Medium-and-long-term wind power generation prediction method and system fusing machine learning model

Abstract

The invention provides a medium-and-long-term wind power generation prediction method and system fusing a machine learning model. The method comprises the steps of eliminating abnormal values of historical power generation data of a wind power plant, and then processing the historical power generation data into monthly average power generation and daily average power generation; based on the monthly average power generation amount and the historical numerical weather forecast data, predicting the monthly average power generation amount of the next month through a machine learning model; performing variational mode decomposition on the daily average power generation amount, and predicting a daily mean through a machine learning model in combination with historical numerical weather forecastdata; and adding the predicted day distance to the predicted monthly average power generation amount to obtain a predicted value of the monthly average power generation amount. Meteorological factorssuch as wind speed of an existing medium-and-long-term numerical forecasting product are used, a machine learning model is added for forecasting, and the accuracy of medium-and-long-term generating capacity forecasting of the wind power plant is improved.

Description

technical field [0001] The invention belongs to the field of new energy wind power generation, and in particular relates to a medium- and long-term wind power prediction method and system that integrates machine learning models. Background technique [0002] With the rapid increase of wind power grid connection, the power system has put forward new requirements for the accuracy of medium and long-term wind power generation forecasting. For new power plants, mid- and long-term power generation forecasts can evaluate local wind resources, help to make overall decisions and evaluate future power generation conditions; for established power plants, mid- and long-term forecasts are of guiding significance for the formulation of future power generation plans. In addition, the dispatch plan of the power grid, the configuration of energy storage equipment, and the power market after the power reform also require more accurate power generation forecasting capabilities. Wind power ge...

AI Technical Summary

Problems solved by technology

[0003] The prediction of wind speed by traditional numerical forecasting models cannot be directly connected with wind power generation. The grid scale of common numerical models for medium and long-term forecasting is relatively large. The accumulation of errors and the inaccuracy of dynamic downscaling in space, on the other hand, are limited by computing resources

In addition to numerical forecasting models, traditional statistical methods in the application of power generation forecasting have problems such as insufficient consideration of meteorological elements, inability to effectively predict seasonal and annual changes in wind speed and power generation fluctuations caused by its variability, etc.

Images