Harmonic Decomposition of Transient Disturbances Using the Ls Prony and Esprit-based Methods

The modeling and analysis of electric power disturbances affecting power quality has emerged as an important field in the changing electric power industry. These disturbances differ in characteristics, and their occurrence can be attributed to various reasons. Tools capable of classifying these disturbances have been investigated and many are under development. This paper illustrates the harmonic decomposition of disturbances, focusing on transient disturbances. The approaches presented are based on modeling the disturbance as a series of damped exponentials, extracting waveform descriptors unique to each disturbance. The two processing methods used are the Least-Squares Prony method and an ESPRITbased approach (Estimation of Signal Parameters via Rotational Invariance Techniques). The approaches are compared by processing transient disturbances corrupted with measurement noise, which is modeled as zero-mean white Gaussian noise, at different voltage deviation levels. Results indicate that the ESPRIT-based method outperforms the Least-Squares Prony method in estimating the waveform descriptors in the presence of measurement noise. The transient disturbances used were created at the Center for Electric Power Engineering (CEPE), Drexel University.