Out-of-plane strain in planse strain element

[harshit175]

I am running a 2D model with CPE8P elements. These are 8 noded elements, plane strain, having pore pressure degree of freedom. Ideally there shouldn't be any out of plane (OOP) strain (E33) in these elements. But I am observing OOP strain. Can anyone provide an explanation as to why this is happening.

Additional Information (not absolutely needed to answer the question):

A horizontal well is modelled. A vertical plane for modelling is considered.

The model has the first step as the geostatic step, wherein initial stresses get established in the rock mass.

In the second step, a borehole is made by removing a semicircular portion. And in its place, a casing is introduced.

 

[Erik Panos Kostson]

That is strange, not sure why. The only thing I can think of is that somehow plane stress elements are used instead where we can get E33 strains in theory.

Not that it is very likely, but I would also look at the results in the default Cartesian system, just in case there is something going on with the transformation of results (unlikely I think though).

Unfortunate that's all I have.

 

[harshit175]

Plane stress elements have displacement (U3) in the third direction. The elements I am using don't have that displacement, so they are definitely plane strain. I queried the element type as well from the ODB file. It is CPE8P. It's a 2D model, and I haven't applied any coordinate. transformation.

I have attached the input file as well.

If anyone could run the file, then it can be seen that out of plane strain from second step onward are non zero.

Any insight would be really appreciated.

 

[Erik Panos Kostson]

Well the UCS (from the image you attached) you are using shows R,T,Z which is a cylindrical one I believe.

If you look on the local element system, or in the Cartesian (default), and you still have E33, I would try to reduce and simplify the problem.

Try remove all of the staging, and just run the model, with rock properties (linear elastic) without removing/adding any elements, so as simple as possible, with a single step and load (say gravity, and no initial in situ stresses if possible).
If that is OK, then it might be something with the staging when elements are added/removed (not morphed, but with strains).

 

[Mustaine3]

The E33-strain comes completely from the inelastic portions (See IE and PE), so I would check if the plasticity model creates those strains to work properly.

 

[harshit175]

Oh, yes you are right. But the OOP strain was exact same even without using those transforms. The z coordinate was kept the same in both cartesian and cylindrical coordinates.

 

[Erik Panos Kostson]

Yes the z coincides.

@ Mustaine3: I never seen that happening (do you have a reference/case when that has happened); I do not know how abaqus does the plastic calculations and return mapping update for mohr coulomb plasticity. It could be the case though here, so that this is caused by the mohr cloulomb (MC) material, since the part that has a zero E33, is either rock or foam that are linear elastic and isotropic, and the part with non zero E33 is the soil/rock with MC.

One way to try it out is to just use linear elastic for the soil/rock as well.

 

[harshit175]

Eric and Mustaine,

I will try rerunning the model with the suggestions you have made. I will get back in 1-2 days.