Intelligent Buildings: Energy Monitoring/Management, and Related Impact on Utility Market Dynamics

Timespan: 
2010-2012
Principal Investigator: 

Despite the full automation of the energy-metering infrastructure at the building level, today there is little knowledge about where and how energy is internally consumed in large buildings and how this consumption corresponds to occupant productivity and comfort needs.  In this project we will investigate new methods for determining electric load decomposition without requiring exhaustive building sub-metering. Using low-cost, lightweight sensor and actuator retrofits, our goal is to develop an infrastructure that better quantifies usage and determines which loads are interruptible and which are not while also identifying waste. Our research will be driven by an experimental deployment on the Yale campus that will provide insights on electricity consumption in large buildings.

The collected measurements and load disaggregation results will also be used as guidelines towards the development of a new set of metrics for evaluating building energy performance. Unlike existing metrics that measure performance based on intensity per unit area, a practice that tends to favor low-occupancy buildings, our work will concentrate on metrics that also account for occupants and occupant productivity. These per capita metrics will provide a more realistic view of building performance and will drive emissions reductions on a per capita basis.

The derived knowledge about the properties of electric loads will allow us to examine the building’s potential for providing building-side demand/response services to the grid. If such services are achievable, these will help the building to reduce its net energy costs in a way that inherently promotes greener sources by optimally making use of intermittent generation from wind solar and other sources. The findings from the areas described above will provide a wealth of information for influencing the formation of new markets and driving new policies for energy management in large buildings.

Andreas Savvides, Departments of Electrical Engineering and Computer Science