To improve efficiency, manufacture of the RDCB test piece is performed in a single lay-up procedure using a multi-specimen moulding plaque, shown in Figure 7. A four-layer ‘sandwich’ containing metal plates 175 mm x 110 mm is constructed, with adhesive applied between each layer. The starter crack length, a0 , and the central (test) bond thickness are controlled using a PTFE mask. After curing, the four individual RDCB test pieces are carefully machined from the moulded plaque (cut lines are indicated dotted in Figure 7).
Figure 7: Schematic showing 175 mm x 110 mm plaque area,
The definition of an adhesive system relates to a known set of parameters for the RDCB test piece. Changing any parameter results in the creation of a different adhesive system. Parameters include:
The loading conditions, including gripping of the test piece, determine the mode of loading imposed at the crack tip. Three different loading conditions can be applied to the basic RDCB test configuration (as illustrated in Figure 8) so as to impose:
Figure 8: Three modes of loading imposed on the universal test piece (arrows indicate direction of displacement, hatched area represent built-in supports)
The test piece is cycled under displacement control to generate crack propagation within the central adhesive layer. Typically, testing is started at high strain energy release rates, and as the crack grows the G value reduces progressively. This allows a single test to generate the entire plot of G versus da/dN .
Multi-specimen testing is required using, ideally, a multi-station servo-hydraulic test machine. An example of such a machine is shown in Figure 9. Control and measurement is performed using a single dedicated software program, enabling constant monitoring of the test condition and crack propagation.
Figure 9: Six-station MERL servo-hydraulic fatigue test machine
The aim of this test method is to determine the bond durability of a specific adhesive system. The deliverable from the test is a plot showing the relationship between strain energy release rate G and crack growth rate da/dN . A typical plot is shown schematically in Figure 10.
Figure 10: Schematic showing typical da/dN versus G curve
From this plot the Paris Law constants (A and B ) can be determined over the linear part of the G versus da/dN relationship, when plotted on a log-log scale. This relationship forms the fatigue model required for FEA life predictions .
The threshold value of strain energy release rate Gth is where possible, measured and used for FE analysis as a limiting design criteria .
Due to thresholds requiring long test times, it is often appropriate to calculate thresholds values based on projections of the Paris Law fit. Upper and lower projected thresholds (G|10-6 and G|10-7 ) can be calculated by extrapolating the fit line to G values corresponding to crack growth rates of 10-6 and 10-7 mm per cycle respectively. The use of G|10-7 would be recommended as a conservative value for design purposes.
Next: Test Methods
