The Design And Implementation Of A Wind Power Forecasting System For Wind Farm In Gansu Province

We investigated the present situation and future development about wind power forecasting at home and aboard and introduced the technical principle and model in wind power forecasting system in the thesis. A wind power forecasting system combined numerical weather prediction with statistical forecast was built to meet wind power industry requirements in Gansu Province.The wind power forecasting system includes two subsystems: numerical weather prediction and wind power prediction. The meteorological field prediction system uses the mesoscale numerical prediction model to achieve downscaling and difference processing. The resolution of numerical prediction model is 3*3km. It is improved to 1*1km by the dynamic downscaling model. It is interpolated to each height of the wind turbine hub to output the wind speed, wind direction, atmospheric pressure, temperature, humidity, and other meteorological data in grid point for the next three days 72 hours 15 minutes. The wind power prediction system collects the speed data of wind turbine through the OPC interface to calculate the power combined with the results of numerical prediction model. By using the historical meteorological data and wind power data based on the power curve of wind turbine, the linear and nonlinear statistical models between meteorological elements and actual output power are established to implement wind power prediction by using physical and statistical methods(artificial neural networks, multiple regression, etc.).The wind power forecasting system was localized in a wind farm in Gansu Province. The test results show that the system is running stable with reliable results after popularization and application. It could solve the unstable problems of electricity grid system caused by the fluctuation of wind power generation.