An advanced method for evaluation of measurement algorithms used in digital protective relaying

Abstract

Input voltage and current signals measured by digital protective relays during fault conditions may contain oscillatory components with different frequencies as well as transient offset. Most of the relays are using only fundamental frequency component to detect a fault condition, and other components are considered nuisance. The sources of these nuisance components could be transients of instrument transformers and various resonant frequencies initiated by a fault. These components might eventually affect performance of measurement algorithm implemented in digital protective relays. The purpose of this paper is to propose a methodology for systematic performance evaluation of measurement algorithms implemented in digital protective relays. The sensitivity to the following nuisance signals have been studied: DC offset and harmonic frequencies; as well as sensitivity to fundamental frequency variations. In this study the performance indices in frequency and time domain are calculated to quantify the effects of nuisance components and frequency variation. Performances of the following measurement algorithms are evaluated: Two-sample method, Sample and first derivative method, First and second derivative method, two versions of Discrete Fourier Transform algorithms (Full cycle DFT and Half cycle DFT), and Cosine algorithm.