arXiv Analytics

Sign in

arXiv:1404.4584 [cond-mat.stat-mech]AbstractReferencesReviewsResources

Fracture strength: Stress concentration, extreme value statistics and the fate of the Weibull distribution

Zsolt Bertalan, Ashivni Shekhawat, James P. Sethna, Stefano Zapperi

Published 2014-04-17, updated 2014-09-04Version 2

The fracture strength distribution of materials is often described in terms of the Weibull law which can be derived by using extreme value statistics if elastic interactions are ignored. Here, we consider explicitly the interplay between elasticity and disorder and test the asymptotic validity of the Weibull distribution through numerical simulations of the two-dimensional random fuse model. Even when the local fracture strength follows the Weibull distribution, the global failure distribution is dictated by stress enhancement at the tip of the cracks and sometimes deviates from the Weibull law. Only in the case of a pre-existing power law distribution of crack widths do we find that the failure strength is Weibull distributed. Contrary to conventional assumptions, even in this case, the Weibull exponent can not be simply inferred from the exponent of the initial crack width distribution. Our results thus raise some concerns on the applicability of the Weibull distribution in most practical cases.

Related articles: Most relevant | Search more
arXiv:1004.1597 [cond-mat.stat-mech] (Published 2010-04-09, updated 2010-07-08)
Entropic aging and extreme value statistics
arXiv:cond-mat/0109313 (Published 2001-09-18)
Extreme Value Statistics and Traveling Fronts: An Application to Computer Science
arXiv:cond-mat/0105599 (Published 2001-05-30)
1/f Noise and Extreme Value Statistics