PSI - Issue 39

Hannes Panwitt et al. / Procedia Structural Integrity 39 (2022) 20–33 Author name / Structural Integrity Procedia 00 (2019) 000–000

32

13

In contrast, the threshold value of the crack width method does neither depend so much on the strain threshold value and crack path detection, nor on the loading type. With the same constant crack width threshold value, calibrated on a SIF-controlled mode I crack, a good agreement with the reference a - N -curves was obtained for all investigated cracks. Therefore, the comparison of the results of both methods can improve the accuracy of the ε1 -method and at the same time overcome the somewhat noisy results of the crack width method. It can be assumed, that the observed relatively noisy a - N -data are a result of the placement of the virtual extensometers (limited by the DIC measurement point grid). However, further investigations regarding this have to be conducted. As a result of the strengths and weaknesses of the presented methods, the following procedure for the evaluation of mixed mode fatigue cracks is recommended: • Calibrate both threshold values ( if necessary non constant for the ε1 -method) on mode I test. If only one test-type • To obtain less noisy data, evaluate with the ε1 -method as well. If the general behavior of the crack growth rate and therefore the correct principal strain threshold value (constant vs. non-constant) is unknown, re-calibrate th ε1 on the curve obtained with the crack width method. The curve obtained by the ε1 -method should be similar, but less noisy compared to the crack width method. • Smooth the crack path, if long crack lengths are investigated. 6. Conclusion Digital image correlation has become an important tool in experimental crack path detection. Unfortunately, its application is strongly dependent on various influencing variables, such as the quality of the speckle pattern or the image noise. Furthermore, the DIC evaluation is often associated with a high manual effort. This is especially true under mixed mode conditions, when a crack does not propagate flatly as under pure mode I, but shows a variable course. A good basis for optimizing the evaluation process is provided by Gehri et al., who developed a tool for automated crack measurements based on DIC data. In the present work, a further development of this tool is presented in order to determine detailed a vs N data for the specimens surface under each type of crack loading. Therefore, two different methods have been introduced as well as calibrated on mode I tests and validated on mixed mode tests. In the first method, the a - N -curves are calculated on the basis of the crack opening, which was already included in the original script. The second method evaluates the crack propagation via the principal strain field. Whereas the crack width method works similarly well with one constant threshold value and is consequently less prone to calibration errors, the properly calibrated ε1 -method results in very detailed and noise free data. With both procedures accurate and more detailed data compared to the marker load technique can be obtained, which shows the potential of the presented methods. In addition to the a - N -curves, the extended ACDM tool also supplies kinking angle data and, furthermore, the possibility to evaluate crack branching. Overall, it can be stated that the enhanced ACDM tool provides a standardizable and comparable determination of the crack path development for both simple, plane mode I cracks and variable, branched mixed-mode fatigue cracks with little manual effort. Additional tests have shown that it is applicable to out-of-phase mixed loading despite the complex processes inside the specimen. In this context, however, further experimental investigations are necessary. Acknowledgements The authors gratefully acknowledge the financial support of this research project SA 960/9-2 provided by the German Research Foundation (DFG) in Bonn, Germany and Gehri et al. for providing their MATLAB script. References Breitbarth E, Strohmann D, Starostin-Penner D, 2021. Experimentelle Ermittlung komplexer Rissfortschrittsdaten mittels smarter Algorithmen und digitaler Bildkorrelation. DVM-Bericht 253, 2021, 97–106. Correlated Solutions, Inc., 2016. VIC-2D v6, Irmo, South Carolina. can be conducted, a test with constant F max is recommended. • Evaluate the mixed mode tests with the crack width method.