Robust Estimation Model of Wind Power Prediction Availability in the Period of Power System Peak Load

The uncertainty of wind power prediction accuracy is caused by the scheduling of the grid because the development plan can not be accurately equivalent to the wind turbine as an adjustable unit, thus the prediction of dispatching and capability of wind power will be inaccurate. Therefore, in the case of the grid peak load periods of different scenarios and the corresponding wind power, we put forward a probability distribution parameter estimation method of wind power prediction credibility under the condition of uncertain power prediction error probability, and established a robust estimation model of wind power power prediction credibility. Firstly, the robust optimization model of uncertainty probability reliability domain for wind power prediction under certain reliability level was established. Then, a model for wind power prediction reliability robust estimation based on the uncertainty of prediction error probability was established according to divergence function between reliability of day-ahead wind power prediction and prediction error, and the robustness assessment method of wind power prediction reliability during the time interval of the known load peak occurrence was proposed; Finally, a wind power accommodation equivalent model of a power grid in Northeast China was established, and the wind power prediction reliability was equated to the virtual unit for day-ahead generation schedule according to the reliability estimation parameters. The simulation results demonstrate that the wind power prediction reliability model can effectively restrain the uncertainty of wind power prediction and improve the capacity of wind power accommodation. © 2019, High Voltage Engineering Editorial Department of CEPRI. All right reserved.