Resilient Power Grids Under Tropical Cyclone Uncertainty Through a Distributionally Robust Optimization Model

Recent tropical cyclones have underscored the catastrophic impact of such events on infrastructures, highlighting the necessity for advanced planning and resilience in power systems. We develop a distributionally robust two-stage optimization model that employs a bi-dimensional Wasserstein-based ambiguity set to consider both physical and disruptive uncertainties of typhoons. This model effectively integrates forecast errors into the decision-making process, reducing biases from inaccurate predictions and addressing both historical and forecast data inaccuracies. By employing a cutting-plane-based algorithm, we solve the optimization problem and validate the approach through case studies in the Hong Kong distribution system. The results demonstrate improved robustness and resilience of grid operations, confirming the model’s efficacy in preparing for and mitigating the impacts of dynamic tropical cyclone disruptions.