Assessment of adhesive bonded repairs

Airworthiness certification of bonded repairs to aircraft primary structures remains unresolved due to the lack of non-destructive inspection techniques to detect weak bonds or a potential deterioration in bond strength caused by environmental degradation from moisture ingress. An alternative approach to the direct detection of weak bonds is to develop a data-driven approach for certification of bonded repairs. Compliance to certification standards can be demonstrated if the probability of a bonded repair failing due to weak bond strength is found to be sufficiently low such that the reliability is not affected. The Adhesive Bonded Repair Assessment Program (ABRAP) was initiated to quantify the environmental durability of adhesive bonded repairs on retired aircraft parts using the Pneumatic Adhesion Tensile Testing Instrument (PATTI). The aims of this research are to (i) interpret the shear strength of the bond from the measured flatwise tensile strength, (ii) to establish the causes for the variation in strength measurements and failure modes observed from the bonds tested and (iii) develop a predictive model to account for the effects of adherend thickness and taper on the pull-strength. Adhesive bonded scarf and single lap-shear joints were used to verify the yielding behaviour of the FM300 adhesive resulting from pure cohesive, mixed cohesive-interfacial and purely interfacial failure modes. Three types of surface preparation techniques and two types of environmental conditioning were used to produce these failure modes. Finite element analyses of the scarf and lap-shear joints were conducted to determine the ratio of shear and tensile stress components affecting their failure. These experiments allowed the effects of the failure mode on the measured pull-off strength to be investigated. Pull-tests were performed on a bonded panel whereby the edges of one of the adherend were milled to form a 3° taper. The purpose of this experiment is to generate strength measurements for pull-tests performed on tapering and constant thickness sections. Numerical models of the pull-test configuration were constructed to evaluate the causes for the variation in strength and fracture mode from the loading configuration aspects. The adhesive layers of these models were assigned with two types of material models. The use of the cap-plasticity material with the maximum principal strain failure criterion at a characteristic distance gives accurate prediction for the strength of the joints and PATTI tests. When using cohesive elements, the best correlation with the experimental results is obtained through the use of a non-softening cohesive law. The plastic yielding behaviour of the FM300 adhesive shows a strong sensitivity to hydrostatic pressure and can be described by the linear Drücker-Prager yield criterion. The shear strength of the joints inferred from the y-intercept of the yield curves was found to be less susceptible to the effects of surface treatment and environmental degradation than the tensile strength. This means that low pull-off strength does not necessarily indicate low shear strength. The application of small off-normal load to the pull-stub and the deflection of thin underlying substrate over which the PATTI pull-tests are performed reduced the tensile