Wind speed point prediction and interval prediction method based on linear prediction model, neural network, and deep learning

With the development of the times, wind energy has increasingly become crucial energy, which has attracted the attention of researchers. To maximize the use of wind energy, the precision of wind speed prediction is very important. Naturally, the calculation of wind speed prediction accuracy has become a critical link in wind energy utilization, and also plays a decisive role in energy supply and management. However, in the process of wind speed prediction, the instability and variability of wind speed make the prediction particularly difficult. Many scholars have done a great deal of research and put forward corresponding methods to improve this situation, but the problem that a single model can solve is limited. In order to improve this vulnerability, this paper proposes a combined model founded on data preprocessing, linear prediction model, neural network, deep learning, and multi-objective optimization to overcome this problem. The predictive results prove that the proposed combined model in this paper has advantages and can effectively improve the prediction accuracy, and it is proved that the combined model proposed can reach Pareto optimal theoretically. To effectively analyze the volatility and uncertainty of the combined model, based on the distribution function theory, an interval prediction model is established in this paper, and the experimental predictive values show that it can increase forecast precision, enhance prediction stability and make up for the shortcomings of the single models.