The existing MATE ("Multi Area Thevenin Equivalent") concept implemented in the real-time simulator OVNI provides a means of partitioning large systems of equations into subsystems connected through links. The subsystems are solved independently, with the overall solution integrated at the level of the links. MATE partitioning enables fast simulation of power system networks in two distinct ways: first, it allows parallel processing in a multi-machine environment and second, it allows integration of different solution techniques for individual subsystems.; In this thesis, we generalize the concept of MATE with the new Multilevel MATE concept, in which each subsystem becomes the basis for a new level of MATE partitioning. The new concept allows subsystems to be further partitioned into sub subsystems that, for example, are of a constant and changing nature. Partitioning at the subsystem level leads to higher overall solution efficiency. Furthermore, power system components can be described using Multilevel MATE by their nodal and/or branch equations on the subsystem level in OVNI. A three-phase induction machine model is an example of a power system component that is naturally described with branch equations. Multilevel MATE also provides a convenient framework for the incorporation of controllers as well as nonlinear elements.; The software implementation of OVNI written in this work is based on the Multilevel MATE algorithm. Models of power system components have been implemented and tested using the newly developed concept. A test case describing a doubly-fed induction generator wind turbine system has been modelled and studied as a practical example of the new solution scheme's capabilities.
展开▼
