13.06.2014 Views

Nonlinear Finite Element Analysis of Concrete Structures

Nonlinear Finite Element Analysis of Concrete Structures

Nonlinear Finite Element Analysis of Concrete Structures

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.

- 140 -<br />

evident contrast to the finding <strong>of</strong> Cedolir. and Dei Poli (1977)<br />

who also investigated beams failing in sheer. As here they analysed<br />

beams tested by Bresler and Scordelis (1963), but their<br />

beams had a shear span ratio = 4 whereas the beams considered<br />

here have a shear span ratio = 5. In their important investigation,<br />

Cedolin and Dei Poli (1977) found an extreme influence<br />

<strong>of</strong> the n-value as n = 0.25 resulted in a failure load twice as<br />

large as that determined when n = 0.025 was utilized. However,<br />

the failure loads as determined by Cedolin and Dei Poli (1977)<br />

were not clearly related to physical phenomena and large differences<br />

between their approach and the present one exist. In<br />

particular, the strain s<strong>of</strong>tening in the post-failure region was<br />

not considered? dilatation and secondary cracking <strong>of</strong> the concrete<br />

was ignored. As here Cedolin and Dei Poli (1977) used constant<br />

strain elements, but no diagonal cracking was determined<br />

at failure. Cedolin and Dei Poli (1977) suggest this to be a<br />

consequence <strong>of</strong> the smeared crack representation. If true this<br />

finding has important consequences, but the present study gives<br />

no support to it as diagonal cracking indeed is determined. As<br />

previously discussed, diagonal cracking follows as a result <strong>of</strong><br />

a strain localization in the region adjacent to the load point<br />

and this strain localization is a consequence <strong>of</strong> strain s<strong>of</strong>tening.<br />

Therefore, modelling <strong>of</strong> strain s<strong>of</strong>tening is considered as<br />

decisive.<br />

The behaviour <strong>of</strong> beams failing in shear is obviously very dependent<br />

on the existence <strong>of</strong> small tensile stresses. However, as<br />

demonstrated by case no. 2 and 4 the choice <strong>of</strong> different realistic<br />

tensile strength values has only a minor influence for a<br />

20% decrease <strong>of</strong> the o.-value results in a decrease in failure<br />

load <strong>of</strong> only J%.<br />

To investigate the importance <strong>of</strong> modelling <strong>of</strong> secondary cracks,<br />

case no. 5 is compared with case no. 2. In case no. 5 the calculations<br />

were terminated before failure was reached. It appears<br />

that modelling <strong>of</strong> secondary cracks is in fact essential. This<br />

conclusion is in accordance with the findings <strong>of</strong> Arnesen et al.<br />

(1979) who also analysed beams failing in shear. Considering<br />

plane stress states Arnesen et al. (1979) also demonstrated that

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

Saved successfully!

Ooh no, something went wrong!