Multi-stage stochastic optimization of islanded utility-microgrids design after natural disasters
Rodney Kizito, Zeyu Liu, Xueping Li, Kai Sun
Abstract
Natural disasters (e.g., hurricanes) can cause widespread power outages within distribution networks and interrupted power supply to critical loads (e.g., grocery stores, hospitals, gas, fire, and police stations) that provide utility services. Microgrids are localized power grids that can incorporate solar/photovoltaic (PV) distributed generators (PV-DGs) and energy storage systems (ESSs) for stand-alone system operations independent of the main grid, known as the island mode. This study investigates a microgrid design problem using PV-DGs and ESSs when facing prolonged power outages in the main grid. We propose a multi-stage stochastic program that holistically considers the techno-economics of microgrid investment and daily operations by optimizing the reliability and resilience of the microgrid during a week-long power outage. The model is designed from a utility perspective that includes budget constraints for investment. Due to the large model size, we develop a nested L-shaped algorithm that solves the problem exactly and analyzes the microgrid’s reliability across different weather scenarios in the entire decision-making horizon. Results from a case study using real-world data show that an islanded utility-scale microgrid can effectively provide uninterrupted power supply to a network of 5 and 10 critical loads, covering 100% and 97% of the demand in all possible future scenarios, with potential investments of $8 million and $15 million, respectively.