11.02.2013 Views

Composite Materials Research Progress

Composite Materials Research Progress

Composite Materials Research Progress

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.

34<br />

Jacquemin Frédéric and Fréour Sylvain<br />

i*<br />

For closed envelopes, the condition − 1 ≤ Fmmnn<br />

≤ 1 has to be satisfied. But a more<br />

detailed theoretical study (see Liu and Tsai, 1998) reduces the admissible range to the domain<br />

1<br />

[-1,0]. The same reference (Liu and Tsai, 1998) advises the choice of F -<br />

2<br />

* I<br />

mmnn = for the<br />

macroscopic interaction term (which corresponds to the generalised Von Mises model), since<br />

this value is reasonable for a wide range of laminates. Taking into account this additional<br />

I I<br />

F = F ensures the<br />

assumption in equation (56), the knowledge of I<br />

F 1111 and<br />

I<br />

determination of the last two missing interaction terms 1122<br />

F and<br />

2222 3333<br />

I<br />

F 2233 , in stress space.<br />

One similar method could be applied in order to determine the macroscopic strength<br />

parameters expressed in strain space from the ultimate strains. Nevertheless, this method is<br />

not useful in practice since uniaxial strains are difficult to apply to a sample. Thus, the<br />

ultimate strains are generally deduced from the ultimate stresses: to reach this goal, one has to<br />

introduce the macroscopic properties, i.e. the stiffness tensor L I , in order to relate both failure<br />

criteria through Hooke’s law (33) expressed at macroscopic scale assuming a purely elastic<br />

load.<br />

5.2.3. Identification of the Microscopic Strength Parameter (of the Matrix Only)<br />

Using an Inverse Method<br />

From the standpoint of the structural designer, it is desirable to have failure criteria which are<br />

applicable at the level of the lamina, the laminate, and the structural component.<br />

Nevertheless, failure at macroscopic scale is often the consequence of an accumulation of<br />

micro-level failure events (Tsai, 1987; Liu and Tsai, 1998). Laminated materials typically<br />

exhibit many local failures prior to rupture. Thus, it is important to build up tools enabling to<br />

enhance the understanding of micro-level failure mechanisms in order to develop higherstrength<br />

materials. The ultimate goal is to have a failure theory that the designer can use with<br />

confidence under the most general structural configuration and loading conditions and that the<br />

developer of materials can use to design and fabricate new products to meet specific needs. In<br />

order to reach this goal, the estimation of microscopic strength criteria would be of a valuable<br />

help.<br />

Since the epoxy resins involved in composite structures generally exhibit an isotropic<br />

hygro-mechanical behaviour, the microscopic strength criterion expressed in terms of stresses<br />

(54) simplifies as follows:<br />

⎧<br />

⎪1<br />

⎨<br />

⎪<br />

⎩<br />

= F<br />

m<br />

1111<br />

+ F<br />

m<br />

11<br />

⎛<br />

⎜σ<br />

⎝<br />

m<br />

2<br />

11<br />

+ σ<br />

+ σ<br />

m m m<br />

( σ + σ + σ )<br />

11<br />

22<br />

m<br />

2<br />

22<br />

33<br />

m<br />

2<br />

33<br />

⎞<br />

⎟ + 2F<br />

⎠<br />

m<br />

1212<br />

σ<br />

m<br />

2<br />

12<br />

+ 2 F<br />

m<br />

1122<br />

m m m m m<br />

[ σ ( σ + σ ) + σ σ ]<br />

Thus, only four strength parameters have to be determined in order to enable failure<br />

m m m m<br />

predictions at microscopic scale:<br />

F1111<br />

, F1212<br />

, F1122<br />

, F11<br />

.<br />

Hypotheses being compatible<br />

with the experimental observations are necessary to build an inverse model enabling the<br />

11<br />

22<br />

33<br />

22<br />

33<br />

(57)

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

Saved successfully!

Ooh no, something went wrong!