Skip to Main Content
 

Global Search Box

 
 
 

ETD Abstract Container

Abstract Header

Development of Building Markers and Unsupervised Non-intrusive Disaggregation Model for Commercial Buildings’ Energy Usage

Hossain, Mohammad Akram
http://rave.ohiolink.edu/etdc/view?acc_num=case1517225790921761

Abstract Details

2018, Doctor of Philosophy, Case Western Reserve University, EMC - Mechanical Engineering.
Energy Diagnostics Investigator for Efficiency Savings (EDIFES) is a scalable data an- alytics tool that uses big data, and rigorous statistical studies to uncover building en- ergy characteristics. To create EDIFES, building energy markers were developed using R and Python functions that compute various types of building identifiers when applied to whole building, 15-minute electricity data, as is that typically collected by the util- ity company. Requisite weather datasets also were analyzed in conjunction with the electricity consumption data. In this study, we developed nine building markers and applied them to 19 commercial buildings located in four different climate zones to com- pare their characteristics. The building markers are: correlation with weather variables, weekday-weekend operational pattern, weekday operational pattern, heating type, sys- tem oversize (heating), system oversize (cooling), HVAC scheduling, HVAC sizing, and baseload. Using the findings from this analysis, we developed a building energy disag- gregation model to further quantify a buildings’ energy usage. Building energy disaggre- gation can identify and estimate equipment-level energy scheduling and consumption which can provide real-time feedback to the customer. The disaggregation tool is unsu- pervised and non-intrusive, and again, uses only whole building electricity and weather datasets for the analysis. Therefore, the disaggregation model can perform the analysis virtually, without installing any sensors/meters in the building. The disaggregation tool is derived from the building markers and by utilizing Bayesian frameworks: the Hidden Markov model and the Factorial Hidden Markov model. The disaggregation tool esti- mates the equipment state with an accuracy of approximately 75% for a scheduled office building. The state of the HVAC can be estimated with the disaggregation tool with an accuracy of approximately 81%. We can conclude that the EDIFES analysis developed to-date and described herein demonstrates an unmatched capability to conduct virtual energy audits.
Alexis Abramson (Committee Chair)
Roger French (Committee Member)
Joseph Prahl (Committee Member)
Mehmet Koyuturk (Committee Member)
Rojiar Haddadian (Committee Member)
Mechanical Engineering
Energy efficiency, Building energy audit, Energy disaggregation

Recommended Citations

Citations

  • Hossain, M. A. (2018). Development of Building Markers and Unsupervised Non-intrusive Disaggregation Model for Commercial Buildings’ Energy Usage [Doctoral dissertation, Case Western Reserve University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=case1517225790921761

    APA Style (7th edition)

  • Hossain, Mohammad. Development of Building Markers and Unsupervised Non-intrusive Disaggregation Model for Commercial Buildings’ Energy Usage. 2018. Case Western Reserve University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=case1517225790921761.

    MLA Style (8th edition)

  • Hossain, Mohammad. "Development of Building Markers and Unsupervised Non-intrusive Disaggregation Model for Commercial Buildings’ Energy Usage." Doctoral dissertation, Case Western Reserve University, 2018. http://rave.ohiolink.edu/etdc/view?acc_num=case1517225790921761

    Chicago Manual of Style (17th edition)

case1517225790921761
315
© 2018, all rights reserved.
This open access ETD is published by Case Western Reserve University School of Graduate Studies and OhioLINK.
Release 3.2.12