skip to main content

SciTech ConnectSciTech Connect

Title: Baselining PMU Data to Find Patterns and Anomalies

This paper looks at the application of situational awareness methodologies with respect to power grid data. These methodologies establish baselines that look for typical patterns and atypical behavior in the data. The objectives of the baselining analyses are to provide: real-time analytics, the capability to look at historical trends and events, and reliable predictions of the near future state of the grid. Multivariate algorithms were created to establish normal baseline behavior and then score each moment in time according to its variance from the baseline. Detailed multivariate analytical techniques are described in this paper that produced ways to identify typical patterns and atypical behavior. In this case, atypical behavior is behavior that is unenvisioned. Visualizations were also produced to help explain the behavior that was identified mathematically. Examples are shown to help describe how to read and interpret the analyses and visualizations. Preliminary work has been performed on PMU data sets from BPA (Bonneville Power Administration) and EI (Eastern Interconnect). Actual results are not fully shown here because of confidentiality issues. Comparisons between atypical events found mathematically and actual events showed that many of the actual events are also atypical events; however there are many atypical events that do notmore » correlate to any actual events. Additional work needs to be done to help classify the atypical events into actual events, so that the importance of the events can be better understood.« less
Authors:
; ; ;
Publication Date:
OSTI Identifier:
1331743
Report Number(s):
PNNL-SA-113634
TE1101000
DOE Contract Number:
AC05-76RL01830
Resource Type:
Conference
Resource Relation:
Conference: CIGRE 2015 Grid of the Future Symposium, October 11-13, 2015, Paris, France
Publisher:
CIGRE, Paris, Fr.
Research Org:
Pacific Northwest National Laboratory (PNNL), Richland, WA (US)
Sponsoring Org:
USDOE
Country of Publication:
United States
Language:
English
Subject:
Situational awareness for complex systems; Multivariate statistical analyses; phasor measurement units (PMU); Baselining analyses