The modernized electric grid, the Smart Grid, integrates two-way communication technologies across power generation, transmission and distribution, in order to deliver electricity efficiently, securely and cost-effectively. On the monitoring and control side, it employs real-time monitoring offered by a messaging-based advanced metering infrastructure (AMI), which ensures the grid?s stability and reliability, as well as the efficient implementation of demand response schemes to mitigate bursts demand. The efficient implementation of these features presents a number of challenges, but also opportunities for technology development in engineering, networking, and data analytics. The intellectual contributions include the development of a framework that encompasses fast signal processing and machine learning algorithms, together with multi-sensor information to assess the health of the network. Specifically, the study will investigate algorithms for adaptive detection over streaming, high-dimensional and potentially missing value data. Furthermore, the project, via industry collaborators and power provisioners, will provide a comprehensive empirical evaluation with real-world data; this includes an open-source proof-of-concept prototype for quickly inferring nefarious activity in home-area networks, and a cloud-based testbed for examining realistic scenarios in a wide-area setting.
TWC: TTP Option: Small: Collaborative: Integrated Smart Grid Analytics for Anomaly Detection