Fast adaptive schemes for tracking voltage phasor and local frequency in power transmission and distribution systems

Real-time measurements of voltage phasor and local frequency deviation find applications in computer-based relaying, static state estimation, disturbance monitoring and control. Two learning schemes for fast estimation of these basic quantities are proposed. The problem was approached from a system identification perspective, in opposition to the well-established extended Kalman filtering (EKF) technique. It is shown that, from a simple nonlinear model of the system voltage which involves only two parameters, the recursive least squares (RLS) and the least mean squares (LMS) algorithms can each provide dynamic estimates of the voltage phasor. The finite derivative of the phase deviation, followed by a moving-average filter, then leads to the local frequency deviation. A constant forgetting factor included in these algorithms provides both fast adaptation in time-varying situations and good smoothing of the estimates when necessary