Abstract
A common method to strengthening or stiffening a weak part of an airplane structure is to adhere a patch over the inferior surface. Typically, this is done in order to prevent a crack from initiating, or to prevent an already existing crack from growing. Evaluation of the efficiency of the patch has traditionally been done with respect to the extent of crack growth, (e.g. Park et al. 1992, and Paul and Jones, 1992), which of course is of crucial interest. However, the integrity of the patched system needs to be considered as well, since the failure of the composite system (formed by the patch and the base structure) may lead to a rapid growth of the preexisting crack in the base structure and may have overall catastrophic consequences. In this study we are therefore interested in investigating the initiation of debonding between the patch and the base structure, as well as the extent and stability of the debonding.
Early studies we conducted with respect to debonding suggested that relative long and relative compliant patches were preferred. Furthermore, an investigation regarding the effects of edge tapering on the debonding behavior showed that there are situations where a beveled edge may increase the propensity for debonding, requiring careful selection to achieve a suitable taper angle. In the present study, we investigate the integrity of the composite system for a base structure made from aluminum, and the patch made from aluminum or fiber reinforced epoxy, where both carbon and glass fiber are studied. In particular, we compare the materials selection in the patch, and for the case of a fiber-reinforced epoxy we also discuss the lay-up sequence.
To model the debonding behavior, an analytical model developed previous is extended to allow for the current materials properties. This model is fully self consistent and includes a Griffith type fracture criteria which yields the condition for the propagating bond zone boundary. The model also considers the unbonded part of the patch, which has earlier been shown to be in either of three configurations: (i) full sliding contact between the unbonded part of the patch and the base structure, (ii) only the edge of the patch remains in sliding contact with the base structure, or (iii) the patch has totally lifted of the base structure.
Results for both flat and curved structures are presented, as well as for a range of loading and boundary conditions. Among other results, it is seen that the degree of tapering is a more important parameter than the stacking sequence is with respect to the initiation and extent of debonding. Furthermore, a simplified testing method is discussed. In this method, the critical load for a case of simple boundary and loading conditions for a particular material system can be directly translated to the critical load for a more complicated structure.
