Large-scale wind power integration imposes technical challenges to the power system operation and scheduling. Traditional deterministic generation dispatch methods cannot ensure system operation security because they don't incorporate wind power uncertainties appropriately. A new robust dispatch scheme based on the wind power prediction interval is introduced for intra-day active power dispatch

in which wind farms are controlled with the maximum-power-point-tracking strategy mostly. This scheme is featured with flexibility and practicality. In combination with this scheme

a robust interval economic dispatch model is formulated as a robust optimization method and further simplified as a quadratic programming

taking the response of AGC regulation into consideration. Finally

Monte Carlos simulations are conducted on a modified IEEE 24-bus system to compare performance of the proposed approach and the traditional one

and the robustness

reliability and economy of the robust dispatch scheme is verified.