Publikationen

DGLR-Publikationsdatenbank - Detailansicht

Titel:

Selection of Critical Load Cases using an Artificial Neural Network Approach for Reserve Factor Estimation

Autor(en): R. Nazzeri, M. Haupt, F. Lange, C. Sebastien
Zusammenfassung: During aircraft preliminary design the selection of load conditions can affect the results of the structural sizing and therefore affect the component weights. After the computation of ground, maneuver and gust cases based on a 1-dimensional condensed aircraft model the most critical load conditions are selected using external load distribution. These external loads are computed at discrete load monitoring stations thanks to the response of the aircraft model. The external loads are then distributed to the loadable grids of a global finite element model for stress analysis, structural sizing and weight estimation. In order to reduce the amount of load cases for analysis the most relevant load cases are selected based on external load criteria. The choice of these criteria is then essential for the design of a safe structure. The present study proposes a selection of load conditions which is based on reserve factor values which are available at the sizing process. The iterative sizing process includes an update of the structural properties for the strength analysis during each loop. In the scope of this study one set of structural parameters together with existing reserve factor values are taken from one sizing loop. Thus the strength analysis is replaced by a surrogate model which is constructed using an artificial neural networks approach. The construction and accuracy of the surrogate model is discussed and a comparison of both load case selection approaches is performed. The study is performed for the wing covers of a backward swept metallic wing. The results show that a further reduction of the selected load cases by SMT values is possible when taking the reserve factor values into account. The constructed surrogate model is a first step for future surrogate based optimization by which the sizing and weight estimation of the wing covers can be performed with cheap function evaluations.
Veranstaltung: Deutscher Luft- und Raumfahrtkongress 2015, Rostock
Verlag, Ort: Deutsche Gesellschaft für Luft- und Raumfahrt - Lilienthal-Oberth e.V., Bonn, 2015
Medientyp: Conference Paper
Sprache: englisch
Format: 21,0 x 29,7 cm, 10 Seiten
URN: urn:nbn:de:101:1-201510306374
Stichworte zum Inhalt: load case selection, surrogate model
Verfügbarkeit: Download - Bitte beachten Sie die Nutzungsbedingungen dieses Dokuments: Copyright protected
Kommentar:
Veröffentlicht am: 30.10.2015