Enhancing Power System Protection with Prediction of Inertia Constant and Frequency Deviations using Machine Learning

This paper explores how machine learning can improve power system protection by predicting critical parameters such as inertia constants and frequency deviations. In the current era, various power systems are consistently integrating renewable energy sources, hence, maintaining the stability of the system has become complex. This study analyzes two configurations i.e. a conventional grid and a microgrid. Machine Learning models are applied to predict frequency deviations and inertia variations under different operating conditions. By forecasting system responses to disturbances, the proposed approach aims to provide timely insights into potential frequency shifts and stability needs. This predictive framework aims to equip power systems with greater resilience and adaptability, addressing the challenges posed by modern grid demands and the variable nature of renewable energy sources.

I. Introduction

The increasing acceptance of renewable energy sources like solar, wind, and other sustainable mechanisms is causing a transformation in the conventional power systems. Dissimilar to the conventional grids which rely mainly on large, inertia-rich generators, renewable-based systems lack the same level of inertia, which is a critical stabilizing factor for frequency control. Inertia in a power system refers to the resistance of rotating mass (like that of generators) to changes in frequency [1]. High inertia allows the grid to naturally buffer against sudden fluctuations, providing time to control systems to restore balance. However, as more renewable energy sources come online, the reduced inertia introduces challenges in maintaining stable frequency, making modern power systems more susceptible to disturbances and rapid frequency deviations [2], [3]. As a result, there is an increasing need for advanced protection techniques which would be adaptable to these changes and would help in maintaining the reliability of the grid.