Linear programming algorithm applied to power system measurement for frequency relaying

A new application for linear programming (LP) algorithm to power system measurements for frequency relaying is presented. The technique is based on the minimization of the sum of the absolute value of the error of the digitized bus voltage at the relay location. The proposed technique is superior to the well known least error squares (LS) technique, when the digitized voltage samples are contaminated with gross errors. It has the ability, in most cases, to reject these bad measurements in contrast to the LS which is adversely affected by the presence of bad data. Two models are used to estimate the frequency of a power system from the digitized bus voltage at the relay location. The first, is the constant-frequency model (CFM), which is used to estimate the steady state frequency deviation, while the second is the variable frequency model (VFM) which is used to estimate the transient frequency and the rate of change of frequency. Effects of sampling rate, data window size, and time-reference-location on the behaviors of the algorithm are examined. Numerical results are reported in this paper, where we compare the proposed LP algorithm, and the LS algorithm. The proposed technique can easily be applied to a microprocessor based algorithm for digital frequency relaying.