The University of Southampton
University of Southampton Institutional Repository

From annual benchmarks to dynamic behaviour: a scalable multi-level analysis framework for energy performance in buildings

From annual benchmarks to dynamic behaviour: a scalable multi-level analysis framework for energy performance in buildings
From annual benchmarks to dynamic behaviour: a scalable multi-level analysis framework for energy performance in buildings
Attaining net-zero emissions targets for a building stock requires rigorous evaluation of building performance to establish tailored decarbonisation strategies. This paper introduces a multi-level, top-down modelling framework for evaluating individual building energy performance through three distinct temporal scales: yearly benchmarking, monthly interval regressions, and sub-hourly interval analysis for occupancy-related energy consumption. The framework is demonstrated using the Highfield Campus at the University of Southampton (UK), analysing data from 35 buildings monitored from 2017 to 2023, categorised across 9 end-use benchmarks. Yearly analysis utilising CIBSE benchmarks evaluated performance variability Pre-, During-, and Post-COVID pandemic, indicating the impact of occupancy pattern changes after weather normalisation. Monthly interpretable change-point regression model analysis was employed to identify heating balance points and temperature dependence. Observed changes in balance point temperature were attributed to reduced internal gains from lower occupancy. Finally, sub-hourly electricity data provided deeper insights into behavioural and operational variability, capturing the impact of hybrid working on building usage. To illustrate the framework, two buildings are discussed. Yearly data suggested no pandemic-related impact, yet granular analysis revealed otherwise. Monthly results Post-COVID showed heating-related reductions, with the thermal balance point decreasing by 1.8 °C and 2.5 °C respectively. Sub-hourly electricity usage Post-COVID during occupied hours was 107 points lower and 53 points lower respectively, with greater variability (around 5 points). Key findings underscore that granular analysis provides valuable insights missed by yearly benchmarks, enabling targeted efficiency measures. Further research should focus on integrating these into BMS to support decarbonisation strategies.
Data-driven energy modelling, Decarbonisation, Energy Analytics, Energy Efficiency, Interpretable machine-learning, Top-down energy modelling
0378-7788
Gonzalez Carreon, Karla M.
83974d01-7128-4e9a-9895-edff89cd9a31
Manfren, Massimiliano
f2b8c02d-cb78-411d-aed1-c4d056365392
Ridett, Ellis
ed32cf3e-5ce3-4596-ba28-0e7bc3794b18
James, Patrick A.B.
da0be14a-aa63-46a7-8646-a37f9a02a71b
Gonzalez Carreon, Karla M.
83974d01-7128-4e9a-9895-edff89cd9a31
Manfren, Massimiliano
f2b8c02d-cb78-411d-aed1-c4d056365392
Ridett, Ellis
ed32cf3e-5ce3-4596-ba28-0e7bc3794b18
James, Patrick A.B.
da0be14a-aa63-46a7-8646-a37f9a02a71b

Gonzalez Carreon, Karla M., Manfren, Massimiliano, Ridett, Ellis and James, Patrick A.B. (2026) From annual benchmarks to dynamic behaviour: a scalable multi-level analysis framework for energy performance in buildings. Energy and Buildings, 353, [116894]. (doi:10.1016/j.enbuild.2025.116894).

Record type: Article

Abstract

Attaining net-zero emissions targets for a building stock requires rigorous evaluation of building performance to establish tailored decarbonisation strategies. This paper introduces a multi-level, top-down modelling framework for evaluating individual building energy performance through three distinct temporal scales: yearly benchmarking, monthly interval regressions, and sub-hourly interval analysis for occupancy-related energy consumption. The framework is demonstrated using the Highfield Campus at the University of Southampton (UK), analysing data from 35 buildings monitored from 2017 to 2023, categorised across 9 end-use benchmarks. Yearly analysis utilising CIBSE benchmarks evaluated performance variability Pre-, During-, and Post-COVID pandemic, indicating the impact of occupancy pattern changes after weather normalisation. Monthly interpretable change-point regression model analysis was employed to identify heating balance points and temperature dependence. Observed changes in balance point temperature were attributed to reduced internal gains from lower occupancy. Finally, sub-hourly electricity data provided deeper insights into behavioural and operational variability, capturing the impact of hybrid working on building usage. To illustrate the framework, two buildings are discussed. Yearly data suggested no pandemic-related impact, yet granular analysis revealed otherwise. Monthly results Post-COVID showed heating-related reductions, with the thermal balance point decreasing by 1.8 °C and 2.5 °C respectively. Sub-hourly electricity usage Post-COVID during occupied hours was 107 points lower and 53 points lower respectively, with greater variability (around 5 points). Key findings underscore that granular analysis provides valuable insights missed by yearly benchmarks, enabling targeted efficiency measures. Further research should focus on integrating these into BMS to support decarbonisation strategies.

Text
Ener_Build_Gonzalez_2026_author_submission - Accepted Manuscript
Restricted to Repository staff only until 21 December 2027.
Request a copy

More information

Accepted/In Press date: 19 December 2025
e-pub ahead of print date: 20 December 2025
Published date: 15 February 2026
Keywords: Data-driven energy modelling, Decarbonisation, Energy Analytics, Energy Efficiency, Interpretable machine-learning, Top-down energy modelling

Identifiers

Local EPrints ID: 509459
URI: http://eprints.soton.ac.uk/id/eprint/509459
ISSN: 0378-7788
PURE UUID: 2e233d65-090a-4230-875a-611b7ba3ec29
ORCID for Massimiliano Manfren: ORCID iD orcid.org/0000-0003-1438-970X
ORCID for Ellis Ridett: ORCID iD orcid.org/0000-0002-1903-7175
ORCID for Patrick A.B. James: ORCID iD orcid.org/0000-0002-2694-7054

Catalogue record

Date deposited: 23 Feb 2026 17:54
Last modified: 24 Feb 2026 03:07

Export record

Altmetrics

Contributors

Author: Karla M. Gonzalez Carreon
Author: Ellis Ridett ORCID iD

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of http://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×