Wind Speed Interval Prediction Model Based on Adaptive Decomposition and Parameter Optimization

Wind energy is a renewable and clean energy source, and it is one of the important ways to transform the energy structure and achieve sustainable development. The random fluctuation of wind speed greatly increases the risk of wind power integration. Therefore, achieving accurate wind speed prediction is the key to improve the efficiency of wind power utilization. In this paper, the three aspects of wind speed feature extraction, model parameter optimization and wind speed uncertainty prediction are improved to forecast wind speed. Firstly, using the complete ensemble empirical mode decomposition with adaptivenoise method to extract wind speed features; then, Combining ARMA model and neural network as prediction model, combined with the beetle antennae search optimization algorithm to realize the optimization of model parameters; lastly, the obtained error sequence is estimated by kernel density, and the corresponding interval prediction results are obtained according to the kernel density function quantile points and different confidence levels. The prediction results show that (1) the model obtained by the complete ensemble empirical mode decomposition with adaptivenoise method has less error and better stability than the traditional neural network prediction model; (2) the beetle antennae search optimization algorithm is used to change the initial weights of the model in order to avoid the model getting into local minimum, and the improved model has better prediction results; (3) the kernel density method is used to achieve the uncertainty prediction of wind speed. The Gaussian kernel function is used to fit the error probability density function, and the prediction interval established by the kernel density quantile almost covers the actual wind speed. The model proposed in this paper realizes the deterministic and uncertainty prediction of wind speed, which greatly improves the prediction accuracy and provides a basis on wind power dispatching for the power department.