Automatic Quality Assessment of FE Draping Simulation Results

Developing and optimizing manufacturing processes for composite components
is commonly supported by finite element (FE) simulations. Initial concepts are
modeled and parametric studies are conducted to determine optimum process
parameters. Built-in application programming interfaces (APIs) typically allow
for a script-based automation of systematic model modifications for many FE
solvers. However, the evaluation of simulation results typically depends on a manual
inspection by experts, despite its repetitive nature. This study aims to use
APIs to develop a fully-automated method for identifying and evaluating critical
features and thus, the quality of draping simulations. The focus thereby lies on
providing a validated framework for the automated evaluation of draping simulation
results, rather than claiming perfect virtual representation of experimental
draping. Three different metrics (deviations of fiber angles, the boundary contour
and topological defects) are used to determine the overall draping quality of simulation
results. With the developed routine, all quality metrics can be estimated
quite well for simulation results. The routine is designed to be extended for the
use with experimental data for a reliable real-life quality assessment.