Fast Power System Frequency Estimation by Shape Class Approximation for Synthetic Inertia Provision by Battery Energy Storage Systems

Abstract

This paper presents a new power system frequency estimation method suitable for a fast synthetic inertial response time. The method utilizes the shape error of a predefined shape class. The shape class of a fundamental waveform enables a fast frequency estimation that requires five consecutive samples. Considering a sampling rate of 2 MHz, the proposed method estimates the system frequency in around 2.5 μs after a step, for undisturbed voltages. To cope with distorted voltages, the proposed method needs a filtering stage, which introduces a delay of at least 30.6 ms. Both the delay and the accuracy of the estimation vary with the level of distortion. Ultimately, the simulation results suggest that the proposed frequency estimation method provides fast estimations at an accuracy that is competitive with other estimation methods.