27.12.2012 Views

ARUP; ISBN: 978-0-9562121-5-3 - CMBBE 2012 - Cardiff University

ARUP; ISBN: 978-0-9562121-5-3 - CMBBE 2012 - Cardiff University

ARUP; ISBN: 978-0-9562121-5-3 - CMBBE 2012 - Cardiff University

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

In the Element-Free Galerkin (EFG) method [16], only nodal data and boundary<br />

descriptions are required to formulate the discrete Galerkin equations. Basically, the<br />

main difference between the EFG method and FEM is how the shape functions are<br />

constructed. The construction of the shape functions in the EFG method is based on a<br />

connectivity-free technique known as the moving least square (MLS) approximation<br />

[17]. One of the advantages of MLS approximants is that their continuity is related to<br />

the continuity of the weight function. A background-cell structure is employed for<br />

evaluating the integral expression. More information about meshfree methods can also<br />

be found in many literatures [18, 19, 20].<br />

In this research, crack propagation is simulated by cohesive modeling (Fig. 2a) that<br />

captures the nonlinear behavior of the process zone, which cannot be modeled by linear<br />

elastic fracture mechanics [21]. It represents the physical processes occurring in the<br />

vicinity of a propagating crack by a simplified traction-displacement relationship. Two<br />

properties of the softening curve are most important: the tensile strength (σc) and the<br />

energy release rate (Gc), which are introduced in the cohesive crack model as constant<br />

material parameters. Several types for this softening function have been developed. So<br />

that, a linear traction-displacement curve was employed in current work and the<br />

cohesive model parameters, energy release rate and critical stress are obtained from the<br />

experimental results reported in the literature for human cortical bone [8] – see table 1.<br />

Table 1<br />

Elastic and cohesive properties of the osteonal cortical bone tissue in transverse direction.<br />

Model<br />

Osteon<br />

Interstitial matrix<br />

Cement line<br />

Elastic modulus<br />

(GPa) [11,12]<br />

Poisson’s<br />

ratio[1]<br />

Normal strength<br />

(MPa) [8]<br />

Fracture toughness<br />

(N/mm) [8]<br />

In EFG cohesive fracture methods, in contrast to cohesive element method, there are no<br />

interface elements involved. It uses cohesive zone to simulate the fracture surface and δu<br />

is the displacement jump across the crack surface. The representation of crack in the<br />

meshfree method is depicted by the so-called “visibility” criterion [16, 22]. Since the<br />

growth of crack surfaces can be reflected without changing a connectivity of nodes in<br />

such a method, consequently it is possible to simulate crack propagation without<br />

remeshing and it is an excellent choice for solving these problems. The presence of the<br />

crack, a surface of discontinuity, only affects how nodes influence the displacement at a<br />

given point. A visibility criterion [16] can be used to select these points and<br />

constructing shape functions near a crack. In this method, the crack boundary is<br />

considered to be opaque (Fig. 2b). Nodes that are on the opposite side of the crack are<br />

excluded in the approximation of the displacement field, during the construction of<br />

weight functions [19].<br />

(a)<br />

12<br />

15<br />

6<br />

Cohesive traction<br />

Crack<br />

Fracture process zone<br />

0.3<br />

0.3<br />

0.4<br />

43.5<br />

43.5<br />

0.71<br />

0.71<br />

Fig. 2. (a) Cohesive crack model, (b) Domain of influence near a crack for the visibility criterion[23]<br />

(b)<br />

Domain of influence of node I<br />

Crack

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!