Robust Generation Dispatch with Purchase of Renewable Power and Load Predictions
From MaRDI portal
Abstract: Robust scheduling enables the operator to determine an optimal strategy that can withstand all scenarios in an uncertainty set. However, with the proliferation of distributed energy resources, there are rising uncertainties from distributed agents of which the operator has little data. This makes it difficult to build an appropriate uncertainty set, leading to inaccurate, or even infeasible, scheduling strategies. This paper proposes a novel two-stage robust scheduling model that allows the operator to purchase predictive information from distributed agents to make better scheduling decisions. An improved uncertainty set with a smaller variation range is used, which is developed by combining the forecasts from the operator and distributed agents. The proposed model is a case of robust optimization with decision-dependent uncertainty (DDU) since the improved uncertainty set is influenced by the first-stage information purchase related decisions. An adaptive column-and-constraint generation (C&CG) algorithm is developed to solve the problem within a finite number of iterations. The potential failures of traditional algorithms in detecting feasibility, guaranteeing convergence, and reaching optimal strategies under DDU are successfully circumvented by the proposed algorithm. Case studies demonstrate the effectiveness, necessity, and scalability of the proposed method.
This page was built for publication: Robust Generation Dispatch with Purchase of Renewable Power and Load Predictions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6412581)
