Wind speed forecasting based on grey predictor and genetic neural network models

Due to high penetration of wind generation in modern power systems, the influence of wind power production on the efficient operation of the power system is increasingly important. Wind speed forecasting is great significant to the connection of wind farms to the electric power system. To solve an increasing interest in forecasting accuracy of wind speed, this paper proposes a new method to improve wind speed prediction accuracy based on different grey models and the genetic neural network. The prediction of wind speed is achieved in two stages. In the first stage, wind speed is predicted by using the different grey models. In the second stage, wind forecasting models is established based on the genetic algorithm neural network (GANN), which of input data based on the results using the different grey prediction models. The weights and thresholds of the neural network are adjusted based on the genetic algorithm. The proposed model is verified by using actual wind speed data, and is compared different grey prediction models in terms of the mean absolute error and the mean relative error. It is concluded that the new model is of high precision. The analysis and simulation results demonstrate that the proposed approach gives better performance.