This study enhances regional integrated energy systems by proposing a Stackelberg planning–operation model with seasonal hydrogen storage, addressing source–network separation. An equilibrium algorithm is developed that integrates a competitive search routine with mixed-integer optimization. In the price–energy game framework, the hydrogen storage operator is designated as the leader, while energy producers, load aggregators, and storage providers act as followers, facilitating a distributed collaborative optimization strategy within the Stackelberg game. Using an industrial park in northern China as a case study, the findings reveal that the operator’s initiative results in a revenue increase of 38.60%, while producer profits rise by 6.10%, and storage-provider profits surge by 108.75%. Additionally, renewable accommodation reaches 93.86%, reflecting an absolute improvement of 20.60 percentage points. Total net energy imbalance decreases by 55.70%, and heat-loss load is reduced by 31.74%. Overall, the proposed approach effectively achieves cross-seasonal energy balancing and multi-party gains, providing an engineering-oriented reference for addressing energy imbalances in regional integrated energy systems.

The rising penetration of electric vehicles is driving huge demand for lithium batteries, making low-carbon manufacturing a critical objective. This goal is challenged by insufficient production scheduling flexibility and the neglect of carbon-reduction technologies. To address these challenges, this paper develops a low-carbon planning methodology for lithium battery plant energy systems by leveraging manufacturing chain flexibilities. First, a lithium battery energy–carbon material modeling approach is developed that accounts for process production delays and intermediate product storage to capture schedulable process energy consumption patterns. A nitrogen–oxygen coupling production framework is introduced to facilitate oxygen-enriched combustion technology application, while energy recovery pathways are incorporated given the high energy consumption of the formation stage. Subsequently, a process scheduling-driven planning model for lithium battery industrial integrated energy systems (IIES) is developed. Finally, the planning model is validated through four contrasting case studies and systematically evaluated using multi-criteria decision analysis (MCDA). The results demonstrate three principal conclusions: (1) incorporating process scheduling effectively enhances process energy flexibility and reduces total system costs by 19.4%, with MCDA closeness coefficient improving from 0.257 to 0.665; (2) oxygen-enriched combustion increases maximum combustion and carbon capture (CCS) rates from 90% to 95%, reducing carbon tax to 40.5% of the baseline; (3) energy recovery on the basis of process scheduling further reduces costs and carbon emissions, with battery recovery achieving an additional 30.2% cost reduction compared to 24.1% for heat recovery, and MCDA identifies this integrated approach as the optimal solution with a closeness coefficient of 0.919.

As the global energy transition accelerates, distribution systems are integrating increasing shares of inverter-interfaced renewables, making reliable voltage support a key operational requirement. In grid-connected microgrids, especially weak radial feeders in rural and remote areas, voltage-reactive power (Volt/Var) control must coordinate multiple inverters under uncertainty from photovoltaic (PV) intermittency, load volatility, and point-of-common-coupling (PCC) disturbances. Existing droop, model-based optimization, and non-graph reinforcement learning (RL) approaches often rely on fixed rules or do not explicitly exploit electrical topology, which limits adaptive coordination. To address this gap, we propose a topology-aware graph reinforcement learning framework for voltage-reactive power control in grid-connected microgrids under uncertainty. The method encodes node states with a graph convolutional network (GCN) and learns coordinated PV/storage reactive-power actions via proximal policy optimization (PPO) with a multi-objective reward balancing voltage quality, control effort, and action smoothness. In a controlled comparison against a multilayer perceptron (MLP)-PPO baseline with identical action space, reward, and PPO objective, our method reduces voltage violation rate (VVR) from 0.0316 ± 0.0086 to 0.0048 ± 0.0019. Additional validation on a modified IEEE 33-bus feeder further reduces VVR from 0.00726 for MLP-PPO and 0.02999 for Droop control to 0.00095, supporting the effectiveness of topology-aware state representation on a larger radial benchmark feeder.