A Bio-inspired Dendritic Topology Optimization for CO₂ Capture, Utilization and Storage Hubs: Integrating Microfluid Dynamics with Macro-scale Infrastructure Design

Abstract Large-scale CCUS systems require optimized pipeline networks; however, conventional approaches oversimplify geo-engineering interactions and impose rigid topologies. Here, we introduce a bio-inspired optimization framework merging pore network fluid dynamics with Hebbian learning to generate cost-optimal dendritic topologies and CCUS clusters on standard hardware. Nationwide simulations show this framework achieves 5.7 Gt/a (75%) for standalone projects and 6.5 Gt/a (85%) for hub clusters at <$90/ton levelized costs. Pipeline networks exhibit bimodal centroid-governed topologies, routing between emission and sink centroids while avoiding high-cost zones. Regional architectures vary: radial networks in northwestern basins, north-south patterns in southern China, multi-centric layouts in the northeast, and west-to-east routes with land-ocean integration in the east. Ningxia validation achieves > 80% emission reduction at <$50/ton, outperforming industry standards. The framework delivers four breakthroughs: physical fidelity via micro-scale flow laws, continuous-variable simulation, 32k-grid calculations in 5 minutes (vs. days conventionally), and quantitative trade-offs via a "carbon abatement matrix." Aligning microfluid mechanics with macro-decarbonization, this scalable approach redefines CCUS infrastructure design for global net-zero targets.