Smart Energy Management System: A Comparative Study of Energy Consumption Prediction Algorithms for a Hotel Building

With the increase of smart buildings in the world mainly hotels with a quantity of 700.000 hotels in the world, an energy management approach must be developed in order to prevent high energy consumption and put the electrical grid in unfavorable conditions. Therefore, energy researchers and managers are studying these building energy demand profiles in order to optimize and improve the energy efficiency in the building firstly and secondly in the grid. The goal of this paper is to provide different algorithms that predict the electrical energy consumption of hotels building, a case study of a hotel building in Shanghai, using machine learning algorithms in RapidMiner software, such as Support Vector Machine (SVM), Artificial Neural Network (ANN), Decision Tree (DT) and Random Forest (RF), describing the methodology required and the comparative study between them. By finding the best-fit algorithm, more investigation would be taken place for improvement in future works. These algorithms can firstly be used in simulation software as a data-driven model, such as Matlab, ETAP, DTS, PSCAD, and others. Secondly, energy managers can use them to estimate the annual and monthly cost of energy consumption which will allow them to find better solutions for power shedding and integrating Distributed Energy Resources (DER).