Variable fidelity methods and surrogate modeling of critical loads on X-31 aircraft

The thesis develops computational tools for early stages of the aircraft design process. The work focuses on a framework which allows several design teams concurrently to develop a baseline concept into a configuration which meets requirements and whose aerodynamics has been assessed by flight simulation. To this end, a data base format suggested by the German Aerospace Center DLR was adopted in the CEASIOM system, developed in the EU 6th Framework Program, enabling more accurate transonic analysis and tabulation of forces and moments as well as control surface authority assessment. Results from simple, fast models are combined with computationally expensive full CFD results by co-Kriging to speed up productionof the aero-data for flight simulation.Non-linear optimization methods in wing design play an increasingly important role together with computational aerodynamics. High performance computing enables the use of high-fidelity non-linear flow predictions in optimization loops. It is argued that the optimization tools should allow the engineer to influence the process by setting up suitable target pressure distributions for the shape to approach, combined with steps to minimize drag under suitable constraints on geometry, forces, and moments. The simulation framework incorporated into CEASIOM was applied to a number of configurations, conventional as well as un-conventional, such as an a-symmetric twin prop, a canard-configured transonic cruiser, and a novel chinrudder concept for transonic airliners. Aerodynamic shape design by the developed methods was applied to the standard M6 benchmark wing, a joined-wing concept, a wing-tip, and a blended wing-body.