Operational evaluation of a meso-scale weather and outage prediction service for electric utility operations

The distribution system of an electric utility, particularly with an overhead infrastructure, can be highly sensitive to local weather conditions. Power outages caused by severe weather events can have major impacts on such operations. Hence, the ability to predict specific events or combination of weather conditions that can disrupt that distribution network with sufficient spatial and temporal precision, and lead time has the potential to enable proactive allocation and deployment of resources (people and equipment) to minimize time for restoration.

Previously, we determined the viability of this idea by building upon the efforts at the IBM Thomas J. Watson Research Center to implement and apply an operational mesoscale prediction system to business problems, dubbed “Deep Thunder”. We extended our capabilities in the New York City metropolitan area in several key ways, working with the emergency management group at a major utility company in the northeastern United States. The first was implementing a number of customized visualizations available via a web browser, which are automatically updated for each forecast cycle. They reflect the relevant weather conditions and predicted threats, and incorporate information concerning the utility's infrastructure and resource response thresholds. This was part of an overall effort to tailor the underlying software and hardware system to meet the geographic, throughput and dissemination requirements of the utility operations. Data from a local, relatively dense mesonet operated by AWS Convergence Technologies were utilized to validate and improve the forecasts as well as model tuning. Historical observations from the mesonet were analyzed along with data from the utility concerning the characteristics of outage-related infrastructure damage from weather events, as well as the information about the distribution network and local environmental conditions. This enabled the development of a statistical representation of such damage for predictive purposes. This damage estimation model was then coupled to weather forecast information derived from the Deep Thunder modeling capability as well as in real-time from observations available from the aforementioned mesonet. This approach also incorporates uncertainties in both the weather and outage data. Therefore, the model provides probabilities of outage restoration jobs per substation area and their likely cause, which is associated with visualizations that explicitly depict the uncertainty. The initial focus of the numerical weather prediction component was to provide 24-hour forecasts at resolutions as high as one km for portions of the utility's service territory. This was replaced with a more sophisticated weather model covering a broader geographic scale, including all of the utility's service territory and that of its subsidiaries. This newer model incorporates data assimilation of observations from the mesonet for improved initialization and operates at two-km resolution to produce 84 hour forecasts. Retrospective analysis of key historical events that impacted the utility's operations was introduced to enable the configuration and validation of both the weather and damage models.

We will discuss the on-going work, the overall approach to the problem, some specifics of the solution, and lessons that were learned through the development and deployment. We will present how the content is being used and how we are evaluating its quality with respect to specific weather events that affect utility operations. This includes the validation methodology used for the weather and damage forecasts and deployment of customized visualizations. We will also discuss the overall effectiveness of our particular approach for these and related applications, issues such as calibration of data and quantifying uncertainty as well as recommendations for future work.