Automated Identification of Weather-Related Utility Outages
Lead PI: Dr. Kara Sulia, UAlbany
Co-PI: Dr. Nick Bassill, UAlbany
Industrial Relevance & Need:
As a standard, utility electric outages are typically communicated via customer reports. More sophisticated systems incorporate outage or non-response detection within circuitry components to provide alerts. As a result, in real-time, outages are reported based on the location of the outage, customers affected, and most utilities provide an automated (and rudimentary) method of predicting an estimated time of restoration.
The responsibility of the utility company is to resume normalcy to their grid as quickly and safely as possible. While more sophisticated outfits may have the capability to perform advanced data reporting, most utilities do not contain the resources to make advanced weather-related outage tracking a priority.
While weather identification can be a component of outage reporting by field technicians or later by analysts, this number is inherently flawed and likely underreported, for a number of reasons, including, but not limited to:
Subjective weather identification via responding field technicians.
The direct cause of an outage (e.g., falling branches) can many times be a result of an unreported weather-related event (e.g., wind gusts).
Customer-reported outages can, at times, be delayed by hours (or days) depending on work/sleep/travel schedules, deeming real-time weather identification irrelevant.
Project Goals:
We propose to develop a tool to automatically identify weather associated with utility outages, with the following goals:
Provide utilities with weather information for each outage through automated real-time identification
To remove human bias associated with identification of weather associated with an electrical outage.
To create a tool usable by utilities and researchers to access weather information associated with real-time or retroactive outages.
Objectives:
Develop an API with relevant weather data sources designed for expandability.
Develop an online interface to easily search, filter, sort, and export data.
Work closely with utility partners to identify useful datasets and to test ergonomics of interface.
Work closely with other WISER projects to identify API additions relevant to their research objectives to alleviate data access burdens.