Development of intelligent, robust and nonlinear models in dynamic equivalencing for interconnected power systems

The stability analysis of interconnected power systems is a difficult key task on one hand because of the large number of network components that need to be considered such as machines, governors, exciters, loads, and converters, among others forming a complex dynamic system and on the other hand due to limited data exchange between energy utilities preserving economic interests in the increasingly liberalized power markets. This work presents innovative non-linear models of complex dynamic systems and their methods in form of simplified power supply models which can be utilized in the fields of network reliability, management and planning, especially to manage blackouts situations and to affront new technical circumstances of the deregulated electricity market. The novel methods are developed involving intelligent systems, such pattern recognition algorithms, Fuzzy concepts, non-linear modelling and identification of dynamic systems using robust dynamic recurrent artificial neural networks (DANN) with locally distributed dynamics. These robust methods are suitable alternatives to the classical dynamic equivalencing with high accuracy and applicability. In order to verify their effectiveness and to compare their properties, case studies, such as the interconnected European power system UCTE/CENTRAL, among others are tested.