Disaggregating power outages from disconnections: A Northern Illinois case study.

Electricity service disruptions, in the form of power outages or involuntary disconnections due to nonpayment, compromise households' ability to meet their energy demand and increase energy insecurity. Achieving equitable electricity access requires understanding how these disruptions vary across demographics. However, spatially granular disruption data are limited, making demographic analysis of disruptions challenging. We present a method of disaggregating power outages from disconnections in residential, unlabeled energy usage data and validate our methodology using the Commonwealth Edison utility service territory as a case study. We find that the regions with the longest and most frequent outages are in rural counties, and regions with a higher percentage of low-income residents and minorities often have higher outage durations and frequencies than the rest of the service territory. Our proposed method enables more granular metrics for electricity reliability, highlights gaps in energy accessibility, and informs the identification of energy-insecure communities.