| 摘要 |
PROBLEM TO BE SOLVED: To quicken prediction by eliminating the need for reviewing the prediction conditions of out-of-step even time, when the configuration and the operation of a power system are changed in the out-of-step prediction of a power system in a middle region. SOLUTION: Condition of steadiness of a time-series model are applied, and the time-series data of the internal phase angle of a generator start to be read from one second before a fault occurs, when the occurrence of the fault is detected (2). Then, two parameters (a1, a2) of an AR model are estimated by the least square method, using the time-series data (3). Furthermore, a characteristic root is obtained from a characteristic equation that is constituted of the two estimated parameters (4), prediction is started from a point when an oscillation waveform after the fault occurs passes one cycle, and then it is decided that a power system is stable and out-of-step, when the size of the absolute value of the characteristic root is less than 1 (6: YES) and equal to or more than 1 (6: NO), respectively.
|
