Following a storm, line and tree (e.g. mutual aid) crews can be called in from across the country and across international borders to help restore power.
The Edison Electric Institute coordinates the Mutual Aid Network, a volunteer network of investor-owned utilities that helps coordinate the allocation of mutual aid crews during natural and man-made emergencies.
After a storm, utility crews work with emergency first responders to address safety locations (g., downed wires) and to clear blocked roads. At the same time, utility crews are working to restore critical facilities (police and fire departments, hospitals and town buildings), and the largest number of customers as quickly and safely as possible to expedite the restoration for all customers.
Over 90 percent of power outages during storm events in Connecticut are tree related.
UConn has developed an Outage Prediction Model that integrates high-resolution weather predictions (winds, heavy rain, ice, snow, flooding) with vegetation characteristics (height, density, leafs on/off) and other geographic data to accurately predict damages on the electric grid and to prepare for and respond to damages.
About 1.8 million customers across three geographic regions are benefiting from UConn’s Outage Prediction Model. Currently, UConn is working with Eversource Energy in Connecticut and western Massachusetts, and United Illuminating in central coastal Connecticut.
The Outage Prediction Model performs accurately for a range of storms including tropical systems (tropical storms and hurricanes) and winter storms (excludes ice storms).
Our weather forecast model, the Weather Research and Forecasting Model (WRF), runs on seven Haswell nodes (Intel x64; 24-core nodes with 128GB RAM each) with a total scratch storage of 620TB. This computing power is equivalent to 168 laptop computers all performing calculations at the exact same time, and the scratch storage is large enough to store 10,540,000 hours of music!
High-Resolution Weather Forecasting
Our current experience demonstrates an advanced window of three to five days is an effective timeframe for reliable weather forecasts. Some variables are easier to forecast in advance, such as temperature.
Forecast accuracy also depends on weather observations, where models are updated at six-hour intervals. Some additional uncertainty arises from the unevenly distributed geographic location of weather stations across our region.
A numerical weather prediction method that attempts to generate a representative sample of possible future weather is termed “ensemble weather forecasting.” This consists of various parallel forecasts based upon differing initial conditions, or multiple models and/or configurations to account for uncertainties in imperfect data and methods. The spread among ensemble members represents the forecast uncertainty associated with each set of conditions, and the average of all members is generally more accurate than the forecast from one single member.