Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings

Hannah C. Fontenot; Bing Dong; Zhi Zhou

doi:10.1007/978-981-19-9822-5_156

Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings

Hannah C. Fontenot, Bing Dong, Zhi Zhou

Department of Mechanical & Aerospace Engineering

Research output: Chapter in Book/Entry/Poem › Conference contribution

Abstract

Building model predictive control (MPC) relies on a white- or grey-box model that can require significant time and domain knowledge to develop and calibrate, often on an ad hoc basis. Black-box models can be developed and trained with limited domain knowledge and are easily transferable. In this study, a deep neural network is trained to predict indoor temperature response using a few easily obtained predictors. The trained network is embedded within an MPC framework, replacing a grey-box model, and this data-driven predictive controller (DDPC) is implemented in a facility consisting of two side-by-side identical office spaces. One office is controlled by DDPC; the other remains under automated control. Experimental results show that DDPC reduces energy consumption by up to 30% compared to baseline control while maintaining indoor temperature throughout the day. DDPC presents a scalable solution to the challenges associated with developing and implementing building MPC on a large scale.

Original language	English (US)
Title of host publication	Proceedings of the 5th International Conference on Building Energy and Environment
Editors	Liangzhu Leon Wang, Hua Ge, Mohamed Ouf, Zhiqiang John Zhai, Dahai Qi, Chanjuan Sun, Dengjia Wang
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	1509-1518
Number of pages	10
ISBN (Print)	9789811998218
DOIs	https://doi.org/10.1007/978-981-19-9822-5_156
State	Published - 2023
Event	5th International Conference on Building Energy and Environment, COBEE 2022 - Montreal, Canada Duration: Jul 25 2022 → Jul 29 2022

Publication series

Name	Environmental Science and Engineering
ISSN (Print)	1863-5520
ISSN (Electronic)	1863-5539

Conference

Conference	5th International Conference on Building Energy and Environment, COBEE 2022
Country/Territory	Canada
City	Montreal
Period	7/25/22 → 7/29/22

Keywords

Data-driven predictive control
Deep neural networks
HVAC energy savings
Model predictive control

ASJC Scopus subject areas

Environmental Engineering
Information Systems

Access to Document

10.1007/978-981-19-9822-5_156

Cite this

Fontenot, H. C., Dong, B., & Zhou, Z. (2023). Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings. In L. L. Wang, H. Ge, M. Ouf, Z. J. Zhai, D. Qi, C. Sun, & D. Wang (Eds.), Proceedings of the 5th International Conference on Building Energy and Environment (pp. 1509-1518). (Environmental Science and Engineering). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-19-9822-5_156

Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings. / Fontenot, Hannah C.; Dong, Bing; Zhou, Zhi.
Proceedings of the 5th International Conference on Building Energy and Environment. ed. / Liangzhu Leon Wang; Hua Ge; Mohamed Ouf; Zhiqiang John Zhai; Dahai Qi; Chanjuan Sun; Dengjia Wang. Springer Science and Business Media Deutschland GmbH, 2023. p. 1509-1518 (Environmental Science and Engineering).

Research output: Chapter in Book/Entry/Poem › Conference contribution

Fontenot, HC, Dong, B & Zhou, Z 2023, Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings. in LL Wang, H Ge, M Ouf, ZJ Zhai, D Qi, C Sun & D Wang (eds), Proceedings of the 5th International Conference on Building Energy and Environment. Environmental Science and Engineering, Springer Science and Business Media Deutschland GmbH, pp. 1509-1518, 5th International Conference on Building Energy and Environment, COBEE 2022, Montreal, Canada, 7/25/22. https://doi.org/10.1007/978-981-19-9822-5_156

Fontenot HC, Dong B, Zhou Z. Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings. In Wang LL, Ge H, Ouf M, Zhai ZJ, Qi D, Sun C, Wang D, editors, Proceedings of the 5th International Conference on Building Energy and Environment. Springer Science and Business Media Deutschland GmbH. 2023. p. 1509-1518. (Environmental Science and Engineering). doi: 10.1007/978-981-19-9822-5_156

Fontenot, Hannah C. ; Dong, Bing ; Zhou, Zhi. / Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings. Proceedings of the 5th International Conference on Building Energy and Environment. editor / Liangzhu Leon Wang ; Hua Ge ; Mohamed Ouf ; Zhiqiang John Zhai ; Dahai Qi ; Chanjuan Sun ; Dengjia Wang. Springer Science and Business Media Deutschland GmbH, 2023. pp. 1509-1518 (Environmental Science and Engineering).

@inproceedings{aaaab9cb4f1c4442866acb5eabf067f8,

title = "Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings",

abstract = "Building model predictive control (MPC) relies on a white- or grey-box model that can require significant time and domain knowledge to develop and calibrate, often on an ad hoc basis. Black-box models can be developed and trained with limited domain knowledge and are easily transferable. In this study, a deep neural network is trained to predict indoor temperature response using a few easily obtained predictors. The trained network is embedded within an MPC framework, replacing a grey-box model, and this data-driven predictive controller (DDPC) is implemented in a facility consisting of two side-by-side identical office spaces. One office is controlled by DDPC; the other remains under automated control. Experimental results show that DDPC reduces energy consumption by up to 30% compared to baseline control while maintaining indoor temperature throughout the day. DDPC presents a scalable solution to the challenges associated with developing and implementing building MPC on a large scale.",

keywords = "Data-driven predictive control, Deep neural networks, HVAC energy savings, Model predictive control",

author = "Fontenot, {Hannah C.} and Bing Dong and Zhi Zhou",

note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.; 5th International Conference on Building Energy and Environment, COBEE 2022 ; Conference date: 25-07-2022 Through 29-07-2022",

year = "2023",

doi = "10.1007/978-981-19-9822-5_156",

language = "English (US)",

isbn = "9789811998218",

series = "Environmental Science and Engineering",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "1509--1518",

editor = "Wang, {Liangzhu Leon} and Hua Ge and Mohamed Ouf and Zhai, {Zhiqiang John} and Dahai Qi and Chanjuan Sun and Dengjia Wang",

booktitle = "Proceedings of the 5th International Conference on Building Energy and Environment",

address = "Germany",

}

TY - GEN

T1 - Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings

AU - Fontenot, Hannah C.

AU - Dong, Bing

AU - Zhou, Zhi

PY - 2023

Y1 - 2023

N2 - Building model predictive control (MPC) relies on a white- or grey-box model that can require significant time and domain knowledge to develop and calibrate, often on an ad hoc basis. Black-box models can be developed and trained with limited domain knowledge and are easily transferable. In this study, a deep neural network is trained to predict indoor temperature response using a few easily obtained predictors. The trained network is embedded within an MPC framework, replacing a grey-box model, and this data-driven predictive controller (DDPC) is implemented in a facility consisting of two side-by-side identical office spaces. One office is controlled by DDPC; the other remains under automated control. Experimental results show that DDPC reduces energy consumption by up to 30% compared to baseline control while maintaining indoor temperature throughout the day. DDPC presents a scalable solution to the challenges associated with developing and implementing building MPC on a large scale.

AB - Building model predictive control (MPC) relies on a white- or grey-box model that can require significant time and domain knowledge to develop and calibrate, often on an ad hoc basis. Black-box models can be developed and trained with limited domain knowledge and are easily transferable. In this study, a deep neural network is trained to predict indoor temperature response using a few easily obtained predictors. The trained network is embedded within an MPC framework, replacing a grey-box model, and this data-driven predictive controller (DDPC) is implemented in a facility consisting of two side-by-side identical office spaces. One office is controlled by DDPC; the other remains under automated control. Experimental results show that DDPC reduces energy consumption by up to 30% compared to baseline control while maintaining indoor temperature throughout the day. DDPC presents a scalable solution to the challenges associated with developing and implementing building MPC on a large scale.

KW - Data-driven predictive control

KW - Deep neural networks

KW - HVAC energy savings

KW - Model predictive control

UR - http://www.scopus.com/inward/record.url?scp=85172739416&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85172739416&partnerID=8YFLogxK

U2 - 10.1007/978-981-19-9822-5_156

DO - 10.1007/978-981-19-9822-5_156

M3 - Conference contribution

AN - SCOPUS:85172739416

SN - 9789811998218

T3 - Environmental Science and Engineering

SP - 1509

EP - 1518

BT - Proceedings of the 5th International Conference on Building Energy and Environment

A2 - Wang, Liangzhu Leon

A2 - Ge, Hua

A2 - Ouf, Mohamed

A2 - Zhai, Zhiqiang John

A2 - Qi, Dahai

A2 - Sun, Chanjuan

A2 - Wang, Dengjia

PB - Springer Science and Business Media Deutschland GmbH

T2 - 5th International Conference on Building Energy and Environment, COBEE 2022

Y2 - 25 July 2022 through 29 July 2022

ER -

Data-Driven Predictive Control (DDPC) with Deep Neural Networks for Building Energy Savings

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this