Carnegie Mellon Electricity Industry Center

2003-04 Seminars

Statistical Characteristics of Cascading Failures in Power Networks

Huaiwei Liao
CMU ECE

Abstract
Electric power systems consist of thousands of interacted components, distributed geographically. On the one hand, it is the interactions among components that realize long-distance economic transmission of electrical energy and complete the service to millions of customers. On the other hand, the interactions may introduce unexpected consequences to the interconnected power systems under certain unfavorable circumstances, such as defected structure of the system, extreme weather, abnormal load variations, hidden failures of protection systems. In some cases, it causes collapse of the power systems. In this talk, after briefly reviewing the scenarios of August 2003 Blackout in US, some statistical characteristics of the cascading failures of power systems are investigated in a regular resistor network and a 3357-bus practical power system respectively by extensive computer simulation. To represent the cascading sequences and the relationship among element failures, a graphical representation based on Boolean networks is used, which could be a promising way to depict the vulnerability of power networks under adversary disturbances.