ArticlePDF Available

Topo-facade: envelope design and fabrication planning using Topological mesh representations


Abstract and Figures

Computational design tools based on Autodesk’s DesignScript language have been used with geometry and topology modelling techniques in the design of a climatised free-form building envelope. This project involves structural and performance analysis tools applied to structural engineering, façade engineering and fabrication planning. The project has progressed from concept through tender phases. The particular geometry presented unique conditions that required non-standard solutions to be used; to this end DesignScript was introduced to allow the design and engineering team to build a number of scripted topological façade models that explored alternative façade configurations. This paper combines a discussion about the specific fabrication project with a more generalised discussion of the role of computational tools in design and fabrication. The main interest is to explore the two-way relationship between practice and tool building by considering how computation can contribute to a practical fabrication project and equally important, how computational tools can be tested and refined by being used in practice on demanding projects.
Content may be subject to copyright.
Fi. 1: Carrier frame fabrication concept.
 
  
Gustav FaGerström, Buro H appold, erik verBoon, Buro Happold, roBert aisH
uni ver sity oF BatH
Computational desin tools based on Autodesk’s DesinScript lanuae have been used with eometry and topoloy modellin
techniques in the desin of a climatised free-form buildin envelope. This project involves structural and performance analysis tools
applied to structural enineerin, façade enineerin and fabrication plannin. The project has proressed from concept throuh
tender phases. The particular eometry presented unique conditions that required non-standard solutions to be used; to this end
DesinScript was introduced to allow the desin and enineerin team to build a number of scripted topoloical façade models that
explored alternative façade confiurations. This paper combines a discussion about the specific fabrication project with a more
eneralised discussion of the role of computational tools in desin and fabrication. The main interest is to explore the two-way
relationship between practice and tool buildin by considerin how computation can contribute to a practical fabrication project and
equally important, how computational tools can be tested and refined by bein used in practice on demandin projects.
 
The architectural concept used in this paper is based on a
sculptural approach in which lass joints alternate between
uniquely anled concave and convex relationships between
adjacent panels (fi. 2).
The self-weiht of the lare insulated lass units (IGU) de-
mands a support stratey where the ede of each panel should
be continuously supported. This requires that a strict eomet-
ric relationship be maintained between the lass and the sup-
port structure. Furthermore, the eometric conditions around
each node are unique, bein the simultaneous meetin point
for both concave and convex lass panels. Consequently, each
node, while based on a common topoloical principle, has a
unique eometric confiuration, and therefore requires the
development of a unique fabrication eometry.
As a base constraint, the architect had instructed that a
point-supported approach was undesirable and expressed a
preference for the use of rectanular or plate primary struc-
tural elements as opposed to the more traditional round hol-
low section and spherical node approach often found in struc-
tures of this type. The subsequent studies looked at both
structural approaches as a system that was offset from the Fi.2: Site Context. (Imae courtesy of Robert A. M. Stern Architects)
lazin line. The method for supportin the lass to the pri-
mary structure utilised continuous anles, or ‘carrier frames’
that followed and were structurally attached to the lass ed-
es via structural silicone sealant. Periodic steel plates struc-
turally linked the lass to the primary structure, while also ad-
dressin the chanin distance and anle between the two
systems. The multiple anled relationships between lass pan-
els required that the IGU’s have steppin or cantilevered inner
or outer lihts in order to maintain a consistent external joint
DesinScript (as the interation of lanuae, eometry, topol-
oy and plu-ins) allowed the enineerin team to assess the
eometric feasibility of the architectural concept by buildin a
number of alteratively scripted topoloical façade models. This
approach enables the team to model the correspondence be-
tween the façade topoloy and the physical components of the
facade: lass panels as topoloical faces, structural members
as topoloical edes and node connectors as topoloical verti-
ces. DesinScript topoloy classes reveal the underlyin func-
tionality of the Autodesk Shape Manaers (ASM) via its API ¹
(Aish, 2012).
The sinle topoloical mesh model allows each of the con-
stituent components (face, ede or vertex) to make topoloi-
cal and eometric queries to the adjacent components, for ex-
ample, the computation of averae vertex normals, and ede
bisectors. Additionally, DesinScript is interated with Robot
Structural analysis and Performative Desin and reveals us-
er-oriented API’s directly to the enineers usin DesinScript.
This allows the sinle topoloical mesh model to be direct-
ly analysed both structurally and environmentally, while the
mesh also forms the basis for related fabrication models.
application: ca se study
Typically, facades are modelled as meshes usin the architect-
established desin surfaces (here represented by the front of
lass (FOG). The structural support system is typically defined
as an offset mesh from this definin mesh. The resultin struc-
ture is more easily realised if the definin mesh has ‘torsion-
free’ nodes. This means that the vertex normals at the end of
each ede are coplanar and the ede members are planar.
In some cases, a mesh with non-torsion-free nodes can be
optimised by movin the vertex positions ² (Wallner and Pott-
mann, 2011). This approach is more appropriate where the
mesh represents a smooth surface and the chanes in vertex
position (and hence the shape of the façade panels) is not visu-
ally apparent. However, the desin intention for this façade is
to create a very specific faceted confiuration, which could not
be optimised in this way.
In a non-torsion-free facade, the ede normal (as the bi-
sector of the ede’s adjacent faces) and the vertex normals at
either end of the ede are not coplanar. If the ede members
are planar and based on their respective ede normal, then
the ede members meetin at a common vertex will not in-
tersect alon a common vector (fi. 3). Alternatively, if it is re-
quired that all ede members intersect at a common vector at
each vertex, then the structural system has to resolve the twist
alon the ede members.
 
A carrier frame and offset structure was considered. If the off-
set structure is based on a uniform offset from the face of the
definin façade, then the edes of the offset may not lay on the
face bisectors, and the relationship between the carrier frame
and the offset structure may have to be desined to accommo-
date such deviations.
  
While these 2D studies were conceptually useful, a 3D ap-
proach was necessary to address the multiple unique condi-
tions imposed by the eometry. Buildin on this exploratory
work, a scripted approach was developed, harnessin mesh to-
poloy and allowin for the automated creation of panels from
mesh faces, structural members from mesh edes and connec-
tor nodes from mesh vertices (fi. 4).
N0 - Average normal
Separate intersection
point, >2’
B - Bisectors
N - Normals
Fi. 3: Characteristics of non-torsion-free eometry,
averae node vertex normal principle and ede face bisectors.
Fi. 4: Desin eometry expressed as topoloy mesh.
Fi. 5: The mesh topoloy used to define the façade.
Orane lines: the edes of the primary façade mesh
(front of lass)
Red lines: the averae vertex normals
Cyan lines: the weihted averae vertex normals
(weihed by the face areas)
Maenta lines: the offset mesh (used to define
the offset structure) (based on offsettin the vertices
of the primary mesh alon the weihted averae
vertex normals)
 
The ede-based structural system has to support one or two
planar sheets of lass and the twist between the vertices at
its ends. The question remains: Should there be a sinle ede
member that combines all these roles or a carrier frame to sup-
port the planar lass linked to a separate structural member to
accommodate the twist?
Four structural hypotheses were considered:
Carrier Frame and Plate oriented alon the ede normal (fi. 6)
Carrier Frame and Plate twistin to accommodate both end
points’ vertex normals (fi. 7)
Chamfered tube with chamfer axis usin the averae vertex
normal (fi. 8)
– Offset structure and carrier frame, with offset node con-
nectors based on the averae vertex normal (fis. 9– 10)
Fi. 6: Structural Hypothesis 1 – Carrier frame and plate, with the plate
oriented alon ede normal.
Fi. 7: Structural Hypothesis 2 – Carrier frame and Plate, with the plate
twisted between end points’ non co-planar vertex normals.
Fi. 8: Structural Hypothesis 3 – Chamfered tube with chamfer axis usin
the averae vertex normal
Fi. 9: Structural Hypothesis 4 – Offset structure and carrier frame,
with the offset node connectors based on the averae vertex normal.
    
The different test models were built on a simple hand-coded
test mesh (fis. 5–8). DesinScript allowed the test mesh to be
swapped out for the full mesh in order to build the complete
facade (f is. 9–10).
Fis. 10–12: Desin development model with offset structure
and carrier frame fabrication concept.
   
DesinScript provides a familiar ‘data flow’ approach to desin
computation and makes the creation and execution of desin
loic accessible to desiners with little or no prorammin ex-
perience. In this project, data from the input nodes in the top
left part of the raph ‘flows’ via intermediate mesh modelin
nodes to create the facade in the bottom riht part of the raph
(f i. 1 4).
The ‘data flow’ approach works well with simple models.
However, usability issues bein to emere when the problem
bein addressed ets more complex and there are many more
nodes to consider. This issue is addressed throuh the ‘node to
code’ functionality in DesinScript which automatically trans-
lates the user’s data flow diaram into an associative script
(fi. 12). DesinScript also includes support for reular impera-
tive scriptin usin conventional ‘for’ loops (for iteration) and
if’ statements (for conditionals) ³ (Aish 2013).
Bracketlength Carrierangle Position
7.204577 112 A111
9.277795 114 A112
8.996179 140 A113
8.973133 52 B71
6.642246 83 B72
8.632262 22 B73
6.056679 85 B21
6.814823 154 B22
7.22186 40 B23
8.648544 131 A91
8.918658 71 A92
6.064782 47 A93
6.664838 152 A81
7.675841 37 A82
6.426951 102 A83
9.645642 48 T21
6.832239 26 T22
7.152264 133 T23
7.25188 126 T91
7.055309 43 T92
6.64823 27 T93
7.916496 46 T21
8.305637 40 T22
9.694806 18 T23
8.837067 52 C51
6.535089 78 C52
6.896258 133 C53
8.967901 42 D51
7.42581 114 D52
8.955504 141 D53
Fi. 14: Diaram outlinin DesinScript data flow raph used as a visual prorammin interface (top)
as well as its ‘node to code’ functionality (bottom) allowin the desiner to selectively replace all
or part of a raph node diaram with the correspondin code, thus makin it possible to reduce visual
clutter and proress to a more succinct form of desin computation.
Fi. 13: umerical output complementin or replacin
traditional shop drawins.
 
Movin forward into fabrication plannin with hypothesis4
(above), the process was reversed with respect to that out-
lined in fi. 4. The topoloically represented structure is now
the source of an additional level of information describin the
carrier brackets’ lenth, shape, anle and position within the
overall assembly. The resultin information packae (fi. 13)
can be used in conjunction with – or entirely in lieu of – tradi-
tional shop drawins.
This project demonstrates that script-driven topoloy can be
used as the central representation for eometric assessment,
structural analysis, performative analysis, fabrication plan-
nin and component enineerin, ultimately providin an ef-
fective way to realise a challenin façade.
More enerally, it is reconised that computation is drivin
more aspects of contemporary architectural and enineerin
practice. The contribution of DesinScript is to unify compu-
tation, eometry and topoloy with alternative prorammin
interfaces (both visual and textual) and thereby support differ-
ent levels of computational skill.
It is important to reflect on the results of this work. At one
level, it is the physical buildin. At another level, it is the op-
portunity this project provided to test and refine a new en-
eration of computational desin tools. But maybe the most im-
portant result is the acquisition of knowlede and skills made
by the practitioners. All three results have the potential to con-
tribute to even more challenin projects.
DesinScript Development Team, Autodesk Sinapore Research
Development Centre
Andrew Marsh for the Performative Desin plu-in for DesinScript
Al Fisher and Buro Happold Bath for the Autodesk Robot Structural
Analysis plu-in for DesinScript
Buro Happold ew York Structures and Facades roups
Robert A.M. Stern Architects
DesinScript is available at Autodesk Labs,
Additional information and plu-ins for DesinScript are available atnscript.or/
1 Robert Aish and Aparajit Pratap, ‘Spatial Information Modelin of
Buildins usin on-Manifold Topoloy with ASM and DesinScript’,
in Proceedings of Advances in Architectural Geometry (Paris:
Spriner 2012), pp. 25–36.
2 Johannes Wallner and Helmut Pottman, ‘Geometric Computin
for Freeform Architecture’, Journal of Mathematics in Industry 1,
no. 4 (2011).
3 Robert Aish, ‘DesinScript: A Learnin Enviroment for Desin
Computation’, in Proceedings of the Design Modelling Symposium
(Berlin: Spriner, 2013).
... This concept was previously explored by the authors in the context of energy analysis, façade design, and additive manufacturing of conformal cellular structures (Jabi 2016;Fagerström, Verboon, and Aish 2014;Jabi et al. 2017). ...
Conference Paper
Full-text available
Buildings enclose and partition space and are built from assemblies of connected components. The many different forms of spatial and material partitioning and connectedness found within buildings can be represented by topology. This paper introduces the ‘Topologic’ software library which integrates a number of architecturally relevant topological concepts into a unified application toolkit. The goal of the Topologic toolkit is to support the creation of the lightest, most understandable conceptual models of architectural topology. The formal language of topology is well-matched to the data input requirements for applications such as energy simulation and structural analysis. In addition, the ease with which these lightweight topological models can be modified encourages design exploration and performance simulation at the conceptual design phase. A challenging and equally interesting question is how can the formal language of topology be used to represent architectural concepts of space which have previously been described in rather speculative and subjective terms?
... In this project, the use of parametric tools allowed a continuous upstream/downstream flow of information during the design sequence as fabrication constraints were discovered [91]. We found Parametric Co-rationalization to be the most commonly used of all the rationalization strategies for architectural projects appearing in the review, especially in the industry [32,48,54,[92][93][94][95][96][97][98][99][100][101][102][103][104][105][106][107][108][109][110]. ...
Rationalization is widely recognized as an important design strategy in contemporary architectural projects, especially in projects with complex geometries, built using digital fabrication processes. However, an up to date review of the rationalization strategies used in these projects, their place in the design sequence and their relation to digital fabrication processes has not been conducted. The purpose of this review is to identify the rationalization strategies used in architectural projects in the practice and the academia. This paper presents the results of a systematic review of over 500 papers describing rationalization and digital fabrication in contemporary architecture. Using the data gathered in the review, we show that the capabilities of the fabrication machinery used are the most frequently encountered rationalization constraint in realized architectural projects. Additionally, we describe a new taxonomy for rationalization strategies, which incorporates functional information with the temporal information described by traditional classifications. Using this taxonomy, we identify trends within the industry and the academia and point to the growing popularity of parametric co-rationalization approaches. We conclude by discussing promising rationalization approaches for future research.
... Formal complexity often has implications for structural typology and conversely structural typology impacts formal complexity (Tomlow 1989). In many cases, formal complexity results in custom components that exhibit a high level of functional complexity (Fagerstrom et al 2014). Grid-based structure typologies are rapidly developed from simplicity and regularity to complexity and irregularity (Fig.01). ...
Full-text available
This paper describes the abilities of parametric iterative design with collaboration of robotic fabrication workflow in structural optimization of the nodes (joints) of special grid-based structure. Experimental structure built in robotic fabrication workshop – (the Dynamo-BUILD workshop at the 2016 International Conference on Robotic Fabrication in Architecture, Art, and Design Conference in Sydney, Australia)-is taken as a case study. In this study, the complexity of structure form combined of joints and members is resolved and developed through parametric design algorithms. Focusing on joints, the case gives workflow structure methods of design and fabrication that transfer the level of mass simplicity production to iterative complexity production .Furthermore, these methods also respect the manufacturing processes and material properties of nodes. The structure was fabricated using robotic fabrication techniques after design optimization using parametric computationally driven manufacturing processes. In order to move from the computational design environment to joint fabrication, custom robotically process was developed to assemble making full structure series of nodes which saved time; cost; and exerted effort if compared to the traditional mass production processes.
... Of particular interest are face and edge offsetting techniques, which give layers or thickness to facial or edge elements respectively. Offsetting the vertices of a mesh is also possible, although it typically introduces a 'twist' in the structural support members connecting the original mesh with the offset mesh, which may present additional fabrication challenges (Aish, Verboon, and Fagerström 2014). Some significant approaches to the face-and edge-offsetting problem have focused on characterizing the mesh geometry that produces 'well-behaved' offset outcomes (Pottmann et al. 2007;Pottmann and Wallner 2008;Wang and Liu 2010). ...
Conference Paper
Full-text available
Architectural designs are frequently represented digitally as plane-faced meshes, yet these can be challenging to translate into built structures. Offsetting operations may be used to give thickness to meshes, and are produced by offsetting the faces, edges or vertices of the mesh in an appropriately defined normal direction. In a previous paper, we described a face-offsetting algorithm for resolving the revised combinatorics of the offset mesh produced by face-offsetting (Ross & Hambleton, 2015). That is, given an input mesh with no design constraints, the algorithm computes the exact offset by determining the new geometric and combinatorial structure of the offset mesh. One of the design freedoms available in that method is the opportunity to specify different offset distances on a per-face basis. In the present paper we consider the implications of this freedom. One question of particular interest is: under what conditions does an offset mesh produced by variable rate face-offsetting also have a uniform distance edge-offset? To physically realise a mesh as a built structure usually requires that the mesh edges are used as the basis for structural members, with some structural depth. Therefore, given a mesh M it is particularly desirable to find an offset mesh M' in which the edges of M' are at a uniform perpendicular distance d from their corresponding edge in M. We present a description of meshes that admit uniform distance edge offsets as a consequence of a variable rate exact face offset, based on a graph-theoretic analysis of the underlying dual mesh. The potential advantage of this approach is that it can provide an opportunity to rationalise the physical realisation of the mesh as a constructible structure where all edge based members have the same depth.
This paper aims to build a theoretical foundation for parametric design thinking by exploring its cognitive roots, unfolding its basic tenets, expanding its definition through new concepts, and exemplifying its potential through a use-case scenario. The paper focuses on a specific type of topological parameter, called non-manifold topology as a novel approach to thinking about designing cellular spaces and voids. The approach is illustrated within the context of additive manufacturing of non-conformal cellular structures. The paper concludes that parametric design thinking that omits a definition of topological relationships risks brittleness and failure in later design stages while a consideration of topology can create enhanced and smarter solutions as it can modify parameters based on an accommodation of the design context.
This project-based paper describes the iterative design, structural optimization, and fabrication of the experimental grid shell structure developed for the MASS Lo-Fab pavilion. In this case, formal complexity is resolved through functional complexity that emerges in both elements of the structural system—the node and the strut—that each maintain a level of simplicity appropriate to respective manufacturing processes and material properties. The structure was fabricated using state-of-the art collaborative robotic fabrication techniques and a combination of traditional craftsmanship and computationally driven manufacturing processes. In order to move from the computational design environment to one of material, the team worked in collaboration with AutodeskTM to develop a novel design-to-robotic fabrication workflow using the emerging visual scripting interface Dynamo. A custom robotically assisted welding process was developed to assemble 1880 steel parts making up 376 nodes that saved over 3 weeks of labor when compared to traditional processes.
ResearchGate has not been able to resolve any references for this publication.