Predicting power outages before the storm

JOHNS HOPKINS/TEXAS A&M (US)—Using data from Hurricane Katrina and four other storms, researchers have created new computer models to help utilities better forecast hurricane-caused power outages in advance.

“The goal is to restore power faster and save customers money,” says Seth Guikema, assistant professor of geography and environmental engineering at Johns Hopkins University.

More accurate models will “provide a much better basis for preparing for restoring power after the storm,” he explains.

Electric companies can use the models to determine in advance how many repair crews to mobilize and where to place them to get electricity restored as quickly as possible after a storm, the researchers say, saving both utilities and customers time and money.

The research, published in the journal Risk Analysis, focused on two common challenges. When a hurricane is approaching, an electric power provider must decide how many repair crews to request from other utilities, a decision that may cost the provider millions of dollars.

The utility also must decide where to send these crews to enable the fastest and most efficient restoration of service after the hurricane passes.

Having more accurate estimates, prior to the storm’s arrival, of how many customers will lose power and where those outages will occur will allow utilities to better plan their crew requests and crew locations, the researchers say.

“If the power company overestimates, it has spent a lot of unnecessary money,” says Steven Quiring, assistant professor of geography at Texas A&M University, who collaborated on the study.

“If it underestimates, the time needed to restore power can take several extra days or longer, which is unacceptable to them and the people they serve. So these companies need the best estimates possible, and we think this study can help them make the best possible informed decision.”

What makes the research team’s computational approach unique and increases its accuracy over previous forecasts, Guikema and Quiring say, is the combination of more detailed information about the storm, the area it is impacting and the power system of the area, together with more appropriate statistical models.

In developing their computer model, the researchers looked at damage data from five hurricanes: Dennis (1995), Danny (1997), Georges (1998), Ivan (2004), and Katrina (2005). In the areas studied, Ivan created 13,500 power outages; Katrina, more than 10,000; Dennis, about 4,800; Georges, 1,075; and Danny, 620.

For the worst of these storms, some customers were without power for up to 11 days. The research team collected information about the locations of outages in these past hurricanes, with an outage defined as permanent loss of power to a set of customers due to activation of a protective device in the power system.

The researchers also included information about the power system in each area (poles, transformers, etc.), hurricane wind speeds, wetness of the soil, long-term average precipitation, the land use, local topography, and other related factors.

The data was then used to train and validate a statistical regression model called a generalized additive model, a particular form of model that can account for nonlinear relationships between the variables.

Johns Hopkins University news:
Texas A&M University news: