Computational Design and Robotic Fabrication of Dry-Stacked non-standard spanning Limestone Assemblies

Natural stone, an abundant and low-carbon material, often generates substantial unusable waste through quarrying and masonry processes, including rubble, offcuts, and irregular fragments. While there is growing interest in repurposing these byproducts, their non-standard and irregular geometries pose major challenges for integration, restricting their use in architecture primarily to cladding, flooring, and other non-structural applications. Finding innovative applications for this waste could help revive the use of stone in construction while further lowering its environmental impact. This study introduces a workflow to optimize the geometry of non-standard stone blocks through minimal modifications, enabling construction with highly variable, locally sourced material. The proposed methodology consists of a heuristic generative algorithm based on a library of 3D scanned limestone blocks and guided by a pre-defined global geometry. This enables a data-driven model for the targeted modification of joints between blocks through robotic machining. In turn, robotic machining is thereby optimized to save energy and time. This approach was validated through the fabrication of 18 stones, which were assembled into a three-legged arch, demonstrating the feasibility of using non-standard stone for architectural structural applications.