R. Zahn, V. Völkl, M. Zieher, C. Breitsamter

This paper presents a reduced-order modeling (ROM) approach based on a hybrid neural network in order to calculate wing buffet pressure distributions due to structural eigenmode-based deformations. For this hybrid ROM a convolutional autoencoder (CNN-AE) and a long short-term memory (LSTM) neural network are connnected in a serial fashion. The NASA Common Research Model (CRM) with the FERMAT structural model is used for forced-motion computational fluid dynamics (CFD) simulations at transonic buffet conditions. Aerodynamic responses are obtained as a result of the eigenmode-based deformations. As eigen shape the first symmetric wing bendig mode is selected. The unsteady simulations are carried out with the triangular adaptive upwind (TAU) solver of the German Aerospace Center (DLR) and the hybrid ROM is trained with this data. When investigating the prediction capability of the hybrid ROM a high accuracy with respect to the forced-motion buffet loads is indicated.

Deutscher Luft- und Raumfahrtkongress 2023, Stuttgart

Deutsche Gesellschaft für Luft- und Raumfahrt - Lilienthal-Oberth e.V., Bonn, 2023

Conference Paper

englisch

21,0 x 29,7 cm, 10 Seiten

urn:nbn:de:101:1-2023111011540577424731

10.25967/610103

Deep Learning, Convolutional Autoencoder (CNN-AE), Long Short-Term Memory Neural Network (LSTM), hybrid Reduced Order Model (ROM), Wing Buffet Aerodynamics, Forced Structural Vibrations, NASA Common Research Model (CRM)

Download - Bitte beachten Sie die Nutzungsbedingungen dieses Dokuments: Copyright protected

Zahn, R.; Völkl, V.; et al. (2023): Transonic Wing Buffet Load Prediction at Structural Vibration Conditions. Deutsche Gesellschaft für Luft- und Raumfahrt - Lilienthal-Oberth e.V.. (Text). https://doi.org/10.25967/610103. urn:nbn:de:101:1-2023111011540577424731.

10.11.2023