Comparison of model-based and data-driven approaches for modeling energy and comfort management systems, with a case study

In the last decades, the field of buildings had a great impact on the final energy consumption (up to 45%). Several efforts have been devoted to achieve the energy saving and optimization challenges that the European Union proposed. New approaches arose in the field of modeling and simulation. In this paper we propose a comparison between a standard physical white box modeling approach relying on the use of the well-known software TERMOLOG and an innovative data-driven black box modeling approach based on a Random Forest model. We present the results of the application of both methodologies on a case study. The aim of this work was to simulate the energy consumption together with some environmental signals such as the room temperature, the humidity and the CO2 concentration, providing the weather forecast and the comfort levels definition (desired room temperature and building usage habits). In the end, a deep analysis led to the statement of the advantages and limitations of both approaches.