Medium Fidelity Aeroelastic Wingbox Optimisation

Medium Fidelity Aeroelastic Wingbox Optimisation

Hertgers, S.W.

De Breuker, R. (mentor)
Werter, N.P.M. (mentor)

Aerospace Engineering

Aerospace Structures and Materials

Aerospace Structures and Computational Mechanics

2015-11-26

In aeroelastic research the interaction between the structural wing and the aerodynamics is investigated. Currently the wingbox structures, as other components, are manufactured out of composite materials to replace the metallic airplanes. Composites have the ability to be designed by defining the orientation and thickness of a laminate. This property can be used in a process to minimize the wingbox weight under aeroelastic loads which is called aeroelastic tailoring or optimisation. The main advantage of a lighter wing(box) design is the lower fuel consumption of the aircraft during its lifetime. Werter and Breuker (2015) have developed the Proteus tool to perform aeroelastic tailoring on a wingbox structure. An interface is created to export the wingbox model to Nastran using ModGen (Software developed by DLR). The Nastran model is further enhanced by using the internal gradient based optimizer to tailor the wingbox structure. Goal of this thesis is twofold. First the 1D Proteus model is verified using the extension to Nastran. Secondly the influence of engine position and sweep on the optimized wingbox weight of the One Engine Reference Model (OERM) is investigated. The designed Nastran interface performs the optimisation. In the Nastran model the skin and spars are optimized for. Ribs should be present such that the cross-section does not deform too much. The ribs themselves are not optimized. No topology optimisation is performed, the dimension and locations of spars and ribs are inputs to the model. A linear analysis is performed in Nastran. Four different analysis types are defined: aeroelastic deflection, flutter, divergence and static point deflection. The tailoring will be performed under different optimisation constraints for each type of analysis. For the first type the strains should be below the allowable and the tip rotation may not exceed 12 $deg$. The second and third condition refer to flutter and divergence which may not occur for the given flight condition. The last case limits for example the engine displacement during landing. The strains are also limited for this load case. To verify the correct implementation of the model a verification procedure has been performed. For a simple rectangular beam the deformation is compared to elementary mechanics. In the next verification step a wingbox structure is considered and the results are compared to Proteus (Werter and Breuker, 2015). For similar flight conditions under a given Mach number similar results are obtained. The divergence pressure between the two models is comparable. To investigate the influence of engine location and sweep the One Engine Reference Model (OERM) is taken as a reference. The Proteus model predicts analogous aeroelastic behaviour compared to the Nastran model. The Proteus model therefore is verified. For the optimisation six load cases are defined. These are derived from the flight envelope as set by airworthiness authorities. The load cases consist out of four aeroelastic flight conditions: cruise, symmetric push down and two different symmetric pull up conditions. Next a stability load case is added as no divergence and flutter may occur. The various engine locations on the OERM are optimized with and without landing load case to see the influence of landing on the engine location. The lightest wingbox design was found for an engine located around 70\% of the span. This design violates the landing load case. Considering the landing load case the lightest wingbox design is found at 50\% of the span. Compared to the baseline model a wingbox weight reduction of 18\% is achieved. If the engine is placed further towards the tip a lot of additional material is required for the engine to not hit the runway. An engine location further down the span is preferred as the engine has a relief effect on the wing lowering the (upward) deflection and thereby lowering the critical strains. The thickness and stiffness distribution creates a tip down deformation of the wing. Next the quarter chord sweep of the OERM is varied. For a swept forward wing the wingbox weight is higher compared to the swept backward wingbox design. Additional stiffness is required to prevent wing divergence. The optimum wingbox weight is found for a sweptback wing of 50 $deg$. Comparing the design to the baseline model a weight reduction of only 4.3\% is achieved. However sweeping the wing comes at the cost of a higher angle of attack during cruise as the produced lift reduces for higher sweep angles. As with the optimized engine location wingbox the stiffness and thickness distribution is as such that a tip down deformation of the wing results during the aeroelastic load cases. Performing an aeroelastic gradient based optimisation using Nastran is possible. Great caution is advised with the Nastran optimizer as the end result is dependent on the type of optimisation performed. The model can be further extended to take ribs and buckling into account. An external optimizer could be considered to have larger control over the optimisation process.

aeroelastic tailoring
Nastran
aeroelasticity
optimisation
wingbox
sweep
engine

http://resolver.tudelft.nl/uuid:f5d74084-bbe0-4c48-9548-f8c1a4715751

Student theses

master thesis

(c) 2015 Hertgers, S.W.