Estimating Power Outage Duration in the Event of Hurricane Land-Falls in the U.S.

• Activity Number: 469
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Roshanak Nateghi*+ and Seth David Guikema and Steven Quiring
• Companies: The Johns Hopkins University and The Johns Hopkins University and Texas A & M University
• Keywords: power outage duration ; hurricanes ; data mining ; power systems ; Random forest

Abstract

Each year, hurricanes cause widespread power outages in the United States that adversely affect industries and millions of customers, leading to large economic losses. Being able to model power outage durations accurately can facilitate more efficient coordination of recovery efforts and, as a result, help in minimizing the restoration periods. The objective of this paper is to develop statistical models to accurately estimate power outage duration prior to hurricane land-falls, using a wide range of covariates such as the hurricane's estimated wind speed and central pressure, the number of transformers, the number of customers, the number of poles, and a variety of geographic and development variables in the region of interest. Our models were developed using the data available from the outages caused by Hurricane Ivan in 2004. The results of this model were validated against the outage durations associated with Hurricane Katrina and Hurricane Dennis. Regression analysis together with data mining techniques were used to develop our statistical models, and we compared the models based primarily on their predictive accuracy.