Automated Analytics for High Performance Buildings

PI: Robert Cox, University of North Carolina at Charlotte

 

Executive Abstract

Modern energy management systems have the potential to obtain significant amounts of data. Often, however, building operators have neither the training nor the time to investigate this vast treasure trove of information. Recent efforts have focused on attempts to provide users with an automated dashboard that summarizes this information. Even still, such dashboards tend to provide relatively high-level details, such as trends in power consumption or EUI. With appropriate analytical tools and data mining, however, systems could begin to provide much richer and more meaningful metrics. Studies suggest, for instance, that 10 to 20% of building energy is wasted by equipment that is either faulty or improperly operated. Such inefficiencies can often be detected using appropriate automated analysis. Furthermore, data can be used to compare real-time performance to expected performance. The proposed research will investigate automated procedures for real-time energy-performance monitoring and diagnostics in commercial facilities, with an eye on providing actionable information if there is a change in building performance (i.e. caused by a fault, etc.).
 

Objective 1: Develop an energy-performance model for smaller commercial facilities
Objective 2: Field test the energy-performance model, perhaps in the presence of faults
Objective 3: Begin to detail what is needed to scale up to larger commercial facilities
Objective 4: Provide feedback on energy performance in the actual field sites selected for study