Modeling and estimation of harmonic signals using ellipsoidal set theory

The objective of this paper is to demonstrate the use of ellipsoidal set theoretic methods for modelling and estimation of power system harmonic signals. The voltage or current signal to be estimated is assumed to be corrupted by measurement and process noises. Also, the magnitude of each harmonic component is assumed to be unknown and time-varying. Three ellipsoidal set theoretic estimation techniques are presented to estimate the harmonic signal when corrupted by noise. These methods are then compared with the optimal Kalman filter based on an equivalent stochastic model using Monte Carlo techniques.