| 摘要 |
PURPOSE:To extract the characteristic value of a power system in a power fluctuating mode at a high speed so as to monitor the steady-state stability of the power system by discriminating the steady-state stability by inputting the phase and absolute value of each generator on its complex plane found in advance through off-line simulation performed on each system cross section to a neural network. CONSTITUTION:Learning of a neural network 6 is performed in advance based on the results of off-line simulation by using the phase and absolute values of characteristic generators in a complex plane as the input layer of the network 6 and the kind of power fluctuating mode as an output layer. Then a data collecting section collects system connecting states of generators, loads, substations, etc., and the operating states of a power system, such as the power generating amount, load, voltage, etc., from the power system and a characteristic value/vector calculating section finds a characteristic value or vector based on the data collected by the collecting section and inputs calculated results to the network 6 of a section 40 for extracting the characteristic value of a power fluctuating mode. The network 6 then extracts the characteristic value of the power fluctuating mode. |
