Master of Science Thesis Optimization strategy for conceptual airplane design P.T. Vasseur B.Sc. May 9, 2014 Faculty of Aerospace Engineering · Delft University of Technology Optimization strategy for conceptual airplane design Master of Science Thesis For obtaining the degree of Master of Science in Aerospace Engineering at Delft University of Technology P.T. Vasseur B.Sc. May 9, 2014 Faculty of Aerospace Engineering · Delft University of Technology Copyright ' P.T. Vasseur B.Sc. All rights reserved. Delft University Of Technology Department Of Flight Performance and Propulsion The undersigned hereby certify that they have read and recommend to the Faculty of Aerospace Engineering for acceptance a thesis entitled “Optimization strategy for conceptual airplane design” by P.T. Vasseur B.Sc. in partial fulfillment of the requirements for the degree of Master of Science. Dated: May 9, 2014 Head of department: prof.dr.ir. L.L.M. Veldhuis Supervisor: dr.ir. R. Vos Reader: dr.ir. H.G. Visser Abstract Due to the ever growing demand for more efficient aircraft novel aircraft concepts have to be explored. By improving design tools the potential of unconventional configurations can be further studied. This requires improvement of conceptual design tools such that more knowledge can be gathered on alternative solutions as early in the design process as possible. Multidisciplinary design optimization (MDO) can support this process by providing an environment in which the various disciplines can be designed and optimized concurrently, while a certain level of consistency is maintained. An optimization design tool has been created to assess the potential performance gains of novel aircraft configurations. It connects with the Initiator design tool, which is a conceptual design framework. As such, it can also be used as a means to expose any analysis or design issues that may exist in the Initiator. With the optimizer tool the following four case studies were performed: a conventional AirbusA320,aforward-sweptcanardaircraft,athree-surfaceaircraftandanoval-fuselage aircraft. For this purpose the genetic algorithm, sequential quadratic programming al- gorithm and a hybrid genetic algorithm were used. From the case studies followed that large improvements can be obtained with unconventional aircraft configurations when compared to the initial aircraft design proposed by the Initiator design tool. Up to 20% improvement was found with the three-surface and canard aircraft. The oval-fuselage aircraft could be improved by a solid 10%, while the lowest improvement was attained with the conventional A320. Among all cases the most contributing factors were the wing longitudinal position, sweep angle and wing aspect ratio. There is a tendency towards lower sweep angles due to the positive effect on the weight of the wing and an underestimation of the drag rise. With the forward-swept canard relatively high sweep angles were found, which contradicts the findings of the aft-swept wings. Therefore, the aerodynamics routine needs further investigation. Fromthehighlyswept, highaspectratiowingsoftheforward-sweptcanard aircraft followed that the weight penalty of forward swept wings is underestimated. In three cases the fuselage fineness ratio was involved in the optimization. The results showed that changing the fineness ratio offered some reduction in fuselage weight due to a more favourable structural loading. v vi Abstract The sizing routine of the control surfaces is found to be inadequate, since the Initiator derives most parameters directly from the wing and does not properly take into account controlandstabilityrequirements. Resultshaveshownthatthismainlyregardsthesweep and dihedral angle. Especially, the sweep angle is of concern, since it changes the lift- curveslopeandthereforealsostallcharacteristics. Thesesizingissuesalsoaffectthestatic margin. It was found that class II design information was not fed back to the control surface sizing. Other discrepancies were found with the wing dihedral. Due to a lack of lateral stability analysis of the Initiator the dihedral was driven by the lift-to-drag ratio rather than its stabilizing effect. As a result a lower dihedral was observed among the cases. From the used optimization algorithms can be concluded that the gradient algorithm was the least effective. It had difficulties with the uncertainties in the computed results of the Initiator. It sometimes stopped prematurely or started oscillating. This was alleviated by increasing the step size of the algorithm, but at the expense of accuracy. The genetic algorithm was found to be the best option since it proved to be very robust. It is far less sensitivity to noise, because it does not use gradient information. Its computational cost could be significantly reduced by applying parallel optimization and using a caching mechanism. The hybrid algorithm was found to be too computational expensive. The obtained increase in objective value did not outweigh the added cost. Acknowledgements This thesis marks the final step in completing the Master Program in System Engineering and Aircraft Design at the faculty of Aerospace Engineering. I had a great time studying at Delft University of Technology, with the company of my fellow students and friends. The research presented in this report would not have been possible without the support of a number of people whom I hereby would like to thank. Fist of all, I would like to express my gratitude to my supervisor dr.ir. Roelof Vos for his feedback, support and valuable insights during the course of my master thesis. I would alsoliketothankthemembersofmycommittee, prof.dr.ir. LeoVeldhuisandDriesVisser for their time to assess my work. I want to express my appreciation to Reno Elmendorp for providing feedback on the optimizer tool and helping me out with integration in the Initiator. Last but not least, I would like to thank my parents for their continuous support and encouragement throughout my study. Delft, The Netherlands P.T. Vasseur B.Sc. May 9, 2014 vii viii Acknowledgements