Stress analysis without meshing - damage tolerance assessment directly from CAD

Submitted by Stephane Bordas on Wed, 01/08/2014 - 10:34

research

stress analysis; meshing; isogeometric analysis; boundary elements; IGA; damage tolerance assessment; CAD; IGA

links to papers for download:

http://orbilu.uni.lu/handle/10993/12159

http://orbilu.uni.lu/handle/10993/14135

http://orbilu.uni.lu/handle/10993/13847

http://orbilu.uni.lu/handle/10993/12157

http://orbilu.uni.lu/handle/10993/11850

movies:

http://www.youtube.com/watch?v=RV0gidOT0-U

http://www.youtube.com/watch?v=cYhaj6SPLTE

---

Thank you so much for your remark Roberto. This makes a lot of sense. In
fact, we arrived, through a different route, to similar conclusions. By
trying to use model reduction techniques (algebraic such as POD) or
multiscale (homogenisation), you can see that the fractured region is
non-reducible, for example here:

http://orbilu.uni.lu/handle/10993/12024

http://orbilu.uni.lu/handle/10993/12012

(for homogenisation)

and here

http://orbilu.uni.lu/handle/10993/10207

http://orbilu.uni.lu/handle/10993/12454

http://orbilu.uni.lu/handle/10993/12453

http://orbilu.uni.lu/handle/10993/14475

(for algebraic reduction)

with the most recent paper here:

http://orbilu.uni.lu/handle/10993/10206

This work was perfrormed by Dr Pierre Kerfriden who's been leading the
model reduction and multiscale group in my team over the last 4 years. I
am sure he would be keen to exchange ideas on this.

All the best once again, from Luxembourg/Cardiff,

Stéphane

Roberto Ballarini

cannot homogenize heterogenous structures to determine s.i.f.

Must be careful when calculating s.i.f.

Wed, 01/08/2014 - 20:32 Permalink

Stephane Bordas

yes, completely agreed. in

yes, completely agreed. in our case, the structures are homogeneous, but indeed, there are special cases when the domain integral domain crosses the boundary of the domain, and where the domain integrals then have to be treated cautiously... thanks for your comment.

--
Stéph
Stéphane P.A. Bordas, Professor
Director, Institute of Mechanics & Advanced Materials (IMAM)
ERC Starting Grant Research Group Leader (RealTCut)
School of Engineering
Cardiff University, Queen's Buildings
The Parade, CARDIFF CF24 3A

Wed, 01/08/2014 - 21:00 Permalink

Roberto Ballarini

stress intensity factor

We have shown that one cannot accurately determine the stress intensity factor at the tip of a crack in a heterogeneous structure (laminated, aggregate of inhomogeneities, etc.) by replacing the whole structure with a homogeneized medium and calculating the s.i.f. of the crack in that medium. This simplification is simply wrong.

What must be done is to retain the crack tip region explicitly (with the actual microstructure), and at best replace the surrounding region with the homogeneized medium. In fact this simplification produces very accurate results. For example, if a crack lies within an inhomogeneity that is surrounded by a random distribution of other inhomogeneities, the simplest model is of a crack lying inside the inhomogeneity, and this inhomogeneity is surrounded by the associated homogeneized medium.

More can be learned from:

Wang, Ballarini and Rodin, ASCE J. of Engineering Mechanics 2008

Wang and Ballarini, Meccanica 2004

Jha, Charalambides and Ballarini, IJSS 1997

Ballarini, Islam and Charalambides IJF 1995

Thu, 01/09/2014 - 04:15 Permalink

Stephane Bordas

lack of separation of scales in fracture mechanics

Thank you so much for your remark Roberto. This makes a lot of sense. In fact, we arrived, through a different route, to similar conclusions. By trying to use model reduction techniques (algebraic such as POD) or multiscale (homogenisation), you can see that the fractured region is non-reducible, for example here: