Buildings Energy Efficiency Analysis and Classification Using Various Machine Learning Technique Classifiers

Energies

César Benavente-Peces¹  Nisrine Ibadah² 

[1] ETS Ingeniería y Sistemas de Telecomunicación, Universidad Politécnica de Madrid, Calle de Nikola Tesla sn, 28031 Madrid, Spain;LRIT Laboratory, Associated Unit to CNRST (URAC 29), IT Rabat Center, Faculty of Sciences, Mohammed V University, Rabat 1014 RP, Morocco;

关键词: buildings energy efficiency;    smart cities;    smart buildings;    sustainability;    ICT;    machine learning;   

DOI  :  10.3390/en13133497

来源: DOAJ

【 摘 要 】

Energy efficiency is a major concern to achieve sustainability in modern society. Smart cities sustainability depends on the availability of energy-efficient infrastructures and services. Buildings compose most of the city, and they are responsible for most of the energy consumption and emissions to the atmosphere (40%). Smart cities need smart buildings to achieve sustainability goals. Building’s thermal modeling is essential to face the energy efficiency race. In this paper, we show how ICT and data science technologies and techniques can be applied to evaluate the energy efficiency of buildings. In concrete, we apply machine learning techniques to classify buildings based on their energy efficiency. Particularly, our focus is on single-family buildings in residential areas. Along this paper, we demonstrate the capabilities of machine learning techniques to classify buildings depending on their energy efficiency. Moreover, we analyze and compare the performance of different classifiers. Furthermore, we introduce new parameters which have some impact on the buildings thermal modeling, especially those concerning the environment where the building is located. We also make an insight on ICT and remark the growing relevance in data acquisition and monitoring of relevant parameters by using wireless sensor networks. It is worthy to remark the need for an appropriate and reliable dataset to achieve the best results. Moreover, we demonstrate that reliable classification is feasible with a few featured parameters.

【 授权许可】

Unknown