Green building energy analytics: Challenges and opportunities

Green building applications need efficient and finegrained determination of power consumption pattern of a wide variety of consumer-grade appliances through non-intrusive load monitoring (NILM) techniques. Fine-grained monitoring of everyday appliances can provide better feedback to the consumers and motivate them to change behavior in order to reduce their energy usage. It also helps to detect abnormal power consumption events, long-term appliance malfunctions and potential safety concerns. Commercially available plug meters can be used for individual appliance monitoring but for an entire house, each such individual plug meters are expensive and tedious to setup. In practice, deploying smart plug based NILM and acquiring the low-level power measures of a large number of devices is often difficult or impossible due to the deployment complexity and varying characteristics of devices and thus must instead be employed at the circuit or house-level and inferred through the incorporation of novel usage-based measurement and probabilistic level-based disaggregation algorithm. But the challenges in deploying non-intrusive load monitoring algorithm involve disaggregating individual devices consumption from the aggregate power measurement, as well as modeling and incorporating the usage based prediction. In this talk, I will discuss on advanced machine learning and data analytics algorithms that capture the measurement based approach and circuit level NILM with the autonomous profiling and prediction logic, and the significant practical impact of intelligent use of such profiling techniques for green building applications. Our approach help improve the performance of energy disaggregation algorithms and provide critical insights on appliance longevity, abnormal power consumption, consumer behavior and their everyday lifestyle activities. The performance of our proposed algorithms on real data traces will be presented. I will conclude this talk with our ongoing research pro- ects in this area and future research directions.