Abstract
The wind power prediction system is composed of a meteorological forecasting module, calculation module of wind power output and HMI(Human Machine Interface) visualization system. The final information from this system is a short-term (6hr ahead) and mid-term (48hr ahead) wind power prediction value. The meteorological forecasting module for wind speed and direction forecasting is a combination of physical and statistical model. In this system, the WRF(Weather Research and Forecasting) model, which is a three-dimensional numerical weather model, is used as the physical model and the GFS(Global Forecasting System) models is used for initial condition forecasting. The 100m resolution terrain data is used to improve the accuracy of this system. In addition, optimization of the physical model carried out using historic weather data in Jeju. The mid-term prediction value from the physical model is used in the statistical method for a short-term prediction. The final power prediction is calculated using an optimal adjustment between the currently observed data and data predicted from the power curve model. The final wind power prediction value is provided to customs using a HMI visualization system. The aim of this study is to further improve the accuracy of this prediction system and develop a practical system for power system operation and the energy market in the Smart-Grid.