PSI - Issue 28

Anja Gosch et al. / Procedia Structural Integrity 28 (2020) 1184–1192 Anja Gosch/ Structural Integrity Procedia 00 (2019) 000–000

1186

3

about the crack tip based on this data. For comparison, the crack length determined via the automated methods (IRT and DIC) was correlated with the results obtained with a travelling microscope. The described crack tip detection methods were verified on three polymers with different mechanical behavior. The polymers chosen display differences in both, the base mechanical properties (from stiff to compliant material response) and also in the fracture behavior (small to large plastic deformations in front of the crack tip). These polymers were chosen to provide a good overview regarding the applicability and accuracy of automated crack length detection methods. Subsequently, not only the measured crack lengths of the different methods, but also the calculated crack growth kinetic curves were compared. Finally, a suggestion for a preferred testing set-up to measure the crack growth in polymers during fatigue testing automatically is given. 2. Experimental 2.1. Material and specimen preparation In order to verify the general applicability of the different crack length measurement methods, three different polymers with various material properties were analyzed: unplasticized Polyvinylchloride (PVC-U), Polymethylmethacrylate (PMMA) and Polyoxymethylene (POM). Compact tension (CT) specimens with a ligament length (W) of 40 mm and with all other dimensions standardized (International Standard ISO 15850, 2002b) were used for all materials. The only difference between the tested materials was the specimen thickness (PVC-U and PMMA specimens were 10 mm thick, those of POM 4 mm). Due to limitations in sample availability, not all tests could be carried out for PMMA. In fracture mechanical tests, polymers are highly sensitive to the quality of the notch tip (Salazar et al. , 2015). To assure the highest quality possible, different procedures were used to create a sharp notch tip for the materials examined. The PVC-U specimens were pre-notched by broaching using a microtome from Leica Microsystems GmbH. In the case of POM specimens, the notch was generated via pushing and the PMMA specimens were pre notched by razor blade tapping. For the crack length detection via travelling microscope, no further specimen preparation was necessary. In the case of the IRT measurements, reflections on the specimen surface had to be avoided. Therefore, the specimens were softly sanded before testing using sandpaper (grit number CAMI 800). DIC measurements require a fine speckle pattern on the specimen surface, which was achieved by spraying white and black graphite speckle onto the surface. Depending on the material, a polymer primer was also used beforehand. 2.2. Testing procedure The crack length measurements were done in fatigue tests in load-controlled mode on a servo-hydraulic testing machine of the type MTS 858 TableTop System (MTS Systems Corporation, Mineapolis, USA). All tests were conducted at room temperature with different load levels. A constant testing frequency of 10 Hz was used and the R ratio (F min /F max ) was chosen as 0.1. The testing program included force controlled stops to determine the crack length via the travelling microscope and to take a picture via the DIC or IR camera. Overall, three testing set-ups were used: (a) two travelling microscopes, (b) a travelling microscope and the IR camera and (c) a travelling microscope and the DIC camera. To compare two different methods, it is recommended to install the testing equipment on the same specimen side. In the case of the aforementioned testing set-ups (DIC and IR camera) this was not possible. Thus set up (a) was included to study differences within each sample (Figure 1). The thermographic analysis was executed using the IR camera ImageIR 8300 and the software IRBIS 3 (InfraTec GmbH, Dresden, Germany). This set-up is shown in Figure 1 as “set-up 2”. The IR pictures were made during testing immediately before the test procedure was stopped for the optical crack length measurement to avoid loss of information due to specimen cooling. For the DIC measurements, a Prosillica GT 6600 camera (26.6 megapixels), a Tokina ATX-PRO D lens and a computer equipped with a Mercury RT software (Sobriety s.r.o., Kurim, Czech Republic) were used (shown in Figure 1 as “set-up 3”).