Skip to main content
Infrastructure Resilience Conference 2018

Full Program »

Learning from Failures: Analysis of Trends for Resilience of Power Systems

In this paper, we conduct a comprehensive empirical evaluation on the resilience of the U.S. power grid based on the database of the Electric Disturbance Events. In particular, we look into the learning effect from each black-out and recovery as this is a key consideration in modeling systems resilience. Three key components associated with learning from power failures are the time between disruptions, the performance loss of each disruption and time needed for recovery. One way of quantifying the learning effect is thus to develop a combined measure for frequency of occurrence, performance loss and recovery time, and monitor the trend of this combined measure. First of all, we consider trend detection in the disruption intensity. The intensity reflects the resistance of the system and thus it is an important component in most definitions of resilience. In addition, we look into the severity of each disruption, which represents the ability of a system to absorb and recover from disruptions. Both the performance loss in terms of service deliveries and the recovery duration are extracted from the database to calculate the severity of a disruption. We further define the cumulative severity to capture the overall resilience of a power system, which is a combination of the disruption intensity and the disruption severity. A modified Lewis-Robinson test is developed for trend detection in the cumulative severity function. For future research, our method could be applied to other infrastructure systems such as water supply systems, transportation systems, and communication systems.

Lijuan Shen
National University of Singapore
Singapore

Beatrice Cassottana
National University of Singapore
Singapore

Loon Ching Tang
National University of Singapore
Singapore

 

Powered by OpenConf®
Copyright ©2002-2016 Zakon Group LLC