Hybrid Tower 2 by CITA

CITA: Centre for I T and Architecture

15 November 2016

HYBRID TOWER 2

Hybrid Tower 2 is a continuation of an inderdisciplinary research collaboration between Centre for Information Technology and Architecture (CITA) at the Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation in Copenhagen (DK), the Department for Structural Design and Technology (KET), University of Arts Berlin (DE), Fibrenamics, Universidade de Minho, Guimaraes (PT), Essener Labor fuer Leichte Flaechentragwerke, University of Duisburg (DE) and the Portuguese textile company AFF a.ferreira & filhos, Caldas de Vizella (PT).

CONTENT

13 INTRODUCTION

TWO TOWERS

CHALLENGE 03

CHALLENGE 02

79 CHALLENGE 05

CHALLENGE 01

CHALLENGE 04

95 RESULTS

REFERENCES

INTRODUCTION RETHINKING TRADITIONAL APPROACHES IN ARCHITECTURE Traditional thinking in architecture and engineering alike is to understand the built environment as static, unaffected by changes in their environment. Buildings are designed for permanence and thought as stable and unchangeless. Tower explores the idea of a moving architecture, a resilient structure that adapts under environmental changes. Current architectural research practice is investigating the design and making of new material systems in which advanced CNC fabrication technologies allow for the precise control of material performance (Oxman 2007, Menges 2012, Palz 2009, Gramazio & Kohler 2008). This interest interfaces the design and fabrication of materials with that of buildings, allowing the conceptualisation of new structural systems that optimise material use and enable the realisation of lighter and smarter buildings (Nicholas 2013).

Left page: Digital excercise on Tower Design, 2014.

A RESILIENT TOWER TYPOLOGY Vertical structures have to cope with gravity and especially dynamic loads from wind. In order to answer this nature, mandkind have developed approaches, which focus on lightness and flexibility. Tower investigates strategies, which allow textile structures to deal with the challenges of the vertical. Hybrid Tower 2 is a continuation of an inderdisciplinary research collaboration between Centre for Information Technology and Architecture (CITA) at the Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation in Copenhagen (DK), the Department for Structural Design and Technology (KET), University of Arts Berlin (DE), Fibrenamics, Universidade de Minho, Guimaraes (PT), Essener Labor fuer Leichte Flaechentragwerke, University of Duisburg (DE) and the Portuguese textile company AFF a.ferreira & filhos, Caldas de Vizella (PT). Left page: Hybrid Tower 2 on the Placa De Toural in Guimaraes. Photo by A.Ingvartsen Top: Shuhov Tower, Moscow 1922 Bottom: Palm trees under heavy wind load

HYBRID SYSTEM: COMBINING TENSION AND COMPRESSION The structural system in Tower is a hybrid structureemploying bending active elements and tensile membranes with bespoke material propertiesand detailing. Hybrid structures are defined as a combination of two or more structural concepts and materials in order to create stronger whole. Hybrid structural systems promise to overcome the limitations of the prevalent form-active approaches towards structures, which are compression or tension only. The shape of these structures is optimised to result in only one type of internal forces. These are transported to the boundaries and are countered by equal reaction forces, often leading to massive support structures. Hybrid structures can per definition work withtension and compression forces. This allows them to cope with the vast amount of building typologies, which deal with tension and compression forces. Left page: Deep Surface Structure employing discontinuous elements restrained by the membrane Top: Philip Block, Thin-tile Vault, 2010 Bottom: Frei Otto, German Pavilion EXPO 1967

TWO

TOWERS

FIVE CHALLENGES FACING THE CHALLANGES TO IMPROVE THE OUTCOME The first demonstrator â&#x20AC;&#x153;Hybrid Towerâ&#x20AC;? was realised in April 2015. The project initialised research into the design and specification of hybrid structural systems that combine membrane structures with bending active members and bespoke knit. The main contribution of the project is the exploration of an active design environment that enables the dynamic incorporation of light weight simulation of hybrid structures in which members of tension and compression interact. The Hybrid Tower 1 demonstrates that knitted fabrics can be principally used as membrane materials. Motivated by the lack of precision in the first attempt, the design team decided to rebuild Tower for the CONTEXTILE conference, Guimaraes in 2016. The project focused on material level and aimed to enable convergence of a material behavior within development and the final built, as well as convergence of the simulated and real behaviour of the structure. Left page: Comparison of Hybrid Tower 1 (2015) and Hybrid Tower 2 (2016) Right page: Close up view on the both towers

DESIGN ITERATIONS EXPLORATION OF MESH DISCRETISATION Design development started with the global design explorations. In particular, the ways the mesh can be discretised into the patches that will be able to be produced on the knitting machine. The limitations of the knitting machine in width but not in lenght were taken into considiration while working on the design approach. Being imposed to produce designs with the linear patch elements, the first set of digital experiments were resulted in multiple supports tower with the developed strips deep surface skin. For those reasons the mesh walker method was used to tessellate the mesh, and the complexity of the patches was derived from the initial mesh geometry: more simple the geometry more straight is the patch. Even though it was a promising approach, but was abandoned on the early stage of development due to the material redudancy in the system withoug achieving extra stiffness. Left page: Three support tower proposal and mesh walker surface development Right page: Variations of Mesh Walker surface developing

tower design iteration 1 double layer solution

tower design iteration 2 stucking of arches solution

tower design iteration 3 discontinous twisted loops solution

tower design iteration 4 stucking of arches with tension cables towards the middle

tower design iteration 5 continuous vertical intercrossing arches

tower design iteration 6 stucking of arches through the lower level

DESIGN PROCESS TOWER

PHYSICAL ITERATING THE TOWER DESIGN In 2015 the project was focusing a lot on the design phase, where the team investigated the design strategies for a resillient tower. Numerous prototypes were produced in order to gain the understanding of the material performance.The idea to create a double layer surface of the tower was underlying principle in the early stages of desing. In all our iterations we tested out the opprotunities of the discontinuous integration of the active bending members into the textile.

Left page: Prototypes for the tower material and structure systems Right page: Internal tensioning proposal for tower stability

3 supports, 1 level input

3 supports, 1 level relaxation

3 supports, 1 level stabilisation

3 supports, 1 level settled

3 supports, 1 level membrane

3 supports, 1 level membrane relaxation

3 supports, 1 level membrane stabilisation

3 supports, 1 level membrane settled

3 supports,2 levels input

3 supports, 2 levels relaxation

5 supports, 2 levels input

5 supports, 3 levels input

5 supports, 3 levels relaxation

5 supports, 3 levels stabilisation

5 supports, 3 levels membrane input

5 supports, 3 levels membrane settled

DISCONTINUOUS ELEMENTS TO FORM TOWER LEVELS In order to create informed architectural scale design models that engage material performance, we need to develop a new class of simulation models. Where simulation in the garment industry is limited to small patches, the design and specification of textile structures in architecture has to tackle a wider geometrical range, larger scales and be able to consider the textile as part of a structural continuum. The ability to produce bespoke materials, shapes and details with CNC knit demands a tighter coupling of these currently discreet processes.

Left page: Digital pipeline opportunities Right page bottom: 3d Scan of the early physical models

Left page: Form-finding investigation in the diretion of knittable membranes Right page: Transformation from mesh to NURBS

Left and right page: First integration tests of the tower design pipeline

6 14-16

1 bending members input 2 points for the pulling springs 3 pulling springs 4 input geo with extra support leg

5 input geo with mesh 6 input geo with mesh vertices 7 relaxation iter 5 8 relaxation iter 20 9 relaxation 80

10 relaxation iter 150 14 developing of the strip, iter 1 11 relaxation iter 200 15 developing of the strip, iter 12 bending radii analysis 80 13 highlighted one strip of 16, developing of the strip, iter membrane 200

DESIGN ITERATIONS TOWER ITERATING THE TOWER DESIGN The integration of the formfinding process for the bending active structure of the tower into the design environment allowed for an extensive exploration of design options. Within the design phase of the first Tower a workshop with students helped to explore the full depth of design options with the developed parametric system. Already small changes in the length or point of junction between the different layers of the structure gives raise to dramatically different shapes and expressions.

Left page: Comparison of Hybrid Tower 1 (2015) and Hybrid Tower 2 (2016) Right page: Close up view on the towers

GFRP active bending members

Tension cables Bespoke knitted textile membranes

7.000 m

FINAL

DESIGN

DECISION TOWER

ITERATING THE TOWER DESIGN

2.980 m

While the principal design directions have been explored in the Tower 1 phases, the second iteration of the Tower concentrated on the improvement of overall structural performance, precision and assembly. For this the design of the Tower and the design iteration could be improved through a tighter collaboration with engineers and the further integration of structural feedback within the design environment. this included a visual feedback on the stresses in the active bend members and the stressed state of the textile membranes.

Left page: Tower 2 geometry perspective view Right page: Facade and the top view of the tower

CHALLENGE

sewing seam openings

piquet lacoste

tubular jersey interlock

tubular jersey

BESPOKE KNITTED STRUCTURE DEVELOPING KNITTED STRUCTURES WITH BESPOKE BEHAVIOUR piquette lacoste

In order to study the influence of the fabric structure, the patterns Pique Lacoste, Pique Lacoste AFF and Ponto Di Roma were selected as candidates for main area of the membrane. On the edge the more dense structure was used for reinforcing the fabric. The tubular jersey was the solution to accomodate the integration of the active bending members into the fabric. The selection criteria was based on existing knowledge at the fabric producer regarding the elasticity, visula impact - transparency, qualities of the fabric details and haptic properties. The initial sampling showed, that transparency and elasticity are opposing properties for the knit - high transparency with the low elasticity was hard to achieve.

interlock

Left page: Close up view of the differentiated nature of the bespoke textile Right page: Samples of the three main knitting structures

CHALLENGE

Pattern

Unit Cell

Loop Parameters

Pique Lacoste Aff

Starts with a Jersey pattern followed by accumulation on the even needles + normal loop on the unpaired ones then switching to normal loop on the even needles and accumulation on the unpaired needles.

Pique Lacoste

Two sets of front loop on the even needles and accumulation on the unpaired ones followed by two sets of accumulation on the even loops and normal loop on the unpaired needles.

Ponto Di Roma

This structure begins with a double knit (knit the front and back) followed Jersey on the front needle bed and finish with jersey on the rear needle bed.

Samples Graduation (Density of Knit)

31 37

Expected Material behaviour

Very soft, low density, low weight, good transparency, High elasticity, very anisotropic elasticity

Very soft, low density, low weight, little loss of transparency, little loss of elasticity

Higher density, low transparency, higher thickness, low elasticity

DETERMINATION OF

MATERIAL

PROPERTIES CULTURAL INTERFACES There are two cultural contexts for textile design of knitted fabrics; the knitwear garment industry and the technical textiles field. Both sectors define design criteria for textiles mainly through 1:1 sampling, the â&#x20AC;&#x2DC;sampleâ&#x20AC;&#x2122; being a 2D panel that visualises and tests the surface quality (yarn, colour and stitch structure) and behaviour of the fabric. This method has its embedded problems. Theoretically, samples translate to larger fabrics directly. However, a fabric may not behave homogeneously across its entire width. When under tension, the distortion of the stitch loop varies relative to local conditions at the edge of the fabric compared to the centre. Furthermore, small variables in machine set-up, slight changes in needle movement and tensioning can lead to large performative changes across the fabric (Renkens 2010). When scaling materials up to architectural scale, these imprecisions are amplified, which makes them difficult to control. Left page:Properties table of the main tested knitting structures Right page: Exploration and visual analysis of the testing knitted samples

Correlation between measured and calculated stress-strain paths and resulting stiffness parameters for Piquet Lacoste 31

Piquet Lacoste structure

Ponto Di Roma structure Left page: CNC fabric crusiform in the load testing machine, Duisburg, Essen, Germany Bottom: Cruciform samples patterns

TESTING THE KNIT MATERIAL In ‘Tower,’ the challenge is to simulate the high strains and transverse contractions under load in both directions. In order to inform the simulations, we developed new testing methods to determine the material properties of the designed textiles. The testing procedure of MSAJ/M–02–1995 defines five different stress ratios (1:1, 2:1 1:2, 1:0 and 0:1) that are consecutively applied on a cruciform shaped test specimen. The result of this procedure is a stress-strain-diagram (Figures 8, 9). From this complete set of test data, ten stressstrain-paths can be extracted. A single design set of elastic constants from the extracted stress-strain-paths can be determined stepwise in a double step correlation analysis.

Right page: Biaxial material performance test accomplished in Duisburg Essen facilities

CHALLENGE

FE Simulations - first results

Analysis after prestressing the system. Whole system with rods Ø 16mm GFRP circular section. Analysis after prestressing the system. Whole system with rods Ø 16mm GFRP circular section. Max von Mises Stresses (+/-) [Limit] = 250 N/mm² Max von Mises Stresses (+/-) [Limit] = 250 N/mm²

Max von Mises Stresses (+/-) = 250N/mm2 Normal Forces [kN] Max tesion 1.3 kN Max compr. –1.3 kN

Max Stress 125 N/mm2

Loadcase: (deadload + shaping of rods + prestressing of membrane+wind)

normal forces [kN] max tension 1,3 kN max compr. -1,3 kN Max von Mises Stresses (+/-) = 250N/mm2

KET · Tower 2.0 – iter 01 : 8 Sides, 8 Floors· FE Simulations - first results

Max von Mises Stresses (+/-) [Limit Max stress 125 N/mm²

Loadcase: (deadload + shaping of rods + prestr

KET · Tower 2.0 – iter 01 : 8 Sides, 8 Floors· FE Simulations - first results

Analysis after prestressing the system. Whole system with rods Ø 16mm GFRP circular section.Analysis after prestressing the system. Whole system with rods Ø 16mm GFRP circular section. Normal Forces [kN] Max von Mises Stresses (+/-) [Limit] = 250 N/mm²

Max von Mises Stresses (+/-) = 250N/mm2 Normal Forces [kN] Max tesion 2.7 kN Max compr. –2.3 kN Loadcase: (deadload + shaping of rods + prestressing of membrane+wind)

normal forces [kN] max tension 2,7 kN max compr. -2,7 kN

KET KET · Tower · Tower 2.0 –2.0 iter–01iter: 802Sides, : 12 Sides, 8 Floors· 8 Floors· FE Simulations FE Simulations - first -results first results

(deadload + shaping of rods + prestres

: 128Sides, KET · KET Tower· Tower 2.0 – 2.0 iter –01iter : 8 02 Sides, Floors·8 Floors· FE Simulations - first results FE Simulations - first results

Analysis Analysis after prestressing after prestressing the system. theWhole system.system Wholewith system rodswith Ø 16mm rods ØGFRP 16mmcircular GFRP section. circular section. Analysis after Analysis after prestressing system. Whole rodsGFRP Ø 16mm GFRPsection. circular section. prestressing the system.theWhole system withsystem rods Øwith 16mm circular Normal Forces Normal[kN] Forces [kN] Max von Mises(+/-) Stresses (+/-)= 250 [Limit] = 250 N/mm² Max von Mises Stresses [Limit] N/mm²

Normal Forces Normal[kN] Forces [kN] Max tesion Max tesion 2.7 kN 1.1 kN Max compr. Max–2.3 compr. kN -0.8 kN Loadcase:Loadcase: (deadload(deadload + shaping+ ofshaping rods +ofprestressing rods + prestressing of membrane+wind) of membrane+wind)

normal forces [kN] max tension 2,7 kN max compr. -2,7 kN

Max von Mises Stresses (+/-) [Limit]

Max Stress Max stress 183 N/mm² 183 N/mm2 Loadcase:

Max von Mises StressesMax(+/-) Max von Mises(+/-) Stresses (+/-)= 250 [Limit] = 25 von Mises Stresses [Limit] N/mm² stress 75 N/mm² Max stressMax 183 N/mm² = 250N/mm2 Loadcase: Loadcase:

+ shaping of rods + prestressing (deadload +(deadload shaping of rods + prestressing of membra Max Stress 183 N/mm2

SIMULATION OF FOR

KNIT

HYBRID

STRUCTURES INFORMING MATERIAL THROUGH SIMULATION ‘Tower’ investigates knit as a structural membrane in which active bent GFRP rods are embedded into a bespoke knitted textile (Ramsgaard Thomsen 2015). The relationship between skin and structure is a central question in the field of architectural textiles, positioning the textile membrane either as a cladding skin, or engaged in hybrid dependencies in which membrane and scaffold act as an integrated structural system. The latter requires a high degree of control and understanding of the membrane’s material behaviour. ‘Tower’ develops new modelling practices needed to devise hybrid behaviours.

Left page: KET Structural analisys of the Tower 2 design iterations Right page: Pull and Squeeze Structural Test

PHYSICAL PROTOTYPING

FORM FINDING Particle Spring System

Bending Radii Analysis Particle Spring System FEM simulation (external loading)

Superimposition of Stresses

47 Evaluation

Production

Total Stress = [Residual Stress from form-finding] + [membrane pre-stress] + [dead load] + [wind]

Left page: Stages of the FE-simulation setup Right page: FEA-simulation approach

CHALLENGE

INTERFACING DESIGN SPECIFICATION INFORMING THE MATERIAL The ability to design for and with material performance is a core resource for design innovation closely tied to material optimisation. The project introduces three scales of design engagement by which to examine material performance: the structure, the element and the material. The project asks how to support feedback between different scales of design engagement moving from material design, across design, simulation and analysis to specification and fabrication.

Left page: Close up view of the differentiated nature of the bespoke textile Right page: Samples of the three main knitting structures

knitting without shaping

Prototype textile patch for Tower 1, 2015. Strip is knitted to the full width of the machine bed without shaping

knitting with shaping

Right page: Prototype textile patch for Tower 2, 2016. First successful attempt to knit â&#x20AC;&#x153;into shapeâ&#x20AC;?.

2.018 2.018 2.018

2.018

0.978

initial input mesh

details layer

details clustered by pattern type

yuliyas adjustments final

pixelated CNC knitting file

INTERFACE DEVELOPING The development of our interfacing tools span both Stoll and Shima Seiki machines. Developed methods are enable the direct creation of the BASIC machine code consisting of patterns of letters defining the knitting beds, the yarn carries and holding patterns, and thereby controlled the formation of the knitted textile. The interfacing is controlled by the generation of pixel-based files in which the controlled colour coding of each pixel denotes the particular structural and material information for each stitch. The files interface directly with the machine software. Here, tiff files are imported into the knitting software, which allows direct control of the structure, material and shape. The complexity of the design specification is controlled through bespoke definitions in Grasshopper in which matrices are used todefine the linear production of each knitting row. Left page: Sequence of the interface developing: from mesh to TIFF file Top: Patch developing process in Rhino Bottom: Corresponding relation between mesh and the pixelated map

initial dimensioning of the pixel file in accordance to the amount of knitting stitches

definiton of all pixels on the canvas

extraction of a+b pixels

extraction of c+d pixels

extraction of h (lace openings) and t (transfer) pixels

extraction of w (waste yarn) pixels

[]

[a,b,c,d,w,h,t]

initial dimensioning of the pixel file in accordance to the amount of knitting stitches

[a]

definiton of all pixels on the canvas

[b]

extraction of a+b pixels

[c]

[d]

extraction of c+d pixels

[h]

[t]

extraction of h (lace openings) and t (transfer)Left pixels page:

[w]

extraction of w (waste yarn) pixels

r o w 2 6 8

Left page: Pixel based code for the knitting machine, fragment of the Tower patch Right page: Data tree with each branch being a string for a knitting row

CHALLENGE

developed patch for the prototype

top view of Prototype 1

side view of Prototype 1

SCALING

FIRST STEP: BUILD 1:1 SCALE PROTOTYPE In order to test out the digital model and the correlation between the digital information and the physical outcome 2-level prototype was built. The intension of the experiemnt was to see wheather the fabricated â&#x20AC;&#x153;intoâ&#x20AC;&#x153; shape textile membranes are fitting the geometry. Prototyping was also a good approach to test out the textile details, the dimensions and the building sequence. Hybrid surface was assembled on the floor and after wards rolled into a cylinder and tensioned inwards by the tension cables. Overall experience of the prototype was very positive and gave a lot of promises for the large scale tower.

Left page: 1:1 scale prorotype Right page: Digital representation and process of building

DISCONTINUITY REINFORCED BY MECHANICAL CONNECTORS The joints between rods were in the first tower made parallel channels, which hold the rods in the knit. This tensegrity connection demonstrates the capability of the used high tenacity fibres and CNC knit. Modelling this type of joint in FEA would however be a precarious task. We modelled a direct connection of rods in the FEA model of the first Tower, which contributed to the divergence between model and the demonstrator. In the second tower we set a focus on the convergence of material and simulation and decided to align the design of the joints to the capabilities of the simulation. We developed a set of joints with a puzzle configuration, which can be attached to the beams after they slid through the channels. The parts are sliding towards each other and provide a solid connection between GFRP rods. They rest on metal stoppers, which are fixed to the beam in predefined locations and are able to withstand vertical loads of up to 50kg each. Left page: Interiour of the built Hybrid Tower 2, close up view on the utilisation of the connectors Right page: Tower 1, junction area between levels Right page top: CNC milling path visualisation

Type 1

Type 2

Type 3

initial assembly logic of joints

bolt connection enables axial rotation

deployed state of joints other rod is stopped inside of a joint to be able to have stucking arches

closed state of joints

bolt connection enables axial rotation

top joint 2 rods connected with the joint type 3

cross joint 4 rods connected with the joint type 2

stopper joint 2 rods connected with the joint type 1

Left page: Diagram of the three types of connector details Right page: The indication diagram of the connectors distribution

FABRICATION OF CUSTOM JOINTS Connectors needed to be developed and produced custom due to the absence of a ready solution for our particular case on the market. The joints needed to meet certain set of requirements: to prevent rods from depositioning, to be able to be installed without having access to the end of the rod (â&#x20AC;&#x153;click-inâ&#x20AC;?) and to consist of strong but not brittle material. The joints development ended up on a puzzle solution for the joint, which can allow installation and removal of the joints after the rods are already integrated into a textile.

Left page: Puzzle connectors parts Right page bottom: Custom fabricated HDPE connectors for GFRP in a puzzle configuration Right page: CNC.milling process

FABRICATING AND PREPARING TEXTILES FOR LARGE ASSEMBLY Textile skin is knitted as set of patches that later need to be prepared for the assembly. It is mostly means the integration of reinforced elements such as metal rings in the pulling details as well as tiyng ropes and seaming patches togethr into a single large membrane. The entire tower skin consisted of 9 patches that were manually sewn together into a single textile membrane. The preparation phase included manual sewing, embedding of plastic reinforcement details (tension rings) and tying ropes. Facilities to complete all those tasks were provided by hosting institute Universidade do Minho â&#x20AC;&#x201C; Fibrenamics.

Left page: Stopmotion of the textile preparation Right page: Seam close up of the Textile Tower

CREATING A HYBRID TOWER SKIN After the textiles were sewn into a single sheet the process of structural elements integration was started. Each GFRP rod was cut into particular length and embedded into the textile membrane row by row. When this job was completed, the hybrid surface was ready to be rolled into a cylinder.

Left page: Stopmotion of the hybrid surface Right page: Tensioning challenge while producing the hybrid surface

ROLLING THE SURFACE INTO A TOWER After the textiles were sewn into a single sheet the process of structural elements integration was started. Each GFRP rod was cut into particular length and embedded into the textile membrane row by row. When this job was completed, the hybrid surface was ready to be rolled into a cylinder.

Left page: Stopmotion of the rolling process Right page bottom: Tensioning of the hybrid surface Right page top: Thin lightweight hybrid surface of Tower 2

TRANSPORTING THE TOWER AND RAISING IT UP ON THE SITE The lightweight structure was assembled on the Plataforma des Artez and easily transported to the installation site without any heavy duty machinery. Seven people could hold it and move through the city. On the Toural Square the hybrid structure was lifted with the ropes and installed onto the prepared steel base.

Left page: Stopmotion of bringing an raising up the tower on site Top: Lightweight tower is being brough by 7 people Bottom: Tower is delivered on site

RESULTS

CONTEXTILE 1:1 PROOF OF CONCEPT Tower 2 was a part of Contextile conference in Guimaraes in August 2016. The exhibiting period was three months and tower performed great in terms of wind and percipitation load. The aim of the research is considered as being reached. The convergence between specified, simulated and real world behavior of a hybrid structure made of bending active GFRP rods and bespoke CNC knitted fabric is found. We were able to find working solutions to overcome the limitations, that plagued the material of the first Tower in 2015. This allowed to achieve high precisions regarding dimensions and material properties in the knit, supported through the developed automated interfacing from the design- to the production environment. The interface can here be considered universal. Minor adjustments of parameters should allow it to produce any type of knit on most single and double bed knitting machines. Left page: Assembled Tower 2 on the Placa De Toural Square in Guimaraes, Portugal. Night view. Right page top: Photo from the local media newspaper. Right page: Poster of Contextile 2016

Left page: Tower 2 during day on the Placa De Toural Square in Guimaraes, Portugal Right page: Interiour of the tower Photos: A.Ingvartsen

Left page: Tower 2 during day on the Placa De Toural Square in Guimaraes, Portugal Right page: Steel base of the Tower 2 Photos: A.Ingvartsen

Left page: View from the bottom on the Textile Tower Top: Transparency of the material Bottom: Beautiful curvature of the pulled cones

Left page: Tower 2 during night on the Placa De Toural Square in Guimaraes, Portugal Right page: Cones pulled inwards, interiour view Photos: A.Ingvartsen

Left page:Facade of the Textile Tower 2. Right page: Tension cable details wit the reinforcing ring Photos: A.Ingvartsen

STRUCTURE (macro)

ELEMENT (meso)

MATERIAL (micro)

materiel prototype

S T R U CTUR A L M EMBRANE S 2015- BARC E L ONA - OCTOBER 2015

Developed Global Modelling Method: • Global geometry from Rhino/GH. • Apply pre-tension (reduce stiffness of tension members/membrane shaping).

digital model

01 5- B A R•C EApply L ONAwind - OCT OBE R 2015 loads.

• Analyse deflections. TOWER • Check utilization.

g Method:

MEMBRANE AND ROD

FABRIC

FEEDBACK

no/GH.

CONCLUSION stiffness of tension members/membrane shaping).

• Superimpose FE results with residual stresses from Rhino/GH.

FURTHER OUTLOOK

Both in tower 1 and tower 2 the projects were accompanied by the monitoring stage, where the diviations and geometrical changes were traced with the photogrametry and the 3dscanning. The differences between the assumed material behaviour and the one, that was measured on the biaxial testing machines demonstrate the need for objective evaluation. The ability to knit directly the cruciform specimens eased the process of testing. residual stresses from Rhino/GH. The interface between design and simulation environment improved and the method for the FEA allows now to simulate structural systems, Ramsgaard Thomsen, et. al(2015) Hybrid Tower, Designing Soft Structures. In Ramsgaard Thomsen, M., Tamke, which consists of 90 or more interacting bendM., Gengnagel, C., Scheurer, F., Faricloth, B.. (Editor) Modelling Behaviour, Springer, Berlin Heidelberg. ing active elements in a reasonable time frame. An important foundation for the fast feedback in the design process of hybrid structures.

A NALY T IC A L : S OFISTI K M O D EL L I N G

C o m put at io nal M o de lli n g Pipeli n e

Left page: 3d Scanned Scene of the Textile Tower 2. Right page top: Multi-Scalar Modelling Diagram Right page bottom: Photogrametry monitoring of the Tower 1

References: 1.

6. 7. 8.

9. 10.

11.

12.

13.

14.

Stanley, K., & Lehman, J. (2015). Why Greatness Cannot Be Planned: The Myth of the Objective (Aufl. 2015 ed.). Cham: Springer International Publishing. Secretan, J., & Beato, N. (n.d.). Picbreeder. Proceeding of the Twenty-sixth Annual CHI Conference on Human Factors in Computing Systems - CHI ‘08. Wagy, M., & Bongard, J. (2015). Combining Computational and Social Effort for Collaborative Problem Solving. PLoS ONE PLOS ONE. Nicholas, P., Stasiuk, D., Nørgaard, E., Hutchinson, C., & Thomsen, M. (n.d.). A Multiscale Adaptive Mesh Refinement Approach to Architectured Steel Specification in the Design of a Frameless Stressed Skin Structure. Modelling Behaviour, 17-34. Palz, N., & Arkitektur, D. (2012). Emerging Architectural Potentials of Tunable Materiality through Additive Fabrication Technologies. København: Kunstakademiets skoler for Arkitektur, Design og Konservering. Chan, M.T.K., Gorbert, R., Beesley, P. and Kulić, D. Curiosity-Based Learning Algorithm for Distributed Interactive Sculptural Systems. Philip Beesley Architect Inc. (2009). Hylozoic Soil [Interactive installation]. Exhibited at Fundacion Telefonica VIDA Matadero, Madrid. Watson, R., Ficici, S., & Pollack, J. (n.d.). Embodied Evolution: Distributing an evolutionary algorithm in a population of robots. Robotics and Autonomous Systems, 1-18. Krijnen, T., & Tamke, M. (n.d.). Assessing Implicit Knowledge in BIM Models with Machine Learning. Modelling Behaviour, 397-406. Patané, G., & Spagnuolo, M. (n.d.). An interactive analysis of harmonic and diffusion equations on discrete 3D shapes. Computers & Graphics,526-538. Lawson, B. (n.d.). Oracles, draughtsmen, and agents: The nature of knowledge and creativity in design and the role of IT. Automation in Construction, 383-391. Risi, S., Lehman, J., D’ambrosio, D., Hall, R., & Stanley, K. (n.d.). Petalz: Search-based Procedural Content Generation for the Casual Gamer. IEEE Trans. Comput. Intell. AI Games IEEE Transactions on Computational Intelligence and AI in Games, 1-1. Clune, J., Chen, A., & Lipson, H. (n.d.). Upload any object and evolve it: Injecting complex geometric patterns into CPPNS for further evolution. 2013 IEEE Congress on Evolutionary Computation. Sims, K. (1997). Galápagos [Interactive media installation]. Exhibited at Intercommunication Center, Tokyo.

Media References: 1.

2. 3. 4. 5. 6. 7.

http://www.guimaraesdigital.com/noticias/64966/fibrenamics-participa-na-contextile-2016-com-torre-textil-no-largo-dotoural http://www.gmrtv.pt/sociedade/27911-fibrenamics-da-uminho-participa-na-contextile-com-projeto-internacional https://www.uminho.pt/EN/follow-uminho/ Pages/event-detail.aspx?Codigo=49054# https://alumni.uminho.pt/pt/news/Paginas/ Not%C3%ADcias%202015/Fibrenamics2. aspx http://sanjotec.com/?p=4946 https://www.youtube.com/ watch?v=drJeNwmxke8 https://www.youtube.com/watch?v=UgdoAFpbG8 (on 0:17sec)

CITA | Centre for Information Technology and Architecture The Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation Philip de Langes AllĂŠ 10 1435 Copenhagen K Denmark http://cita.karch.dk