Online Program Home
My Program

Abstract Details

Activity Number: 609
Type: Contributed
Date/Time: Wednesday, August 3, 2016 : 2:00 PM to 3:50 PM
Sponsor: Section on Statistical Learning and Data Science
Abstract #319931 View Presentation
Title: Identifying Typical Patterns and Atypical Behavior in Copious Amounts of Streaming Data
Author(s): Brett Amidan* and James Follum
Companies: Pacific Northwest National Laboratory and Pacific Northwest National Laboratory
Keywords: Big data analytics ; feature extraction ; multivariate anomaly detection

This presentation looks at applying situational awareness methodologies to identify patterns and unusual behavior in the power grid domain using synchrophasor data. Synchrophasor data consists of hundreds of variables measured 30 or 60 times a second. Methodologies were created using R to handle large amounts of streaming data, filter bad data, and to extract useful features from this data. Multivariate algorithms were employed to establish normal baseline behavior and to identify when grid behavior was atypical. These methods handle random and non-random missing data. Drill down visualizations were used to help identify the specific variables and locations that contributed to the atypical behavior, allowing the system engineers tools to investigate issues concerning the grid. Initial results found that many atypical discoveries correlated to actual known events; however some atypical moments went undetected by the system. Further investigation is needed to determine the importance of these findings, and to help tune the algorithms to discover more meaningful events. The user interface, including displayed results and interactive plots, were developed using the R Shiny package.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

Copyright © American Statistical Association