Prediction of the performance of adhesively-bonded composite joints

• Main Author: Brett, Michael Alexander de Oliveira
• Other Authors: Taylor, Ambrose ; Kinloch, Anthony
• Published: Imperial College London 2012
• Subjects: 620.199
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.550244

The use of adhesively-bonded joints instead of the traditional types of joining can give reduced weight and increased stiffness in a structure. However, most industries have concerns about the use of adhesive joints in anything other than secondary structures, due to uncertainties over the long-term service life. This thesis discusses the prediction of the lifetime of adhesively-bonded composite structures. A fracture mechanics approach was used to characterise the fracture behaviour of an epoxy film adhesive, Cytec FM-300M, mainly using composite substrates prepared using wet peel ply, removing the need for any additional surface treatment. Aluminium alloy substrates were also used for some tests. Tapered double cantilever beam and double cantilever beam specimens were used to determine the mode I critical strain energy release rate, GIC, and end loaded split specimens were tested to obtain the mode II critical strain energy release rate, GIIC. Lastly, fixed ratio mixed mode specimens were used to obtain the relationship between GIC and GIIC when a joint undergoes mixed mode failure. For validation purposes, single lap joint and double scarf joint specimens were also tested. These data were then applied in finite element models using Abaqus. Two different modelling techniques were used, the virtual crack closure technique and cohesive zone modelling, CZM. Simulations of the tests performed were executed, in the process obtaining the CZM fitting parameters. Good agreement with the experimental data was verified for each of the models tested. Fatigue tests were also performed in order to obtain the mode I and mode II threshold values of the fracture energy below which crack growth did not occur, by executing double cantilever beam and end loaded split tests, respectively. For validation purposes, single lap joint fatigue tests were also performed to determine the threshold maximum load the joint could withstand without failure. Finally, using the CZM fitting parameters obtained in the quasi-static tests and the experimentally obtained threshold values of the fracture energy, modelling of single lap and double scarf joints was performed in order to predict the maximum load value for which no failure would occur when subject to cyclic loading. These predictions showed excellent agreement with the experimental results, showing that this simpler model can obtain good results.