Fracture propagation

20 views
Skip to first unread message

Mario Ceresa

unread,
Mar 3, 2017, 1:28:20 PM3/3/17
to sfepy-devel

Dear all,

I would like to ask for advice on a the following problem. I'm modeling tissue destruction in lungs, where tissue is initially modeled as a pre-stressed network of nodes connected by springs at equilibrium. Due to inflammation, some nodes are "lost" and the new equilibrium of the remaining nodes is computed, leading to increase in the stress level that will cause even more ruptures. An introductory review can be found, for example, here:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4041120/


I was wondering which could be the best way to simulate that: I started creating a grid with networkx and discretizing the elastic equations on the nodes but soon I realized that I was replicating existing FEM code. I thought I could use sfepy for that, however, is it possible to remove nodes from the mesh during the computation? I tried to look it up in the documentation but I couldn't find anything explicit.

Thanks for any suggestion you might provide and thanks for this beautiful software!

With best regards,

Mario

Robert Cimrman

unread,
Mar 3, 2017, 3:44:37 PM3/3/17
to sfepy...@googlegroups.com
Hi Mario,

to simulate damage/weakening of a material, in past people have used (with
sfepy) a solution-dependent material parameters, see e.g. [1]. You could, for
example, reduce the stiffness depending on the stress/strain value. I am not
aware of other approaches - sfepy currently does not support damage directly.

r.

[1]
http://sfepy.org/doc-devel/examples/linear_elasticity/material_nonlinearity.html

On 03/03/2017 07:07 PM, Mario Ceresa wrote:
>
>
> Dear all,
> I would like to ask for advice on a the following problem. I'm modeling
> tissue destruction in lungs, where tissue is initially modeled as a
> pre-stressed network of nodes connected by springs at equilibrium. Due to
> inflammation, some nodes are "lost" and the new equilibrium of the
> remaining nodes is computed, leading to increase in the stress level that
> will cause even more ruptures. An introductory review can be found, for
> example, here:
> https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4041120/
>
> <http://journals.plos.org/ploscompbiol/article/figure/image?size=large&download=&id=10.1371/journal.pcbi.1005282.g001>
Reply all
Reply to author
Forward
0 new messages