Vibrational Response Reconstruction and Model Validation

Licentiate thesis, 2008

In this thesis, laboratory response reconstruction and computational model validation is investigated. The former concerns applications of control theory as the objective is to reproduce a certain loading for a known component, while the latter concerns optimization of a parameterized model as the objective is to modify a computational model to mimic a certain component's behavior. These sibling problems are also often referred to as drive signal identification for rig testing and model updating, respectively. They are both inverse problems, as in the case of drive signal identification the objective is to find a suitable input based on knowledge of the system and its output, while the model updating objective is to modify a system based on knowledge of its input and output. These problems are often further classified as inverse problems of the first and second kind, respectively. Aspects of these two neighbor problems, namely controllability considerations in the case of response reconstruction and error metric considerations in the case of model validation, are treated.