An Intelligent Energy Management System with an Efficient IoT based Deep Learning Framework

Efficient and economical energy utilization is ensured using green energy management systems that currently exist. However, integration of this technology with the Internet of Things (IoT) and edge intelligence is not completely explored. A smart energy management system with a deep learning framewo...

Full description

Saved in:
Bibliographic Details
Published in2022 International Conference on Sustainable Computing and Data Communication Systems (ICSCDS) pp. 33 - 37
Main Authors Bazil Wilfred, C., George, Santhi M, Sivaranjani, S., Selvan, Saravana, Feros Khan, J. Mohammed, Beulah David, D
Format Conference Proceeding
LanguageEnglish
Published IEEE 07.04.2022
Subjects
Cloud computing
Deep learning
Edge computing
Energy forecasting
Energy management
Internet of Things
Long short-term memory
Performance evaluation
Predictive models
Servers
Smart grids
Online AccessGet full text
DOI10.1109/ICSCDS53736.2022.9760757

Cover

More Information
Summary:Efficient and economical energy utilization is ensured using green energy management systems that currently exist. However, integration of this technology with the Internet of Things (IoT) and edge intelligence is not completely explored. A smart energy management system with a deep learning framework is presented in this paper to address the requirements of energy management in smart industries, homes and grids. An efficient communication is established between the consumers and energy distributors while predicting the future energy consumptions over short time intervals. With least error rate and reduced time complexity, a smart energy management system with optimal normalization model selection and cloud-based data supervising server for energy forecasting in IoT and edge devices is introduced. Communication between the smart grids and the edge devices in the IoT networks connected to a common cloud server regarding efficient energy demand and response features occur in a secure and continuous manner. Short-term energy requirement forecasting is performed with an efficient decision making algorithm while using various preprocessing techniques to manage the electricity data which is of diverse nature. This model is implemented in resource constrained devices and shows promising outcomes. For commercial and residential datasets, the proposed system offers reduced mean-square error (MSE) and root MSE (RMSE) values.
DOI:10.1109/ICSCDS53736.2022.9760757