Accurate global horizontal irradiance (GHI) prediction is significant for the stability and economy of power system operation. This paper proposes an advanced model Informer to predict GHI and selects Root Mean Square Error (RMSE) as the primary evaluation metric to calculate the error. Through discussing the Pearson correlation coefficient, chooses five strong correlation parameters and selects the optimal input through the experiments of 2 and 3 inputs. The Informer model is compared with five reference machine learning (ML) models, and the performance improvement is over 99% under optimal input, which proves the proposed model's superiority. Finally, the Informer's long series prediction ability is verified, and the results showed that Inforemer can efficiently complete long series prediction tasks without losing accuracy, which has high practical value.
 

Global horizontal irradiance (GHI); forecasting; Inforemer; Optimal input; Long series