Optimization of Multi-layer Perceptron for Wind Power Generation Prediction Based on Improved Grey Wolf Algorithm

Wind power is a renewable energy, and its power prediction plays an important role in operation. The traditional wind power prediction model has a lot of room for improvement. This paper proposes a wind power prediction method based on improved grey Wolf algorithm to optimize multi-layer perceptron, in which multi-layer perceptron is a powerful nonlinear prediction method, aiming to improve the model effect from the perspective of selecting optimization algorithms. In this study, data preprocessing is carried out first, which includes outlier filtering, missing value filling and principal component analysis. After that, the multi-layer perceptron model is constructed and trained. Secondly, considering the limitation of searching ability of standard Gray Wolf algorithm, an improved Gray Wolf algorithm is proposed, the convergence factor is designed as the form of adaptive decline, and the position update formula is improved for parameter optimization of multi-layer perceptron model. Finally, the data set of a wind farm in western China is used to verify the experiment. The results show that the prediction effect of the multi-layer perceptron model optimized based on the improved Gray Wolf algorithm is better than other models on the verification set and the test set, and the average absolute error and the root-mean-square error are significantly reduced, which verifies the validity of the proposed method. The method proposed in this study improves the accuracy of wind power prediction and provides a new technical approach for the development and application of renewable energy.