Prediction of the Electrical Load for Egyptian Energy Management Systems: Deep Learning Approach

Abstract

In the era of the Internet of Things (IoT) and machine learning, energy utilities have now the possibility of providing many benefits to the customers. Using advanced metering infrastructure is an important step to optimize and manage energy consumption. Smart meters are deployed in millions of households worldwide, representing an opportunity for monitoring systems and data analytics. Egypt has started its path into smart metering and needs to use big data analytics to endow the Egyptian grid with intelligence. In this paper, we investigate the role of big data analytics and available machine learning technologies that mine the data to generate insights into smart grid management process. We investigate a deep learning-based methodology such as long short-term memory (LSTM) network model to perform prediction of the electrical load for energy management systems. The LSTM model presents a decrease of 14% in forecasting error compared to deep forward neural networks.