Membrane Parametric Form-finding

Tensile-membrane architecture via parametric form-finding — Kangaroo-based physical simulation in Grasshopper for shell and membrane geometries whose form emerges from force equilibrium rather than top-down authorship.

membrane form 1

Parametric form-finding for membrane structures: geometries that emerge from the equilibrium of forces rather than being explicitly modeled. Built with Kangaroo physics inside Grasshopper, the study explores a family of tensile shells — each one a stable solution to an underlying mesh + force-density problem.

membrane form 2 membrane form 3 membrane form 4

Approach

Kangaroo physics simulation inside Grasshopper for mesh relaxation. Each shell starts as a flat triangulated mesh; springs along edges resolve to a minimal-energy configuration under the imposed boundary conditions
Anchor-point + force-density experiments — pinning, tensioning, and releasing edges in different combinations produces a family of related shells. The designer’s lever is not the geometry directly but the topology and force schedule that lead to it
Form emerging from equilibrium — rather than explicit modeling of the final geometry, the designer sets up boundary conditions and constraints; the shape follows. This inverts the usual CAD workflow: the architect specifies what the structure must do and the simulation reports what it must look like

Why force-driven form

Tensile membranes, cable nets, and shell structures share a property that flat-packed CAD modeling cannot fake: their geometry is the structural diagram. A doubly-curved minimal surface under tension carries load through pure axial force. A funicular shell in compression follows the inverted form of a hanging chain network. The same logic Antoni Gaudí used with weighted strings at Sagrada Familia — except the iteration loop now runs at thirty frames per second.

For the architecture student this means the form-finding step is also the structural-intuition step. By tuning anchor positions, edge tensions, and density, the designer learns to read the relationship between boundary topology and emergent geometry — which is the precondition for designing structurally honest membrane buildings rather than decorating them after the fact.

Outputs

The study produced four representative shell families documented in the gallery: a saddle, a four-point pavilion, a cable-net shading screen, and a tessellated patchwork that demonstrates how multiple smaller form-found pieces can be stitched into a larger composite envelope. Each family is a parameter range, not a single geometry — the underlying Grasshopper definition exposes anchor positions, tension scalars, and mesh density as live sliders.

Independent research at Rice University, 2021–2024, advised by Prof. Juan Jose Castellon as part of the broader CNT-fiber experimental-models lab. Companion to the [[2021-2024-Rice—fiber-based-pavilion|Fiber-based Pavilion]] — both share the parametric form-finding vocabulary, with the pavilion extending the membrane logic into a fabricated full-scale prototype.