A Model for Predicting Wind Speed and the Probability of a Wind Gust

The purpose of this work is to build a model for predicting the speed and probability of a wind gust to create a model for controlling a wind generator with a segment stator and obtaining an optimal characteristic for the return of electricity. The method of implementing the model is the analysis of wind data taken from the Open Weather Map website and the further construction of a forecasting model. The method of obtaining data is described in detail in the article [1]. We will analyze and process data using Excel programs and the Python programming language in the Jupyter Notebook shell. The purpose of data analysis is to determine the dependence of a wind gust on a change in wind direction and speed. The data are samples for a certain period from 05.12.2020 to 08.12.2020 with a sampling frequency of 1 point in 30 minutes, which is not enough for long-term forecasting, but it is suitable for forecasting for the next 6-12 hours. The data set reflects only one time of the year, and to create an accurate model, it is desirable to use historical data to estimate the change taking into account the change of seasons. The main result of the work is a model for predicting wind speed and the probability of a wind gust using the “k” nearest neighbor method-KNN. This method has proven itself well in this work [2]. An additional result is the preparation of data and prerequisites for the creation of a neural network model for controlling the drive of a wind generator segment depending on changes in the wind situation. Since high quality, training of a neural network requires a larger amount of data in terms of both their quantity and duration, this article will only consider the prerequisites and methods.