A Tutorial on Detection and Characterization of Special Behavior in Large Electric Power Systems

The objective of this document is to report results in the detection and characterization of special behavior
in large electric power systems. Such behavior is usually dynamic in nature, but not always. This is also
true for the underlying sources of special behavior. At the device level, a source of special behavior
might be an automatic control system, a dynamic load, or even a manual control system that is operated
according to some sharply defined policy. Other possible sources include passive system conditions, such
as the state of a switched device or the amount of power carried on some critical line.

Detection and characterization are based upon “signature information” that is extracted from the behavior
observed. Characterization elements include the signature information itself, the nature of the behavior
and its likely causes, and the associated implications for the system or for the public at large. With
sufficient data and processing, this characterization may directly identify a particular condition or device
at a specific location. Such conclusive results cannot always be done from just one observation, however.
Information environments that are very sparse may require multiple observations, comparative model
studies, and even direct testing of the system.

Download