A Review of Time- and Space-Differentiated Dynamic Electricity Carbon Factor Calculation Based on Distribution Network Carbon Flow Tracking and Park-Level Source-Load-Storage Carbon-Electricity Co-Optimization
Abstract: [Objective] Under the transition toward new-type power systems, carbon accounting on the electricity consumption side is shifting from annual regional averages to dynamic electricity carbon factors with high spatiotemporal resolution. However, the mechanism of carbon flow tracking at the distribution-network level, the specification of dynamic factor calculation, and their closed-loop application in park-level source-load-storage co-optimization remain unsystematically examined. Following a carbon-flow-theory → factor-calculation → co-optimization → application chain, this paper surveys and synthesizes representative literature on carbon emission flow theory, average versus marginal emission factors, spatiotemporal dynamic carbon factors, integrated energy system modeling, and park-level source-load-storage optimization. It distinguishes the mechanisms and applicability boundaries of three families of dynamic factor calculation—carbon flow tracing, sensitivity-based marginal analysis, and data-driven prediction—and analyzes engineering outcomes across typical scenarios including low-carbon industrial parks, data centers, electric vehicle charging, building cooling and heating, and virtual power plants. [Results] The findings show that: (1) electricity carbon factor methods have evolved through four generations—annual regional average, marginal factor, network carbon flow allocation, and data-driven spatiotemporal dynamic modeling—among which the nodal carbon potential model based on the carbon flow matrix has become the dominant route at the distribution scale owing to its physical interpretability and structural isomorphism with network topology; (2) high R/X ratios, three-phase unbalance, and bi-directional flows from distributed generation make loss-carbon responsibility and loop-flow issues prominent in distribution networks, so the proportional sharing assumption inherited from transmission-level practice requires revision in active distribution networks; (3) Embedding dynamic carbon factors as price signals into park-level source-load-storage optimization can jointly reduce electricity costs and carbon emissions, yet the realized abatement is highly sensitive to factor specification, temporal resolution, and accounting boundary; (4) major obstacles remain, including inconsistent factor specifications, insufficient measurement, contentious loss allocation, lack of causality, double claiming of renewable attributes, and ambiguous life-cycle boundaries. [Conclusions] Toward carbon peak and carbon neutrality, future work should prioritize a unified and comparable accounting framework for time- and space-differentiated factors, carbon-aware distribution system state estimation and digital twins, integrated carbon–electricity–green certificate markets, foundation-model-based carbon factor forecasting, and edge–cloud hierarchical distributed control, so as to move carbon-electricity coordination from ex-post accounting toward real-time regulation.
