Transformer-Based Deep Learning Framework for Renewable Integrated Power Demand Prediction in Smart Grids Forecasting for Smart Grids

DOI

10.2991/978-94-6239-654-8_37How to use a DOI?

Keywords

Temporal Fusion Transformers (TFT); Dilated Temporal Convolutional Networks (TCNs); Solar; Load demand; Demand Prediction

Abstract

Currently, the custom of renewable energy to moderate the effects of global warming and its impact on the environment has become increasing day to day. Several nation-states currently have a majority of photovoltaics in use, which suggests that these renewable energy sources can be calculated with significant effect. Short-term power-demand forecasting is essential for renewable integrated smart grids, where the stochastic nature of solar resources complicates real-time balancing. In previous work, Challenges have been faced in accurate load demand predictions for specific solar renewable energy sources, The load demand in smart grids is partially dynamic, requiring estimates or adaptation to real-time operational shifts. This analysis propose a hybrid deep learning approach that integrates Temporal Fusion Transformers (TFT) with Dilated Temporal Convolutional Networks (TCNs) to accurately forecast solar renewable integrated energy needs in smart grids, managing the factors that standard forecasting measures. TCN can extract multi-scale temporal patterns from past power demand (e.g., from historical loads) and renewable energy information. TCNs can be trained both short-term and long-range dependencies and detected patterns were passed through a Temporal Fusion Transformer (TFT), which fuses static and time-varying inputs using attention mechanisms, gating layers, and variable range procedures into a single model. The TFT also displayed point forecasts and probabilistic forecasts featuring quantiles, allowing the smart grids manage to make more accurate estimates RMSE of 1.53 and MAPE of 2.12%. The experimental results validate that the proposed approach enhances load demand efficiency through an experiment conducted in MATLAB.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - R. Raguraman
AU  - K. Sakthivel
PY  - 2026
DA  - 2026/04/24
TI  - Transformer-Based Deep Learning Framework for Renewable Integrated Power Demand Prediction in Smart Grids Forecasting for Smart Grids
BT  - Proceedings of the Global Conference on Sustainable Energy Systems, Smart Electronics and Intelligent Computing (GCSESEIC 2025)
PB  - Atlantis Press
SP  - 457
EP  - 472
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6239-654-8_37
DO  - 10.2991/978-94-6239-654-8_37
ID  - Raguraman2026
ER  -