To effectively handle the multi-source uncertainties and carbon emission constraints caused by high renewable energy penetration in distribution networks under the “dual-carbon” policy

this paper proposes a day-ahead and intra-day two-stage coordinated operation method for distribution energy storage that comprehensively coordinates the carbon emission processes of all components in the distribution network balancing area. Specifically

at the day-ahead stage

a typical scenario-based energy storage operation model is established

wherein medium- to long-term scenario clustering of wind

photovoltaics

and load is utilized to formulate an energy storage charging and discharging strategy that accounts for both carbon quota trading and system economics. Subsequently

at the intra-day stage

a deep reinforcement learning-based scheduling strategy is introduced by leveraging a Markov decision process to tackle real-time fluctuations in wind/solar power

electricity price

and load

further incorporating a carbon quota trading mechanism within the distribution network. Simulation results show that this method can be adopted to achieve effective scheduling of energy storage resources and each component within the distribution network under complex uncertain conditions. Compared with conventional strategies

the proposed approach significantly reduces both operational costs and carbon emissions

providing a viable technical pathway for achieving low-carbon and economic objectives in distribution networks.