Non-Gaussian statistics and extreme events on experimental facture surfaces
Experimental studies show that fracture surfaces exhibit rather remarkable scaling properties characterized by universal roughness exponents,close to ζ = 0.4 in brittle materials and close to ζ = 0.8 in quasi-brittle and ductile materials.In this work,we go beyond the value of the roughness exponent,and focus on the distribution of height fluctuations on the fracture surface of a large range of materials,from brittle to ductile and quasi-brittle solids.At first,we show how damage accompanying crack propagation results on average into deviations to the Gaussian statistics observed on brittle fracture surfaces.Then,we identify on the fracture surface the location of the largest jumps responsible for the fat tails observed on these distributions,and show that these extreme events are actually organized in a network of clusters made of connected events.The statistical analysis of these clusters show many interesting features,including a characteristic sizes reminiscent of the typical size of the damage processes in the material studied,a power law distribution of cluster size for ductile and quasi-brittle fracture surface,while their probability distribution decay exponentially in brittle fracture surface.This new approach in the analysis of the morphology of fracture surface is a first step into a better understanding of the damage processes occurring within the process zone during crack propagation,and open promising perspectives into the description of damage mechanisms in a large range of materials by an unified theory.
Fracture surface roughness statistics scaling behavior
Yuanyuan Cao Stéphane Vernède Laurent Ponson
Institut Jean Le Rond dAlembert(UMR 7190),CNRS-UPMC Université Paris 06,75005 Paris,France EO Technology,Fengrun Building 606,280 Heping North Road,213000 Changzhou,China.;Baseline Consulting
国际会议
北京
英文
1-11
2013-06-16（万方平台首次上网日期，不代表论文的发表时间）