A Multi-stage Coordinated Scheduling Method for Heterogeneous Resources in Distribution Systems

Quantitative characterization of multidimensional uncertainties for heterogeneous renewable energy resources is a challenging problem. Existing methods typically rely on experiential knowledge and predefined policies. However, some temporal and spatial correlation relationships can not be precisely formulated based on experience, potentially leading to suboptimal solutions. Motivated by these challenges, this paper proposes a data-driven budget uncertainty set generation method to formulate the temporal and spatial correlation relationship for heterogeneous resources in distribution systems. The main idea is to iteratively find cutting planes (budget constraints) by solving straightforward optimization problems. The computational complexity is low. Besides, a multi-stage robust optimization framework is developed to derive an optimal coordinated scheduling policy for various renewable energy resources and storage devices. When uncertainties are observed gradually, decisions are adaptively optimized in a rolling horizon manner. No explicit decision assumption is required and therefore the performance is improved compared with existing methods. Numerical tests are implemented on a modified IEEE 33-bus system, verifying the effectiveness of the proposed method.