Constrained DFO for Multidisciplinary Design Optimization

Abstract

The SpaceLiner is a concept study of a hypersonic airplane for ultra-fast passenger transport, developed at the German Aerospace Center (DLR). When designing spacecraft, one of the major issues is to devise its thermal protection systems needed during atmospheric re-entry. As the goal is to find the optimal preliminary SpaceLiner design which considers all main disciplines (geometry, aerodynamics, structural masses and thermal protection system), multidisciplinary design optimization (MDO) techniques have to be applied. In simulation-based engineering applications, oftentimes no derivatives are available. Thus, we would like to present our new optimization algorithm, a model-based trust-region SQP algorithm, to solve equality-constrained optimization problems without derivatives. Finally, we want to present the integration of the described optimization framework into the workflow management system RCE which provides a graphical user interface to couple simulation codes in a collaborative environment.