Dynamic indoor thermal environment using Reinforcement Learning-based controls: Opportunities and challenges

Building and Environment, Journal Year: 2023, Volume and Issue: 244, P. 110766 - 110766

Published: Aug. 26, 2023

Currently, the indoor thermal environment in many buildings is controlled by conventional control techniques that maintain temperature within a prescribed deadband. The latest research provides evidence more dynamic variations of can promote health and trigger positive alliesthesia making changes still comfortable. But such an requires flexible responsive system adapt to real-time. As emerging technique, Reinforcement Learning (RL) has attracted growing interest demonstrated its potential enhance building performance while addressing some limitations other advanced techniques. Thus, comprehensive review explored boundaries possibilities apply RL for controls suitable varying environment. first part discussed studies on permissible limits step acceptable drifts human occupants. It also debated flexibility range comfort adaptation. In next part, HVAC were explored, focusing their application creating different algorithms, systems, co-simulation environment, action spaces, energy-saving potentials discussed. Overall, based review, this work outlined pathway RL-based controller dynamically vary temperature. Suitable environmental parameters be controlled, choice algorithm, space, are

Language: Английский

---

Model-Free HVAC Control in Buildings: A Review

Energies, Journal Year: 2023, Volume and Issue: 16(20), P. 7124 - 7124

Published: Oct. 17, 2023

The efficient control of HVAC devices in building structures is mandatory for achieving energy savings and comfort. To balance these objectives efficiently, it essential to incorporate adequate advanced strategies adapt varying environmental conditions occupant preferences. Model-free approaches systems have gained significant interest due their flexibility ability complex, dynamic without relying on explicit mathematical models. current review presents the recent advancements control, with an emphasis reinforcement learning, artificial neural networks, fuzzy logic hybrid integration other model-free algorithms. main focus this study a literature most notable research from 2015 2023, highlighting highly cited applications contributions field. After analyzing concept each work according its strategy, detailed evaluation across different thematic areas conducted. end, prevalence methodologies, utilization equipment, diverse testbed features, such as zoning utilization, are further discussed considering entire body identify patterns trends field control. Last but not least, based field, provides future directions aspects areas.

Language: Английский

---

Integrating few-shot personalized thermal comfort model and reinforcement learning for HVAC demand response optimization

Journal of Building Engineering, Journal Year: 2024, Volume and Issue: 91, P. 109509 - 109509

Published: May 5, 2024

Language: Английский

---

Systematic Review on Deep Reinforcement Learning-Based Energy Management for Different Building Types

Energies, Journal Year: 2022, Volume and Issue: 15(22), P. 8663 - 8663

Published: Nov. 18, 2022

Owing to the high energy demand of buildings, which accounted for 36% global share in 2020, they are one core targets energy-efficiency research and regulations. Hence, coupled with increasing complexity decentralized power grids renewable penetration, inception smart buildings is becoming increasingly urgent. Data-driven building management systems (BEMS) based on deep reinforcement learning (DRL) have attracted significant interest, particularly recent years, primarily owing their ability overcome many challenges faced by conventional control methods related real-time modelling, multi-objective optimization, generalization BEMS efficient wide deployment. A PRISMA-based systematic assessment a large database 470 papers was conducted review advancements DRL-based different types, directions, knowledge gaps. Five types were identified: residential, offices, educational, data centres, other commercial buildings. Their comparative analysis appliances controlled BEMS, integration, DR, unique system objectives than energy, such as cost, comfort. Moreover, it worth considering that only approximately 11% considers real implementations.

Language: Английский

---

Coordinated voltage regulation of high renewable-penetrated distribution networks: An evolutionary curriculum-based deep reinforcement learning approach

International Journal of Electrical Power & Energy Systems, Journal Year: 2023, Volume and Issue: 149, P. 108995 - 108995

Published: Feb. 23, 2023

Language: Английский

---

Prediction approach on pedestrian outdoor activity preference under factors of public open space integrated microclimate

Building and Environment, Journal Year: 2023, Volume and Issue: 244, P. 110761 - 110761

Published: Aug. 23, 2023

Language: Английский

---

Experimental study on the physiological parameters of occupants under different temperatures and prediction of their thermal comfort using machine learning algorithms

Journal of Building Engineering, Journal Year: 2024, Volume and Issue: 84, P. 108676 - 108676

Published: Jan. 29, 2024

Language: Английский

---

Modeling the mood state on thermal sensation with a data mining algorithm and testing the accuracy of mood state correction factor

New Ideas in Psychology, Journal Year: 2024, Volume and Issue: 76, P. 101124 - 101124

Published: Sept. 25, 2024

Language: Английский

---

Prospects and Challenges of Reinforcement Learning- Based HVAC Control

Journal of Building Engineering, Journal Year: 2024, Volume and Issue: unknown, P. 111080 - 111080

Published: Oct. 1, 2024

Language: Английский

---

Sex-based thermal comfort zones and energy savings in spaces with joint operation of air conditioner and fan

Building and Environment, Journal Year: 2023, Volume and Issue: 246, P. 111002 - 111002

Published: Nov. 1, 2023

Language: Английский

---