Experimental determination of fracture toughness for delamination analysis in structural components

Abstract

Delamination is an important damage mechanism in composite materials. A correct prediction of its behaviour is of great importance in the design of aerospace composite structures. Nevertheless, when using simple numerical models which do not account for the delamination migrating to other interfaces or other intralaminar damage mechanisms, the accuracy of the numerical analyses is restricted by the inadequacy of the standardised experimental determination of the fracture toughness. This interlaminar property is usually measured using unidirectional specimens, which ignores the effect of the lay-up. Furthermore, the changing fracture toughness as the crack propagates, the so-called R-curve effect, is generally ignored as well by only employing the initiation values from the experiments. This leads to highly conservative results as the resistance to delamination is not correctly accounted for. The aim of this MSc Thesis was to contribute to the improvement of the blind simulation of delamination in co-cured composite aerospace structures. In particular, the experimental determination of the fracture toughness in coupons without a 0/0 interface to then be used as parameters in the simulation of one structural component: a single-stringer compression (SSC) specimen.Based on the lay-up formed by the skin and the stringer of the SSC specimen, various layups were defined for DCB and MMB testing. After evaluating them numerically, the fracture toughness was measured experimentally in unidirectional [012//012] and multidirectional [011, 45//-5, 011] coupons, in DCB and MMB for mode mixtures 20% and 50%. The R-curve effect of the multidirectional specimens was far greater than for the unidirectional ones, the steady-state fracture toughness was around three times higher, due to the crack migrating to other interfaces or propagating simultaneously in between different plies.The experimental results were used to improve the numerical analyses of the tests performed. A new approach was presented to model the measured R-curve effect by varying the fracture toughness values according to the distance to the crack tip. The improvements achieved were especially relevant for the multidirectional specimens. This approach and the R-curve from the multidirectional specimen were employed in the analysis of the SSC specimen. The results obtained emphasise the importance of accounting for the effect of the lay-up and the crack growth length on the fracture toughness. Furthermore, it was shown the potential and shortcomings of using an experimentally determined material law, accounting for delamination migration and other damage mechanisms, to improve delamination modelling.