Improving Hurricane Power Outage Prediction Models Through the Inclusion of Local Environmental Factors

Area of Science:

• Electrical Engineering
• Environmental Science
• Disaster Management

Background:

• Tropical cyclones cause significant damage to electrical power systems.
• Accurate spatiotemporal outage forecasts are essential for efficient power restoration.
• Existing models like the Spatially Generalized Hurricane Outage Prediction Model (SGHOPM) have limitations.

Purpose of the Study:

• To enhance the predictive accuracy of the Spatially Generalized Hurricane Outage Prediction Model (SGHOPM).
• To introduce a novel two-step prediction procedure for outage forecasting.
• To incorporate a broader range of environmental variables into the prediction model.

Main Methods:

• Developed a two-step prediction model: first predicting outage occurrence, then predicting outage numbers.
• Expanded predictor variables beyond wind characteristics to include elevation, land cover, soil, precipitation, and vegetation.
• Validated the enhanced SGHOPM against previous versions.

Main Results:

• The new two-step prediction procedure significantly improved outage forecasting.
• Inclusion of additional environmental variables (elevation, land cover, soil, precipitation, vegetation) enhanced model performance.
• The enhanced SGHOPM demonstrated an approximate 17% increase in overall prediction accuracy.

Conclusions:

• The enhanced SGHOPM provides more accurate spatiotemporal forecasts of tropical cyclone-induced power outages.
• The integration of diverse environmental factors is critical for improving hurricane outage prediction models.
• This improved forecasting capability will aid utility companies in optimizing power restoration efforts post-disaster.