Clerc, Gaspard (2020). Analysis of the fatigue performance of adhesively bonded wood joints (Dissertation, Technische Universität München, Fakultät für Bau Geo und Umwelt)
![[img]](https://arbor.bfh.ch/style/images/fileicons/text.png) |
Text
DISSERTATIONsmall.pdf Restricted to registered users only Download (9MB) | Request a copy |
Wood, as natural composite material, is generally considered to have a good resistance against fatigue loading due to its fibrous structure and viscoelastic properties. Nowa- days, timber structures are generally an assemblage of several wood pieces adhesively bonded and the influence of the adhesive on the fatigue performance remains largely unknown. The influence of the adhesive properties (stiffness, presence of fibers, type of adhesive system) on the fatigue performance was investigated in this thesis. The first hypothesis was to investigate if ductile adhesives are able to dissipate a higher amount of energy per loading cycle compared to brittle adhesives. This amount of dissipated energy would not participate to damage accumulation in the bond line. This hypothesis was investigated in paper I on pure adhesive film samples and verified on adhesively bonded wood lap-shear samples in paper II. The results showed that samples bonded with ductile adhesives are able to sustain a higher number of cycles than samples glued with brittle adhesive systems. This is probably due to a more homogeneous stress distribution for the ductile adhesively bonded samples. In paper II, a new fatigue model has been developed to analyze experimental fatigue results. This model is a combination of a physical and a statistical model which allows to describe the behavior at low and high relative strengths, i.e., for the complete lifetime of the specimen. For lap-shear samples, the strength degradation with increasing number of loading cycles takes place through an accumulation of micro-damages. The loading situation, however, is different if a macro crack is present in the sample. Indeed, the modulus and strength degradation during the fatigue loading will be the result of the propagation of cracks. The influence of the adhesive properties on the crack growth was investigated in paper III and paper IV. In both papers, the sample chosen is a 4-point End-Notched Flexure specimen (4-ENF), where the crack propagation occurs in Mode II (shearing). In paper III, it was shown that the crack propagation of adhesively bonded wood joints can be described using the Paris equation. In paper IV, the domain of application of the Paris equation was expanded using the modified Hartman-Schijve equation which, for the first time, was applied to successfully describe the fatigue fracture of adhesively bonded wood joints. It was shown in paper III and paper IV that the brittle adhesive systems have generally a slower speed of crack propagation at similar applied stress level compared to ductile adhesives. The addition of fibers to the adhesive was also shown to increase the performance of the ductile adhesive. The reason for the better performance of the brittle adhesives are investigated in paper V. A new fractography technique combined with an unsupervised pattern recognition of Acoustic Emission sig- nals with source location was developed. With this, it was shown that in samples glued with the brittle adhesive system, the crack does not propagate in the adhesive layer but at the interface with the wood or directly in the wood. Crack propagation hence results in a slower crack speed at higher energy release rate due to the higher wood/adhesive adhesion. Hence, it appears that the development of a high-performance adhesive for fatigue loading is a complex topic as the choice of adhesive properties depend on the presence of cracks and loading situation. Further developments should investigate the modification of the surface properties of the bond line to increase the adhesion of the wood with the adhesive, allowing to combine the advantage of a ductile adhesive with a crack propagation at the wood interface.
Actions (login required)
 |
View Item |