Forecasting Electric Vehicle Charging Demand Using Recurrent Neural Networks and LSTM Models: A Deep Learning Approach to Smart Grid Optimization

Authors

  • Veena More Akkamahadevi Women’s University, Athani Rd, Jnanashakti Campus, Torvi, Vijayapura, Karnataka, India. Author
  • Sowmya Kella Akkamahadevi Women’s University, Athani Rd, Jnanashakti Campus, Torvi, Vijayapura, Karnataka, India. Author
  • Ramesh K Akkamahadevi Women’s University, Athani Rd, Jnanashakti Campus, Torvi, Vijayapura, Karnataka, India. Author

DOI:

https://doi.org/10.47392/IRJAEM.2025.0382

Keywords:

Electric Vehicle Charging, Deep Learning, LSTM Networks, Recurrent Neural Networks (RNN), Smart Grid Management, Load Forecasting

Abstract

The rapid proliferation of Electric Vehicles intro- duces both transforming opportunities and complex challenges to modern power grid infrastructure. This study examines at how deep learning can help predict when and how much EVs will   charge, so that the grid can be better prepared. We employed two types of neural networks—Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) networks—trained on a dataset containing electric vehicle (EV) charging data. The objective was to evaluate the effectiveness of these models in predicting future demand using historical charging patterns. Found that the LSTM model gave more accurate results, especially for longer-term trends. These predictions could be useful for companies to plan energy distribution, avoid overloads, and support renewable energy use. This research adds to the growing effort to make electric grids smarter and more adaptable as EV numbers increase.

Downloads

Download data is not yet available.

Downloads

Published

2025-07-24