Deep Reinforcement Learning Based HVAC Control for Reducing Carbon Footprint of Buildings

Kuldeep Kurte; Kadir Amasyali; Jeffrey Munk; Helia Zandi

doi:10.1109/ISGT51731.2023.10066358

Deep Reinforcement Learning Based HVAC Control for Reducing Carbon Footprint of Buildings

Kuldeep Kurte, Kadir Amasyali, Jeffrey Munk, Helia Zandi

Applied Research for Communities in Extreme Environments

Oak Ridge National Laboratory

Research output: Contribution to conference › Paper › peer-review

2 Scopus Citations

Abstract

In this paper, we present our work on deep reinforcement learning (DRL) based intelligent control of Heating, Ventilation, and Air Conditioning (HVAC) with the goal of reducing carbon emission. We performed this task using 1) Marginal Operating Emission Rates (MOER), where the objective was to shift the demand to the low emission period of the day and 2) Time-Of-Use (TOU) demand-response price where the objective was to shift the demand to low price period of the day. This was achieved by learning an optimal pre-cooing strategy. We found the carbon emission reduction in the range of ≈ 6%-16% depending on the opportunity presented by the MOER signal. Similarly, we observed the carbon emission reduction in the range of ≈23%-29% during the peak price period when TOU price was used. The results clearly demonstrated the applicability of our approach in reducing the carbon footprint of the building.

Original language	American English
Number of pages	5
DOIs	https://doi.org/10.1109/ISGT51731.2023.10066358 https://doi.org/10.1109/ISGT51731.2023.10066358
State	Published - 2023
Event	2023 IEEE Power and Energy Society Innovative Smart Grid Technologies Conference, ISGT 2023 - Washington, United States Duration: 16 Jan 2023 → 19 Jan 2023

Conference

Conference	2023 IEEE Power and Energy Society Innovative Smart Grid Technologies Conference, ISGT 2023
Country/Territory	United States
City	Washington
Period	16/01/23 → 19/01/23

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

NREL Publication Number

NREL/CP-5600-86288

Keywords

carbon emission
deep reinforcement learning
HVAC control

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

Cite this

Kurte, K., Amasyali, K., Munk, J., & Zandi, H. (2023). Deep Reinforcement Learning Based HVAC Control for Reducing Carbon Footprint of Buildings. Paper presented at 2023 IEEE Power and Energy Society Innovative Smart Grid Technologies Conference, ISGT 2023, Washington, United States. https://doi.org/10.1109/ISGT51731.2023.10066358, https://doi.org/10.1109/ISGT51731.2023.10066358

@conference{1ffb592303ef49179f8c47b9f4959dd2,

title = "Deep Reinforcement Learning Based HVAC Control for Reducing Carbon Footprint of Buildings",

abstract = "In this paper, we present our work on deep reinforcement learning (DRL) based intelligent control of Heating, Ventilation, and Air Conditioning (HVAC) with the goal of reducing carbon emission. We performed this task using 1) Marginal Operating Emission Rates (MOER), where the objective was to shift the demand to the low emission period of the day and 2) Time-Of-Use (TOU) demand-response price where the objective was to shift the demand to low price period of the day. This was achieved by learning an optimal pre-cooing strategy. We found the carbon emission reduction in the range of ≈ 6%-16% depending on the opportunity presented by the MOER signal. Similarly, we observed the carbon emission reduction in the range of ≈23%-29% during the peak price period when TOU price was used. The results clearly demonstrated the applicability of our approach in reducing the carbon footprint of the building.",

keywords = "carbon emission, deep reinforcement learning, HVAC control",

author = "Kuldeep Kurte and Kadir Amasyali and Jeffrey Munk and Helia Zandi",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 IEEE Power and Energy Society Innovative Smart Grid Technologies Conference, ISGT 2023 ; Conference date: 16-01-2023 Through 19-01-2023",

year = "2023",

doi = "10.1109/ISGT51731.2023.10066358",

language = "American English",

}

Kurte, K, Amasyali, K, Munk, J & Zandi, H 2023, 'Deep Reinforcement Learning Based HVAC Control for Reducing Carbon Footprint of Buildings', Paper presented at 2023 IEEE Power and Energy Society Innovative Smart Grid Technologies Conference, ISGT 2023, Washington, United States, 16/01/23 - 19/01/23. https://doi.org/10.1109/ISGT51731.2023.10066358, https://doi.org/10.1109/ISGT51731.2023.10066358

TY - CONF

T1 - Deep Reinforcement Learning Based HVAC Control for Reducing Carbon Footprint of Buildings

AU - Kurte, Kuldeep

AU - Amasyali, Kadir

AU - Munk, Jeffrey

AU - Zandi, Helia

PY - 2023

Y1 - 2023

N2 - In this paper, we present our work on deep reinforcement learning (DRL) based intelligent control of Heating, Ventilation, and Air Conditioning (HVAC) with the goal of reducing carbon emission. We performed this task using 1) Marginal Operating Emission Rates (MOER), where the objective was to shift the demand to the low emission period of the day and 2) Time-Of-Use (TOU) demand-response price where the objective was to shift the demand to low price period of the day. This was achieved by learning an optimal pre-cooing strategy. We found the carbon emission reduction in the range of ≈ 6%-16% depending on the opportunity presented by the MOER signal. Similarly, we observed the carbon emission reduction in the range of ≈23%-29% during the peak price period when TOU price was used. The results clearly demonstrated the applicability of our approach in reducing the carbon footprint of the building.

AB - In this paper, we present our work on deep reinforcement learning (DRL) based intelligent control of Heating, Ventilation, and Air Conditioning (HVAC) with the goal of reducing carbon emission. We performed this task using 1) Marginal Operating Emission Rates (MOER), where the objective was to shift the demand to the low emission period of the day and 2) Time-Of-Use (TOU) demand-response price where the objective was to shift the demand to low price period of the day. This was achieved by learning an optimal pre-cooing strategy. We found the carbon emission reduction in the range of ≈ 6%-16% depending on the opportunity presented by the MOER signal. Similarly, we observed the carbon emission reduction in the range of ≈23%-29% during the peak price period when TOU price was used. The results clearly demonstrated the applicability of our approach in reducing the carbon footprint of the building.

KW - carbon emission

KW - deep reinforcement learning

KW - HVAC control

UR - http://www.scopus.com/inward/record.url?scp=85151520469&partnerID=8YFLogxK

U2 - 10.1109/ISGT51731.2023.10066358

DO - 10.1109/ISGT51731.2023.10066358

M3 - Paper

AN - SCOPUS:85151520469

T2 - 2023 IEEE Power and Energy Society Innovative Smart Grid Technologies Conference, ISGT 2023

Y2 - 16 January 2023 through 19 January 2023

ER -

Deep Reinforcement Learning Based HVAC Control for Reducing Carbon Footprint of Buildings

Abstract

Conference

Bibliographical note

NREL Publication Number

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this