Scalable optimization approaches for microgrid operation under stochastic islanding and net load

This paper focuses on operating a microgrid when uncertain islanding events and net load are stochastic, from the perspective of a microgrid operator. To derive effective and reliable decisions that can adapt when these uncertain factors sequentially reveal in a given planning horizon, multistage stochastic optimization models that address the dynamics and probabilistic nature of uncertainty can be used. However, since they suffer from the curse of dimensionality in that the number of scenarios needed grows exponentially with regard to the number of time periods, it is not practical to use the models directly. To overcome this drawback, this paper proposes scalable optimization approaches that efficiently derive good-quality solutions that are adaptable to sequential realization. In particular, integrated optimization models are presented to deal with both uncertain factors based on the combination of the proposed models for each of the two, which is scalable compared to the standard multistage model used in the literature. Numerical experiments demonstrate that practical-sized instances can be efficiently solved using the proposed models, whereas the standard multistage model in the literature, despite well-known decomposition methods being applied, cannot provide solutions within a reasonable time. The results also demonstrate the effectiveness of the solutions from the proposed models compared with the standard multistage model.