Identification of Low-Frequency Oscillations Modes Based on Fuzzy Clustering and Dynamic Partition
|School||Tohoku Electric Power University|
|Course||Proceedings of the|
|Keywords||Low-frequency oscillation Fuzzy Clustering Dynamic segmentation theory Dominant oscillation mode Prony's method|
The large regional power grid interconnection is prone to low-frequency oscillation of a factor can not be ignored . The oscillation characteristics of the online analysis is an important theoretical basis for online monitoring and wide area damping control power system low frequency oscillation . Based on the development and application of the Global Positioning System (GPS) , Wide Area Measurement System (WAMS) provides a new opportunity for online analysis of low frequency oscillation mode even control . The oscillation curve analysis system oscillation Information has been widely used. Interconnected multi-machine power system , there are a variety of oscillation modes , the need to focus on the area between the dominant weak damping or negative damping mode . Due to outside interference by many factors , such as noise and other factors , the actual multi-machine system to conventional signal recognition inconvenience . Oscillation curve extraction of multi-machine oscillation information , you need to select the appropriate curve , in order to quickly get the the system led oscillation mode . First, identify the angle from the coherency of the nature and cause of disturbance next big regional power grid low frequency oscillation . In this paper, some basic assumptions and do the necessary verification system based on a fuzzy clustering method based on fuzzy by iterative self-organizing data analysis technique (ISODATA) to form the cohomology recognition fuzzy sets , fuzzy clustering iterative clustering , until you get a satisfactory classification results . Second , in a multi-machine system is cohomology grouping on the basis of dynamic segmentation theory . This paper introduces the basic concepts of the theory of dynamic segmentation points without cutting method , and then propose to choose the low-frequency oscillation signal curve . The theory reveals a weak link between the network , looking for a suitable oscillation curve provides an important theoretical value . Finally , the use of Prony analysis algorithm to accurately extract the dominant oscillation mode from the selected curve . Further verify the correctness of the proposed method , through the example of EPRI 8 machine 36 - node system simulation , to verify the feasibility and effectiveness of the method .