International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 06 | Jun 2025
p-ISSN: 2395-0072
www.irjet.net
A REVIEW OF COMPARATIVE ANALYSIS OF AI-POWERED RENEWABLE ENERGY FORECASTING TECHNIQUES Sri Nivas Singh1, Mrs. Arifa Khan2 1Master of Technology, Computer Science and Engineering, Lucknow Institute of Technology, Lucknow, India 2Assistant Professor, Department of Computer Science and Engineering, Lucknow Institute of Technology,
Lucknow, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - This trend of sustainable development in the
thus make the process of operating the power systems and long-term planning rather difficult (Hong, Pinson and Fan, 2014).
world has placed more pressure on the effective utilization of renewable power which is mainly sourced on solar, wind and hydro. Accurate prediction of renewable generation is critical to the stability of the grid, optimal energy storage and cost transgression of operations. However, the inherent variability and non-continuity of these sources makes it difficult to carry out conventional forecasting algorithms. In the recent past, the role played by Artificial Intelligence (AI) in eliminating such challenges has been resounding due to the utilization of advanced data-driven techniques. The current review is a comparative analysis of some AI-based forecasting systems, including machine learning systems Support Vector Machines, Random Forests - and deep learning systems, including Artificial Neural Networks, Convolutional Neural Networks and Long Short-Term Memory networks. The analysis evaluates their accuracy in a wide range of forecasting horizons and energy types of interest and brings to attention key performance indicators that include RMSE and MAE. The outcomes characterize the personal advantages, weaknesses and ideal uses of the wellknown methods and thus confirm the heightened potential of AI in increasing the precision and reliability of renewable energies forecasting in addition to serving as an instructional indication on future study and implementation.
Grid stability, economic dispatch and economic energy storage deployments thus require accurate prediction of renewable- energy generation. Due to inaccurate forecasts, the use of backup systems driven by fossil fuels is likely to be enhanced, and thus sustainability goals are at risk (Mandal, 2012; Falvo, Zaninelli and Lamedica, 2016). Therefore, there is an immediate need of accurate and flexible forecasting models which would be able to cope with the variability of renewable production.
1.2 Motivation for AI in Forecasting The peculiarities of renewable energy data rarely prevail in conventional prediction techniques, which mostly include statistical models, as well as time series forecasting, such as ARIMA (Taylor, 2010). The common methods are usually based on the assumption of linearity and stationarity and hence are inappropriate in the highly fluctuating and non-linear nature of data specific to renewable sources (Raza, Khosravi, & Nahavandi, 2020). One of the alternatives that have risen is the Artificial Intelligence (AI), due to its ability to represent a complex non-linear relationship and to learn with large amounts of data. Support Vector Machines (SVM), Random Forests (RF) and ensemble methods are examples of machine learning tools that have demonstrated a better performance in forecasting performance in various time horizons. Similarly, the structures of deep-learning networks namely, Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), and Long ShortTerm Memory (LSTM) are invariably better at accommodating influences of time and hierarchies (Ahmed & Khalid, 2019; Yagli, Ozdemir, & Teke, 2021). These models are ad hoc, accommodating, and can be able to create themselves whenever more data is present.
Key Words: Artificial Intelligence (AI), Renewable Energy Forecasting, Machine Learning, Deep Learning, Hybrid Models, Smart Grid Integration.
1.INTRODUCTION 1.1 Background It has been constantly observed that renewable energy is one of the key action areas in terms of climate change reduction and reliance on fossil fuels. Solar, wind, and hydro-based technologies have resulted in widespread attention due to their similar benefits to the environment and natural sustainability (REN21, 2023; IEA, 2022). However the nature of these sources which are variable and intermittent in nature, provides challenges to grid integration. Renewable sources of electricity generation are extremely vulnerable to the variable nature of phenomena like the solar irradiance and wind velocity and
© 2025, IRJET
|
Impact Factor value: 8.315
More high-resolution weather and energy sensors are also deployed around the globe, also being possible thanks to the spread of Internet of Things (IoT) devices and smartgrid technologies, which only adds to the value of AI in forecasting (Riahi et al., 2022). These models together
|
ISO 9001:2008 Certified Journal
|
Page 1225