An official website of the United States government
We present the design and implementation of RECA, a novel human-centric recommender system for co-optimizing energy consumption, comfort and air quality in commercial buildings. Existing works generally optimize these objectives separately, or by only controlling energy consuming resources within the building without directly engaging occupants. We develop a deep reinforcement learning architecture based on multitask learning, demonstrate how it can be used to jointly learn energy savings, comfort and air quality improvements for different actions, and build a recommender system with humans-in-the-loop. Through real deployments in multiple commercial buildings, we found that RECA has the potential to further reduce energy consumption by up to in energy-focused optimization, improve all objectives by in joint optimization, and improve thermal comfort by up to in comfort and air quality focused optimization, over existing solutions.
more »
« less
A Deep Reinforcement Learning Based Recommender System for Occupant-Driven Energy Optimization in Commercial Buildings
In this work, we present recEnergy, a recommender system for reducing energy consumption in commercial buildings with human-in-the-loop. We formulate the building energy optimization problem as a Markov Decision Process, show how deep reinforcement learning can be used to learn energy saving recommendations, and effectively engage occupants in energy-saving actions. is a recommender system that learns actions with high energy saving potential, actively distribute recommendations to occupants in a commercial building, and utilize feedback from the occupants to learn better energy saving recommendations. Over a four week user study, four different types of energy saving recommendations were trained and learned. improves building energy reduction from a baseline saving (passive-only strategy) of 19% to 26%.
more »
« less
- PAR ID:
- 10168836
- Date Published:
- Journal Name:
- IEEE Internet of Things Journal
- ISSN:
- 2372-2541
- Page Range / eLocation ID:
- 1 to 1
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Large-scale concrete 3D printing and digital construction has brought enormous potential to expand the design space of building components (e.g., building envelope) for the integration of multiple architectural functionalities including energy saving. In this research, a modular 3D printed vertical concrete green wall system – namely the 3D-VtGW, was developed. The 3D-VtGW envelope was assembled with prefabricated (3D printed) multifunctional wall modular elements, which serves as the enclosure of the building as well as the backbone for a green wall system to improve building’s energy efficiency. Using this design concept and large-scale concrete 3D printing, a prototype commercial building was built in Nanjing, China. To quantify the energy-saving potential of the 3D-VtGW system, a thermal network model was developed to simulate the thermal behavior of buildings with 3D-VtGW system and for thermal comfort analysis. Whole-building energy simulation was carried out using Chinese Standard Weather Data (CSWD) o Nanjing, China. The simulation results indicate that the building with 3D-VtGW exhibited prominent potential for energy saving and improved thermal comfort. The integrated greenery system in 3D-VtGW largely reduces wall exterior surface temperature and through-wall heat flux via the combined effects of plant shading, evapotranspiration, and heat storage from soil. This study presents the immense opportunities brought by digital fabrication and construction to extend the design space and function integration in buildings.more » « less
-
null; null (Ed.)“Smart” buildings that can sense and detect people’s presence have been in use for the past few decades, mostly using technologies that trigger reactive responses such as turning on/off heating/ventilating, lighting, security, etc. We argue that to be considered truly smart, buildings must become “aware” about the locations and activities of their inhabitants so they can proactively engage with the occupants and inform their decision making with respect to which actions to execute, by whom and where. To help assess the potential impact of “aware” buildings on their occupants, we are developing a multi-agent simulation-powered building management system that can sense human and building assets, extrapolate patterns of utilization, simulate what-if scenarios and suggest changes to user activities and resource allocation to maximize specific Key Performance Indicators (KPIs). The system is able to evaluate the implications of potential conflict resolution strategies and account for individual and collaborative activities of different types of users in semantically rich environments. Sensing in our case is based on Visible Light Communication (VLC) technology, embedded in a building’s LED lighting system. It can detect the actors, where they are located and what they do. To understand what happens in each space at any given time the information derived from the VLC system is combined with models of users’ activity schedules, profiles, and space affordances. We demonstrate our approach by hypothetically applying it to a Cardiac Catheterization Laboratory (CCL). The CCL is high-intensity hospital unit, second only to the Emergency Department in terms of the urgency of the cases it must handle. An aware building will help both patients and staff to allocate their (always scarce) resources more efficiently, saving time and alleviating stress.more » « less
-
null (Ed.)One of the major barriers to closing the energy efficiency gap is the failure to successfully inform the population about measures to conserve energy. This paper introduces the design of a mobile application developed to improve energy conservation of residential buildings by informing occupants of transferrable energy efficient green features in a green-certified, non-residential building. The application was developed to investigate dissemination of transferable energy saving practices to explore spillover effects from non-residential to residential buildings. Our research aims to capitalize on such spillover effects to narrow the energy efficiency gap.more » « less
-
In recent years, a new line of research has taken an interventional view of recommender systems, where recommendations are viewed as actions that the system takes to have a desired effect. This interventional view has led to the development of counterfactual inference techniques for evaluating and optimizing recommendation policies. This article explains how these techniques enable unbiased offline evaluation and learning despite biased data, and how they can inform considerations of fairness and equity in recommender systems.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1109/JIOT.2020.2974848
