C. Calì

Three-dimensional crack growth: Numerical evaluations and experimental tests

Abstract Experimental observations of three- and two-dimensional fatigue crack growth are compared to numerical predictions from the computer code BEASY. The two dimensional propagation occur in a Multiple Site Damage (MSD) scenario created on a pre-notched specimen, undergoing a traction fatigue load as defined by a general load spectrum. Experimental analyses on a fatigue machine were carried out in order to validate the numerical simulation and to provide the necessary material fatigue data for the aluminium plates. The numerical code adopted (BEASY) is based on Dual Boundary Element Method (DBEM). General modelling capabilities are allowed by this approach, with the allowance for general crack front shape and a fully automatic propagation process. By means of a non-linear regression analysis, applied on in house obtained experimental data, the material parameters for the NASGRO 2.0 crack propagation law were defined, capable to effectively keep into account the threshold effect and the unstable final propagation (the crack closure option was switched off). A satisfactory agreement between numerical and experimental crack growth rates was obtained, even starting from a complex MSD scenario, created by the presence of three holes in the plate. Moreover the load introduction to the specimen was monitored by strain gauge equipment. The numerical simulation include also the through the thickness propagation, corresponding to 3D part-through cracks; in this case some specimen were pre-notched by a corner crack on one of the holes and the 3D experimental crack propagation monitored.

DBEM Crack Growth Simulation for a Riveted Aeronautic Reinforcement under Non-linear Contact Conditions

A special specimen was created cutting a rectangular notched area from the surrounding of the upper left corner of a wide body aircraft door. Then a fatigue traction load was applied in order to induce an MSD crack initiation and propagation. An innovative DBEM (Dual Boundary Element Method) modelling approach was devised, capable of explicitly modelling the different test article layers with their rivet connections even in a 2d approach. The rivets that are close to the propagating crack are coupled with the corresponding holes by non linear contact conditions, and the accuracy improvements are assessed in comparison with a previous linear analysis, in which traction and displacements continuity conditions on the hole-rivet interface had been imposed. The importance of such influence on the simulation precision need to be assessed due to the strong impact that a non linear analysis produces on computational times. For such a complex problem (three different panels, made of different materials, each one with a variable thickness and connected by numerous rivets), experimental crack propagation data were available for the numerical-experimental comparison. With such non linear approach, a significant improvement on the growth rate correlation is obtained, that justify the increased computational effort.