Research on Wind Power Prediction Based on Time Series

Abstract

Due to the uncontrollable characteristics of wind, such as randomness and volatility, the uncertainty of wind power production will be caused to some extent. In order to minimize the influence of wind farm on the stability and safety of the power grid, the prediction of wind power becomes particularly important. In this paper, considering many factors affecting power generation, the ARIMA models based on unary and multivariate time series are used for modeling and analysis of multiple input sequences and output sequences respectively. After comprehensive comparison and analysis, the optimal prediction method is obtained, which provides a new idea for wind power prediction.