These slides were presented at the NAFEMS World Congress 2025, held in Salzburg, Austria from May 19–22, 2025.

Abstract

Ever-increasing legal requirements and the pursuit of more energy-efficient washing machines make lightweight construction indispensable in this sector. A current challenge arises from the requirements for environmental labeling, which were tightened by the EU in March 2021. To receive the best energy label in the future, washing machines must be designed for higher spin speeds. The higher mechanical requirements associated with this can no longer be met with short-fiber-reinforced thermoplastic (SFT) components. Long-fiber-reinforced thermoplastics (LFT) are a promising alternative with better resistance to fatigue damage. We will present a multiscale simulation approach for the fatigue design of components made of fiber-reinforced plastics. This approach only requires the following measurement effort on tensile bars taken from plates for calibration: 1. CT image to determine the fiber orientation (across the thickness of the plate). 2. Incineration of the fiber-reinforced plastic to determine the fiber length distribution. 3. Determination of S-N curves in which the drop in dynamic stiffness was also recorded during the measurement. Using fiber orientation and fiber length distribution, a virtual multilayer microstructure model is constructed that allows us to use elastic FFT-based full-field simulations to inversely calibrate the Young’s modulus of the plastic at the beginning of the fatigue measurement. Subsequently, we perform FFT-based fatigue simulations on virtual long-fiber-reinforced volume elements with different fiber orientations and (mean) fiber lengths. Using model order reduction methods, we obtain an effective material map for the fatigue behavior of the fiber-reinforced plastic that considers the local fiber structure. The recorded decreasing dynamic stiffness during the measurement of the S-N curves allows us to calibrate the fatigue speed of this material map. After additional calibration of a failure criterion at the S-N curves for notched and unnotched tensile bars, the time and location of failure of a component made of this fiber-reinforced plastic can also be efficiently predicted in three steps: 1. Run an injection molding simulation, e.g. with Moldflow. 2. Transfer the fiber orientations and (mean) fiber lengths to the FE mesh. 3. Run the fatigue simulation using the effective material card as UMAT in Abaqus. We will carry out this procedure for the leach container as an example and compare the predictions with component measurements. In addition, we will present the computational effort and the scalability of the individual steps.