Spring Break
Hopkins researchers predict power outages
Issue date: 11/5/09
Researchers from Hopkins and Texas A&M University believe that they can accurately predict the number of power outages that occur in the aftermath of major hurricanes.
After Hurricane Katrina, 1.7 million people in the Gulf Coast lost power and for some it took weeks to restore it.
In preparation for the power outages that are a certainty after a powerful hurricane, electricity providers use estimates to determine how many emergency crews to dispatch and to what areas. These crews are borrowed under agreements with other utility providers.
Overestimation of the number of outages is extremely costly, running into the millions of dollars because of the unneeded labor. Underestimation leads to extended periods of time before full power can be restored to all affected customers. Not only does this create disgruntled customers and a damaged public image for the provider, but quickly restoring basic utilities can also be crucial for rebuilding other damaged areas.
Generally, existing models have been unable to accurately estimate the number of power outages a major hurricane can cause in a particular area.
According to the team's work, which was published in the current issue of the journal Risk Analysis, the failure of past models to predict these hurricane-related power outages lies in their assumption that the number of outages is directly related to several factors, mainly the intensity of the hurricane that causes them.
The team also found that such models, commonly referred to as "generalized linear models" (GLMs), are not credible because they are unable to accurately account for population density: In urban areas, the models substantially overestimate the number of outages, and in rural areas, power losses can be underestimated.
Additionally, GLMs do not account for nonlinearity or variation in the relationship between the elements of the hurricane (such as strength and variations in the path of the storm) and the number of outages.
After Hurricane Katrina, 1.7 million people in the Gulf Coast lost power and for some it took weeks to restore it.
In preparation for the power outages that are a certainty after a powerful hurricane, electricity providers use estimates to determine how many emergency crews to dispatch and to what areas. These crews are borrowed under agreements with other utility providers.
Overestimation of the number of outages is extremely costly, running into the millions of dollars because of the unneeded labor. Underestimation leads to extended periods of time before full power can be restored to all affected customers. Not only does this create disgruntled customers and a damaged public image for the provider, but quickly restoring basic utilities can also be crucial for rebuilding other damaged areas.
Generally, existing models have been unable to accurately estimate the number of power outages a major hurricane can cause in a particular area.
According to the team's work, which was published in the current issue of the journal Risk Analysis, the failure of past models to predict these hurricane-related power outages lies in their assumption that the number of outages is directly related to several factors, mainly the intensity of the hurricane that causes them.
The team also found that such models, commonly referred to as "generalized linear models" (GLMs), are not credible because they are unable to accurately account for population density: In urban areas, the models substantially overestimate the number of outages, and in rural areas, power losses can be underestimated.
Additionally, GLMs do not account for nonlinearity or variation in the relationship between the elements of the hurricane (such as strength and variations in the path of the storm) and the number of outages.
Viewing Comments 1 - 1 of 1
New Jersey movers
posted 11/11/09 @ 5:57 PM EST
Quote:
"The team also found that such models, commonly referred to as "generalized linear models" (GLMs), are not credible because they are unable to accurately account for population density: In urban areas, the models substantially overestimate the number of outages, and in rural areas, power losses can be underestimated. (Continued…)
Post a Comment