An Artificial Intelligence Energy Management System For An Education Building

Abstract

The study of energy demand and consumption has become a topic of increasing importance as a result of the growing interest in energy sustainability. In the present energy crisis of South Africa, where Eskom is currently the only electricity provider, it cannot always meet the electricity needs of the customers. Therefore, the consideration of electrical power savings in education buildings can play a huge role, which was implemented in this study. University campuses represent a specific group of various buildings with significant energy demands and consumptions. Due to the various buildings, a University campus can be seen as a small town on its own, therefore, it offers an excellent test bed to monitor energy consumption and to understand the demand for electricity of the different buildings. In addition, it was possible to predict with an artificial intelligence concept using different prediction models when peak load will occur and to determine a maximum demand. A suitable database for the Engineering Technology Building (ETB) at the Central University of Technology (CUT), Free State, was created and available data of the electricity energy usage were collected and analysed for this purpose, with the aid of utilising methods, namely Moving Average, Straight Line and Kalman Filter. The available data were tested and evaluated by the switchgear, according to the priority list, and it proved to be successful. There was a total saving of approximately 5.83% on the power consumption per day. This proved to work well and a greater percentage of savings could be achieved by switching another circuit breaker according to the priority list.