Slow sampling on-line optimization approach to estimate power system frequency

The paper presents a real-time optimization approach based on the Newton-type algorithm (NTA) and the least squares (LS) method for power system frequency estimation. A non-linear non-recursive Newton algorithm is utilized to track the modulation effect of frequency variation on the on-line estimation of the phase angle. The LS curve fitting technique extracts the instantaneous power system frequency from the time-varying phase angle estimation by the NTA. A very low sampling rate is adopted to realize the introduced NTA-LS optimization technique and compare its performance with the decomposition of single-phase into orthogonal components (DSPOC) algorithm and the NTA. The presented slow sampling NTA-LS structure is a very efficient real-time algorithm and promises to reduce the hardware complexity, alleviate the runtime jams in digital signal processors (DSPs), and rectify the need for wide bandwidth sensors. The performance of the proposed method is tested by presenting simulations in MATLAB-Simulink which prove the robustness and accuracy of the NTA-LS method under time-varying conditions.