Use machine learning algorithms to predict turbine power generation to replace renewable energy with fossil fuels

Abstract

Recently, power systems have faced the challenges of growing electricity demand, reducing fossil fuels, and exacerbating environmental pollution due to carbon emissions from fossil fuel-based power generation. Integrating low-carbon alternative energy, renewable energy sources (RES), is becoming very important for energy systems. Effective management of the integration of the production capacity of RES is as important as the production capacity of wind farms with the production capacity of fossil fuel power plants. This article analyzed 850,660 data recorded by a wind farm from March 01, 2020, 00:00:00 to December 31, t2020, 23:50:00 were analyzed. And by using machine learning and extra tree, light gradient boosting machine, gradient boosting regressor, decision tree, Ada Boost, and ridge algorithms, the production power of the wind farm was predicted. The best performance predicting the turbine production power was assigned to extra tree, and the worst performance was related to the Ridge algorithm.