| Autor: |
A.I. Megahed, Nourhan M. Ibrahim, Nabil H. Abbasy |
| Rok vydání: |
2021 |
| Předmět: |
|
| Zdroj: |
2021 5th International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT). |
| DOI: |
10.1109/ismsit52890.2021.9604650 |
| Popis: |
Forecasting of energy use in smart buildings is crucial in operations, demand response methods, and the installation of distributed generation systems, as being critical components of smart grid energy management. Numerous methods have been tried to forecast energy consumption at different levels of distribution and transmission networks. Apart from aggregated load on a large scale, forecasting the electric demand of a single energy consumer is challenging owing to the inherent unpredictability and uncertainty. Recurrent neural networks have been widely utilized as forecasting models for energy because they outperform the majority of machine learning methods in time series forecasting. This paper explores the forecasting issue using deep learning recurrent neural network with Long Short-Term Memory LSTM units that fits both aggregated load and individual household loads. The proposed LSTM deep learning model is evaluated using datasets from individual household smart meters and aggregated electrical demand statistics derived from publicly accessible sources. The performance of the proposed model is extensively compared to that of various benchmark forecasting methods. The comparison demonstrates that the suggested LSTM model outperforms other prediction methods for household level forecasting. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
