Project

Back to overview

Constructive Variational Freeform Surfaces for Architectural Design

English title Constructive Variational Freeform Surfaces for Architectural Design
Applicant Pauly Mark
Number 137626
Funding scheme Project funding (Div. I-III)
Research institution Laboratoire d'informatique graphique et géométrique EPFL - IC - ISIM - LGG
Institution of higher education EPF Lausanne - EPFL
Main discipline Information Technology
Start/End 01.08.2012 - 31.07.2015
Approved amount 461'255.00
Show all

Keywords (3)

Architectural Geometry; Geometric Modeling; Design and Optimization

Lay Summary (English)

Lead
Lay summary

Recent advances in construction and material technology have enabled the use of freeform surfaces as striking elements in contemporary architecture. Today's Computer Aided Design software provides intuitive interfaces for the design of arbitrarily complex freeform surfaces. The construction of freeform surfaces, however, requires a segmentation of the surface into many small pieces, called panels, which are then separately manufactured and mounted onto a support structure. Despite its high practical relevance for freeform architecture, this rationalization process remains a challenging problem with many fundamental questions still unsolved. 


The main goal of the paneling process is to maximize rationalization quality while minimizing total cost. The rationalization quality depends on various measures, including how well the panels approximate the surface, how well the neighboring panels fit together, and how smooth the inter-panel connection lines are. Production and assembly cost is directly related to the complexity of the panels (simpler panel shapes should be preferred over more sophisticated ones) and mold reuse (the same mold should be used to produce multiple panels). This diverse set of opposing objectives leads to a highly complex optimization problem. Except for very small or very simple surfaces, the manual layout of panels is infeasible. Computational tools are required.


The objective of this project is to investigate mathematical concepts, robust algorithms, scalable geometric optimization techniques, and flexible data structures to form a comprehensive toolset for constructive variational freeform surface design. Our results will allow architects to design a freeform surface and obtain immediate feedback on its rationalization. We will formulate a novel optimization approach for paneling with single and doubly curved panels, including mold reuse and the automatic selection of the simplest possible panel shapes. In parallel, we will investigate new complexity reduction techniques to enable interactive, construction-aware design that bridges the gap between design and production. 
The computational tools that we plan to develop will enable the construction of freeform surfaces at a whole new level of geometric complexity, thus pushing the limits of what is currently possible in architecture and design.


Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants

Employees

Publications

Publication
Interactive Design Exploration for Constrained Meshes
Deng Bailin, Bouaziz Sofien, Deuss Mario, Kaspar Alexandre, Schwartzburg Yuliy, Pauly Mark (2015), Interactive Design Exploration for Constrained Meshes, in Computer-Aided Design, 61, 13-23.
Interactive Design of Probability Density Functions for Shape Grammars
Dang Minh, Lienhard Stefan, Ceylan Duygu, Neubert Boris, Wonka Peter, Pauly Mark (2015), Interactive Design of Probability Density Functions for Shape Grammars, in ACM Trans. Graph., 34(6), -.
Real-Time Deformation of Constrained Meshes Using GPU
Kaspar Alexandre, Deng Bailin (2015), Real-Time Deformation of Constrained Meshes Using GPU, in Cai Yiyu (ed.), Springer, Singapore, 15-34.
Iso-level tool path planning for free-form surfaces
Zou Q., Zhang J., Deng B., Zhao J. (2014), Iso-level tool path planning for free-form surfaces, in Computer-Aided Design , 53, 117-125.
Local barycentric coordinates
Juyong Zhang, Bailin Deng, Zishun Liu, Giuseppe Patane, Sofien Bouaziz, Kai Hormann, Ligang Liu (2014), Local barycentric coordinates, in ACM Transactions on Graphics, 33(6), 1.
Wire Mesh Design
Garg A., Sageman-Furnas A. O., Deng B., Yue Y., Grinspun E., Pauly M., Wardetzky M. (2014), Wire Mesh Design, in ACM Trans. Graph., 33(4), 66.
Exploring Local Modifications of Constrained Meshes
Deng Bailin Bouaziz Sofien Deuss Mario Zhang Juyong Schwartzburg Yuliy Pauly Mark (2013), Exploring Local Modifications of Constrained Meshes, in Computer Graphics Forum, 32(2pt1), 11-20.
Assembling Self-Supporting Structures
Deuss M., Panozzo D., Whiting E., Liu Y., Block P., Sorkine-Hornung O., Pauly M., Assembling Self-Supporting Structures, in ACM Trans. Graph., 33(5).
Shapeop - a robust and extensible geometric modelling paradigm
Deuss Mario, Deleuran Anders, Bouaziz Sofien, Piker Daniel, Pauly Mark, Shapeop - a robust and extensible geometric modelling paradigm, in Faircloth Billie, Scheurer Fabian, Tamke Martin, Gengnagel Christoph, Ramsgaard Thomsen Mette (ed.), Springer, Switzerland, 505.

Collaboration

Group / person Country
Types of collaboration
CITA, Copenhagen Denmark (Europe)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
- Exchange of personnel
Interactive Geometry Lab, ETH Zurich Switzerland (Europe)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
Discrete Differential Geometry Lab / University of Goettingen Germany (Europe)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
Computer Graphics Group, Columbia University United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results
- Publication

Scientific events

Active participation

Title Type of contribution Title of article or contribution Date Place Persons involved
Design Modelling Symposium Talk given at a conference ShapeOp - A Robust and Extensible Geometric Modelling Paradigm 30.09.2015 Copenhagen, Denmark Pauly Mark;
Design Modelling Symposium Talk given at a conference Keynote Address: Think local, Act Global 29.09.2015 Copenhagen, Denmark Pauly Mark;
Advances in Architectural Geometry Talk given at a conference Keynote: Computer Science & Architecture 18.09.2014 London, Great Britain and Northern Ireland Pauly Mark;
ACM SIGGRAPH Conference Talk given at a conference Wire Mesh Design 10.08.2014 Vancouver, Canada Deng Bailin;
International Workshop on Geometry Talk given at a conference Interactive Design Exploration for Constrained Meshes 28.08.2013 Strobl, Austria Deng Bailin;
Eurographics Conference Talk given at a conference Exploring Local Modifications for Constrained Meshes 06.05.2013 Girona, Spain, Spain Deng Bailin;


Abstract

Recent advances in construction and material technology have enabled the use of freeform surfaces as striking elements in contemporary architecture. Today's computer aided design software provides intuitive interfaces for the design of arbitrarily complex freeform surfaces. The construction of freeform surfaces, however, requires a segmentation of the surface into many small pieces, called panels, which are then separately manufactured and mounted onto a support structure. Despite its high practical relevance for freeform architecture, this rationalization process remains a challenging problem with many fundamental questions still unsolved. The main goal of the paneling process is to maximize rationalization quality while minimizing total cost. The rationalization quality depends on various measures, including how well the panels approximate the surface, how well the neighboring panels fit together, and how smooth the inter-panel connection lines are. Production and assembly cost is directly related to the complexity of the panels (simpler panel shapes should be preferred over more sophisticated ones) and mold reuse (the same mold should be used to produce multiple panels). This diverse set of opposing objectives leads to a highly complex optimization problem. Except for very small or very simple surfaces, the manual layout of panels is infeasible. Computational tools are required.During the last few years, researchers in architectural geometry, a new emerging field at the interface of architecture, design, mathematics, and computer science, have recognized this challenge and introduced several important computational techniques. Current algorithms, however, only address certain special cases, e.g. paneling with planar or single curved panels. No solution exists today for the general problem of paneling with doubly curved panels or with mold reuse. Moreover, even for simple surfaces the existing paneling process is an involved optimization procedure usually performed after the surface has been designed. Thus, design information is not construction information and the realization of architectural freeform surfaces involves costly and time-consuming iterative cycles from design to production.The objective of this project is to investigate mathematical concepts, robust algorithms, scalable geometric optimization techniques, and flexible data structures to form a comprehensive toolset for constructive variational freeform surface design. Our results will allow the architect to design a freeform surface and obtain immediate feedback on its rationalization. We will formulate a novel optimization approach for paneling with single and doubly curved panels, including mold reuse and the automatic selection of the simplest possible panel shapes. In parallel, we will investigate new complexity reduction techniques to enable interactive, construction-aware design that bridges the gap between design and production. Agreements are in place to obtain architectural design data from our close industry contact Evolute. This gives us the invaluable opportunity to evaluate our findings on real-world projects.This proposed research has far-reaching practical, theoretical, and educational value. The computational tools that we plan to develop will enable the construction of freeform surfaces at a whole new level of geometric complexity, thus pushing the limits of what is currently possible in architecture and design. We will significantly advance the young field of architectural geometry and inspire future work in this area. Several master and semester thesis will be offered to students in computer science and architecture in the scope of this project. Moreover, since our algorithms operate on a level of abstraction independent of the architectural context, other applications and research involving geometric optimization, such as car and airplane design, medical simulations, machine manufacture, computational biology, character modeling and animation, will directly benefit from our findings.
-