Hourly solar irradiance prediction based on enhanced incremental extreme learning machine

As more and more photovoltaic power generation systems are connected to the grid, the intermittency and fluctuation of their output power have placed a great burden on the grid. Solar irradiance forecasts are critical for planning and managing photovoltaic system power generation. In order to improve the prediction accuracy of solar irradiance, this paper introduces an hourly solar irradiance prediction method. The enhanced incremental extreme learning machine algorithm is used to establish a solar irradiance prediction model. Only the historical irradiance data is used as the Input features, and output the hourly average solar irradiance at the time to be predicted. The experimental results of this paper show that, compared with the original incremental extreme learning machine, the solar irradiance prediction model established by the enhanced incremental extreme learning machine algorithm can obtain smaller prediction errors. At the same time, the enhanced incremental extreme learning machine can obtain more stable and accurate prediction results with fewer nodes.