Fast Accurate Phasor Estimation in Less than One Cycle using Neural Networks

Estimation of phasors holds significant importance within power system protection applications. Due to the inher-ent inductive characteristics of electric power systems, fault currents often exhibit transient oscillations, such as decaying DC offset (DDC), higher-order harmonics, and noise. Fourier analysis stands out as a widely adopted technique for phasor estimation calculations. Nevertheless, when dealing with decaying DC currents, which may persist for up to 5 cycles (in the case of a 50/60Hz system), the accuracy of Fourier analysis diminishes significantly. An inaccurate phasor estimation could lead to malfunction or misoperation of protective devices, most notably distance protection relays, an undesirable outcome. Therefore, it becomes imperative to address the issue of DC offset signal mitigation. While various analytical methods have been proposed in recent years, there exists a pressing need to explore less complicated data-driven approaches. In this paper, we aim to employ two neural network methods to estimate the current in less than one cycle, taking into account second-order DDC and system frequency deviations. Compared with other presented data-driven methods, the proposed scheme is only uses a half cycle of the fault current signal with less number of trained parameters and without applying encoder/decoder methods.

I. Introduction

A broad spectrum of analytical techniques has been pro-posed to tackle precise phasor estimation within power system applications. Four primary categories, namely pre-discrete Fourier transform (DFT), post-DFT, Least Square (LS)-based, and Artificial Intelligence (AI)-based methods, have been explored in prior studies. In pre-DFT approaches, the initial step involves the removal of DDC, followed by the application of the DFT for phasor computation. DFT exhibits distinctive attributes such as simplicity, immunity to high-order harmonics, and rapid response, rendering it an optimal choice. In [1], an integral-based method is introduced for DDC estimation preceding DFT. Notably, the integral of sinusoidal components during one cycle equals zero, resulting in the preservation of only the integral of the DDC parameter. [2] leverages this principle, along with full cycle DFT, to derive the DDC parameters. This method employs the first two terms of the Taylor series expansion for DDC approximation. Additionally, in a study outlined in [3], an auxiliary signal is generated, functioning akin to a high-pass filter for DDC calculation. This approach involves a hierarchical process, encompassing the summation of both FC and an auxiliary signal, DFT error assessment, and DDC mitigation.