Interpretable Data-Driven Methods for Building Energy Modelling—A Review of Critical Connections and Gaps

Massimiliano Manfren (), Karla M. Gonzalez-Carreon and Patrick A. B. James
Additional contact information
Massimiliano Manfren: Faculty of Engineering and Physical Sciences, University of Southampton, Boldrewood Innovation Campus, Burgess Rd., Southampton SO16 7QF, UK
Karla M. Gonzalez-Carreon: Faculty of Engineering and Physical Sciences, University of Southampton, Boldrewood Innovation Campus, Burgess Rd., Southampton SO16 7QF, UK
Patrick A. B. James: Faculty of Engineering and Physical Sciences, University of Southampton, Boldrewood Innovation Campus, Burgess Rd., Southampton SO16 7QF, UK

Energies, 2024, vol. 17, issue 4, 1-22

Abstract: Technological improvements are crucial for achieving decarbonisation targets and addressing the impacts of climate change in the built environment via mitigation and adaptation measures. Data-driven methods for building performance prediction are particularly important in this regard. Nevertheless, the deployment of these technologies faces challenges, particularly in the domains of artificial intelligence (AI) ethics, interpretability and explainability of machine learning (ML) algorithms. The challenges encountered in applications for the built environment are amplified, particularly when data-driven solutions need to be applied throughout all the stages of the building life cycle and to address problems from a socio-technical perspective, where human behaviour needs to be considered. This requires a consistent use of analytics to assess the performance of a building, ideally by employing a digital twin (DT) approach, which involves the creation of a digital counterpart of the building for continuous analysis and improvement. This paper presents an in-depth review of the critical connections between data-driven methods, AI ethics, interpretability and their implementation in the built environment, acknowledging the complex and interconnected nature of these topics. The review is organised into three distinct analytical levels: The first level explores key issues of the current research on the interpretability of machine learning methods. The second level considers the adoption of interpretable data-driven methods for building energy modelling and the problem of establishing a link with the third level, which examines physics-driven grey-box modelling techniques, in order to provide integrated modelling solutions. The review’s findings highlight how the interpretability concept is relevant in multiple contexts pertaining to energy and the built environment and how some of the current knowledge gaps can be addressed by further research in the broad area of data-driven methods.

Keywords: energy transitions; energy modelling; building performance; data-driven methods; interpretability; explainability; digital twins (search for similar items in EconPapers)
JEL-codes: Q Q0 Q4 Q40 Q41 Q42 Q43 Q47 Q48 Q49 (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/1996-1073/17/4/881/pdf (application/pdf)
https://www.mdpi.com/1996-1073/17/4/881/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jeners:v:17:y:2024:i:4:p:881-:d:1338711

Access Statistics for this article

Energies is currently edited by Ms. Agatha Cao

More articles in Energies from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().