More Efficient Fatigue Evaluation of Composite Structures by Leveraging Machine Learning and Surrogate Modelling

Fatigue evaluation of composite structures requires experimental data for each possible lay-up, making it costly. Consequently, companies choose to reduce design freedom by only using a few qualified lay-ups. Alternatively, understanding of fatigue is only built up in a late stage of development, when design changes are expensive. In this presentation, the D-STANDART project will present how leveraging new experimental techniques, machine learning, and surrogate modelling, can enable cheaper and more accurate fatigue evaluation at an early design stage.