Back to Search
Start Over
Forecasting Electricity Consumption for Accurate Energy Management in Commercial Buildings With Deep Learning Models to Facilitate Demand Response Programs.
- Source :
-
Electric Power Components & Systems . 2024, Vol. 52 Issue 9, p1636-1651. 16p. - Publication Year :
- 2024
-
Abstract
- In the context of rapidly increasing energy demands and environmental concerns, optimizing energy management in commercial buildings is a critical challenge. Smart grids, empowered by advanced Energy Management Systems (EMS), play a pivotal role in addressing this challenge through Demand Side Management (DSM). However, the efficiency of DSM-based building EMS is often limited by the accuracy of load forecasting. This paper addresses this gap by exploring load forecasting models within DSM-based building EMS, focusing on a case study in a commercial building in Ankara, Turkey. Employing Deep Learning (DL) models for load forecasting, we provide inputs for rule-based controllers to enhance energy efficiency. Our major contribution is the development of the ANFIS-IC algorithm, aimed at maximizing demand response participation in commercial buildings. ANFIS-IC, integrating ANFIS controllers with LSTM-based load consumption forecasts, leads to a 33.14% reduction in energy consumption and a 39.22% decrease in energy costs, surpassing the performance of rule-based controllers alone which achieve reductions of 25.34% in energy consumption and 34.03% in energy costs. These findings not only highlight the potential of integrating rule-based controllers with deep learning algorithms but also underscore the importance of accurate load forecasting in improving the effectiveness of DSM-based building EMS. [ABSTRACT FROM AUTHOR]
Details
- Language :
- English
- ISSN :
- 15325008
- Volume :
- 52
- Issue :
- 9
- Database :
- Academic Search Index
- Journal :
- Electric Power Components & Systems
- Publication Type :
- Academic Journal
- Accession number :
- 176405421
- Full Text :
- https://doi.org/10.1080/15325008.2024.2317353