Journal
APPLIED ENERGY
Volume 301, Issue -, Pages -Publisher
ELSEVIER SCI LTD
DOI: 10.1016/j.apenergy.2021.117502
Keywords
Building energy retrofit; Measurement and verification; Data driven approach; Generalized additive models; Building energy performance; Energy savings estimation
Categories
Funding
- European Commission through the H2020 project SENSEI [847066]
- H2020 Societal Challenges Programme [847066] Funding Source: H2020 Societal Challenges Programme
Ask authors/readers for more resources
The study proposes a novel data-driven methodology for measuring energy efficiency savings in commercial buildings through clustering techniques and innovative technology to provide accurate dynamic energy savings estimates.
Methods to obtain accurate estimations of the savings generated by building energy efficiency interventions are a topic of great importance, and considered to be one of the keys to increase capital investments in energy conservation strategies worldwide. In this study, a novel data-driven methodology is proposed for the measurement and verification of energy efficiency savings, with special focus on commercial buildings and facilities. The presented approach involves building use characterization by means of a clustering technique that allows to extract typical consumption profile patterns. These are then used, in combination with an innovative technique to evaluate the building's weather dependency, to design a model able to provide accurate dynamic estimations of the achieved energy savings. The method was tested on synthetic datasets generated using the building energy simulation software EnergyPlus, as well as on monitoring data from real-world buildings. The results obtained with the proposed methodology were compared with the ones provided by applying the time-of-week-and-temperature (TOWT) model, showing up to 10% CV(RMSE) improvement, depending on the case in analysis. Furthermore, a comparison with the deterministic results provided by EnergyPlus showed that the median estimated savings error was always lower than 3% of the total reporting period consumption, with similar accuracy retained even when reducing the total training data available.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available