Reinforcement Learning for Mixing Loop Control with Flow Variable Eligibility Trace

Anders Overgaard; Brian Kongsgaard Nielsen; Carsten Kallesøe; Jan Dimon Bendtsen

doi:10.1109/CCTA.2019.8920398

Reinforcement Learning for Mixing Loop Control with Flow Variable Eligibility Trace

Anders Overgaard, Brian Kongsgaard Nielsen, Carsten Kallesøe, Jan Dimon Bendtsen

Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

6 Citations (Scopus)

Abstract

Mixing Loops are often used for proper pressurization and temperature control in building thermal systems. Optimal control of the mixing loop maximizes comfort while minimizing cost. To ensure optimal control for mixing loops in a wide range of different buildings with different load conditions, a self learning controller is here proposed. The controller uses Reinforcement Learning with flow variable eligibility trace. The controller is shown to improve performance of the mixing loop control compared to state of the art reinforcement learning and industrial grade controllers. The controller is tested on a hardware in the loop setup for rapid testing of mixing loop control used in building heating.

Original language	English
Title of host publication	IEEE Conference on Control Technology and Applications (CCTA)
Number of pages	6
Publisher	IEEE
Publication date	2019
Pages	1043-1048
Article number	8920398
ISBN (Print)	978-1-7281-2768-2
ISBN (Electronic)	978-1-7281-2767-5
DOIs	https://doi.org/10.1109/CCTA.2019.8920398
Publication status	Published - 2019
Event	2019 IEEE Conference on Control Technology and Applications (CCTA) - Hong Kong, China Duration: 19 Aug 2019 → 21 Aug 2019

Conference

Conference	2019 IEEE Conference on Control Technology and Applications (CCTA)
Country/Territory	China
City	Hong Kong
Period	19/08/2019 → 21/08/2019

Access to Document

10.1109/CCTA.2019.8920398

Cite this

@inproceedings{f2a515c9a2294c5f9a3b4563c7a6f23d,

title = "Reinforcement Learning for Mixing Loop Control with Flow Variable Eligibility Trace",

abstract = "Mixing Loops are often used for proper pressurization and temperature control in building thermal systems. Optimal control of the mixing loop maximizes comfort while minimizing cost. To ensure optimal control for mixing loops in a wide range of different buildings with different load conditions, a self learning controller is here proposed. The controller uses Reinforcement Learning with flow variable eligibility trace. The controller is shown to improve performance of the mixing loop control compared to state of the art reinforcement learning and industrial grade controllers. The controller is tested on a hardware in the loop setup for rapid testing of mixing loop control used in building heating.",

author = "Anders Overgaard and Nielsen, {Brian Kongsgaard} and Carsten Kalles{\o}e and Bendtsen, {Jan Dimon}",

year = "2019",

doi = "10.1109/CCTA.2019.8920398",

language = "English",

isbn = "978-1-7281-2768-2",

pages = "1043--1048",

booktitle = "IEEE Conference on Control Technology and Applications (CCTA)",

publisher = "IEEE",

address = "United States",

note = "2019 IEEE Conference on Control Technology and Applications (CCTA), CCTA ; Conference date: 19-08-2019 Through 21-08-2019",

}

Overgaard, A, Nielsen, BK, Kallesøe, C & Bendtsen, JD 2019, Reinforcement Learning for Mixing Loop Control with Flow Variable Eligibility Trace. in IEEE Conference on Control Technology and Applications (CCTA)., 8920398, IEEE, pp. 1043-1048, 2019 IEEE Conference on Control Technology and Applications (CCTA), Hong Kong, China, 19/08/2019. https://doi.org/10.1109/CCTA.2019.8920398

Reinforcement Learning for Mixing Loop Control with Flow Variable Eligibility Trace. / Overgaard, Anders; Nielsen, Brian Kongsgaard; Kallesøe, Carsten et al.

IEEE Conference on Control Technology and Applications (CCTA). IEEE, 2019. p. 1043-1048 8920398.

Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

TY - GEN

T1 - Reinforcement Learning for Mixing Loop Control with Flow Variable Eligibility Trace

AU - Overgaard, Anders

AU - Nielsen, Brian Kongsgaard

AU - Kallesøe, Carsten

AU - Bendtsen, Jan Dimon

PY - 2019

Y1 - 2019

N2 - Mixing Loops are often used for proper pressurization and temperature control in building thermal systems. Optimal control of the mixing loop maximizes comfort while minimizing cost. To ensure optimal control for mixing loops in a wide range of different buildings with different load conditions, a self learning controller is here proposed. The controller uses Reinforcement Learning with flow variable eligibility trace. The controller is shown to improve performance of the mixing loop control compared to state of the art reinforcement learning and industrial grade controllers. The controller is tested on a hardware in the loop setup for rapid testing of mixing loop control used in building heating.

AB - Mixing Loops are often used for proper pressurization and temperature control in building thermal systems. Optimal control of the mixing loop maximizes comfort while minimizing cost. To ensure optimal control for mixing loops in a wide range of different buildings with different load conditions, a self learning controller is here proposed. The controller uses Reinforcement Learning with flow variable eligibility trace. The controller is shown to improve performance of the mixing loop control compared to state of the art reinforcement learning and industrial grade controllers. The controller is tested on a hardware in the loop setup for rapid testing of mixing loop control used in building heating.

UR - http://www.scopus.com/inward/record.url?scp=85077787911&partnerID=8YFLogxK

U2 - 10.1109/CCTA.2019.8920398

DO - 10.1109/CCTA.2019.8920398

M3 - Article in proceeding

SN - 978-1-7281-2768-2

SP - 1043

EP - 1048

BT - IEEE Conference on Control Technology and Applications (CCTA)

PB - IEEE

T2 - 2019 IEEE Conference on Control Technology and Applications (CCTA)

Y2 - 19 August 2019 through 21 August 2019

ER -

Reinforcement Learning for Mixing Loop Control with Flow Variable Eligibility Trace

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this