Writing scalable building efficiency applications using normalized metadata

Abstract

Extracting meaningful information from a building's sensor data, or writing control applications using the data, depends on the metadata available to interpret it, whether provided by novel networks or legacy instrumentation. Commercial buildings comprise large sensor networks, but have limited, obscure 'tags' that are often meaningful only to the facility managers. Moreover, this primitive metadata is imprecise and varies across vendors and deployments. This state-of-the-art is a fundamental barrier to scaling analytics or intelligent control across the building stock, as even the basic steps involve labor intensive manual efforts by highly trained consultants. Writing building applications on its sensor network remains largely intractable as it involves extensive help from an expert in each building's design and operation to identify the sensors of interest and create the associated metadata. This process is repeated for each application development in a particular building, and across different buildings. This results in customized building-specific applications which are not portable or scalable across buildings. We have developed a synthesis technique([2]) that learns how to transform (normalize) a building's primitive sensor metadata to a common namespace(similar to[1]) by using a small number of examples from an expert, such as the building manager 1(Figure 1). Once the transformation rules are learned for one building, it can be applied across buildings with a similar metadata structure. This common and understandable namespace automatically yields semantic relationships between sensors, which enable analytics applications that do not require apriori building-specific knowledge. In this demonstration, we present three efficiency and analytics applications - (a) Identifying rogue thermal zones, i.e zones that require constant heating or cooling (b) Identifying stuck air flow dampers, and (c) Identifying the presence of night-time setbacks - all built against our common namespace. We were able to apply these applications unmodified to more than 10 commercial buildings on the University of California, Berkeley campus. These buildings comprised sensor networks which were commissioned by two different vendors at different points in time over the last two decades. The applications helped identify candidate efficiency and comfort improvements in each of these buildings without any manual inspection.