AUTOMATIC DETECTION OF GEOMETRIC DEFECTS IN MACROSCOPIC FE DRAPING SIMULATION RESULTS

Publikation: KonferenzbeitragPapierBegutachtung

Abstract

A common tool assisting the development of new composite manufacturing processes is numerical simulation using finite elements. Trial-and-error experiments can be reduced by iteratively simulating different scenarios and optimizing parameters beforehand. However, the optimization still needs a lot of loops entailing manual evaluation of each iteration. Due to its repetitive nature, this task is quite time-consuming and error prone. To simplify the evaluation process and reduce manual evaluation time, we investigated the automatic evaluation of topological defects in macroscopic finite element simulation results, especially focusing on bridging and wrinkles emerging in draping simulation results. The developed routine enables an automated classification of defects, by comparing individual results to a pre-defined reference component. We used varying draping results from a parametric study to develop the routine. To determine the prediction accuracy of the routine, we compared the automatic evaluation of 23 simulations with manual evaluations. The resulting defect classifications were in good agreement. Compared to manual evaluation, the routine tends to slightly underestimate the bridging size. Wrinkles are classified quite similarly to the manual evaluation. The developed routine provides a robust method to reduce manual defect classification in draping simulation results.
OriginalspracheEnglisch (Amerika)
Seiten183-189
Seitenumfang7
PublikationsstatusVeröffentlicht - Juli 2024
Veranstaltung21th European Conference on Composite Materials (ECCM21) - Nantes, Frankreich
Dauer: 2 Juli 20245 Juli 2024

Konferenz

Konferenz21th European Conference on Composite Materials (ECCM21)
Land/GebietFrankreich
OrtNantes
Zeitraum02.07.202405.07.2024

Zitieren