해외사례 고찰을 통한 풍력 예측모형의 과거와 현재

Abstract

The importance of wind energy has been increasing. Most of countries have studied for generating effective forecasting models about wind power prediction. This study follows up research about statistical wind forecasting models from 2000 to 2016 and compares each model in several aspects. Variety of estimators measure such as mean squared error and correlation coefficient that used to compare different models are defined. Most of methods use wind speed for dependent and independent variable. The extended application of artificial neural network and ARIMA model have mainly used to predict wind power in the past. The state-of-the-art of focuses on nonlinear regression, feature selection and Ensemble methods which are different from classical ANN and ARIMA.