skip to main content
10.1145/3563357.3566140acmconferencesArticle/Chapter ViewAbstractPublication PagesbuildsysConference Proceedingsconference-collections
research-article

XENIA: eXplainable ENergy Informatics and Attributes for building energy benchmarking

Published: 08 December 2022 Publication History

Abstract

Benchmarking energy usage help identify operational and strategic best practices suitable for an establishment while creating awareness of energy consumption. Therefore in this work, we present XENIA, a data-driven energy benchmarking methodology for buildings in Singapore using a public dataset of building attributes. We develop an ensemble tree model to predict energy consumption using the building attributes as predictors. Symmetric mean absolute percentage error of these models for hotel and retail buildings is 5.15% and 5.02%, respectively. A benchmark grade is then assigned to each building using the actual and predicted energy consumption. To interpret the model, we provide a global explanation using the partial dependence function to show the effect of building attributes on energy consumption. For local explanation, i.e., for a specific building, we use the SHAP value to show the influence of each building attribute in the prediction model. The results for hotels and retail buildings show that change in AC and non-AC floor has the highest positive impact on energy consumption.

References

[1]
Pandarasamy Arjunan, Kameshwar Poolla, and Clayton Miller. 2020. EnergyStar++: Towards more accurate and explanatory building energy benchmarking. Applied Energy 276 (2020), 115413.
[2]
Pandarasamy Arjunan, Kameshwar Poolla, and Clayton Miller. 2022. BEEM: Data-driven building energy benchmarking for Singapore. Energy and Buildings 260 (2022), 111869.
[3]
Building and Construction Authority. Singapore. 2018. BESS (Building Energy Submission System). Retrieved August 9, 2022 from https://www.bca.gov.sg/BESS/BenchmarkingReport/BenchmarkingReport.aspx
[4]
Jerome H. Friedman. 2001. Greedy Function Approximation: A Gradient Boosting Machine. The Annals of Statistics 29, 5 (2001), 1189--1232. http://www.jstor.org/stable/2699986
[5]
Antonio Galli, Marco Savino Piscitelli, Vincenzo Moscato, and Alfonso Capozzoli. 2022. Bridging the gap between complexity and interpretability of a data analytics-based process for benchmarking energy performance of buildings. Expert Systems with Applications 206 (2022), 117649.
[6]
Matheus Soares Geraldi and Enedir Ghisi. 2022. Data-driven framework towards realistic bottom-up energy benchmarking using an Artificial Neural Network. Applied Energy 306 (2022), 117960.
[7]
Matheus Soares Geraldi, Veronica Martins Gnecco, Antonio Barzan Neto, Bárbara Augusta de Mafra Martins, Enedir Ghisi, Michele Fossati, Ana Paula Melo, and Roberto Lamberts. 2021. Evaluating the impact of the shape of school reference buildings on bottom-up energy benchmarking. Journal of Building Engineering 43 (2021), 103142.
[8]
Veronica Martins Gnecco, Matheus Soares Geraldi, Michele Fossati, and Maria Andrea Triana. 2022. Comparison between national and local benchmarking models: The case of public nursery schools in Southern Brazil. Sustainable Cities and Society 78 (2022), 103639.
[9]
Xiaoyu Jin, Fu Xiao, Chong Zhang, and Ao Li. 2022. GEIN: An interpretable benchmarking framework towards all building types based on machine learning. Energy and Buildings 260 (2022), 111909.
[10]
Xue Liu, Hao Tang, Yong Ding, and Da Yan. 2022. Investigating the performance of machine learning models combined with different feature selection methods to estimate the energy consumption of buildings. Energy and Buildings 273 (2022), 112408.
[11]
Scott M. Lundberg and Su-In Lee. 2017. A Unified Approach to Interpreting Model Predictions. In Proceedings of the 31st International Conference on Neural Information Processing Systems (Long Beach, California, USA) (NIPS'17). Curran Associates Inc., Red Hook, NY, USA, 4768--4777.
[12]
MATHWORKS. 2022. Ensemble Algorithms. Retrieved July 7, 2022 from https://in.mathworks.com/help/stats/ensemble-algorithms.html
[13]
Sokratis Papadopoulos and Constantine E. Kontokosta. 2019. Grading buildings on energy performance using city benchmarking data. Applied Energy 233--234 (2019), 244--253.
[14]
United Nations Environment Programme. 2021. 2021 Global Status Report for Buildings and Construction: Towards a Zero-emission, Efficient and Resilient Buildings and Construction Sector. Nairobi. https://globalabc.org/sites/default/files/2021-10/GABC_Buildings-GSR-2021_BOOK.pdf
[15]
Irfan Qaisar and Qianchuan Zhao. 2022. Energy baseline prediction for buildings: A review. Results in Control and Optimization 7 (2022), 100129.

Cited By

View all
  • (2024)Explaining Deep Learning-based Anomaly Detection in Energy Consumption Data by Focusing on Contextually Relevant DataEnergy and Buildings10.1016/j.enbuild.2024.115177(115177)Online publication date: Dec-2024

Recommendations

Comments

Information & Contributors

Information

Published In

cover image ACM Conferences
BuildSys '22: Proceedings of the 9th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation
November 2022
535 pages
ISBN:9781450398909
DOI:10.1145/3563357
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]

Sponsors

Publisher

Association for Computing Machinery

New York, NY, United States

Publication History

Published: 08 December 2022

Permissions

Request permissions for this article.

Check for updates

Author Tags

  1. energy benchmarking
  2. ensemble model
  3. explainable AI
  4. machine learning
  5. partial dependence
  6. shapley value

Qualifiers

  • Research-article

Conference

BuildSys '22
Sponsor:

Acceptance Rates

Overall Acceptance Rate 148 of 500 submissions, 30%

Contributors

Other Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

  • Downloads (Last 12 months)17
  • Downloads (Last 6 weeks)2
Reflects downloads up to 18 Feb 2025

Other Metrics

Citations

Cited By

View all
  • (2024)Explaining Deep Learning-based Anomaly Detection in Energy Consumption Data by Focusing on Contextually Relevant DataEnergy and Buildings10.1016/j.enbuild.2024.115177(115177)Online publication date: Dec-2024

View Options

Login options

View options

PDF

View or Download as a PDF file.

PDF

eReader

View online with eReader.

eReader

Figures

Tables

Media

Share

Share

Share this Publication link

Share on social media