Sustainable Energy Forecasting: A Machine Learning Framework for India's Power Sector
Year: 2025
Type: article
Abstract: This study presents a comprehensive analysis of India's electricity generation trends, focusing on oil, gas,and wind energy sources, using advanced machine learning models. The research leverages data from theIndia Climate and Energy Dashboard (ICED) to explore state-wise temporal trends in electricitygeneration, aiming to provide insights into generation patterns, demand forecasting, and resourceutilization. Three datasets—wind, oil, and gas electricity generation—were analyzed using ARIMA,LSTM, and Gaussian Mixture Models (GMM) to capture temporal dependencies, seasonal trends, andregional disparities in energy production. The ARIMA and LSTM models demonstrated strongperformance in forecasting electricity generation, with LSTM showing particularly high accuracy incapturing complex temporal patterns. GMM clustering revealed significant regional disparities,highlighting states like Gujarat and Maharashtra as leaders in operational capacity. The study underscoresthe growing importance of renewable energy, particularly wind, while also emphasizing the challenges ofgrid integration and variability in generation. The findings offer valuable insights for energy planners andpolicymakers, supporting informed decision-making in resource allocation, grid management, andsustainable energy development. The research contributes to the broader goal of achieving energy securityand sustainability in India, providing a robust methodological framework for future studies in energyforecasting and optimization.
Author: Omkar Singh, Sherilyn Kevin, Shreya Muley
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6968574331
FWCI:
Cites:
Cited by:
Related to: