Predicting the Impact of Storms on Utility Customers With Physics-Informed AI Using Weather Data
In an era marked by increasing storm severity and aging utility infrastructure, the ability to predict and mitigate the impact of weather-related events on electrical utility services is more critical than ever. The science required to develop predictive algorithms to address this complex forecasting problem requires a synthesis of multiple
areas of physical and data science. A variety of algorithms can be applied to leverage weather forecast data in predicting damage to the electrical grid. This presentation demonstrates use of non-linear regression, decision trees, CNN, and LSTM methods to develop physics-informed AI to predict the number and location of electrical outages from damaging weather events.
areas of physical and data science. A variety of algorithms can be applied to leverage weather forecast data in predicting damage to the electrical grid. This presentation demonstrates use of non-linear regression, decision trees, CNN, and LSTM methods to develop physics-informed AI to predict the number and location of electrical outages from damaging weather events.
About the Presenter
With a PhD in statistical astrophysics, David Corliss is a Senior Data Scientist at DTE Energy where he develops machine learning and AI applications to drive business value. He is active in the American Statistical Association, serving on their board as the representative for local chapters in the Midwest region. His research focuses on industry applications in time series analysis and physics-informed AI, building bridges between academic and industry, and ethical best practices in data and statistical practice. Dr. Corliss is the founder of Peace-Work, a volunteer cooperative of statisticians, data scientists and other researchers applying analytics in issue-driven advocacy.
|