PSI - Issue 57

Sanjay Gothivarekar et al. / Procedia Structural Integrity 57 (2024) 487–493 S. Gothivarekar et al./ Structural Integrity Procedia 00 (2023) 000 – 000

490

4

2.2. Elastic heterogeneity

During every loop of the automated Python script, a new model is created from a copy of the reference model. After Voronoi partitioning and cutting the grains are individually assigned with elastic material properties according to a normal distribution of the Young’s modulus based on 30WG1600 electrical steel [2]. A range is defined from 165 GPa to 195 GPa in steps of 3 GPa, resulting in a total of 10 different material sections with a median elasticity of 180 GPa [7]. Homogenous solid sections with plane stress thickness of 0.3 mm are automatically defined and assigned. Figure 3 (b) illustrates four different examples of uniquely generated microstructures with normally distributed elastic properties. The distribution shows that the reference stiffness of 180 GPa has the largest probability rate of 20%, as found in a large number of grains in the Voronoi part. Further, a linear elastic assumption is made as the maximum target stress should not exceed the yield strength of 440 MPa of the material.

Fig. 4. Elastic heterogeneity, color coding and normal distribution of material properties for four different models.

2.3. Fatigue simulation and estimation To develop a numerical data-set that represents a fatigue test campaign, 4 different stress levels are modelled with maximum stress ranging from 380 to 440 MPa in steps of 20 MPa and load ratio of 0.1. For every load, 5 repetitions were modelled where every repetition corresponds with a new iteration of the modelling loop, exemplified in Figure 4. This leads to a total of 20 different simulations. Fatigue loading is applied using a tensile load combined with a cyclic amplitude where the increments of the sine wave and step time are synchronized [8]. A constant amplitude sine wave with frequency of 2 Hz is defined and the step time is set to 0.25s with increments of 0.025s , resulting in 10 frames. Using the Smith-Watson-Topper (SWT) criterion for fatigue damage [9], a lifetime calculation was performed according to: ( 1 ∆ 1 2 ) = ′ ′(2 ) + + ′ 2 (2 ) 2 (1) Here, the fatigue properties for the investigated steel were fitted to the test data obtained from literature [2] and are presented in Table 1. The left-hand side determines the maximum product of the principal stress and principal strain range. Whereas, on the right-hand side of the equation, ′ and are the fatigue strength coefficient and exponent, respectively. And, ′ and represent the fatigue ductility coefficient and exponent, respectively. More details on the application and implementation of this fatigue life criterion can be found in [8, 10]. Table 1. Fatigue properties of 30WGP1600 steel [2]. ′ ′ 1218 0.225 -0.115 -0.65