An enhanced adaptive data processing Technique for computing synchrophasor measurements under power system frequency variation

This work proposes a frequency-adaptive data processing technique for reliable synchrophasor measurements estimation under power system frequency variation. The proposed technique consists of four steps in which frequency estimation, filtering process, fast linear interpolation, and Radix-2 decimation-in-time (DIT) fast Fourier transform (FFT) approach are applied in a synergistic manner to provide accurate synchrophasor estimates. The sampling rate remains unchanged and the frequency measurement is estimated from the classical weighted least-squares approach, filtering process is executed by a Finite Impulse Response (FIR) filter bank, composed by lowpass and bandpass filters, whose goal is to separate all input signal harmonic components. The digital filter bank's size is limited by the desired harmonic component imposed by the 3-dB cut-off frequency, related to the analog anti-aliasing lowpass filter, and an optimized design has been employed to obtain symmetric coefficients and odd length (type 1 FIR filter). Fast linear interpolation based on trigonometric identities, using adaptive-angles related to the estimated frequency, is applied at each filter output data whose outcomes are used by the half-cycle Radix-2 DIT FFT approach to provide raw data in which the corrected synchrophasor measurements are obtained after the filter phase and amplitude compensations. The proposed technique aims also to evaluate the impact of the frequency variation on each harmonic component present in the input signal. Test signals with ±3 Hz offset-nominal corrupted by harmonics components are applied to evaluate the technique's performance and the results demonstrate its capability in dealing efficiently with signals corrupted by spurious frequencies, being envisaged for potential applications involving M-Class PMU model.