Identification of system characteristics of a power system with time series data. Identification of frequency fluctuation characteristics of a small-scale isolated system

Understanding actual characteristics of a power system with recorded time series data is of great importance, for example, in improving the performance of the system. Although system identification is a well-known technique to achieve this goal, its applicability to a certain system should be examined for the particular case because its accuracy highly depends on the inherent characteristics of the system. While many papers have discussed application of a system identification technique to a power system, few papers have examined its applicability to the actual data of a power system. This paper presents a new system identification method to estimate characteristics of a power system while using output of intermittent generators or fluctuating loads as an external disturbance. The method employs cross spectra and coherence as a key factor in the identification; it estimates a transfer function of a power system, contribution of observed disturbance to total disturbance, and so on. The method is applied to time series data of two model systems: simulation results and measured data of an isolated power system with diesel generators. The study gives satisfactory results; implication on the accuracy of the method is discussed through the sample studies.