Power optimization based on fuzzy inference system and IoT. Case study: University of Rwanda, College of Science and Technology, AGACIRO block building

Abstract

Electricity plays an outstanding role in daily life of a human being, from lightened homes, lightened roads, offices, schools, working and recreational areas. For electricity efficiency, measures and techniques are deployed to manage its usage. With the rise of new technologies and concepts, IoT and Machine learning concepts contributed a lot in optimizing many functions and systems and same techniques are hereby deployed to optimize power consumption. This research was conducted in University of Rwanda, College of Science and Technology, Agaciro Block building and it deploys an IoT system which yields the data of total room occupancy, the presence of electric current and voltage, the ambient room temperature and humidity, the presence and quantity of light and the presence of vibrations. The system has another part of Machine Leaning based Fuzzy Logic controller which is trained by using the fetched sensory data, hence a Fuzzy Inference System having Sets and Rules. The Agaciro block building has offices, conference rooms, computer labs, Server room, mechanical & chemical and soil testing laboratories, but this research only considers a single room as its scope. With the help of sensory prototype results, Matlab fuzzy logics based simulations, and the Matlab Simulink experimental results, this research proves its ability of optimizing the overall power consumption, as there is an an automation of existing campus power system to work under both supervised and unsupervised learning as the sensory data are used to train the machine learning based Fuzzy Logic Controller. The evaluation results show that the proposed optimization system save energy compared to the existing current power systems. Keywords: Electricity optimization, IoT, Fuzzy Logics, Fuzzy Inference System, Supervised learning, Unsupervised Learning