Human decision making during eco-feedback intervention in smart and connected energy-aware communities
Human decision making during eco-feedback intervention in smart and connected energy-aware communities
复制标题
DOI：
10.1016/j.enbuild.2022.112627
复制
发表时间：
2022-11-16
影响因子：
6.7
通讯作者：
Braun，James E.
中科院分区：
文献类型：
--
作者：
Kim，Huijeong;Bilionis，Ilias;Braun，James E.
供暖和制冷 (HC) 能源使用量约占美国普通家庭年能源消耗总量的 42%，并且受居民能源相关行为的显着影响。在本文中，我们的目标是为具有能源意识的社区实现一种新范例，利用智能生态反馈设备和社交游戏让居民了解并减少家庭 HC 能源使用。为了实现这一目标，我们提出了一种基于效用理论的新社会技术建模方法，以揭示人类决策中的因果效应，并推断在生态反馈干预期间影响每个家庭恒温器调节行为的属性。我们的方法 1) 基于实用模型，根据与热环境和生态反馈设计相关的决策属性，量化居民对室内温度的偏好；2) 纳入确定每个家庭独特行为特征的潜在参数。对于参数学习，我们开发了一个分层贝叶斯模型，该模型具有非中心参数化，使用从印第安纳州韦恩堡的多单元住宅社区收集的现场数据进行校准。数据集包括两部分；没有任何行为干预的基线期，以及通过具有智能恒温器控制功能的居民参与设备（包括壁挂式平板电脑和智能扬声器）部署个性化生态反馈和社交游戏的干预期。通过模型校准，我们量化了生态反馈对家庭恒温器调节行为的影响。我们建议，这项工作中开发的实用模型可以作为分析具有生态反馈节能项目的互联住宅社区居民行为的基础。
Heating and cooling (HC) energy use is responsible for about 42% of the total annual energy consumption of the average household in the U.S and it is significantly affected by residents’ energy-related behavior. In this paper, our goal is to realize a new paradigm for energy-aware communities that leverages smart eco-feedback devices and social games to engage residents in understanding and reducing their home HC energy use. Towards this goal, we present a new sociotechnical modeling approach based on utility theory to reveal causal effects in human decision-making and infer attributes affecting the thermostat adjustment behavior of each household during an eco-feedback intervention. Our approach 1) is based on a utility model that quantifies residents’ preferences over indoor temperatures given decision attributes related to their thermal environment and eco-feedback design and 2) incorporates latent parameters that determine the unique behavioral characteristics of each household. For parameter learning, we develop a hierarchical Bayesian model with non-centered parameterization calibrated using field data collected from a multi-unit residential community located in Fort Wayne, IN. The dataset comprises two parts; a baseline period without any behavioral intervention, and an intervention period, where personalized eco-feedback and social games are deployed through resident engagement devices with smart thermostat control capabilities, including a wall-mounted tablet and smart speaker. Through the model calibration, we quantify the impact of the eco-feedback on households’ thermostat-adjustment behaviors. We propose that the utility model developed in this work can serve as the foundation for analyzing resident behavior in connected residential communities with eco-feedback energy-saving programs.