A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Design and implementation of an occupant-centered self-learning controller for decentralized residential ventilation systems
https://orcid.org/0000-0003-4015-236X
Abstract Mechanical ventilation systems acquired relevance in the past decades as a solution to guarantee air exchange in residential energy retrofit. The individualization of the control strategies to the user preferences could enhance the acceptance of innovative technologies and meet the energy efficiency targets for buildings. In this publication, a novel occupant-centered self-learning control strategy for decentralized ventilation systems is introduced. The controller has a standard-based default control profile, which learns the occupants’ preferences using a classification algorithm. The proposed controller is first simulated and compared to other available mechanical ventilation control strategies. A stochastic model for the manual operation of ventilation systems based on artificial comfort profiles is implemented in the simulation model. Results show that the proposed controller improves the indoor environmental conditions (measured with the relative humidity and CO2 concentration) without disregarding energy efficiency. The self-learning controller can successfully adapt itself to these profiles. The proposed solution was implemented in a real apartment. Measurements were carried out for three months, where two occupants operated the ventilation devices. The installed ventilation concept succeeded in maintaining an acceptable indoor environment, while the distinctive user profiles were learned in each room. The bedrooms were shifted towards lower air exchange rates, while in the rest of the rooms higher air exchange rates were preferred. This implementation confirms the individualization potential of this controller. Finally, the novel self-learning strategy provides an adaptive solution for residential decentralized ventilation controllers, targeting the occupant preferences regarding comfort, health and energy efficiency.
Highlights Decentralized ventilation systems are a suitable technology to guarantee air exchange rate in renovated multifamily buildings A novel occupant-centered control strategy for decentralized ventilation is proposed, based on the learning of the occupant preferences towards the relative humidity and CO2 concentration The simulation results show that the proposed controller has the potential to save energy and keep the indoor environmental quality in the desired range The self-learning control strategy was successfully implemented in a real building, where the mechanical ventilation operation behavior was recorded The self-learning controller could successfully adapt itself to the occupant preferences in every room, after learning from the occupants' manual operation.
Design and implementation of an occupant-centered self-learning controller for decentralized residential ventilation systems
https://orcid.org/0000-0003-4015-236X
Abstract Mechanical ventilation systems acquired relevance in the past decades as a solution to guarantee air exchange in residential energy retrofit. The individualization of the control strategies to the user preferences could enhance the acceptance of innovative technologies and meet the energy efficiency targets for buildings. In this publication, a novel occupant-centered self-learning control strategy for decentralized ventilation systems is introduced. The controller has a standard-based default control profile, which learns the occupants’ preferences using a classification algorithm. The proposed controller is first simulated and compared to other available mechanical ventilation control strategies. A stochastic model for the manual operation of ventilation systems based on artificial comfort profiles is implemented in the simulation model. Results show that the proposed controller improves the indoor environmental conditions (measured with the relative humidity and CO2 concentration) without disregarding energy efficiency. The self-learning controller can successfully adapt itself to these profiles. The proposed solution was implemented in a real apartment. Measurements were carried out for three months, where two occupants operated the ventilation devices. The installed ventilation concept succeeded in maintaining an acceptable indoor environment, while the distinctive user profiles were learned in each room. The bedrooms were shifted towards lower air exchange rates, while in the rest of the rooms higher air exchange rates were preferred. This implementation confirms the individualization potential of this controller. Finally, the novel self-learning strategy provides an adaptive solution for residential decentralized ventilation controllers, targeting the occupant preferences regarding comfort, health and energy efficiency.
Highlights Decentralized ventilation systems are a suitable technology to guarantee air exchange rate in renovated multifamily buildings A novel occupant-centered control strategy for decentralized ventilation is proposed, based on the learning of the occupant preferences towards the relative humidity and CO2 concentration The simulation results show that the proposed controller has the potential to save energy and keep the indoor environmental quality in the desired range The self-learning control strategy was successfully implemented in a real building, where the mechanical ventilation operation behavior was recorded The self-learning controller could successfully adapt itself to the occupant preferences in every room, after learning from the occupants' manual operation.
Design and implementation of an occupant-centered self-learning controller for decentralized residential ventilation systems
https://orcid.org/0000-0003-4015-236X
Abstract Mechanical ventilation systems acquired relevance in the past decades as a solution to guarantee air exchange in residential energy retrofit. The individualization of the control strategies to the user preferences could enhance the acceptance of innovative technologies and meet the energy efficiency targets for buildings. In this publication, a novel occupant-centered self-learning control strategy for decentralized ventilation systems is introduced. The controller has a standard-based default control profile, which learns the occupants’ preferences using a classification algorithm. The proposed controller is first simulated and compared to other available mechanical ventilation control strategies. A stochastic model for the manual operation of ventilation systems based on artificial comfort profiles is implemented in the simulation model. Results show that the proposed controller improves the indoor environmental conditions (measured with the relative humidity and CO2 concentration) without disregarding energy efficiency. The self-learning controller can successfully adapt itself to these profiles. The proposed solution was implemented in a real apartment. Measurements were carried out for three months, where two occupants operated the ventilation devices. The installed ventilation concept succeeded in maintaining an acceptable indoor environment, while the distinctive user profiles were learned in each room. The bedrooms were shifted towards lower air exchange rates, while in the rest of the rooms higher air exchange rates were preferred. This implementation confirms the individualization potential of this controller. Finally, the novel self-learning strategy provides an adaptive solution for residential decentralized ventilation controllers, targeting the occupant preferences regarding comfort, health and energy efficiency.
Highlights Decentralized ventilation systems are a suitable technology to guarantee air exchange rate in renovated multifamily buildings A novel occupant-centered control strategy for decentralized ventilation is proposed, based on the learning of the occupant preferences towards the relative humidity and CO2 concentration The simulation results show that the proposed controller has the potential to save energy and keep the indoor environmental quality in the desired range The self-learning control strategy was successfully implemented in a real building, where the mechanical ventilation operation behavior was recorded The self-learning controller could successfully adapt itself to the occupant preferences in every room, after learning from the occupants' manual operation.
Design and implementation of an occupant-centered self-learning controller for decentralized residential ventilation systems
Carbonare, Nicolas (author) / Pflug, Thibault (author) / Bongs, Constanze (author) / Wagner, Andreas (author)
Building and Environment ; 206
2021-09-19
Article (Journal)
Electronic Resource
English
Occupant-centered control strategies for decentralized residential ventilation
| BASE | 2021
Occupant-centered control strategies for decentralized residential ventilation
| UB Braunschweig | 2021
Occupant-centered control strategies for decentralized residential ventilation
| TIBKAT | 2021