PSI - Issue 13

Christopher Schmandt et al. / Procedia Structural Integrity 13 (2018) 799–805 C. Schmandt, S. Marzi / Structural Integrity Procedia 230 (2018) ECF22

801

3

Fig. 2. (a) DCB specimen with measuring flags in clamping devices with rotary encoders; (b) Test setup with line sensor and camera system.

In a first step, monotonic loading was supplied by controlling the testing machine on various constant crack opening velocities v . Tests were driven until crack propagation took place, because external controlling on v is only senseful if the line sensor is focused on current crack tip. The J -Integral according to Rice (1968) for mode I loading was calculated directly from measured data using an approach as proposed by Anthony and Paris (1988), � � �� � � (1) with b as specimen’s width. All tractions except those in mode I direction were assumed to be zero, which is not entirely correct regarding distinctive necking in the adhesive layer during tests. The mode I fracture energy J c was chosen as maximum of measured J . Applying the crack model proposed by Barenblatt (1962) and Dugdale (1960), Rice (1968) showed that the cohesive stress �� � � � � � (2) at the location of crack tip can be obtained by deriving J with respect to u . The cohesive strength σ 0 was calculated by averaging over the resulting plateau within σ - u -curves. The initial slope of the so delivered traction separation-laws yielded the cohesive stiffness �� � � � �� � � . (3) The recorded data was low-pass-filtered prior to derivation to reduce effects of noise in the resulting derivatives. In a second step, monotonic loading was supplied by controlling the testing machine on various constant cross head velocities v ch to investigate the dependency of fracture behavior on loading rate. This was done by comparing pictures recorded from specimen’s edge with corresponding J -time-curves while crack was propagating. 3. Results and Discussion Experimental results are presented in this section regarding the determination of characteristic material parameters for cohesive law and the rate dependency of fracture process. Phenomenological dependencies of fracture energy and cohesive strength on crack opening velocity are described by analytical approaches. At first, it must be stated that the investigated adhesive is very sensitive to deviations from the prescribed manufacturing process. Especially the primer layer should be applied with care as thin as possible within one step only. Too thick primer layers or several times applied primer layers can cause huge adhesively failed regions. For that case, crack propagation takes place within primer layer instead of adhesive layer. Furthermore, there should be a curing time of approximately two hours after primer application, which must be used to warm up the substrates