Fig. 11: Patch detachment 1 and 2 , crack initiation and final failure
Again, the images show the region around the hole which is covered by the CFRP patch and whose woven structure can be identified. For each event three consecutive images are shown that are found to be taken close to the characteristic events.
The two points of patch detachment ( and ) show a slight diagonal increase in temperature around the central hole. In the first row two diagonal lines can be seen and in the second one only one line. It seems that these temperature increase can be related to metallic slip bands. But it actually cannot clearly be said whether these lines of increased temperature result from diagonal slip bands or from the diagonal structure of the patch.
Considering crack initiation () and final failure (), an increase in temperature can be seen through the patch in the specimen centre. As the patch is already detached around the hole, outlines are blurred and only a non-specific field of increased temperature can be made out. The temperature difference here lies at around \(T=45\) at the patch surface, compared to a \(T=7.1\)K for the temperature increase measured at the metallic surface. The very last picture covers the moment of patch detachment. The actual temperature increase around the hole can be seen at the right-hand side where the metal sheet is not covered by the patch. While the temperature difference is more than\(T=8\)K at the edge of the metallic part, on the patch surface only difference of around \(T=2.59\)K is detectable.
3.2. Fatigue Loading
Compared to the quasi-static tests, under cyclic loading different conditions apply. Fatigue crack growth differs from static failure. The main difference is that cracks grow at lower stresses and, more important here, lower strains. This means that also temperature evaluation is less than for static damage. Further, due to the continuous loading and unloading, temperature does not solely increase because of plastic deformation at the stress peaks, but also globally due to internal friction. Cyclic loading leads to molecular movements inside the material. Thereby, the entire specimen heats up over testing time. Therefore, temperature differences \(T\ \)are expected to be small. On the other hand, due to smaller deformations, sudden patch detachment is not expectable. A proper bonding leads to improved visibility of the subsurface phenomenae.
Thermal images of one exemplarily chosen unpatched specimen, shown in Fig. 12, depict the temperature development during crack propagation for four exemplarily chosen numbers of load cycles N .