Comprehensive carbon accounting for power systems considering hybrid power trading mode

The carbon emission flow theory has played a crucial role in assessing the carbon emission responsibilities of electricity consumers, fostering the sustainable transition to low-carbon power systems. However, applying existing theories to carbon emission accounting in new power systems often fails to ensure equitable low-carbon benefits for both sustainable energy generation and consumption. To address this, we propose an innovative comprehensive carbon emission accounting method for the “source-grid-load” process, accommodating both bilateral and pool trading modes amid the increasing integration of sustainable energy. Our approach begins by establishing a dynamic carbon emission factor model of generation units based on the Informer model. We then formulate carbon emission accounting for bilateral sustainable energy supply and consumption, treating it as a maximum flow carbon tracing problem between the source and load. Employing deep reinforcement learning techniques, we continuously solve for carbon emission flow under the uncertainties of “source-load” impacts, culminating in a holistic carbon emission assessment for the new power system. Through a case study using the IEEE 14-node system and real power system in a region, we demonstrate the rationality and fairness of our proposed accounting methods for hybrid trading mode.