Parametric Design

Parametric design is a way of modelling in which the geometry is not drawn directly but produced by a set of rules operating on input values. A wall is not a polyline; it is the result of a curve, a thickness, a height, and a rib spacing. Change any of those values and the wall regenerates. The output is still ordinary geometry — what changes is that the design intent, rather than the final shape, is what has been recorded.

How it works

A parametric model is a directed graph. Inputs are numbers, curves, points, or data lists — a wall length, a number of divisions, a site boundary. Operations transform them: offsetting, dividing, extruding, filleting, arraying, evaluating a surface. Outputs are the resulting geometry, and often also derived data such as material quantity or element count. Because the graph is explicit, the relationship between an input and the result is traceable, which is not true of a drawn model.

In practice this is built in a visual programming environment. Grasshopper, running inside Rhino, is the common choice in fabrication work: components are placed on a canvas and wired together, and the geometry updates live as sliders move. The same logic can be written as a script; the visual form simply makes the dependency structure legible.

For fabrication, the important step is embedding constraints in the definition rather than checking them afterwards. If the printer has a 30 mm nozzle, wall thickness should be expressed as a multiple of bead width, not as a free number. If the minimum printable radius is bounded by bead width and acceleration, the definition should refuse to produce a tighter curve. Once the constraints live in the model, an out-of-range design cannot be produced by accident.

The definition can also generate the toolpath directly. For printing, that means slicing the geometry into contours at the chosen layer height, ordering them, inserting travel and start–stop positions, and exporting machine code. Doing this inside the same definition keeps geometry and toolpath consistent: there is no possibility of printing a file that no longer matches the design.

Key parameters

The value of the approach comes from relating these groups. Texture amplitude is not a free aesthetic choice on a printed wall: beyond a certain per-layer offset, the bead overhangs its support and collapses. A well-built definition ties amplitude to layer height so that the aesthetic parameter cannot violate the fabrication one. That single relationship prevents a large share of failed prints.

Applications

In printing work, parametric definitions are used for wall and shell geometry with variable section, surface texture and relief patterns derived from layer offsets, structural optimisation of ribs and cavities, families of related products such as planters or benches in many sizes, adaptation of a standard design to site-specific dimensions, and direct toolpath generation. Outside printing, the same approach drives CNC cutting layouts, panelisation of complex surfaces, facade systems, and quantity extraction.

Its clearest commercial use is mass customisation: one definition, many variants, each individually generated without redesign — which is exactly the economic model that additive fabrication supports.

Advantages

Variation becomes cheap. Once the definition exists, producing twenty related designs costs little more than producing one. Changes propagate: adjusting a base curve updates the geometry, the quantities, and the toolpath together. Constraints can be enforced rather than checked, which removes a class of errors before they reach the machine. The design logic is explicit and reusable across projects. And the model can output data — quantities, areas, print time estimates — that would otherwise be measured by hand.

Limitations

The upfront effort is real: building a robust definition takes longer than drawing a single object, and it is wasted if only one object is needed. Definitions become opaque as they grow, and an undocumented one is often easier to rebuild than to understand. They are brittle outside their intended input range, failing in ways that are hard to diagnose. There is also a temptation to generate complexity because it is easy, rather than because it serves anything. And parametric fluency does not substitute for construction knowledge: a definition that does not encode real fabrication limits will happily generate unbuildable geometry with great precision.

Concreef context

Concreef uses parametric modelling as the front end of its printing work: geometry is defined with rules that reference the actual machine and material constraints of the Crane WASP printer in its Sofia workshop, and the toolpath is generated from the same definition rather than exported and re-processed separately.

Those constraints are not theoretical. They come from the test walls printed in the workshop — roughly one metre in height — and from the material experiments run alongside them, which is how values such as usable layer height, stable bead width, and the per-layer offset at which a texture stops standing up were arrived at. Concreef has no completed buildings, no delivered client projects, and no legal entity yet, so the parametric work is aimed at element-scale objects: furniture, planters, panels, and prototypes, where a definition can be validated against something physically printed rather than against an assumption.