Optimizing the dynamic behavior of structures using substructuring and surrogate modeling

During a design process, analysis and, when necessary, modification of the vibration characteristics of a structure are important. Thanks to the developments in computer technology and in numerical analysis methods, particularly the Finite Element Method, a structure can be analyzed extensively in the computer environment long before its first prototype is built. To improve its design and to find a globally optimal configuration, in theory it is possible to directly couple a Finite Element (FE) model with a global optimization method. However, in practice this may not be feasible for complex structures due to the required number of the FE calculations and the corresponding computational costs. Analysis time grows rapidly with the amount of details in the FE model. If the vibration characteristics of a structure need to be improved by modifying the design of the detailed sections, long running analyses are a bottleneck in optimization. The objective of this thesis is to develop an efficient strategy for optimizing the dynamic behavior of complex structures. The strategy is required to be robust, accurate and able to provide a solution which is as close to the global optimum as possible.