Prediction of Dynamic Carbon Emission Factors Based on the Informer Model

With the continuous promotion of the dual carbon strategic goal, how to accurately predict the dynamic carbon emission factors in real time has become a key problem. Based on this, this paper proposed a dynamic carbon emission prediction based on Informer model. Firstly, the differences of dynamic carbon emission factors of different power equipment were analyzed, and thermal power units with the largest contribution to carbon emission among traditional energy sources represented by fossil energy were selected for study. Secondly, the Informer model used for long time series forecasting tasks was transferred to the fitting model of dynamic carbon emission factor in unit operation. The nonlinear relationship between the unit operation climb power and carbon emission factor was extracted by using the efficient extraction ability of the encoder's time series features, and the dynamic carbon emission factor in the future was predicted. Finally, based on the data collected by a regional power grid company, error evaluation metrics and correlation evaluation metrics are used to verify the effectiveness of the algorithm.