skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Scalable tuning of building models to hourly data

Abstract

Energy models of existing buildings are unreliable unless calibrated so they correlate well with actual energy usage. Manual tuning requires a skilled professional, is prohibitively expensive for small projects, imperfect, non-repeatable, non-transferable, and not scalable to the dozens of sensor channels that smart meters, smart appliances, and cheap/ubiquitous sensors are beginning to make available today. A scalable, automated methodology is needed to quickly and intelligently calibrate building energy models to all available data, increase the usefulness of those models, and facilitate speed-and-scale penetration of simulation-based capabilities into the marketplace for actualized energy savings. The "Autotune'' project is a novel, model-agnostic methodology which leverages supercomputing, large simulation ensembles, and big data mining with multiple machine learning algorithms to allow automatic calibration of simulations that match measured experimental data in a way that is deployable on commodity hardware. This paper shares several methodologies employed to reduce the combinatorial complexity to a computationally tractable search problem for hundreds of input parameters. Furthermore, accuracy metrics are provided which quantify model error to measured data for either monthly or hourly electrical usage from a highly-instrumented, emulated-occupancy research home.

Authors:
 [1];  [2]
  1. Jacksonville State Univ., Jacksonville, AL (United States)
  2. Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
Publication Date:
Research Org.:
Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
Sponsoring Org.:
USDOE Office of Energy Efficiency and Renewable Energy (EERE), Building Technologies Office (EE-5B)
OSTI Identifier:
1185507
Alternate Identifier(s):
OSTI ID: 1247854
Grant/Contract Number:
AC05-00OR22725; BT0201000
Resource Type:
Journal Article: Accepted Manuscript
Journal Name:
Energy (Oxford)
Additional Journal Information:
Journal Name: Energy (Oxford); Journal Volume: 84; Journal ID: ISSN 0360-5442
Publisher:
Elsevier
Country of Publication:
United States
Language:
English
Subject:
32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATION; 97 MATHEMATICS AND COMPUTING; autotune; energy plus; calibration; optimization; evolutionary computation

Citation Formats

Garrett, Aaron, and New, Joshua Ryan. Scalable tuning of building models to hourly data. United States: N. p., 2015. Web. doi:10.1016/j.energy.2015.03.014.
Garrett, Aaron, & New, Joshua Ryan. Scalable tuning of building models to hourly data. United States. doi:10.1016/j.energy.2015.03.014.
Garrett, Aaron, and New, Joshua Ryan. Tue . "Scalable tuning of building models to hourly data". United States. doi:10.1016/j.energy.2015.03.014. https://www.osti.gov/servlets/purl/1185507.
@article{osti_1185507,
title = {Scalable tuning of building models to hourly data},
author = {Garrett, Aaron and New, Joshua Ryan},
abstractNote = {Energy models of existing buildings are unreliable unless calibrated so they correlate well with actual energy usage. Manual tuning requires a skilled professional, is prohibitively expensive for small projects, imperfect, non-repeatable, non-transferable, and not scalable to the dozens of sensor channels that smart meters, smart appliances, and cheap/ubiquitous sensors are beginning to make available today. A scalable, automated methodology is needed to quickly and intelligently calibrate building energy models to all available data, increase the usefulness of those models, and facilitate speed-and-scale penetration of simulation-based capabilities into the marketplace for actualized energy savings. The "Autotune'' project is a novel, model-agnostic methodology which leverages supercomputing, large simulation ensembles, and big data mining with multiple machine learning algorithms to allow automatic calibration of simulations that match measured experimental data in a way that is deployable on commodity hardware. This paper shares several methodologies employed to reduce the combinatorial complexity to a computationally tractable search problem for hundreds of input parameters. Furthermore, accuracy metrics are provided which quantify model error to measured data for either monthly or hourly electrical usage from a highly-instrumented, emulated-occupancy research home.},
doi = {10.1016/j.energy.2015.03.014},
journal = {Energy (Oxford)},
number = ,
volume = 84,
place = {United States},
year = {Tue Mar 31 00:00:00 EDT 2015},
month = {Tue Mar 31 00:00:00 EDT 2015}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Citation Metrics:
Cited by: 3 works
Citation information provided by
Web of Science

Save / Share: