NESL Technical Report #: 2011-9-1
Although the HVAC (Heating, Ventilation and Cooling) systems dominate the total power consumption of modern buildings, the IT equipments from PCs to network and server equipments are becoming an increasing contributor to energy bills and carbon emissions. Recent study also shows that much of the energy consumption is due to the IT infrastructure. In a typical office where desktop computers make up a substantial component of working area, the energy savings would have been potentially remarkable if users could be made aware of energy using as well as wasting, and an effective strategy was employed to intelligently control these computers.
In this thesis, we propose a monitoring and actuation platform to address this issue. Our platform is essentially a system that incorporates several function modules including data sampling, data visualization, notification, analysis and control. The contribution of our platform can be divided into two aspects. One is that we organize meter/sensor data stream efficiently and present them in a way that clients can easily access them. The other is that we analyze the meter/sensor data and build an effective profile model for each user to predict the subjects activities, and further control subjects computer accordingly to reduce unnecessary energy waste. We demonstrate that our approach achieve a 12% energy savings on average. In addition, the rate of accuracy for prediction of occupant activities using our model reaches approximately 90%.
Publication Forum: MS Thesis, UCLA Electrical Engineering Department
Page (Count): 50
Place: Los Angeles
Public Document?: Yes
NESL Document?: Yes
Document category: Report