Two soils parts with a contact surface

[xtrata]

Hello I have two soil blocks: one is large (clay) and the other is small (gravel).

The gravel block is embedded in the clay block. Since I am using a half space the clay and gravel blocks have a free face which is restrained in the x direction. The sides surfaces of the clay block are restrained in the x,y direction and the bottom is retained in the x,y and z.

My surface contact (between clay and gravel) is specified as:

*Surface Interaction, name=LC_BK
1.,
*Friction
0.,
*Surface Behavior, pressure-overclosure=HARD

*Contact Initialization Data, name=LC_BK

*Contact Inclusions
LClay_Backfill , Backfill_LClay

*Contact Property Assignment
, , LC_BK

*Contact Initialization Assignment
LClay_Backfill , Backfill_LClay , LC_BK

with a Geostatic step of

*Geostatic, utol
1e-05, 1e-05, 1e-10, 1e-05

with a gravity load of

*Dload
, GRAV, 9.81, 0., 0., -1.

All I want to do is bring the mesh to equilibrium with the gravity load. However after spending 2 weeks on this I have not been successful. I have attached the .inp file.

My status file shows negligble vertical displacements on the top node of the gravel block (top surface of the model) 3.906e-008.

But I get the following errors:

1. 49-51 NEGATIVE EIGENVALUES.
2. ***WARNING: SOLVER PROBLEM. NUMERICAL SINGULARITY WHEN PROCESSING NODE HALF_LCLAY-1.73 D.O.F. 8 RATIO = 8.16516E+12. (I don't have any PWP running so this is ok)
3. DISPLACEMENT INCREMENT FOR CONTACT IS TOO BIG (reads like a contact problem, but in all increments my constraint errors are within tolerances)
4. FORCE EQUILIBRIUM NOT ACHIEVED WITHIN TOLERANCE. (The largest residual force occurs in Dof 3 but average force of 83.1)


The final warning are (apart from Vol flux):
1. Displacement increment for contact is too big.
2. The strain increment is so large that the program will not attempt the plasticity calculation at 416 points
3. The plasticity/creep/connector friction algorithm did not converge at 2 points


With the crashing error being:
Too many attempts made for this increment
Any thoughts?

Thanks

 

[AnnaG]

Hello,

Is it really necessary to use two blocks of soil? May be it will be possible to use one part divided into two partitions with different sections. Like this You will avoid this error.

Anna

 

[xtrata]

Hi Anna

I am running a 3D half space analysis (should have mentioned), as far as I can see I can only partition edges,faces and cells.

The reason I need to use parts is that in the end I want to run a foundation test modelling the actual foundation as it has a complex geometry so merely applying a stress/velocity won't replicate actual soil movement.

I confirmed that my boundary conditions are OK by running them on a uniform block of soil.

Taking your idea I got the two soil blocks and merged their geometry (retaining the boundary) and then meshed the new part. But it is currently not converging, I am going through this process (

http://www.eng-tips.com/viewthread.cfm?qid=191374)

The error is THE PLASTICITY/CREEP/CONNECTOR FRICTION ALGORITHM

1. check your unit systems are correct.
2. start with an elastic analysis and introduce one non-linear function at a time (plasticity/creep/contact/etc)- this way you can see what is causing the problems.
3. Check your mesh quality for sharp discontinuities (these can cause problems when plasticity and creep are introduced)
4. Reduce your time increment (or at least the smallest allowable time increment, if automatic) [DONE THIS]
5. De-bug your model with 'Job Diagnostics' tool to see where convergence fails
6. refine your non-linear behaviour - add hardening to plasticity
7. Check your loads and boundary conditions aren't the problem - is it load or displacement controlled? are the units and magnitudes correct? [ONLY LOAD IS GRAV AND MY BCS ARE OK]

Even though the geometry is merged are contact conditions still required? Between the two soil blocks?

.INP attached.

 

[xtrata]

So I have gone back to a static/elastic analysis. I am able to apply a surface load of 9.81kpa but whenever I try to run gravity it just shuts me out.

See attached with fake gravity