Measurement and verification of building systems under uncertain data: A Gaussian process modeling approach

Michael C. Burkhart, Yeonsook Heo, Victor M. Zavala

Research output: Contribution to journalArticle

27 Citations (Scopus)

Abstract

Uncertainty in sensor data (e.g., weather, occupancy) complicates the construction of baseline models for measurement and verification (M&V). We present a Monte Carlo expectation maximization (MCEM) framework for constructing baseline Gaussian process (GP) models under uncertain input data. We demonstrate that the GP-MCEM framework yields more robust predictions and confidence levels compared with standard GP training approaches that neglect uncertainty. We argue that the approach can also reduce data needs because it implicitly expands the data range used for training and can thus be used as a mechanism to reduce data collection and sensor installation costs in M&V processes. We analyze the numerical behavior of the framework and conclude that robust predictions can be obtained with relatively few samples.

Original languageEnglish
Pages (from-to)189-198
Number of pages10
JournalEnergy and Buildings
Volume75
DOIs
Publication statusPublished - 2014 Jan 1
Externally publishedYes

Fingerprint

Sensors
Costs
Uncertainty

Keywords

  • Data uncertainty
  • Expectation maximization
  • Gaussian process modeling
  • Measurement and verification

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Building and Construction
  • Mechanical Engineering
  • Electrical and Electronic Engineering

Cite this

Measurement and verification of building systems under uncertain data : A Gaussian process modeling approach. / Burkhart, Michael C.; Heo, Yeonsook; Zavala, Victor M.

In: Energy and Buildings, Vol. 75, 01.01.2014, p. 189-198.

Research output: Contribution to journalArticle

@article{fdae0e77e24a478fa5adf19c20798671,
title = "Measurement and verification of building systems under uncertain data: A Gaussian process modeling approach",
abstract = "Uncertainty in sensor data (e.g., weather, occupancy) complicates the construction of baseline models for measurement and verification (M&V). We present a Monte Carlo expectation maximization (MCEM) framework for constructing baseline Gaussian process (GP) models under uncertain input data. We demonstrate that the GP-MCEM framework yields more robust predictions and confidence levels compared with standard GP training approaches that neglect uncertainty. We argue that the approach can also reduce data needs because it implicitly expands the data range used for training and can thus be used as a mechanism to reduce data collection and sensor installation costs in M&V processes. We analyze the numerical behavior of the framework and conclude that robust predictions can be obtained with relatively few samples.",
keywords = "Data uncertainty, Expectation maximization, Gaussian process modeling, Measurement and verification",
author = "Burkhart, {Michael C.} and Yeonsook Heo and Zavala, {Victor M.}",
year = "2014",
month = "1",
day = "1",
doi = "10.1016/j.enbuild.2014.01.048",
language = "English",
volume = "75",
pages = "189--198",
journal = "Energy and Buildings",
issn = "0378-7788",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - Measurement and verification of building systems under uncertain data

T2 - A Gaussian process modeling approach

AU - Burkhart, Michael C.

AU - Heo, Yeonsook

AU - Zavala, Victor M.

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Uncertainty in sensor data (e.g., weather, occupancy) complicates the construction of baseline models for measurement and verification (M&V). We present a Monte Carlo expectation maximization (MCEM) framework for constructing baseline Gaussian process (GP) models under uncertain input data. We demonstrate that the GP-MCEM framework yields more robust predictions and confidence levels compared with standard GP training approaches that neglect uncertainty. We argue that the approach can also reduce data needs because it implicitly expands the data range used for training and can thus be used as a mechanism to reduce data collection and sensor installation costs in M&V processes. We analyze the numerical behavior of the framework and conclude that robust predictions can be obtained with relatively few samples.

AB - Uncertainty in sensor data (e.g., weather, occupancy) complicates the construction of baseline models for measurement and verification (M&V). We present a Monte Carlo expectation maximization (MCEM) framework for constructing baseline Gaussian process (GP) models under uncertain input data. We demonstrate that the GP-MCEM framework yields more robust predictions and confidence levels compared with standard GP training approaches that neglect uncertainty. We argue that the approach can also reduce data needs because it implicitly expands the data range used for training and can thus be used as a mechanism to reduce data collection and sensor installation costs in M&V processes. We analyze the numerical behavior of the framework and conclude that robust predictions can be obtained with relatively few samples.

KW - Data uncertainty

KW - Expectation maximization

KW - Gaussian process modeling

KW - Measurement and verification

UR - http://www.scopus.com/inward/record.url?scp=84897428690&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84897428690&partnerID=8YFLogxK

U2 - 10.1016/j.enbuild.2014.01.048

DO - 10.1016/j.enbuild.2014.01.048

M3 - Article

AN - SCOPUS:84897428690

VL - 75

SP - 189

EP - 198

JO - Energy and Buildings

JF - Energy and Buildings

SN - 0378-7788

ER -