II-New: A Programmable Data-Driven Testbed for Sustainable Buildings Research

今天的建筑消耗了美国近40%的总能源和75%的电力。因此，利用计算、传感和通信技术，通过减少能源使用、成本和占地面积，使建筑物变得更加智能和绿色，变得越来越重要。 当使用计算技术使建筑物更智能和更环保时，出现了许多研究问题。 例如，我们如何通过将太阳能和风能等可再生能源纳入建筑物能源管理系统来“绿化”建筑物？我们如何通过自动化容易出错的手动任务来减少能源使用，例如关闭恒温器？ 不幸的是，目前的研究是这一领域面临着两个关键的障碍：缺乏从真实的建筑物的能源使用数据的访问，以及由于安全和舒适的原因，无法在现场建筑物进行实验研究。 该项目旨在通过建立基础设施，从真实的家庭中收集匿名能源使用数据用于研究目的，以及建立可编程模拟建筑，以实现智能能源管理算法的实验，从而减轻这些障碍。该项目将建立一个可编程数据驱动的测试台，以实现一系列以可持续建筑为重点的研究和教育活动。我们的试验台将由两个组件组成：数据收集组件，用于对大量各种类型的真实的建筑物进行检测（例如，单户住宅、多户住宅、公寓、公共住房和小型办公楼），以收集匿名的细粒度实际能源使用和大规模运行数据;以及可编程智能建筑组件，其包括可编程电源和电气负载，以使各种能源管理机制能够在“模拟”智能建筑内以安全和可重复的方式进行评估。该试验台将实现各种新颖的研究和教育活动，如设计智能负载管理技术和采用此类技术的强烈用户激励;将可再生能源和储能整合到智能建筑的能源管理机制中;使用学习和建模来推动能源数据的分析，以获得丰富的见解，以及智能电网技术的隐私机制的设计，如智能电表;以及，利用该平台进行实践教育和推广到广泛的社区。该领域的感兴趣的研究人员将有机会获得真实的数据集和一个试验台，与当地电力公司合作进行实证研究。 因此，该测试平台的研究有可能直接影响真实的公用事业的网格管理的操作和效率。有一个重要的社区组件，可以与更广泛的研究社区共享来自我们的仪器建筑的匿名数据以及设计规范。

Today's buildings consume nearly 40% of the total energy, and 75% of the electricity, produced by the United States. Consequently, the use of computing, sensing, and communication to make buildings smarter and greener by reducing their energy usage, costs, and footprint has become increasingly important. Many research questions arise when using computational techniques for making buildings smarter and greener. For example, how do we ``green'' buildings by integrating renewable sources, such as solar and wind, into the building energy management system? How do we reduce energy usage by automating error-prone manual tasks, such as turning down a thermostat? Unfortunately, current research is this area faces two key barriers: lack of access to energy usage data from real buildings, and an inability to conduct experimental research in live buildings due to reasons of safety and comfort. This project focuses on alleviating such barriers by building infrastructure to collect anonymous energy usage data from real homes for research purposes, as well as building a programmable mock building to enable experimentation with smart energy management algorithms.This project will build a programmable data-driven test-bed to enable a range of research and education activities focused on sustainable buildings. Our test-bed will be comprised of two components: a data collection component to instrument a large number of real buildings of various types (e.g., single family homes, multifamily dwellings, apartments, public housing and small office buildings) to gather anonymous fine-grained actual energy usage and operational data at scale; and a programmable smart building component that comprises programmable power sources and electrical loads to enable a variety of energy management mechanisms to be evaluated in safe and repeatable manner inside a ``mock" smart building. The test-bed will enable a variety of novel research and education activities like designing intelligent load management techniques and strong user incentives for adopting such technologies; integrating renewables and energy storage into a smart building's energy management mechanisms; using learning and modeling to drive analytics of energy data to derive rich insights, as well as the design of privacy mechanisms for smart grid technologies, such as smart metering; and, using the platform for hands-on education and outreach to a broad community.Interested researchers in this field will have access to realistic data sets and a test-bed to conduct empirical research in collaboration with local electric utility. As a result, the research enabled by the test-bed has the potential to directly impact the operations and efficiency of a real utility's grid management. There is a significant community component to share anonymized data from our instrumented buildings, as well as the design specifications with the broader research community.