A Hybrid Neuro-Fuzzy & Bootstrap Prediction System for Wind Power Generation

Abstract

Increasing integration of wind turbines in the electrical grid creates more challenges daily because of the unstable power of these units because they rely mainly on the wind. One of the solutions used to deal with these challenges is to predict the power produced from the wind turbine for a short or long term. Prediction of the power to be produced from the wind turbines gives a future vision of how to deal with these sources. Many techniques are used to improve the prediction of wind turbine energy. In this paper, a bootstrap hybrid neuro-fuzzy short-term prediction system for wind turbine power is presented. The bootstrap technique is used to increase the accuracy of the predicted value by creating multiple small datasets with the same features of the main dataset. A neuro-fuzzy model is created for each dataset using the hybrid optimization agent in MATLAB. The output of each neuro-fuzzy model is a prediction of wind turbine power. Therefore, the output of all models is combined using the combination model to calculate the final predicted wind turbine power. The system is simulated using MATLAB and the result show that the bootstrap hybrid neuro-fuzzy system predicts the wind turbine power for the next 24 h with accuracy 94.3%. The relative percentage error and average error for the system outputs are calculated and showed that the error is minimized to ±5.7%. Therefore, this prediction system could be useful for the challenges of integration wind turbine in electrical grid