Software Defined Buildings

Abstract:

The SDB project seeks to design, engineer, and evaluate the foundational information substrate for cyberphysical systems in a concrete, canonical form - creation of efficient, agile, model- driven, human-centered building systems. Modern commercial buildings provide increasingly integrated Building Management Systems, but are typically closed or based on proprietary interfaces, are difficult to extend, and it is expensive to add new capabilities. Key underlying technology trends driving building system design are the rapidly declining costs of new cyber technologies, e.g., ubiquitous wireless communications, mobile devices, cheap embedded processing, and scalable processing and storage in the Cloud. The research challenge is how to extend the building "machine," to harness these trends, making it better, more capable, and more efficient. Our approach is to develop software-defined buildings, to shatter existing stovepipe architectures, dramatically reduce the effort to add new functions and applications without "forklift upgrades," and expand communications and control capabilities beyond a single stand- alone building to enable groups of buildings to behave cooperatively and in cooperation with the energy grid. We will investigate how such Software-Defined Buildings can be founded on a flexible, multi-service and open Building Operating System Services (BOSS). BOSS seeks to allow 3rd party applications to run reliably in a safe. sandboxed environment. It supports sensor and actuator access, access management, metadata, archiving, and discovery, as well as multiple simultaneously executing programs. Like a computer OS, the programs run at a variety of privilege levels, from critical to desirable but deferrable. Such programs access different resources, yet share the underlying physical resources. BOSS can extend to the Cloud or to other buildings, outsourcing expensive or proprietary operations as well as load sharing, but does so safely with fail-over to local systems when connectivity is disrupted. Building operators retain supervisory management, controlling the separation physically (access different controls), temporally (change controls at different times), informationally (what information leaves the building), and logically (what actions or sequences thereof are allowable). We will construct, deploy, and demonstrate the capabilities of a prototype BOSS in the context of university, residential buildings and closely related industrial processes. This poster presents some of the results of the first two years of the project, including the outline of BOSS, its use in three building applications (Auto Demand Response, Demand Controlled Ventilation, and Personalized Environmental Control) along with foundations in Model Predictive Control, Historian, and Metadata acquisition and organization. It expands more fully on active learning-based software synthesis techniques for automated metadata acquisition in order to enable fully hands-off building energy analysis, including rogue zone detection, over- cooled zones, stuck dampers, and inadequate setbacks, in a manner that can potentially scale to the millions of existing commercial units in the building stock. It also expands on a full scale study to create an open source building automation system that integrates HVAC, lighting, energy management and appliance control for small to medium scale commercial building using BOSS, but simplified wherever possible, and harnessing modern networked thermostat, lighting, an general control devices that form the emerging Internet of Things.