On the Forecast of the Voltage Sags Using the Measurements in Real Power Systems

The availability of measured sags in real power systems opens the appealing option of forecasting the occurrence of voltage sags rather than estimating the average performance, as traditionally done in the literature. The voltage sags measured in real electric power systems can be divided into two main categories, i.e., rare voltage sags and clusters of voltage sags. Rare voltage sags meet the conditions of a Poisson process; instead, the presence of clusters brings the phenomenon far from the conditions of the Poisson model. Very recently, the authors of this paper demonstrated that the forecast of the number of rare voltage sags of a system or a part of is viable using only three years of measurements with acceptable prevision errors. If the clusters are included in all the measured sags, a stochastic model is needed to account for the time dependence of the sags. In this paper, using the recorded measurements of three years at the sites of a real regional electric system, the voltage sags comprehensive of clusters were modelled to forecast the number of sags in the incoming year. Not completely satisfactory results on all the system suggested to deep analyze in few sites the rare sags separately from the sags comprehensive of clusters. The intermittence indices, previous proposed by the authors, proved to be an effective tool to discriminate the forecast model to use.