Estimation Method for Voltage Sag Occurrence Frequency Based on Calculation of Protection Operating Time Characteristics

Voltage sag occurrence frequency (VSOF) estimation includes the prediction of residual voltage and duration. The former has been well studied, whereas the latter presents many challenges, and is the main focus of this study. A method for estimating the VSOF based on the calculation of the protection operating time characteristics is proposed. First, this article adopts a recognition method for protection types based on monitored historical data. The protection type is recognized by dividing the fault electrical quantity (FEQ) into several intervals and calculating their mean standard deviation. Second, this article proposes a calculation method for the protection operating time characteristics. The FEQ and operating time samples are clustered based on the proposed algorithm. The density-based spatial clustering of applications with noise (DBSCAN) algorithm is enhanced by the Mahalanobis distance and K-nearest-neighbor (KNN) method. The least-square fitting and segmentation calculation of the operating time curve is performed in accordance with the clustering results. Finally, this study proposes a method to predict VSOF by considering operating time curves, different fault resistances, and generator scheduling, whereas the IEEE-30 test system is used to verify this method.