Day-Ahead Energy Power Forecasting under Carbon Neutrality Based on XGBoost

An essential solution to global warming is to achieve carbon neutrality. The excessive use of fossil fuels has caused many greenhouse gas emissions. Converting fossil fuel energy into clean energy is an essential strategy for carbon neutrality. The accurate forecast of the power generation of new energy based on wind energy and solar energy can help power companies study load planning and use energy rationally. This paper introduces the development of a forecasting model for the power generation of wind power plants and photovoltaic power plants within 24 hours. The purpose is to obtain accurate wind power and photovoltaic power forecasting using a gradient boosting decision tree (XGBoost). This paper proposes adding time information to the input variables of photovoltaic power generation forecast. Compared with the forecast results obtained by traditional physical methods, the proposed method can improve the relevance between the forecast and the ground truth. Accurately forecasting the generated power is helpful for the study of load planning, which achieves the effect of neither wasting energy nor affecting the regular use of electricity. Our research plays a vital role in advancing carbon neutrality plans by increasing the utilization rate of new energy.