TPMS-Optimized Cement Geometries for Carbon-Efficient Structural Performance

Abstract

Cement production generates 8% of global anthropogenic CO₂ emissions, and concrete's low tensile strength (1/10 of its compressive strength) leaves structures vulnerable to seismic failure. This study shows that topology-optimized Triply Periodic Minimal Surface (TPMS) geometries can simultaneously reduce cement consumption and improve structural resilience without altering material chemistry. We generated and computationally evaluated 8,750 unique TPMS configurations, two orders of magnitude larger than prior studies, across the full structural porosity range (0-70%) by varying seven topological families, wall thickness, radial porosity gradients biomimicked from trabecular bone, and span. All structures were simulated under Mazars continuum damage mechanics; top performers then underwent Newmark-β seismic simulation, with physical validation via a calibrated shaker table. Retaining 93% of the solid's compressive strength, Lamella was the optimal topology, reducing mass by 32% and achieving 139.9% of the control's strength-to-weight ratio, along with an 180x improvement in tensile strength-to-weight ratio. All seven topologies achieved 85-98% seismic response amplification reduction and 50-88% peak stress reduction, shifting failure mode from brittle crushing to distributed energy absorption. Dual machine learning frameworks enabled predictive optimization (R² = 0.987-0.999) and generative synthesis of hybrid topologies achieving 104% of the best classical compressive performance. These results offer a component-level CO₂ reduction pathway requiring no supply-chain or workforce changes, while democratizing seismic resilience at the material level for earthquake-prone, low-income communities.