Identifying the Class of Disturbance Events Using Recurrence Quantification Analysis

Classifying disturbance events in a power system is a challenging task. This paper proposes an online application of event detection and classification using the Recurrence Quantification Analysis (RQA). Nonlinear measures based on recurrence plots are examined to assess the characteristics of location and type of disturbance events. The Principle Components Analysis is used for dimensionality reduction and an unsupervised clustering method is applied to identify two types of disturbance events, namely, short circuit faults and generator trips. Simulations are conducted on a 29-machine 179-bus reduced WECC system to validate the proposed algorithm. Results show that the RQA measures might be an effective tool to identify the location and type of disturbance events.