Data-driven Identification of Occupant Thermostat-Behavior Dynamics

2019-12-13

Building occupant behavior drives significant differences in building energy use, even in automated buildings. Users' distrust in the automation causes them to override settings. This results in responses that fail to satisfy both the occupants' and/or the building automation's objectives. The transition toward grid-interactive efficient buildings will make this evermore important as complex building control systems optimize not only for comfort, but also changing electricity costs. This paper presents a data-driven approach to study thermal comfort behavior dynamics which are not captured by standard steady-state comfort models such as predicted mean vote. The proposed model captures the time it takes for a user to override a thermostat setpoint change as a function of the manual setpoint change magnitude. The model was trained with the ecobee Donate Your Data dataset of 5 min. resolution data from 27,764 smart thermostats and occupancy sensors. The resulting population-level model shows that, on average, a 2{\deg}F override will occur after ~30 mins. and an 8{\deg}F override will occur in only ~15 mins., indicating the magnitude of discomfort as a key driver to the swiftness of an override. Such models could improve demand response programs through personalized controls.

📄 PDF Abstract BibTeX arXiv:1912.06705

Code (0)

등록된 구현이 없습니다.

Using Reinforcement Learning for Demand Response of Domestic Hot Water Buffers: a Real-Life Demonstration

2017-03-16 · Oscar De Somer, Ana Soares, Tristan Kuijpers, Koen Vossen 외

This paper demonstrates a data-driven control approach for demand response in real-life residential buildings. The objective is to optimally schedule the heating cycles of the Domestic Hot Water (DHW) buffer to maximize …

Decision Making Model-based Reinforcement Learning reinforcement-learning Reinforcement Learning+2

Identifying Grey-box Thermal Models with Bayesian Neural Networks

2020-09-13

Smart thermostats are one of the most prevalent home automation products. They learn occupant preferences and schedules, and utilize an accurate thermal model to reduce the energy use of heating and cooling equipment whi…

Time Series Analysis

Self-Excitation: An Enabler for Online Thermal Estimation and Model Predictive Control of Buildings

2015-12-27 · Peter Radecki, Brandon Hencey

This paper investigates a method to improve buildings' thermal predictive control performance via online identification and excitation (active learning process) that minimally disrupts normal operations. In previous stud…

Active Learning Model Predictive Control

Stochastic simulation of residential building occupant-driven energy use in a bottom-up model of the U.S. housing stock

2021-11-02 · Jianli Chen, Rajendra Adhikari, Eric Wilson, Joseph Robertson 외

The residential buildings sector is one of the largest electricity consumers worldwide and contributes disproportionally to peak electricity demand in many regions. Strongly driven by occupant activities at home, househo…

Comparative Field Deployment of Reinforcement Learning and Model Predictive Control for Residential HVAC

2025-10-01 · Ozan Baris Mulayim, Elias N. Pergantis, Levi D. Reyes Premer, Bingqing Chen 외 arxiv

Model Predictive Control (MPC) has demonstrated significant performance improvements over today's control methods for residential Heating, Ventilation, and Air Conditioning (HVAC), but deploying MPC often requires substa…

Reinforcement Learning