Modeling Weather Risks: Applying analytics to inform electrical utilities

I presented in the University of Colorado-Boulder Department of Geography Colloquium Series on Friday, September 11. My talk is available on YouTube.

Abstract:

Weather can cause significant damage to the electrical power system, leading to prolonged power interruptions to a large number of customers. The estimated annual cost to the U.S. economy from storm-related power outages is >$20 billion. The number of weather-related outages has increased significantly in recent years. One approach to deal with this problem is to develop predictive techniques for forecasting how storms will impact the power grid hours to days in advance. This information can help utilities, first responders, and emergency managers to better prepare for the outages and more quickly restore power. This presentation summarizes the data-driven power outage models that we have developed for the U.S. Department of Energy and a number of investor-owned electrical utilities in the United States. These models are used to support decision making for near-term events (e.g., pre-storm preparation) and longer-term planning. The development and validation of our models will be presented and our approach for quantifying uncertainty will also be discussed. The talk will also highlight the challenges and successes from recent applications for American Electric Power, FirstEnergy, Southern Company and Southern California Edison.