No announcement yet.

Bulk Viscosity Parameter in Abaqus Explicit

  • Filter
  • Time
  • Show
Clear All
new posts

  • Bulk Viscosity Parameter in Abaqus Explicit

    I am working on a nonlinear contact problem with Abaqus. One of the parts
    involved in the complete model is very small and as its behaviour needs to
    be studied when it interacts with other parts, it is modelled with pretty
    smaller elements as compared to other parts. When I do a Dynamic explicit
    analysis then the time increment which Abaqus calculates from an Element by
    element estimate is of the order 1e-8 due to that very small part. I did
    mass scaling such that the increment becomes 1. e-7 (by DT= 1. E-7) and
    used selective subcycling to complete my simulation. The percentage change
    of mass in this process is 2.6 % which I think is pretty acceptable. But
    there is one problem now. When I observe the transient response of my
    simulation then I see that the part which is very small and of material
    rubber is kind of jittering a lot as though it is modelled as fluid.
    However, on the positive note this jittering is almost dying out at around
    half-time of the simulation. I think that as the time increment size is very
    less i.e. 1e-7 it could be high frequency vibrations that I see at the start
    of the simulation. I tried increasing the default Bulk Viscosity parameter
    to (0.25, 4.0) and I do observe that the high frequency jittering has
    reduced at the start of the simulation itself. But I want to completely get
    rid of with such high frequency vibrations (if I am right that they are the
    same)...Is increasing the Bulk Viscosity parameter more further until I get
    rid of them is the solution? Also there are limits to this value as
    increasing the Bulk Viscosity parameter reduces the stable time increment
    and affects accuracy also.

    Thanks for your help.


  • #2
    You are exactly running into the (perhaps) most common problem with Explicit FE simulations. I don't know if a quick fix, try different parameter combinations until you get the response that you can live with...

    Jorgen Bergstrom, Ph.D. PolymerFEM Administrator


    • #3
      Have any one know how to fix it?
      I modeled a frictional contact between 2 blocs of concrete using Abaqus explicit for quasi-static problem (1st photo).
      When printing the friction-displacement of contact, I observed the oscillation response (2nd photo). While the mean value of friction seems correct (Friction = 0.2*2MPa*surface = 0.016) and cinematic energy is negligible in comparison with the external work (3nd photo).
      I tried to modify mesh, time step and viscous parameter, but it does not help.
      If you have any experience to deal with the oscillation, could you please let us to know?
      Thank you very much

      Click image for larger version

Name:	test2.jpg
Views:	1
Size:	13.7 KB
ID:	12203 Click image for larger version

Name:	test.jpg
Views:	1
Size:	15.5 KB
ID:	12204 Click image for larger version

Name:	test3.jpg
Views:	1
Size:	8.5 KB
ID:	12205


      • #4

        I have a general question concerning the linear and bulk viscosity parameters. Is there a criterion to determine a reasonable range which both parameters should be in?

        Is it quite common to set parameters like (0.25, 4.0) or should these values already be considered "extreme"?

        It there's easy accessible literature (apart from the Abaqus User's Manual) dealing with this topic?

        Thank you: