Developing a Power Outage Prediction Model (OPM) for New Hampshire

Weather conditions often determine the severity of power outage events. Overhead power lines commonly experience damages due to high winds, thunderstorms, and winter precipitation. The ability to accurately predict power outage events would allow power companies to reduce outage times for their customers and thereby improve electrical grid reliability. In recent years, the Eversource Energy Center at the University of Connecticut has developed an outage prediction model (OPM) for Eversource service territories in Connecticut. The purpose of this project is to improve outage modelling capabilities, with a focus on winter weather forecasting, for New Hampshire service territories by expanding upon the existing OPM. The first of two research objectives was to establish a numerical relationship between Eversource’s two outage reporting methods. The current OPM in New Hampshire utilizes data from the Outage Management System (OMS) after January 1, 2016 for model development. A comparison between the current OMS and the older Trouble Report and Unsatisfactory Performance of Equipment Reports (TRUPERs) was conducted during the overlap period of the two datasets (2016-2017). With the exception of a few outlier events, results of the comparison show a high correlation (R² = 0.8393) between the two datasets. Identifying major power outages from previous years will expand the OPM event catalog and improve model performance. The second objective was to complete a Weather Research and Forecasting (WRF) Model analysis for all New Hampshire METAR locations by comparing WRF forecasted values of 18 model variables with observed values. WRF simulations were run for a sample of 12 rain/wind and 13 winter outage events that occurred from 2016-2017. The final results of this analysis will allow the performance of Plymouth’s WRF model to be assessed, and will determine if Plymouth’s WRF model is compatible with the current OPM.