2:00  -  3:00

May 7, 2019

Harris

302: Network Event Analysis Center

Session Category :  Foundational Analytics 

The Network Event Analysis Center (NEAC) equips AEP Transmission Operations staff with a tool to automatically detect a subset of the root causes of transmission faults. This will allow AEP to resolve system issues more efficiently, optimize asset renewal strategy, and improve service. The NEAC applies pattern recognition to assign causes based on oscillography data stored by relays. Advanced visualization techniques were used to help label historical events for classification. A random forest model was utilized to assign confidence levels to different causes. A feedback loop is also employed to support ongoing machine learning to further improve accuracy.

Attendees will learn how AEP Transmission has accomplished the following:

  • Automated oscillography data collection from relays
  • Labeling and training of a random forest to detect a subset of Transmission line fault types
  • Incorporation of a feedback loop to facilitate machine learning to improve the model moving forward