# Parametric Design and Computational Tools Parametric design is a methodology in which the form, performance, and organisation of a building are driven by defined parameters and their relationships — rather than by fixed geometric decisions. The designer establishes rules, constraints, and objectives; the computational tool generates, evaluates, and optimises solutions within that design space. For the practicing architect, parametric and computational design tools extend the designer's capacity to explore complex geometries, optimise performance, manage large datasets, and produce fabrication-ready outputs — but they demand a fundamentally different workflow from conventional drawing-based design. --- ## Table of Contents - [What is Parametric Design](#what-is-parametric-design) - [Key Software Platforms](#key-software-platforms) - [Visual Programming](#visual-programming) - [Generative Design and Optimisation](#generative-design-and-optimisation) - [Performance-Driven Design](#performance-driven-design) - [Digital Fabrication](#digital-fabrication) - [Interoperability and Data Exchange](#interoperability-and-data-exchange) - [Practical Applications](#practical-applications) - [Limitations and Considerations](#limitations-and-considerations) - [See Also](#see-also) --- ## What is Parametric Design In conventional design, the architect draws a wall at a specific location, length, and height — fixed values. In parametric design, the wall's position, length, and height are defined as parameters linked to other design variables (room area, structural grid, daylight target, budget). Changing one parameter propagates through the model, updating all dependent geometry and data. **Core concepts**: | Concept | Definition | Example | |---------|-----------|---------| | **Parameter** | A variable that controls geometry or behaviour | Floor-to-floor height; column spacing; facade module width | | **Constraint** | A rule that limits parameters | Maximum span; minimum daylight factor; maximum cost | | **Relationship** | A mathematical or logical connection between parameters | Facade panel width = bay width / number of divisions | | **Algorithm** | A sequence of operations that generates form or data | Voronoi tessellation; shortest path; structural optimisation | | **Design space** | The range of all possible solutions defined by parameters and constraints | All possible floor plate shapes that fit the site and achieve the brief | **Parametric ≠ curvy**: Parametric design is not inherently about complex geometry. A rigorously parametric model of a rectilinear office building — where structural grid, core position, facade module, and floor plate depth are all linked and adaptable — is as valid as a parametric doubly-curved shell. --- ## Key Software Platforms | Platform | Environment | Primary Use | Scripting/Visual Programming | |----------|------------|------------|-------------------------------| | **Rhino + Grasshopper** | Rhinoceros 3D | Geometry generation; analysis; fabrication | Grasshopper (visual); Python; C# | | **Revit + Dynamo** | Autodesk Revit | BIM automation; data management; geometry | Dynamo (visual); Python | | **ArchiCAD + Param-O** | Graphisoft ArchiCAD | BIM parametric objects | Param-O (visual); GDL scripting | | **Houdini** | SideFX | Complex procedural modelling; VFX-quality | VEX; Python; node-based | | **Processing / p5.js** | Open source | Generative art; data visualisation | Java / JavaScript | | **Blender + Geometry Nodes** | Open source | Free-form modelling; visualisation | Geometry Nodes (visual); Python | | **CATIA** | Dassault Systèmes | Complex surface modelling; aerospace-grade | CATIA scripting | ### Grasshopper Plugin Ecosystem | Plugin | Function | |--------|---------| | **Ladybug / Honeybee** | Environmental analysis (solar, daylight, energy, comfort) | | **Kangaroo** | Physics simulation (form-finding, structural behaviour) | | **Karamba** | Structural analysis and optimisation | | **Galapagos / Octopus** | Evolutionary optimisation (single/multi-objective) | | **Lunchbox** | Panelling; tessellation; data management | | **Elefront** | Rhino attribute management; BIM data handling | | **Weaverbird** | Mesh manipulation; subdivision | | **Heteroptera** | Logic; data flow; user interface components | --- ## Visual Programming Visual programming (node-based) environments allow architects to build algorithms by connecting graphical components rather than writing text-based code: **How it works**: 1. **Input nodes** define parameters (sliders, points, curves, data) 2. **Processing nodes** perform operations (move, scale, divide, evaluate) 3. **Output nodes** generate geometry, data, or actions (bake to model, export to file) 4. Connections between nodes represent data flow **Advantages**: - Accessible to designers without programming background - Visual feedback — see the effect of parameter changes in real time - Modular — components can be reused and recombined - Transparent logic — the algorithm is visible as a graph **Limitations**: - Complex algorithms become visually cluttered ("spaghetti" graphs) - Slower than text-based code for large datasets - Debugging can be difficult without structured programming practices - Temptation to over-parameterise (making everything variable when design decisions should be fixed) --- ## Generative Design and Optimisation ### Optimisation Methods | Method | Description | Application | |--------|------------|------------| | **Evolutionary algorithm (GA)** | Population-based search inspired by natural selection | Multi-variable; multi-objective; form-finding | | **Simulated annealing** | Probabilistic search accepting worse solutions early, converging later | Single-objective; large design spaces | | **Gradient descent** | Iterative movement toward local optimum | Smooth objective functions; structural sizing | | **Topology optimisation** | Material distribution for minimum weight/maximum stiffness | Structural form-finding; node design | | **Machine learning** | Neural networks predict performance from parameters | Rapid surrogate evaluation; pattern recognition | ### Multi-Objective Optimisation Real architectural problems involve competing objectives (minimise cost AND maximise daylight AND minimise energy). Pareto-optimal solutions represent the best trade-offs — no objective can be improved without worsening another. **Tools**: Octopus (Grasshopper); Galapagos (single-objective); Wallacei (multi-objective with visualisation); Opossum (surrogate-based). --- ## Performance-Driven Design Computational tools enable the integration of performance simulation directly into the design loop: | Performance Domain | Simulation Tool | Integration | |-------------------|----------------|-------------| | **Solar radiation / daylight** | Ladybug / Honeybee; Radiance; DIVA | Grasshopper; real-time feedback | | **Energy** | Honeybee (EnergyPlus); DesignBuilder | Grasshopper; parametric energy modelling | | **Thermal comfort** | Honeybee; Ladybug (UTCI, PMV) | Outdoor comfort mapping; indoor analysis | | **Wind / CFD** | Butterfly (OpenFOAM via Grasshopper); Autodesk CFD | Wind comfort around buildings | | **Structural** | Karamba; Grasshopper-SAP2000 link | Structural sizing; form-finding | | **Acoustic** | Pachyderm (Grasshopper); CATT-Acoustic | Room acoustic simulation | | **Pedestrian flow** | SpaceSyntax; PedSim (research) | Circulation optimisation | | **View analysis** | Ladybug View Factor; isovist analysis | View quality; privacy | **Workflow**: Define design parameters → generate geometry → simulate performance → evaluate against targets → optimise parameters → iterate. --- ## Digital Fabrication Parametric design generates geometry that can drive fabrication machinery directly: | Fabrication Method | Input | Application | |-------------------|-------|------------| | **CNC milling** | 3-5 axis toolpaths from 3D model | Formwork; moulds; timber panels; stone carving | | **Laser cutting** | 2D DXF/DWG profiles | Sheet materials; steel; acrylic; card models | | **3D printing (additive)** | STL/3MF mesh files | Models; prototypes; concrete printing (emerging) | | **Robotic fabrication** | Custom toolpaths; KUKA/ABB programs | Bricklaying; welding; timber assembly; winding | | **Wire cutting (hot wire)** | 2D profiles swept through EPS | Formwork for complex concrete geometry | | **Water jet cutting** | 2D profiles | Stone; metal; glass; precision cutting | | **Folding / bending** | Fold patterns from unrolled surfaces | Sheet metal; perforated panels | **File-to-factory**: The parametric model generates fabrication data (nesting, toolpaths, labelling) directly — reducing manual translation errors and enabling mass customisation (every panel different, efficiently produced). --- ## Interoperability and Data Exchange | Format | Use | Standard | |--------|-----|---------| | **IFC** | BIM model exchange (open standard) | ISO 16739 | | **gbXML** | Energy model exchange | Green Building XML | | **STEP / IGES** | CAD geometry exchange | ISO 10303 | | **DXF / DWG** | 2D/3D CAD exchange | Autodesk (de facto standard) | | **OBJ / FBX / glTF** | Mesh geometry; visualisation | Various | | **CSV / JSON** | Data exchange (schedules, parameters, results) | Open | | **Speckle** | Open-source BIM data exchange and collaboration | Speckle | | **Rhino.Inside** | Embed Rhino/Grasshopper within Revit | McNeel | **Rhino.Inside.Revit** is particularly significant: it allows Grasshopper parametric logic to drive native Revit BIM elements — combining the geometric freedom of Rhino/Grasshopper with the documentation and data capabilities of Revit. See [[BIM Fundamentals and LOD]] and [[IFC and Open BIM]] for BIM data management. --- ## Practical Applications | Application | Parametric Approach | Example | |------------|-------------------|---------| | **Facade panelling** | Parametric subdivision; panel sizing; pattern generation | Variable-density perforations; solar-responsive louvres | | **Structural form-finding** | Physics-based optimisation of shell/gridshell geometry | Funicular shells; tension membranes; minimal surfaces | | **Floor plate optimisation** | Parametric testing of core position, depth, and shape | Maximise NIA; optimise daylight; minimise structure | | **Site massing** | Parametric massing studies varying height, setbacks, orientation | Maximise development yield within planning constraints | | **Daylighting** | Parametric window sizing and shading geometry | Achieve target daylight factor while controlling glare | | **Timber fabrication** | Unique joint angles; CNC-cut connections | Mass timber structures with non-orthogonal geometry | | **Acoustic optimisation** | Parametric ceiling/wall geometry for diffusion/absorption | Concert hall shaping; auditorium reflector design | --- ## Limitations and Considerations | Issue | Response | |-------|---------| | **Tool obsession** | Parametric tools are means, not ends — the quality of the question matters more than the sophistication of the tool | | **Over-parameterisation** | Not everything should be variable — good design involves making decisions and fixing parameters | | **Black box risk** | Understanding the algorithm is essential — optimisation without understanding produces unintelligent results | | **Skill dependency** | Computational design skills are scarce; practice investment in training is essential | | **Fabrication reality** | Computationally elegant geometry may be unbuildable or unaffordable — test against construction constraints early | | **Interoperability friction** | Moving between parametric environments and BIM/documentation tools remains imperfect | | **Maintenance** | Parametric definitions require documentation and management — an undocumented Grasshopper file is useless to anyone except its author | --- ## See Also - [[BIM Fundamentals and LOD]] - [[IFC and Open BIM]] - [[Modular and Prefabricated Construction]] - [[Structural Analysis Fundamentals]] - [[Daylighting Design Principles]] - [[Biophilic Design]] --- #digital #parametric #computational #grasshopper #optimisation #fabrication