Hello Everyone,
I am trying to simulate a fiber filament (with a diameter 1,5E-5 m and 0.5 m long) under gravity. It supposed to be a really simple one and is thought to be a precondition for step 2 (which has "dynamic, temp-disp, explicit" step procedure).The purpose is the fiber that being fixed on one side, should be leaning on the floor before step 2 starts.
Because of the step 2, I could only use "dynamic, explicit" step procedure for step 1. I assigned the fiber as a beam element, but it won't work because abaqus/explicit 6.13 doesn't support anisotropic elasticity with truss, rebar, pipe, and beam elements (see the manual "22.2.1 linear elastic behavior").
I have also tried using "engineering constant" elasticity and beam method, but the following error occured: "Anisotropic material properties without a local orientation system have been defined for 10 elements." Even after I defined the material orientation (in module property), the same error still occurs. The problem still persists even using a truss element. I attach as well my input file.
Could anyone advise me how to simulate a fiber fillament that being attached on one side and is under gravity stress?
It has to be with step "dynamic, explicit" step. I could simulate the same simulation using "static, general" step method, but unfortunately this step method doesn't compatible with the step 2 (dynamic, temp-disp, explicit).
Thank you for any suggestion :-)
Best regards
Gunardilin
Let L be of the length of the filament, A be the area of cross section.
Create a finite element model with, let us say, 20 elements each of length L/20.Fix one of the nodes in the direction of gravity.
Now apply nodal loads calculated as (Mass Density*g*A*L/20)/2 at the fixed end node and free end node. Apply nodal loads calculated as Mass (Density*g*A*L/20) at all other 19 intermediate nodes. All the nodal loads are in the direction of gravity.
Run the finite element analysis as a static analysis.
The deflection at the free end should be mass density*g*Length*Length/2E
Plot the deformation, stress and strain plots.
Increase the number of nodes for more accurate results.
For more info refer to problem 2.4 in the book "A Primer on Finite Element Analysis" by Anand V. Kulkarni, Venkatesh K. Havanur.
Hope this helps.
Thank you Koratagere Venkatesha for your interest in my problem. But still I can't solve my problem. I think I have to define the material with a subroutine VUMAT.
Have Anyone experience with VUMAT subroutine for Beam elements? An example or some feedbacks would be much appreciated.
Best regards,
Gunardi
In case Abaqus does not support anisotropic material properties for truss and beam element you have no other choice as to model the fiber by three-dimensional elements. Thus, obviously the numerical costs will be much higher.
Another approach would be two write your own FE code or adapt existing ones such as CALFEM (https://sourceforge.net/projects/calfem/) or FEAP (http://www.ce.berkeley.edu/projects/feap/).
There are not many other choices if commercial codes do not provide you with the capabilities you need.
Best regards
Thank you Sascha Duczek for your feedbacks.
I built my filament model already with a 3D-elements (i.e. 3D beam elements). But the problem arised after defining the material properties as it is not supported for beam elements in Abaqus.
What do you think if I define the material with VUMAT subroutine? Do you think it will solve my problem? I am struggling right now to understand the coding of VUMAT. I have the elastic tensors for the material ready. I am just wondering, if this "path" is the right one to solve my problem.
For your suggestion about writing my own FE-Code, I don't think I have that capicity and enough time to do that.
Best regards,
Gunardi
With three-dimensional elements I was referring to continuum elements and not to three-dimensional beam elements. With continuum elements there would be no problem with your modelling approach.
Using a vumat you can principally implement the constitutive equations of your choice. So in theory it should work. But I am not really familiar with the Abaqus implementation of beam elements and derived restrictions. That is why I suggested continuum elements as this will definitely work. As mentioned before the numerical costs will be quite high considering your model dimensions.
A beam element would be the model of choice and I do not understand why Abaqus does not permit anisotropic material laws for beams. Did you try or have access to a different FE software? Otherwise implementing your own user material would be the computationally cheapest choice. From my point of view implementing a vumat for beam elements should work, if Abaqus has no other restrictions.
Considering your own FE-Code CALFEM is for example a open Matlab code which offers beam elements. Probably there is a chance to adapt the existing code for your purposes.
- Badges
- Science topic