Papers by Author: A.D. Crocombe

Abstract: Triaxiality function (Rv) has been known as one of the important factors that responsible for damage initiation in adhesive bonding. Damage evolution law for low cycle fatigue (LCF) is function of Rv, von Mises equivalent stress (Seqv) and number of cycles (N). From previous research, it was found that the Rv values of two cases: bulk adhesives and single lap joint (SLJ), were close to unity. Those values are uncontrollable. Meanwhile, the damage equation for general solution contains Rv as an independent variable. There is need to choose another joint type that can characterise Rv as an independent variable. This paper presents the choice of scarf joint as specimen that can simulate variation of Rv. Several types of adhesive joints have been modelled and analysed using ANSYS as finite element analysis (FEA) tool. In ANSYS, Rv values were calculated directly from direct output results: von Misses equivalent stress and Hydrostatic stress. From FEA, it was shown that Rv changed as a function of adhesive bondline angle of the scarf joint. The values of Rv are constant along adhesive line except at the free edges. This choice is better than Cleavage joint where the values of Rv are not constant along adhesive line due to the presence of bending moment.

Abstract: Adhesive in joints will have complex stress state rather than bulk adhesives. This will lead to the assumption behind bulk adhesive that triaxiality function (Rv) is equal to one (uni-axial stress state) is not valid anymore. In this paper, new procedure to find damage parameters α and β for single-lap joints has been developed based on global damage of adhesive joints. With this procedure, damage parameters α and β have been found. Validating the procedure by calculating the number of cycles to failure (Nf) has been performed successfully. The accuracy of the damage evolution equation is less than 2 %.

Abstract: This paper describes a dynamic test carried out on intact and damaged FRP composite beams with fixed-fixed boundary condition. Hammer excitation is used to excite the beam at fixed locations. The modal parameters are extracted from the time response using a time domain analysis, i.e. the stochastic subspace identification technique. In order to introduce damage, two sections of the beam are bonded together using an epoxy adhesive, and then a static test is carried out. For the static test, a 3-point bending -configuration is used, i.e. the beam is fixed at both ends and a static load is gradually applied in the middle of the beam using a screw jack. Different static load steps, and in turn different damage stages, are considered. After each load step, dynamic measurements are carried out. The results obtained from both tests are presented and analysed.

Abstract: In this paper, the characterisation of damage in an epoxy adhesive has been investigated. Bulk adhesive samples were used in this study for two reasons; firstly the stress distribution in the bulk adhesive sample is simpler than that in a joint, secondly, the specimen’s dimensions meet fatigue test specimen standards. Low cycle fatigue (LCF) tests with a load ratio of 0.1 and a frequency of 5 Hz were performed on bulk adhesive dumbbell specimens. Damage curves, relating damage in the specimen to number of cycles to failure, were plotted using an isotropic damage equation in which damage is a function of stress, which decreases as damage progresses. The damage curves were then fitted using a LCF damage evolution law. This equation was derived from a dissipation potential function using Continuum Damage Mechanics (CDM) theory. Curve fitting was performed using a Robust Least Square technique rather than ordinary linear least square because the damage curve has extreme points (usually at the breaking point). Two damage parameters α and β were found from the curve fitting process. This process resulted in different values of α and β for different stress levels. The logarithmic α and β points were then plotted respect against stress level and linear regression was used to determine α and β as a function of stress. With this function, damage parameters for other stress level can be predicted.