Importance of Tree Species and Precipitation for Modeling Hurricane-induced Power Outages
MetadataShow full item record
Hurricanes can be a major threat to electric power systems, often resulting in costly repairs and lengthy restoration times. In addition, power companies often lack the personnel required to restore power in a timely and efficient manner and must rely on outside assistance from other utility companies. Statistical power outage models, such as the Hurricane Outage Prediction Model (HOPM), provide estimates of outages at least 24 hours before a hurricane makes landfall. These models can greatly benefit utility companies by allowing for better allocation of resources and potentially shortening restoration times. This research will investigate the addition of two new variables, tree species and storm-derived precipitation, to the HOPM. Tree species information was extracted for the service area of a major Gulf Coast utility company. Storm-derived precipitation was also extracted for the service area 24 hours before and after hurricane landfall. The model was then run for the service area with the new variables added, and results were generated that showed the impact of the new predictors on model performance. Of the two predictors, tree species resulted in the greatest model improvement. Certain tree species, such as sweetgum, may be better predictors of outages than others. Storm-derived precipitation was also an important predictor of outages, particularly in urban areas. Precipitation amounts less than about 7 inches had the greatest impact on outages. Inclusion of tree species and storm-derived precipitation in future versions of the HOPM may enhance model performance and, in turn, aid utility companies in their goal of more efficient power restoration.
Subjecthurricane power outage modeling
tree species and power outages
precipitation and power outages
Maderia, Christopher M (2015). Importance of Tree Species and Precipitation for Modeling Hurricane-induced Power Outages. Master's thesis, Texas A & M University. Available electronically from