GRNN Network Based Node Carbon Emission Factor Prediction Algorithm for Carbon Reduction Optimization on the Grid User Side

With the development of society, the energy industry’s carbon emissions are witnessing a growing share attributed to the power industry. Circumventing model rigidity and inefficiency in classic carbon flow analysis, this work employs GRNN for real-time carbon intensity forecasting. A nonlinear relationship between grid load data and carbon emission factor (CEF) has been constructed in this framework, effectively removing the reliance on grid topology information and power flow equations. Besides, extensive tests conducted on two IEEE systems have validated the effectiveness of the GRNN-based approach. Specifically, for the 39-bus IEEE system, the Mean Absolute Percentage Error (MAPE) attained by the proposed approach stands at 2.44%, and it merely takes 0.4058 s for the computation process. Therefore, this approach can achieve real-time carbon footprint tracking and low-carbon dispatching, providing a feasible solution for refined estimation precision and speed of users’ carbon emission measurement and optimizing the operation of the power system.