• From: Vancouver-Canada
  • Registered: 19-05-2007
  • Posts: 60

Topic: Under Integration and Non-Local Averaging

Dear Borek,

I'm running the damage propagation problem with my own material model that uses non-local averaging.
The problem is when I use under integration in my explicit dynamic analysis the damage pattern changes significantly.

Since the results of fully integration totally makes sense both in terms of load-displacement that I get and the damage pattern, I was wondering if you have any idea how under integration would change the response that dramatically?
I have attached a figure showing the damage patterns in the fully integrated (top) and under integrated (bottom).

Thanks for your helps,
Alireza

  • From: Vancouver-Canada
  • Registered: 19-05-2007
  • Posts: 60

Re: Under Integration and Non-Local Averaging

I tried the built-in nonlocal RCSD model and this model also shows a huge dependency of results on the number of integration points.

When I use the single integration scheme, damage initiates and grows much faster compared to fully integrated one.

Please note that my mesh size is about 0.5mm and the non-local radius is 2mm therefore there are plenty of integration points in the averaging zone even in the under integrated one.

I'm wondering if this indicates a problem in the averaging process or not!

Thanks
Alireza

3 Reply by bp

  • bp
  • Administrator
  • Offline
  • From: Prague, Czech Rep.
  • Registered: 02-03-2007
  • Posts: 1,047

Re: Under Integration and Non-Local Averaging

Dear Alireza,

the problem you describe is interesting. However, it is hard to say from the pictures you have provided what is actually happening. We can tell more, but need more details. Could you share input file? This can help a lot.

Borek

bp's Website