Eur. Phys. J. Appl. Phys. 84, 10902 (2018)
https://doi.org/10.1051/epjap/2018180176

Regular Article

Analysis of factors affecting the performance of BIPV panels^★

¹ Department of Génie-Civil, ENS Paris-Saclay, Université Paris Saclay, France
² Department of Architecture and Civil Engineering, City University of Hong Kong, Hong Kong (SAR)

^a e-mail: ericlee@cityu.edu.hk

Received: 12 June 2018
Received in final form: 12 September 2018
Accepted: 13 September 2018
Published online: 28 November 2018

Abstract

We explore different methods of analyzing large and complex datasets related to building-integrated photovoltaics (BIPV). We use the data of the European RESSOURCES project obtained from ETNA, an experimental setup consisting of two full-scale replicas of residential homes featuring a double-skin facade. We show that classic data mining methods such as mutual information can be used to gain a better understanding of the physics behind BIPV systems and to highlight discrepancies between different experimental setups. We then use artificial neural networks to model the airflow inside a double-skin facade and quantify its contribution to the cooling and heating of buildings.