Ultra-short-term wind power prediction method and device and medium

Abstract

The invention discloses an ultra-short-term wind power prediction method and device and a medium, and relates to the technical field of wind power. Obtaining wind measurement data of the target wind power plant and other wind power plants, determining a correlation coefficient of an actual wind speed sequence and an actual wind direction sequence of the target wind power plant and the upwind wind power plant according to the wind measurement data, judging the correlation coefficient, determining whether a reference sequence exists or not, and when the reference sequence exists, judging whether the reference sequence exists or not; related parameters of the wind power plant are determined according to the reference sequence, and finally predicted wind power is determined according to the related parameters. It can be seen that the related parameters of the wind power plant are determined through the reference sequence, the related parameters comprise the wind conditions of the wind power plant, the wind power of the target wind power plant can be predicted according to the specific wind conditions, and the situation that data need to be determined again when problems occur through the data during wind power prediction is avoided; and the complexity and uncertainty of the system during wind power prediction are reduced. In addition, the wind condition of the wind power plant is more in line with the meteorological law, and the accuracy of ultra-short-term power prediction is improved.

Description

technical field [0001] The present application relates to the field of wind technology, and in particular, to a method, device and medium for predicting ultra-short-term wind power. Background technique [0002] As a new energy source, wind power is more suitable for large-scale development. Accurate wind power forecast data can alleviate the pressure of peak regulation and frequency regulation of the power system, and maximize the benefits of electricity spot trading. Reasonable arrangements for fan maintenance to reduce unnecessary downtime and loss of power generation, thereby reducing safety hazards. Therefore, both power grids and power generation companies attach great importance to the forecast of wind power, and the requirements are getting higher and higher. At present, wind farm power prediction systems generally use continuous method, moving average method or machine learning algorithm to establish an ultra-short-term power prediction model based on short-term pr...

AI Technical Summary

Problems solved by technology

[0003] However, model training needs to be performed for each forecast period, that is, if the power is predicted for the next four hours, and the forecast is made every 15 minutes, model training is required every 15 minutes, and the next four hours will appear every 15 minutes When there is a problem with one of the data, all the data needs to be re-determined, which increases the complexity and uncertainty of the system

Images