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.

Context

Period: 2021-2024, Rice University. Role: individual research, 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]] research — both share the parametric form-finding + computational architecture vocabulary, with the pavilion extending the membrane logic into a fabricated full-scale prototype.