Forecasting of Electricity Demand and Renewable Energy Generation for Grid Stability

Abstract

The electricity generation units, network operators, and consumers operate in sync by making commitments based on their projections of the consumption and generation capacities. The balance between generation and demand is very important for smooth operation of the system, reduction in the harmonics and losses, and moreover for the grid stability. This requires accurate forecast of electricity demand along with its self-sufficiency in terms of renewable energy generation. The short-term forecasting of electricity demand and its generation from the renewable sources like solar and wind have been considered in this paper. An hourly forecasting of electricity demand, solar generation, and wind generation has been carried out with 24 and 48 h advance forecasting. To maintain the stability of the grid, it is important to generate the electricity with due consideration to solar and wind generation. We hence present the machine learning-based models to directly forecast the net generation requirement of electricity with solar and wind generation data. We present the exhaustive results to analyze these forecasts with and without the availability of future weather information. For various machine learning algorithms, forecasting accuracy has been compared for different seasons, days of the week, and hours of the day to evaluate the robustness of the algorithms.