Personalizing thermal environment using digital technologies: towards greater occupant wellbeing and decarbonization of buildings

建筑产生了全球近40%的碳排放，这在很大程度上是由空间供暖和制冷的能源需求驱动的，以提供居住者的舒适度。热环境对居住者的舒适、健康和生产效率有很大的影响。因此，为了实现加拿大2050年的净零排放目标，在确保热满意环境的同时减少建筑能耗至关重要。拟议的研究重点是开发研究工具，以显著减少能源使用的个性化供暖和制冷体验。特别是，将研究(1)互联网连接的个人舒适系统（PCS）和(2)个人舒适模型（PCM）在建筑设计和运营中的整合。本提案的范围以办公楼为目标。三个短期目标是：(1)表征影响乘员舒适度、健康和生产力的多模态（感知、行为、生理和认知）热反应；(2)开发可解释的机器学习（ML）方法，提高PCM的透明度和保真度；(3)开发一种数字孪生驱动的仿真工具，以测试和评估集成的HVAC和PCS控制，以实现深度节能。首先，拟议的研究将与人类受试者进行一系列实验室实验，通过各种物联网传感器创建丰富的多模态热响应数据集。该数据集将用于开发一个开源的数字人体模型，用于集成到现有的设计和仿真软件中，以告知与乘员舒适、健康和生产力相关的关键决策。接下来，将开发用于可解释ML的特定领域方法，以提高对热舒适和行业采用数据驱动的环境控制的基本理解。将提出通用指标，以规范热舒适ML模型的评估和基准。最后，将创建传感器、模型和系统的数字双胞胎，以实现集成HVAC和PCS控制策略的实际测试和评估。将确定一套最优控制策略，以促进低能耗、低碳建筑运行，同时提高建筑中的居住者舒适度。上述研究将由申请人牵头，由8名高素质人才（HQP）执行，其中包括3名同等学历的博士、3名硕士和2名本科生。学员将发展(1)对热舒适的基本理解，(2)低能耗建筑技术的实践经验，以及(3)智能建筑分析和控制应用的尖端数字技能。研究结果将有可能大大推进以居住者为中心的建筑设计和运营，减少建筑对环境的影响，同时提高加拿大不同社区的生活质量和繁荣。

Buildings generate nearly 40% of global carbon emissions, which is largely driven by energy demand for space heating and cooling to provide occupant comfort. Thermal environment has a great influence on occupant's comfort, health and productivity. Therefore, to meet Canada's 2050 net zero emissions target, it is critical to reduce building energy consumption while ensuring thermally satisfactory environment. The proposed research focuses on developing research tools to personalize heating and cooling experiences with significantly less energy use. In particular, the integration of (1) Internet-connected personal comfort system (PCS) and (2) personal comfort models (PCM) in building design and operation will be investigated. The scope of this proposal targets office buildings. Three short-term objectives are to: (1) characterize multimodal (perceptual, behavioural, physiological, and cognitive) thermal responses that affect occupant comfort, health, and productivity; (2) develop interpretable machine learning (ML) methods to improve transparency and fidelity of PCM; and (3) develop a digital-twin powered simulation tool to test and evaluate integrated HVAC and PCS controls for deep energy savings. First, the proposed research will conduct a series of laboratory experiments with human subjects to create a rich dataset of multimodal thermal responses via various IoT sensors. The dataset will be used to develop an open-source digital human model for the integration into existing design and simulation software to inform key decisions related to occupant comfort, health, and productivity. Next, domain-specific methods for interpretable ML will be developed to improve fundamental understanding about thermal comfort and industry adoption of data-driven environmental control. Common metrics will be proposed to standardize evaluation and benchmarking of ML models in thermal comfort. Lastly, digital twins of sensors, models, and systems will be created to enable realistic testing and evaluation of integrated HVAC and PCS control strategies. A set of optimal control strategies will be identified to promote low-energy, low-carbon building operation while improving occupant comfort in buildings. The above research will be led by the applicant and executed by 8 highly qualified personnel (HQP), including 3 equivalent PhD, 3 master's, and 2 undergraduate students. Trainees will develop (1) fundamental understanding about thermal comfort, (2) hands-on experience with low-energy building technologies, and (3) cutting-edge digital skills for smart building analytics and control applications. The research results will have the potential to greatly advance occupant-centric building design and operation and reduce environmental impacts of buildings while improving the quality of life and prosperity of diverse communities in Canada.