Design of FIR second-degree differentiator for frequency estimation

Finite-impulse-response (FIR) lowpass second-degree differentiators (with coefficients that are based on an optimal design approach) are presented. By minimising the filter variance, the differentiator rejects noise above the cut-off frequency, with steep roll-off in the transition-band, and showing close-to-ideal magnitude response accuracy in the passband. The proposed FIR differentiators are used to estimate the principal frequency of a non-ideal power system, based on a variant of Sarkar and Sengupta's algorithm. Use of an FIR filter, rather than an infinite-impulse-response (IIR) filter, for fundamental frequency estimation is shown to improve the accuracy of frequency estimation.