Confidence Intervals Estimation in the Identification of Electromechanical Modes From Ambient Noise

Abstract

This paper discusses the estimation of uncertainty intervals associated with the electromechanical modes identified from ambient data resulting from random load switching throughout the day in power systems. A connection between the second order statistical properties, including confidence intervals, of the identified electromechanical modes and the variance of the parameters of a selected linear model is demonstrated. The results of the presented method are compared with respect to the ones obtained from a Monte Carlo technique, showing its effectiveness in reducing the number of trials, which would be beneficial for online power system monitoring, as it can decrease the number of samples, thus ensuring that the system dynamics would not change significantly over the monitoring time window, and yielding more dependable results. Two test cases, namely, the two-area benchmark system and the IEEE 14-bus system, with different orders of the system identification model used, are utilized to demonstrate the effectiveness of the proposed methodology.