Research on wind power prediction based on machine learning

The accuracy of wind and solar power prediction not only affects the assessment of "two regulations" for various wind and solar power plants, but also affects the power generation revenue of wind and solar power plants in the electricity market. Regarding the power prediction of regional wind and photovoltaic power plant clusters, the traditional method of summing up the power of individual stations has low accuracy and efficiency due to different station construction times and uneven single station forecasting accuracy. Therefore, a machine learning method is used to reasonably divide the wind power plants and photovoltaic power station clusters in the region, select representative power stations, and combine historical power data to optimize and correct the numerical prediction elements. Then, a short-term power prediction framework model based on the wind power plants and photovoltaic power station clusters is established using the BP neural network method to improve the prediction accuracy and efficiency.