Creep Behavior of Adhesive Joints

Creep tests have been carried out on single lap joints at different temperatures and loads as indicated in Fig. 33.4. The resulting shear strain y versus time relationships appear as straight lines in a double-logarithmic plot which show the same slope but are shifted by a factor. This indicates that creep data may be represented by a power law of the form of Eq. (2), where t is time and Aq, n and m denote material parameters [9]. 

Deviations from linearity in the appHed shear stress r are described by an exponent n. Fig. 33.4 also indicates that the exponent m is nearly independent of the applied stress. However, in cases where the adhesive is not fully cured, it has been observed that m depends on stress. 

If the exponent n is varied, the fit yields values of n which are shghtly, but significantly, smaller than unity (Table 33.4). Hence, it can be concluded that forcing linearity between the pre-factor and the apphed stress is a good approximation for our adhesives. 

The main advantage of using Eq. (2) is its availabihty in most commercial finite element codes. However, it does not include the reversible part of creep deformation, so a drop to zero of the applied stress due to load variations on the structure will not lead to back-deformation of the adhesive bond in a model. Viscoelastic material models containing time history must be used in such a case. 

---

The time-dependent increase in strain of viscoelastic polymers under static stress conditions is generally referred to as creep or retardation. To describe creep behavior of adhesive joints under static stress conditions using viscoelastic models, it is assumed that a specimen be subjected at a certain time f = 0 to a defined constant level of tensile stress cr = [Pg.879]

---