Finding the optimal shape of an object subjected to a fluid stream is a challenging endeavor, in which many different and complex points need to be addressed. The method of using adjoint based sensitivities for a gradient based optimization shows one of the most promising approaches. Nowadays even home computers have sufficient computing power to solve simple fluid dynamics problems. The effectiveness of the adjoint approach to compute sensitivities makes an optimization possible even for complex shapes. In the following thesis, the optimization using adjoint based sensitivities is presented. In three examples, the difficulties that arise whilst computing adjoint sensitivities are analyzed and evaluated. On the shape of a cylinder the stability of the algorithm to compute the sensitivities is tested for an increasing Reynolds number in a transient flow. For higher Reynolds numbers the turbulences, which presumably cause instabilities, are modeled using the Spalart-Allmaras turbulence model. Thereby it is possible to calculate sensitivities for complex shapes even in a hight Reynolds number regime. Finally, a procedure is attempted, using adjoint based sensitivities to optimize the shape of a Naca0012 airfoil to increase its lift force.     «

Finding the optimal shape of an object subjected to a fluid stream is a challenging endeavor, in which many different and complex points need to be addressed. The method of using adjoint based sensitivities for a gradient based optimization shows one of the most promising approaches. Nowadays even home computers have sufficient computing power to solve simple fluid dynamics problems. The effectiveness of the adjoint approach to compute sensitivities makes an optimization possible even for comple...     »