November 2020
Phase 1: Exploration
1.2 Parametric Mesh Modeling
Objective: Develop a pathway to digital modeling & simulating woven structures.
To explore widely what possible shapes the weave can be applied to, I applied software that I had been introduced to during my Industrial Design Masters at Georgia Tech. I mainly used the digital modeling software Rhinoceros 3D, a nodebased parametric controller called Grasshopper, and its inbuilt physics engine Kangaroo.
There are 3 elements necessary to digitally model a flexible woven surface:

Modeling the unit cell & grid mesh population

Filling pattern into skewed mesh units

Meshes and forces
1. The cell
In a woven surface, the repeated stitch can be thought of as the single unit of a rectangular grid. After modeling this stitch geometry, it can be duplicated onto each cell of a grid of any size.
Because there is just one single programmed geometry for the weave, changing parameters of that geometry affects every cell in the grid mesh.
2. Selfreferencing grid units
Instead of using and XY grid system, in which all units are the same length and on a flat plane, a more flexible approach is one where any surface area can be subdivided rectilinearly, without requiring lines to be parallel or cells to be of equal size. In surface modeling, this is known as a UV grid.
UV grids allow flexible and nonplanar shapes to be subdivided in ways that operate identically to the XY system, with coordinates, vertices, and cell edges. When set up correctly, UV grids can skew out of regularity and planarity, yet utilize familiar regular 2D ordering structures.
3. Meshes and forces
The Kangaroo2 plugin for Grasshopper allows for a grid area to be treated as a flexible mesh. Meshes can be programmed to receive force and anchoring from at specified points and vectors. This allows for simulations of gravity, anchoring, floatation and lateral tidal forces. The result of these simulations retains the data structure of the original grid.
Integration into a single digital model
When a mesh distorts in response to forces (#3), data structure of the mesh remains intact, and only the skew of the mesh lines is changed. By using selfreferential parameters to determine the mesh edges (#2), the boundary and orientation of each face can be filled by the weave geometry (#1). This gives a visualization not only of what a large bobbin lace woven patch would look like, but also an animation of the geometry responding to the forces of gravity and interwoven floatation elements.
This "model" imagines a simple rectangular patch of bobbin lace ground stitches with anchoring bars at either end and flotation beads interwoven in the middle of the design.
Closing Notes:
This section provided a digital pathway for visualizing woven structures with live parameters, allowing for inputting different densities, rope thicknesses, weave widths and counts. To rely on this as a design tool, it's essential that the models produced in this way are accurate representations of what can be physically fabricated. Bobbin lace is woven on pins over a pattern printed onto paper. The pattern is essentially a blueprint for making the woven piece.
Section 1.3 will focus on turning the parametric grid configuration that simulated the digital model into a pattern blueprint for creating physical models with the same simulated behavior.