Extracting resilience metrics from distribution utility data using outage and restore process statistics

Nichelle'Le K. Carrington,Ian Dobson,Zhaoyu Wang
DOI: https://doi.org/10.1109/TPWRS.2021.3074898
2021-04-21
Abstract:Resilience curves track the accumulation and restoration of outages during an event on an electric distribution grid. We show that a resilience curve generated from utility data can always be decomposed into an outage process and a restore process and that these processes generally overlap in time. We use many events in real utility data to characterize the statistics of these processes, and derive formulas based on these statistics for resilience metrics such as restore duration, customer hours not served, and outage and restore rates. The formulas express the mean value of these metrics as a function of the number of outages in the event. We also give a formula for the variability of restore duration, which allows us to predict a maximum restore duration with 95% confidence. Overall, we give a simple and general way to decompose resilience curves into outage and restore processes and then show how to use these processes to extract resilience metrics from standard distribution system data.
Systems and Control
What problem does this paper attempt to address?