Contact and automatic stabilization

[jamescook]

Hi,

I am using automatic stabilization in a static, general analysis step for a model involving contact. I am using the specify the dissipated energy fraction method with adaptive stabilization set to 0.05. From reading the manuals I reduced the dissipated energy fraction until the ratio of energy dissipated by viscous damping (ALLSD) to the total strain energy (ALLIE) was below 5% which was achieved for a value of 6.25E-6. Upon further reading of the manual (v6.8) in the section titled "Ensuring that an accurate solution is obtained with automatic stabilization" it states that "viscous damping energy (ALLSD) should be compared with the total strain energy (ALLSE) to ensure that the ratio does not exceed the dissipated energy fraction or any reasonable amount". Is there a typo in the manual as in a previous section ALLIE was referred to as the total strain energy ?

When comping ALLIE TO ALLSD in my model the ratio is below 5% as previously stated whereas if I compare ALLSD to ALLSE there is enormous fluctuation in the magnitude of the ratio. Also a comparison of viscous forces (VF) to overall forces in the analysis can be used to gauge the accuracy of the solution. Are the overall forces referred to here total forces (TF)?

I am using auto stabilization as my model aborts without it as soon as contact is made so would it be more appropriate for me to use something like contact controls or contact damping instead. I have tried using contact controls stabilize already, varying the default value by a couple of orders of magnitude but to no avail.

Any help would be much appreciated.

 

[Elabbasi]

You should compare ALLSD with ALLIE not ALLSE. Abaqus may have corrected the text in the newer versions because according to the 6.10 documentation, the Abaqus Analysis User's Manual section on "Ensuring that an accurate solution is obtained with automatic stabilization" states that you should "Compare the viscous damping energy (ALLSD) with the total strain energy (ALLIE), and ensure ...". In many cases ALLSE and ALLIE are comparable at the start of an analysis (since there is no plasticity, damage or creep yet). ALLIE is actually the sum of ALLSE and several other energy terms ... the equation is somewhere in the documentation. If you have a big difference between ALLIE and ALLSE it could be due to hourglass control which could cause the convergence problems you're seeing. Yet again it could be something else! Abaqus does provide plenty of output in the .msg and .dat file to help identify the problem. Look also at the number of nodes experiencing severe discontinuities due to contact, as reported in the .msg file.

Nagi Elabbasi
Veryst Engineering