Probabilistic Assessment of Power Distribution Systems Resilience Under Extreme Weather

We present a probabilistic assessment of power distribution systems resilience under different weather conditions. It is based on Monte Carlo simulation considering failures and repairs models obtained in previous studies for the used power distribution system allowing consideration of different weather scenarios during the resilience estimation. Our main contribution is the methodology for assessing power distribution systems under different weather scenarios, which integrates time-to-event models to estimate resilience. The assessment is carried out on distribution networks of a Brazilian power company, and using two resilience perspectives: structural, considering networks connectivity and service, quantifying energy delivered. An investigation of which features are related to the resilience loss in extreme weather is also performed. Results show that all the networks are resilient to a normal weather condition in both perspectives and weather adversities reduce their structural resilience, mainly in large networks. At the same time, service resilience loss is not linearly related to network size, but with how electricity flows over the networks, as the average load passing through vertices and the presence of lateral branches. These findings can support decision making about management and planning of such systems as resilience-oriented modifications, resources allocation, and expansion planning.