof of a primary clarifier. The pump room is constructed of two 6 meter high walls with a 10" slab spanning about 4 meters between walls. incorporated into the slab is a influent trough. the tank walls are subjected to soil, groundwater and surcharge loads, and the roof to snow and basic live load. the model is returning values which compare well with estimated hand calcs with one exception. at the junction between the walls and the roof slab where the elements join, the stress values are reported as 0 or near to. Since the boundary condition between the slab and wall is fixed, i would assume to see stress values at these locations, but as i stated, the values converge to 0. I should note that a short distance from the corner, the values are close to what hand calcs predict.

My question is this, is there a boundary condition problem that is causing this to happen. Any input is very helpful.

 

[pphuddar]

Hi!
Two things I would like to suggest.
1. Try and model the walls if possible
2. Change the End conditions. Because the the connection between wall and slab is not fully fix so u are getting some absurd values.

Regards,
Prashant

 

[corus]

You should have no boundary conditions between the slab and wall as these are simply connected through the nodes.
Check the displaced shape of your model to see if it is as you expect. Usually you can tell what the problem is from that.

corus

http://www.corusresearch.com