Beschreibung
In the presented work, parameter uncertainties are regarded as intervals. Novel control designs are introduced, which are based on a linear matrix inequality approach suitable for uncertain systems. Two extensions to state-of-the-art designs were given; the first with a constant controller gain approach over the complete time horizon and a second using a gain scheduling design over temporal subslices. Here, both rely on iterative solutions in the terms that controller gains are adapted based on the reachability analysis of former simulations. This means, that an efficient application of such methods is only realized with a reliable computation of possible interval enclosures. State-of-the-art enclosure techniques are often subject to overestimation, a possible solution comes in form of so-called cooperative systems. The structure of these systems allows for a separately, point-valued evaluation of the worst-case bounds, while guaranteeing the real value to be insight said bounds. This property can be found in numerous systems, however, exceptions occur especially concerning models from the fields of electrical as well as mechanical applications. To widen the applicability of cooperativity into these fields, this work presents transformation methods to adapt the structure of the treated system in such a way that it becomes cooperative while keeping its original stability properties. Due to the nature of said transformations this is done for systems with purely real eigenvalues and systems including conjugate-complex ones. As a final theoretical contribution, a state estimation is added to the controlled system as a form of fault diagnosis.
Herstellerkennzeichnung:
Shaker Verlag GmbH
Am Langen Graben 15a
52353 Düren
DE
E-Mail: info@shaker.de
