Title:  Large Scale Data Analytics for Energy Networks

Committee: 

Dr. Ji, Advisor     

Dr. Davenport, Chair

Dr. Ganz

Abstract:

The objective of the proposed research is to study the resilience of energy infrastructure and services that are considered a national priority, due to disruptive weather events that have been causing large scale power failures in the US, affecting millions of people for extended time durations. Here resilience is defined as reducing failures (infrastructure resilience) and recovering rapidly (service resilience) upon disruptions, natural or man-made. It is unknown, however, to the research community and general public, how resilient our infrastructure and services really are to a wide range of weather disruptions, over the years and locations. Service resilience, in particular, is insufficiently studied, where systematic analysis is lacking and needed on both the effectiveness and limitations of recovery from disruptions. A key challenge is a lack of detailed and large-scale data as well as analytics for resilience study. This thesis proposal focuses on two aspects. One is to study service resilience using granular field data from multiple provider networks in the US. The other is to develop data analytics, drawing from unsupervised learning and non-stationary random processes.