EAGER: Modeling and Control of COVID-19 Transmission in Indoor Environments

尽管关于COVID-19还有很多未知之处，但观察性研究表明，微小的冠状病毒颗粒可以通过空气传播，因此可以在空中停留几个小时，传播距离远远超过社交距离的6英尺指导。因此，这增加了在封闭的室内空间（如餐厅、办公室和宿舍）暴露于病毒的风险。虽然它们通常不是为大流行情况设计的，但如果设计和控制得当，供暖、通风和空调（HVAC）系统可以显着减轻室内空间的疾病传播。麻省理工学院和三菱电机研究实验室（MERL）之间的合作研究项目协同控制理论，流体动力学和机器学习的研究活动，以实现建筑环境HVAC系统的优化设计和控制，从而最大限度地减少居住者的暴露风险。该项目将建立一个新的理论框架，采用现实的疾病传播的计算模型，以及室内热流体动力学的暖通空调系统的优化控制。这将为业主和建筑物管理人员提供暖通空调操作的一般指导方针。该研究将把拟议的框架与数据驱动的降阶模型相结合，并采用强化学习技术开发计算上易于处理的算法，用于HVAC系统的自适应在线控制，以遏制冠状病毒的传播。 首先，该项目将通过系统研究不同级别的建模复杂性并分析其对载有病原体的液滴和气溶胶传输的影响，解决制定可靠的COVID-19安全指南所需的最低建模复杂性水平的不确定性。其次，该项目将开发一个最佳控制框架，研究病毒传播的流动物理学和室内空间的一般气流。第三，本研究将开发一个数据驱动的自适应控制框架，可用于设计即插即用的HVAC系统控制器。 该研究将为设计特定病例感染控制指南的改进方法奠定基础，从而产生显著的社会和经济效益，这些指南可以通过负担得起的HVAC设备（如空调和风扇）实现。该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的知识价值和更广泛的影响审查标准进行评估来支持。

Although much remains unknown about COVID-19, observational studies have demonstrated that tiny coronavirus particles can become airborne and hence, remain aloft in air for a few hours and travel distances much longer than the 6-feet guidance of the social distancing. This consequently increases the risk of exposure to the virus in confined indoor spaces, such as, restaurants, offices, and dormitories. Although they are generally not designed for pandemic situations, heating, ventilation, and air-conditioning (HVAC) systems, if properly designed and controlled, can significantly mitigate the spread of diseases in indoor spaces. This collaborative research project between MIT and Mitsubishi Electric Research Laboratories (MERL) synergies research activities in control theory, fluid dynamics, and machine learning to enable optimal design and control of HVAC systems for built environments so as to minimize the exposure risk of occupants. The project will establish a new theoretical framework for optimal control of HVAC systems employing realistic computational models of the disease transmission as well as the indoors thermofluid dynamics. This will provide business owners and building managers with general guidelines for the HVAC operations. The research will integrate the proposed framework with data-driven, reduced-order models and employ reinforcement learning techniques to develop computationally tractable algorithms for adaptive, online control of the HVAC systems in order to contain the spread of the coronavirus. First, the project will resolve the uncertainty that exists about the minimum level of modeling complexity required for developing reliable, COVID-19 safety guidelines by systematic study of different levels of modeling complexities and analyzing their effects on the transport of pathogen-laden droplets and aerosols. Second, the project will develop an optimal control framework that studies the flow physics of the virus transmission and the general airflow in the indoor space. Third, this research will develop a data driven, adaptive control framework that can be used to design plug and play controllers for HVAC systems. The proposed research will have significant social and economic benefits by setting the foundation for an improved methodology in designing case-specific infection control guidelines that can be realized by affordable HVAC devices, such as, ventilators and fans.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.