A Wind Power Prediction Method Based on Width Learning

Abstract

The invention discloses a wind power prediction method based on width learning, which obtains the power data of a certain wind turbine for one year through the SCADA system, uses an outlier detection algorithm to detect abnormal points on the power data of one year, removes bad data, and obtains The normal power data of the unit. Using the normal power data, calculate the daily average power data of the wind turbine for one year, group the normal daily average power data, and perform normalization operation on each group of power data. Introduce seasonal information, construct a training set for spring, summer, autumn and winter width learning system, and construct a width learning system for spring, summer, autumn and winter. Using the constructed training set, the width learning system is trained to obtain spring, summer, autumn and winter power prediction models. Use the power forecasting model to forecast wind power 24h. The design process of the invention is simple, the prediction model parameters obtained are few, the prediction accuracy is high, and accurate and effective power prediction information can be provided for wind power grid connection.

Description

technical field [0001] The invention relates to the field of wind power generation grid-connected control, in particular to a wind power prediction method based on width learning. Background technique [0002] With the development of the world economy, people's demand for energy is increasing day by day, and the energy crisis has become increasingly prominent. In order to alleviate the energy crisis, people have turned their attention to renewable energy. As a new energy power generation method with the fastest development and the most mature commercialization in the world today, wind power generation has been favored by researchers. [0003] However, the wind in nature is highly random and intermittent, which will lead to fluctuations and instability of the output power of the wind power system, which will greatly affect the quality of wind power grid connection. Therefore, wind power forecasting has important research significance for improving wind power stability and g...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the problem that the existing wind power prediction method does not consider the outlier detection of historical data and the seasonal information, the prediction accuracy is not high, and the existing method has a complicated training process and many parameters that need to be debugged. The present invention provides a A wind power prediction method with high precision and simple implementation, which can eliminate abnormal data in the historical training data and incorporate seasonal information into the power prediction model, requires less parameters to be debugged and the training process is simple, and can provide reliable, Effective and accurate grid-connected information to improve the quality of wind power and the economic benefits of wind farms

Images