The present work is concerned with the advancement of the knowledge of structural optimization in engineering design while focusing on efficient and easy to use ways of setting-up the required automated processes as well as the problems arising from it. Three industry examples are considered. In the first example a software tool that serves as a hands-on decision guidance for many occurring design situations for structured wall PE pipes is developed. In order to avoid licensing fees only public domain software or in-house code are used. It offers the efficient and automated simulation of the ringstiffness test as well as the most common pipe installation scenarios. In addition, an optimization feature is implemented for the design of optimum pipe profiles with regards to the ringstiffness test. In the second example a framework for the optimum design of carbon fibre mountain bike frames is developed. An extensively parameterized and automated simulation model is created that allows for varying tube shapes, paths and laminate ply thicknesses as well as joint locations. For improved efficiency a decomposition approach has been employed that decomposes the original optimization problem into a size optimization sub problem and a shape optimization top level problem. The former is solved by the built-in optimization tool in OptiStruct and the latter by means of surrogate based optimization where each experiment in the DoE is a full size optimization. The third example is concerned with the optimum design of a blade for a novel vertical axis wind turbine. A design approach similar to those with horizontal axes is chosen. The altered design requirements are accounted for by creating a parameterized simulation model and performing size optimization runs for 32 models with different material settings and shear web locations where the model creation process has been automated.
|Electronic Thesis or Dissertation
Page generated in 0.0165 seconds