PSI - Issue 38

Alexander Erbe et al. / Procedia Structural Integrity 38 (2022) 192–201 Author name / Structural Integrity Procedia 00 (2021) 000 – 000

195

4

2.2. Experimental campaign and load determination procedure In order to systematically determine the influence of different loading conditions in a casted casing, respective principal axis-ratios were identified for testing. Strain-controlled experiments are used to assure the comparability to already available uniaxial LCF-data performed under strain-controlled conditions as well. Furthermore, those experiments address secondary stresses on the component due to e.g. thermal transients during start-up and shut-down processes. In contrast to this, primary stresses occurring during operation in gas-carrying components (a corresponding internal pressure) are further investigated by load-controlled experiments, which are not part of this publication. The experiments were designed with finite element simulations within the software Abaqus TM where an eighth part 3D-model of the cruciform geometry was modeled. As the induction heating leads to a local maximum temperature in the center of the specimen only, the real temperature distribution of the specimen has been captured with a thermal imaging camera. For pre-designing each test an idealized temperature distribution over the specimen has been applied to the simulation assuming a stationary heat flux from the center of the specimen towards the cooled fixtures. The strain controlled experiments were simulated with an appropriate subroutine, controlling the forces A and B iteratively until a defined extensometer-strain in direction A and B is reached. For the sake of simplicity, the variable is introduced defining the ratio of the axis strains controlled through the extensometer right within the center as: Temperature dependent cyclic flow curves at midlife were used to define the elastic-plastic material behavior in a simple manner where the aim of the design was to achieve a defined equivalent von Mises strain. The whole design workflow is briefly outlined in Figure 2, left. As the maximum occurring equivalent von Mises strain varies in position and magnitude depending on the axis-ratio and load amplitude applied, a design point (DP) has been defined as the foremost node right in the center of the FE-model or respectively in the center of the real sample (Figure 2, right). All experiments were designed to meet an equivalent strain range at the design point addressing different loading levels Δε eq,DP . With this, axis ratios of 1.0, 0.0, -0.5, - 1.0 are covered at two isothermal temperature levels of 400 °C and 500 °C. Following the experiments, the maximum equivalent von Mises strains and their positions (HS = Hot-Spot) were re-evaluated and compared to the positions of crack-initiation. The experiments thus permit to verify the suitability of the von Mises strain and its possibilities of describing the lifetime and position of crack-initiation of experiments with a series of systematically different axis ratios. B A    = (2)

Figure 3: left side: design and assess scheme; right side: strain contour-plot with HS- and DP-position, extensometer position and loading axis