Multistage Stochastic Optimization for Microgrid Operation Under Islanding Uncertainty

Microgrids can be distinguished from traditional power systems based on their ability to perform the islanded operation. This study explores the problem of optimizing the microgrid operation in the context of uncertain islanding events. This occurs when the microgrid is isolated from the main grid and it operates as an independent system. A multistage stochastic optimization model that considers multi-occurrence and multi-period islanding events is proposed to optimize the proactive policy. This is achieved by creating parameters for the maximum number of time periods of the islanded operation in the planning horizon. To solve the resulting large-scale mixed-integer program, an efficient algorithm based on the two-stage Benders' decomposition method is proposed. Numerical experiments show that the proposed policy significantly reduces the expected operation costs of a microgrid. This is more cost-effective in comparison to the other reactive policies that prepare a certain reserve level and reschedule the resources after an islanding event occurs. This study demonstrates that the proposed decomposition algorithm can efficiently solve large-scale problems that have longer planning horizons or a larger number of time periods for the islanded operation.