Variance Based Analysis for an Isolated Power System Using Kalman Filter and LQR

Abstract

Automatic Generation Control (AGC) is an important approach to acquire the stability to ensure reliable operation of power systems. For stable operation of power systems, the frequency of the system should be reserved within the nominal value. Towards this, the estimation of states is of paramount importance. In this paper, a comparison is made of (i) the continuous estimation of the states from the measurement of one of the state variable of a power system using Kalman filter, followed by a pole placement to ensure stability and (ii) the optimal control approach viz Linear Quadratic Regulator (LQR) to the Automatic Generation Control (AGC), implementing control from direct measurement of all state variables. The comparison is made on the basis of the mean of variances of frequency estimate of both the approaches under different noise levels on the measurements made, from independent Monte Carlo simulations. The comparison is done on an isolated power system, modelled using Simulink/MATLAB.