Damage-Based Stochastic Restoration Against Wind-Induced Events for a Resilient Distribution System

Rapid increase in weather events suggests the need for an approach to capture and quantify the damage intensity of uncertain events. This paper develops a tool that formulates the problem as a two-stage approach against uncertain wind events. The first stage investigates the dynamic wind impact on vulnerable overhead components and models the component-level fragility curve. This approach overpowers the past fragility modeling by framing a hybrid dataset for a realistic weather-dependent analysis. The second stage formulates the problem mathematically using the markov decision process, where the state transition is illustrated via a directed graph, exhibiting five stages: normal, failure, retrofit, replace, and restore. It depicts the statistically estimated action and optimizes the cost as a reward. The novel event-driven dynamic programming algorithm is devised to minimize the expected cost by deriving an optimal policy for each state-based action. The formulated damage-level assessment matrix evaluates the real-time damage status and suggests a relevant remedial measure. The proposed approach is implemented on modified IEEE 33 bus system with an extensive case study, delivering optimal results for a resilient and cost-effective power distribution system.