抄録
Reactor noise analysis method based on information dimension is applied to the monitoring and diagnosing of power oscillation. The method focuses on the utilization of the slope of the correlation integral (SOCI) which determines the information dimension of attractors. For practical application, the information dimension is expected to be the same as the fractal dimension of attractors ; it can be used to classify different asymptotic regimes of nonlinear dynamical systems.
We examined a real power oscillation using SOCI and the results implied that the oscillation was just a noisy limit cycle, although it is not possible to assert that there is no chaotic character in the oscillation because large oscillatory time-series data sets are not available. In addition, the application of SOCI to the realtime monitoring of power oscillation is proposed and examined.
