Predictive Outage Modeling
How Theorem Geo can help.
As an energy provider, you want to know if there is high probability of an electrical outage and where that outage is most likely to occur. When weather events occur our analytical models utilize past data with present day forecasting to provide you with current outage probabilities in easy to view geographically displayed locator maps. This forecasted information allows managers to make informed decisions about how to deploy assets and when. Before severe weather occurs be prepared to mobilize faster with the appropriate personnel to keep the lights on.
Analyze historic data to understand statistically significant meteorological factors that led to electrical outages.
Power outages caused by strong and severe thunderstorms are a frequent event in the United States. These thunderstorms engender lightning and strong winds that cause whole trees or tree limbs to collapse, which consequently disturb power transmission and distribution lines. The severity of power interruption amplifies with time, as long restoration periods become costly to service providers and hinders customers’ ability to return to normalcy after the storm has passed. Forecasting for these types of events is challenging because it is hard to accurately represent convection due to computational limitations of existing weather models. Even the most sophisticated weather models limit forecasters to producing broad outlooks that leave mitigation decision makers with insufficient knowledge on storm severity and storm type at the correct spatiotemporal scale. Mitigation decision makers need to know rough timing, location, and severity of the storms in order to request and, if possible, pre-position repair crews in order to reduce the amount of time their customers are without power.
Monitor and simulate current information and probabilistic risk areas based on historic model analysis.
What We Can Do
Forensic analysis of thunderstorm-induced power outage events to improve the outage prediction during such events. Meteorological diagnostic/forecast parameters, common to severe storm environments, are investigated on past outage events of varying intensity. Statistical techniques are used to identify which factors are most discriminatory between minor and major power outage events. Parameters with strong probability of detection (POD) and a low probability of false detection (POFD) of power outages are combined to build statistical models that will be initiated by high-resolution numerical weather models to more accurately anticipate the location and number of power outages with a 12-60 hours lead time. Predictions will be deployed via a mobile application designed specifically for your utility that will identify potential staging location options.
Provide probability forecast of areas at risk for electrical outages based on simulation results, geographic features, and varying weather forecasts.
Why Use It
- Help alleviate forecasting errors between power outage causing thunderstorms and ordinary thunderstorms
- Have a graphical output of where and when your distribution grid will be comprised to better prioritize your resources
- Have as few or as many of your resources view probable outage locations
- Preidentified suitable staging locations will be automatically recommended in the forecast high outage density areas
- Save money by not overstaffing on low outage occurring days and not understaffing on high outage occurring days