A Study Of Robust Dispatch For Power System Considering The Predictive Error Of Renewable Energies

In recent years,the installed capacity of wind power,solar and other renewable energies increases rapidly due to the technology development and cost reduction.Renewable energy sources(RES),represented by wind power,are generally characterized by intermittent and volatility.The forecasting errors of RESs are relatively large.Large-scale renewable energy has brought great challenges to power system security.Robust optimization(RO)has attracted extensive attention in the field of power system dispatching.It does not require the distribution function of random variables.Moreover,it avoids generating of multiple typical scenes.A robust unit commitment model considering prediction error of RES is established in this paper.The conservativeness of RO is directly infected by uncertainty set.The fundamental concepts and theory of robust optimization are introduced in the first place.Then we introduce a classical solution algorithm: C&CG algorithm(column and constraint generation).Furthermore,we examine the forecasting error of wind farm output and establish a temporal-correlation constraint based on Pearson correlation coefficient.Considered its discrete character,the constraint is simplified to mixed integer linear constraint.In this way,the temporal-correlation constraint can be applied in robust unit commitment(RUC)problems.We improve the solution of sub problem in Bender's decomposition.Based on C&CG algorithm,we develop a new solution for RO with discrete uncertainty set.The bilinear programming caused by the dual form of the sub problem is transformed into a mixed integer linear programming problem.This algorithm is suitable for discrete uncertainty set.At last,numerical experiments are performed with real-world data of wind farms.The results suggest that the proposed algorithm can improve RUC performance without jeopardizing system robustness.