data management generative input/ output analysis parameters modelling genes parametric evolutionary genetic fitness criteria efficiency performance material systems BIM real-time urban systems simulation optimization robotics cnc laser cutting rapid prototyping
DESIGN
art fashion product interior architecture
&more
algorithms
digital
systems
fabrication
laser sintering aurdino kinect data mapping rhinoceros max revit python nest sketchup maya autocad grasshopper iworks ms office mac adobe pc
rt eecsh en oa lrocghy
by Pradeep Devadass
3D
2D
Design To Fabrication This section of the portfolio illustrates the methodology in exploration of advanced digital fabrication techniques using computational design methods and manufacturing 1:1 scale prototypes which mainly focuses on optimization of material, maximize quality and quantity and production of complex shapes with ease. Although there have been several attempts, industrial robots are still mainly used as tools in the fabrication process. The research also aims to redefine how industrial robots can be innovatively used not only as a fabrication tool in the production but an integral part of the design process. The digital fabrication tools mainly include 3 axis CNC and 6 axis Robotic arm.
Light Vault
Robotic Fabrication Research Project Information:
Name: Light Vault Project type: Research Location: Archi-Union, Shanghai
Project Description:
The research integrates advanced computational design and digital fabrication methods through the project Light-Vault, which addresses the challenges faced by current contemporary practices in development and execution of complex designs. The project shows the development of a vault created through aggregation of several dissimilar components in which the interior volume is carved out leading to formation of ruled surfaces. Porosity of the component is parametrically designed through a developed genetic algorithm controlled by multiple fitness criteria. In parallel, the project explores and implements the potential of robotic technology and develops personalized robotic tools and production techniques, in generating quick shaping volumes through Hot-Wire Cutting process. Algorithms are developed to ensure design thinking and fabrication procedures are simultaneously developed in a non-linear, parallel performance based process. This cumulative cohesive process between advanced digital and physical computation methods is translated through a full-scale built pavilion.
OPTIMIZED MODEL
COMPONENT
GENETIC ALGORITHM
Methodology
PARAMETRIC MODEL SIMULATION PARAMETERS CONTROL POINTS OF CURVES GRASSHOPPER FORMATION OF RULED SURFACE
COMPONENT AGGREGATION Component 1 + Component 2 +
KUKAprc ROBOTIC CODE GENERATOR
+ COMPONENT n
KUKA kr-30
EVALUATION PARAMETERS EXPOSURE ANALYSIS LIGHT-VAULT SYSTEM
VOLUMETRIC ANALYSIS STRUCTURAL ANALYSIS
DIGITAL DEVELOPMENT
HOT-WIRE CUTTING PROCESS
MANUAL ASSEMBLY
SUBTRACTIVE MANUFACTURING
ADDITIVE MANUFACTURING Methodology
FABRICATION DEVELOPMENT
In design practices, the decisions on execution of a project are developed at the final stage of design phase, through a linear process. A parallel thinking methodology is developed to combine design development and fabrication process. Execution of a design has always been an isolated unparalleled process which is developed at the end of the design stage Evans (1997) explains the inevitable gap between drawing, the medium of design and the final outcome in architecture. This distinct unconnected procedure leads to a problematic translation from digital environment into reality. Development of a methodology from design to fabrication through integration of technology in a controlled digital environment is established in the project Light-Vault. A performance based design approach is the key focus in developing the methodology. Also, evaluation is normally generated during the end of the design stage leading to the absence of continuous feedback process for the betterment of design output. Genetic algorithms are developed for optimization of a system, imitating the process of natural selection (Mitchell, 1996).
The project Light-Vault questions rectilinear form and breaks through for an organic morphological design process using freeform development which is defined by algorithms, controlled by parameters. In nature, the growth is governed by a set of instructions influenced by the external stimuli, in which cells are generated and regenerated, leading to an individualistic and characteristic morphological development, thereby making it challenging to realize the intelligence behind the logical approach but easily perceived by the external morphology (Vogel, 2003). This systematic methodology in nature is translated into the project using computational methods integrating performance of the system, materiality, fabrication constraints, component logistics and assembly process.
Structural Formation of the Vault (Left), Optimization of Vault Components (Right Box)
Conclusion
This approach creates a new paradigm in methodology by achieving performance and fabrication through optimization: Optimization of the Design Performance: Development of design and fabrication in a controlled parametric approach and continuous feedback process by synchronization and interlinking of various parameters through a single algorithm gives complete control to the designer and realizes flexibility in optimizing performance at every stage leading to a significant improvement in the quality of the output. Optimization of Fabrication Process: Optimization of design by incorporation of available raw material in fabrication of components results in a cost efficient method. In this methodology, the potentiality, where parameters such as accuracy of the size and shape, production quality, and time duration in the process are individually optimized (making the right choices and creating balance) which can be exploited by commercial offices according to their requirements.
v
v
Urban Perspectives CNC Fabrication
Project Information:
Name: Urban Perspectives, Acrylisize Project type: Commercial Location: RoboFold, London
Project Description
To design an Art installation, which consists of 400 pieces of polished aluminium folded boxes which are mounted on MDF panel. Due to usage of polished aluminium material the box exhibits a reflection on its surface. Each piece is customized and oriented with respect to the location of its focal point(object), which is required to reflect to the observer. The entire panel is split into 5 zones, each zones has its unique focal point. The main challenge was to produce 400 unique pieces, in which each piece varies from another, in dimensions and form
Prototype
v
v
Final Installation
Production with CNC
Design Methodology
The design is controlled by 3 Attractor Points/ Parameters • Object or the focal point • Observer point • Converging point is attractor parameter where the boxes orient towards it to form an overall concave form in that zone
Design development
The form of the box is a result from the orientation of the top surface of the box with reference to the object, observer and converging point by shifting the angle of the surface to equalize the incident and reflected ray.
Observer
Object
Focal Point 2 Zone 2
a° a° Focal Point 3
Mirror Observer
Previous location New location Observer Point
Zone 4
a° a°
Zone 5
Zone 3
Mirror
Observer
attractor pt
Zone 1
Above: Principles of reflection. Below: Orientation of the mirror with respect to the new location of the object
Parameters Object
Focal Point 1
Focal Point 4
=
The entire panel is split into 5 zones. Each zone focuses its = respective object point (In elevation)
Focal Point 5
Various Focal points (In plan)
Zone
Zone development
Converging point
Geometry development Focal point (Object)
Fabrication development
Unfolding surfaces
Observer Entire panel grid - 20 x 20 (Perspective)
Zone 3 (5 x 5 grid) Parameters - Location of points
Incident Ray from Object to Mirror
Reflected Ray from Mirror to Observer
Converging point
Re-arrangement of surfaces on sheets Generation of drawing for cnc Generation of g-code for cnc
formation of geometry
Zone 3: Final geometry (5 x 5 grid, 25 boxes)
Zone 3: Final geometry
Final Geometry
Re-Orienting the top surface to XY Plane
Unfolding geometry with bottom surface split into flaps
Unfolded geometry onto XY Plane
Production Methodology
Unfolded 25 boxes onto XY Plane
Optimise 25 unfolded pieces onto 2 sheets
Screw holes Double score lines Geometry ID Rivet holes Screw holes Cut Final Line Drawing for CNC Milling
Due to the multiple parameters, and application of the similar strategy on 5 zones, a parametrically designed algorithm is developed and tested on Zone 3 using Grasshopper along with various Plug-ins on Rhino platform. The geometry from the final design is obtained and unfolded to a flat surface. The top surface is oriented to the XY Plane while the bottom surface is split into flaps for production. The unfolded surface is then optimized onto sheets (size:1500x2500mm) using Rhino Nest in grasshopper for CNC milling. The unfolded pieces are then automatically drafted using the algorithm which are based on a set of rules which includes • Tolerance required for folding of the material. • Holes on the flaps for rivets and screws which attach to the MDF panel. • Double scoring of material required for 360° fold. • Classification and assortment of layers according to score and cut for CNC G-Code generator.
Project Information:
Name: Material feedback in robotic fabrication Project type: Workshop Location: University of Michigan, Ann Arbor
Project Description
The workshop explores on the production of complex arrangement of components, using a feedback robotic workflow and possible applications. The research focuses on four main topics 1. Materialization of complex surfaces: Complex geometry architecture requires multiple one-off components. In this type of projects, fabrication of large number of unique pieces becomes a major challenge. 2. Material feedback: The mainstream approach to robotic fabrication is based on a linear sequence that extracts information from a three dimensional model, creates the robotic motion and executes it. The introduction of sensors to the robot, combined with realtime streaming of commands to the robot open new opportunities for workflows that make use of information feedback. In this way, the fabrication process becomes a cybernetic mechanism, which adjusts the system in real-time to unpredictable material behavior. 3. Custom tooling: Custom tools are an important aspect of this production concept. Deforming, smoothing, carving or patterning are possible manufacturing processes which require the use of therefor designed tools as well as tailored motion setting
Material Feedback in Robotic Fabrication
Rob|Arch, 2014 - Conference +Workshop
Methodology
Aggregation is widely used method in the construction industry where colossal structures are easily erected through assembly of smaller components. The challenge here is to create an aggregation, dense or sparse arrangement of material/ components not based on aesthetics but on the availability. First, the sensor (Microsoft Kinect) attached to robotic arm (ABB 140) scans the quantity of available material. The algorithm arranges and rearranges the components a given frame according to the fitness criteria set by the user from the scanned information through use of genetic algorithm. Through multiple iterations the final output is then sent to the robotic arm for execution. Applications: The developed methodology can be applied to a re-configurable wall where extreme fluctuation of temperatures are found in various seasons of the year.
Methodology v
v
Kinect scanning v
v
v
Sparse multi layer arrangement of same Dense Single layer arrangement of same quantity of material quantity of material
Project Information:
Name: Adaptive Structural Skins Project type: Workshop Location: University College London, London
Base Diagram for King-Kong
3d Digital Model - Mesh
Simulated 3d Digital Model - Polysurface
Prototype Paper Model
Project Description:
To develop a prototype through Curve Sheet Metal Folding and simultaneously analyse the strength of the system.
Curved Folding
Paper folding
Kangaroo Folding
Akti Model
robofold godzilla
UNiversal Robots
Generation of G-code
CNC Cut aluminium sheet
Robotic Simulation of Digital Models
Aluminium Sheet Folding using Robots
Robofold cam CNC Router
Adaptive Structural Skins
Smart Geometry, 2014 - Conference +Workshop
Final Prototype
Mountain
Valley
Project Information:
Name: King Kong(Grasshopper Plugin), Sheet Metal Folding Project type: Workshop Location: RoboFold, London
Mountain
Inputs for King Kong plugin Paper Model: Manual Folding process
Project Description:
Developing design using King-Kong(a grasshopper plugin developed by RoboFold) and Fabricate using CNC and Robots . The plugin facilitates digital simulation of Curve Sheet Folding into components and arrangement of the developed components onto a surface with a range of attractor points. The Plugin was used to design a Wall Panel consists of leaf shaped panel using King Kong (Grasshopper plugin) and then manufacture prototype using CNC and Robots.
1:1 scale model: Robotic folding process
Output - Polysurface model
Output - Mesh model
Paper Model
King Kong(Grasshopper Plugin), Sheet Metal Folding Software
Workshop + Robotic Fabrication
Final Installation
Robotic Simulation Software
Project Information:
Name: Godzilla (Grasshopper Plugin), Robotic Simulation Software Project type: Software Development Location: RoboFold, London
Project Description:
A software application which simulates the robotic movement in a digital environment to develop and test the designer’s process intuitively before using the robot. The application automatically generates Code for Robotic fabrication
Godzilla(Grasshopper Plugin), Robotic Simulation Software
Workshop + Robotic Fabrication
Computational Design of Material Systems
This section of the portfolio illustrates the methodology in understanding and exploration of Materials used in interconnected systems. The properties of materials are studied and incorporated in the digital environment for designing using computational methods. The research focuses on development of systems through real time optimization and analysis using genetic and evolutionary algorithms.
Project Information:
Name: Adaptive Skins Project type: Academic Research Location: Architectural Association, London
Abstract:
The project investigates responsive building skin systems that adapt to the dynamic environmental conditions to regulate the internal conditions in a habitable space over different periods of time by exhibiting a state of motion and dynamism. Heat and Light are the primary parameters for regulation, leading to energy efficiency and dynamic spatial effects. Passive and active skins using shape memory alloys and pneumatic actuators are developed through investigations of smart systems that integrate smart materials and smart geometries. The precedents in this domain have rarely dealt with individually controlled multiple parameters of heat and light in a single system, which is attempted in this project. Owing to the complexity of the multi-parametric system, genetic algorithms are developed for system optimization and calibrated with physical prototypes at varied scales. The developed systems are tested against two distinct climatic models- New Delhi and Barcelona, and evaluated for performance, based on heat and light, which are quantified as solar gain and illuminance as principles, and daylight factor for evaluation purpose. The use of genetic algorithms makes the problem solving faster and accurate. New tool-sets are developed in the process by combining various digital tools, to create a feedback and memory loop system.
adaptive [skins]
Objectives & Scope
• To design a responsive building skin system that adapts to the environmental changes and regulates the internal conditions in a habitable space by exhibiting a kinetic state. • To develop a system that is energy efficient, by the virtue of its organization and configuration- making use of the component • To embed material intelligence into the system to reduce the amount of energy used to drive the kinetics of the system. • To integrate material intelligence and geometric intelligence, developing a building skin system which can be deployed to multiple contexts (climatic models) at architectural scale of a façade or a roof.
COMPONENT BASED SYSTEM LOCAL Change >>> GLOBAL Effect
adaptive [skins]
MULTIPLE Parameters
>>>
ENERGY EFFICIENCY REGULATING INTERNAL CONDITIONS
Individual Control of Each Parameter using a SINGLE SYSTEM
SPATIAL EFFECTS / AESTHETICS Flowchart showing the aims & objectives of the research
Methods
The two major parameters that drive the research / project are heat and light. These are articulated in terms of building science as SOLAR GAIN and ILLUMINANCE, respectively. These two become the input parameters to develop and evaluate the system. Solar gain is controlled by changing the angle of incidence on a surface, which can change the angle at which sunlight hits the surface, controlling the amount of radiation that passes through. This simple principle is used as a parameter to control solar gain in a space. Illuminance is altered by changing the light infiltration and regulating the amount of direct and diffused light that enters a space. By changing the angle of incidence of a surface, openings can be altered through which light can pass through, either directly as direct exposure or indirectly as diffused light. By aiming for a dynamic system that changes the angle of incidence of the surfaces of a system, solar gain can be controlled by seasonal movements of the system. In summers, solar gain can be reduced to bring the internal conditions to a thermal comfort zone by INCREASING the angle of incidence of the overall surface / skin. In winters, the effect can be reversed by DECREASING the angle of incidence and increasing the solar gain to bring the temperature up to thermal comfort zone. With the changing angle of incidence, direct light can be altered and converted to diffused light by configuring the surfaces in a certain way. Increasing the direct light also contributes to an increase in temperature and this effect can be used in a controlled way depending on the seasonal requirements. Also, direct light leads to glare, which needs to be controlled and changed to diffused light, for a good illuminance in a given space.
Angle of Incidence : LESS Angle of Reflection : LESS Solar Gain : MORE
Angle of Incidence : MORE Angle of Reflection : MORE Solar Gain : LESS SOLAR GAIN
Diagram explaining the principle of solar gain
Size of opening : LESS Illuminance : LESS
Size of opening : MORE Illuminance : MORE Illuminance
Diagram explaining the principle of illuminance
Responsive ARCHITECTURE
Methodology Overview
Dynamic Kinetic
Generations
EVOLUTION
Adaptive Architecture
COMPONENT DEVELOPMENT
BEHAVIOUR
SYSTEM DEVELOPMENT
OPTIMIZATION
FORM
SPATIAL EFFECTS
PHYSICAL EXPERIMENTS GENETIC ALGORITHM
METHODS
MATERIAL COMPUTATION DIGITAL EXPERIMENTS
COMPUTATIONAL TOOLS
GENERATIONS
EVALUATION feedback
PHYSICS ENGINE
EVOLUTIONARY SOLVER
ENERGY EFFICIENCY TOOLS
OUTPUT Flow diagram showing the methods used for the research development
In most of the projects and research that lie within the domain of responsive architecture, dynamic architecture or kinetic architecture, which are all sub-domains of adaptive architecture, there is a non-linear approach towards design. In most of the cases, a component system is developed which is aggregated to form a system (with some modulations in dimensionality and configurations) that responds to external stimuli. Evaluations are usually done during the component development stage and performance is calibrated after the deployment process. There is no feedback involved after the deployment. Due to this, there is a loss of energy, in the process, as well as in the project as a whole, since there are projects which exhibit dynamism but lose a lot of energy for their functionality. To deal with the complexity of projects, algorithmic design is used in most cases. The common approach to develop a responsive system is argued in the research – adaptive [skins] and a new methodology is formulated using the existing computational powers that form the state of the art. Also the approach used in Hygroscope project that deals with material computation is taken as a role model while developing the new methodology. A series of parallel research based on smart geometric systems and smart material systems is carried out in a non-linear way, with continuous feedback networks. Genetic algorithms are used, after understanding the complexity of the multi-parametric problem to build a component based system. Genetic algorithms are used to create a strong feedback network in the algorithm where real-time evaluations inform the initial stages of the algorithm. For embedding real-time evaluations in the algorithm, a new tool set needs to be created by integrating some of the state of the art computational tools to have a single interface where gene pool formation, evaluation, development, feedback, re-development and re-evaluation is done. This is done by creating a work-flow of digital simulations in a computational environment, which is able to simulate physical forces, integrated with physical tests based on materiality, leading to a material computational methodology, which is specific to the project.
Need for a Digital Environment
A single interface digital environment is needed to be developed in order to run real-time evaluations in the system with feedback networks to inform the gene pool used in the genetic algorithm. INPUT: Skin- A simple surface is divided into several parts which exhibits morphological change when force is applied. Tools - The parametric model is developed using Grasshopper in Rhino platform. SIMULATION PARAMETERS: Force- Rotation of individual parts of surface around the central axis of the primary surface. Tools - Physics Engine (Kangaroo for Grasshopper) is used, which simulates the rotation angle of each surfaces individually. The angle of rotation is between +30째 to -30째 around the axis. OUTPUT: New Skin- A number of morphologically different skins are achieved by changing the angle of rotation for different part of surfaces. Evaluation Parameters Two parameters - Heat and Light are quantified in terms of Solar Radiation (leading to Solar Gain) and Solar Exposure (leading to Illuminance) respectively and are considered to evaluate the efficiency of the system. The efficiency of the system varies with the responsiveness of morphology of the system to ever changing climatic factors. During summers, due to high temperatures and excessive glare, the system changes its morphology by two methods: 1. By increasing the angle of incidence 2. By changing the size and angle of the openings Analysis Incident Radiation Analysis - To calculate and understand the Incident Solar Radiation on the skin a colour code is developed which is applied on each individual surface, according to the Angle of Incident solar radiation on that surface, thereby exhibiting different colours on each individual surface, after the force is applied in the simulation. Solar Exposure Analysis - To calculate the Direct Solar Exposure passing through the skin, an exposure tool is developed using scripting, which calculates the direct light passing to a given surface (analysis surface assigned parallel to the skin) through defined skin.
WINTER Input
Min Angle of Incidence Skin/Surface
Solar radiation SUMMER
ol
Max Angle of Incidence
Simulation Parameters / GENES WINTER Output
Min Exposure
Solar exposure SUMMER
Max Exposure
Axis tion a t o R
Individual Surface
Simulation Parameters
Multi-parametric problem solving
Parametric model
Input
grasshopper
Physics engine
Physics engine
kangaroo
kangaroo
gene 2
gene 1
gene n Angle of rotation of 9 individual surfaces
Actuated model
Actuated model
Evaluation Parameters
NEED OF A GENETIC ALGORITHM An optimization technique is necessary to evaluate the efficiency of a system, which can involve multiple simulation parameters and multiple evaluation parameters.
Radiation analysis
Exposure analysis
geco
exposure tool
Solar Exposure Analysis
Output
Analysis Output
Analysis Output
Evolutionary solver
Evolutionary solver
galapagos
galapagos
OptimIZED model Optimized state of skin according to desired Fitness Criteria Flowchart of Genetic Algorithm (Source: Authors)
Incident Radiation Analysis
Solar ray
From 0° to 20° From 20° to 45° From 45° to 60° More than 60°
Amount of Solar Exposure Solar Exposure Analysis
Angle of Solar Radiation
Exposed to light Not exposed to light
1. PARAMETRIC MODEL Parametric model/system is controlled by several simulation parameters which are also called genes. These parameters control the form and function of the system which in-turn affect the efficiency. The parametric model, in this case is developed using software called Grasshopper that runs in Rhinoceros 3D Platform (Robert McNeel& Associates). The Dynamism of the system is controlled by a plug-in for Grasshopper, Kangaroo which is a Physics Engine and can simulate physical forces in a digital medium. 2. REAL TIME ANALYSIS Based on the Evaluation Parameters, also termed as Fitness Criteria, in the language of genetic algorithms, - Solar Exposure and Solar Radiation, two parallel scripts are developed using Python Script in Grasshopper. These are scripted using Solar Exposure Tool (in-built in Grasshopper) integrated with a plug-in called Geco which links Autodesk Ecotect and Grasshopper for Real-Time analysis. To find the efficiency of the system a real time analysis is required which evaluates, in parallel, with the change in genes of the system as outputs which are fed back to the input parameters for continuous iterations in an evolutionary process. 3. EVOLUTIONARY SOLVER Galapagos (plug-in for Grashopper) is integrated in the algorithm due to multiple permutations of output which can be obtained from the simulations. This evolutionary solver is required to collect, store, compare and analyse the large amount of data produced and finally give an optimized result according to the required fitness criteria. To test the algorithm a surface skin is created by dividing a flat surface of a square grid into 9 individual square surfaces. Each grid has the possibility of rotation from +30° to -30° around its central axis. This is controlled by 9 number sliders which act as genes for the galapagos tool to control 9 individual square surfaces. The test is run for various fitness criteria to check the functioning of algorithm on a simple parametric model, before progressing to a complex scenario.
Actuation / Force
Interconnected Systems
To initiate the experiments to develop a system that is energy efficient, a component based approach is chosen. A local change, input as a force or actuation into a system, has the ability to bring about a global effect, if certain conditions of component aggregation are satisfied. These conditions are connectivity, flexibility and lightness (weight). This principle is exploited by developing interconnected systems using a component based systems. A series of experiments are carried out in physical and digital medium, simultaneously, for real-time calibration. This is the chosen workflow of experiments during the course of the research.
LOCAL CHANGES Component 1
+
Component 2
+
Component 3
GLOBAL EFFECT Collective Effect
Tensegrity System
Due to high degree of freedom in the interconnected mesh geometry experiment and no component behaviour present, Tensegrity Systems are explored in this phase of research. Tensegrity systems have the potential to be developed from a component scale, and the principle of interconnectivity can be integrated with it. These are light-weight structural systems that have loads distributed as tension and compression into individual members of the structure. This combines the potentials of the previous systems into a more dynamic yet controlled system with more flexibility and stability. In 1962, Fuller and Snelson patented Tensegrity (Oppenheim and Williams 1997). Tensegrity systems are those composed of two types of discrete members, tension members and compression members that form self-stressing structures. In a tensegrity component the tension members and compression members are in perfect equilibrium. This feature of the system is exploited by adjusting the compression member or tension members to create dynamism. 3 Strut TensegrityComponent The simplest tensegrity structure which consist of three compression members(struts) which are similar in size and thickness are connected by tension members (cables). The cables provide tension and hold the struts in position. 4 Strut TensegrityComponent A similar arrangement as the previous component but with 4 compression members as shown.
Various physical models of tensegrity system (3 strut and 4 strut) developed to study the component behaviour.
+
+
Component n
System Development – Component Scale (Local) Developing a Simple 3 Strut Tensegrity Component into a Dynamic Component: To develop the model into a dynamically controlled system, an actuator is used, by replacing side tension members as shown. The shape memory alloy which in this case is the nitinol spring, has an ability to return to its original shape after deformation due to any force, by heating it to a particular temperature. Using this concept forward, when heat is supplied to the nitinol spring, it compresses and comes back to original shape, and when the heat supply is stopped, due to tension from other tension members, the nitinol spring expands. This returns the component to its original shape thereby maintaining equilibrium in the component. This deformation is studied and developed to control the form of the system which can be further developed into a skin. The nitinol springs are chosen to replace the side tension members which can expand to a maximum dimension of 1 unit and compresses to a length of 0.3 units. This data is fed from the material experiments carried in the previous phases of research and is computed digitally now. Based on this dimension a proportional tensegrity component is built using steel springs as tension members, aluminium struts as compression members and tensile membrane / fabric as the top surface. The springs chosen for the experiment have the ability to compress or return to 1/3rd or 0.3 of its original length on the application of heat. A series of experiments are carried out to understand the behaviour of the component on actuation of each nitinol spring, individually and collectively. The actuation deforms the component as shown. A parallel digital simulation is carried out to calibrate the kinetic motion of the component. System Development – Global Scale Several components are aggregated to form an interconnected system in which each component has 3 actuators. The system is assembled by aggregation of the components on a triangular grid at the base. The components are rotated at 60° leaving a triangular gap on the top side, which is filled up by connecting through small springs in a small triangular grid To understand the collective behaviour of the system, each actuator of each component is actuated simultaneously as shown below. The actuation leads to various forms which are stored in the digital medium and studied. Each type of actuator of each component is collectively actuated to understand and study the morphology exhibited by the system.
Component
Tensile Membrane
Top Springs
3 Strut Tensegrity Component showing different arrangements of actuators (Left); Actuation of nitinol wire (Right) Steel Struts
Nitinol Actuators
Bottom Springs
Building the 3 strut Tensegrity Component using Nitinol Springs as Actuators
Exploded diagram showing parts of 3 Strut Tensegrity Component
Actuator 1
1.0 Actuator 2
0.9
0.8
0.7
0.6
0.5
0.4
0.3
1.0 Actuator 3
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.9
0.8
0.7
0.6
0.5
0.4
0.3
1.0
Study of Digital model of a 3 strut tensegrity component when each actuator is actuated individually through a simulation.
Multiple states of a 3 Strut Tensegrity Component when actuated with heating by a common hair-dryer.
Study of Digital model of a 3 strut tensegritysystem when each actuator of individual components in the system is actuated.
NoN - actuated STATE : direct solar radiation
Actuated STATE : CONTROL OVER DIRECT AND DIFFUSED LIGHT
Tensegrity System as a Skin
When actuated, the system changes its form, where the components rearrange giving rise to openings at different angles by changing the orientation of the surfaces. This collective effect is studied to optimize the system according to requirements. The openings formed by the surface are manipulated in order to increase or decrease diffused light, depending on the local condition requirements for the system being installed. This manipulation is done by altering the angle of the opening. The above effect is also optimized for the incident solar radiation on the surface. The combination of multiple evaluation parameters add complexity to the system
System Development
Comparative Analysis between different options of 3 Strut & 4 Strut Tensegrity Systems 6 different options of 3 and 4 strut tensegrity components are developed based on location and orientation of surfaces, which are digitally analysed based on the evaluation parameter as explained in the methods chapter. For evaluation of the systems developed, the system is analysed in a digital set-up where it is deployed on a rectangular box anchored at 4 points. The box contains an inner exposure analysis surface. (bottom surface). The solar ray is perpendicular to the surface for all the analysis. The surfaces of the skin returns with colour codes chosen withing the algorithm showing the variation in angle of incident radiation which varies with the actuation.
Solar radiation
genes actuators 0.3 1.0 1.0
0.3
1.0
0.3
Min Angle of Incidence
WINTER
Max Angle of Incidence
SUMMER
Final model
PHYSICS ENGINE KANGAROO
EVOLUTIONARY SOLVER GALAPAGOS
Solar Exposure
Min Exposure
SUMMER
Max Exposure
WINTER
Evolutionary Algorithm for Analysis
Algorithm Development – Comparative Analysis
Each developed system is evaluated for various fitness criteria using the Genetic Algorithm. The type of actuators are taken as gene inputs, which are simulated by a physics engine based on the evaluations scripted by the exposure analysis and incident radiation analysis. The evolutionary solver iterates the maximum possibilities and finds out the most optimum state of the system according to the selected fitness criteria. This evaluation is carried out for every system, and outputs are compared and studied.The aim of these simulations is to find the system with maximum possible range of each parameter, within the constraints of their actuation range. The simulation outputs the actuation limits of each system as well. The evaluated results of the various developed systems for every fitness criteria are tested separately, which are compared and studied to find the efficiency of each system within that criterion. The system having the maximum range for a fitness criterion, which would in turn have the maximum flexibility, is chosen, as the system is required to optimize over extreme ranges of climatic conditions over the year.
Incident Radiation Analysis
Solar ray
Angle of Solar Radiation From 0° to 20° From 20° to 45° From 45° to 60° More than 60°
Amount of Solar Exposure Exposed to light Solar Exposure Analysis
Not exposed to light Analysis model Algorithm explaining the genome and fitness criteria
3 Strut tensegrity 4 Strut tensegrity
Conclusion
The evaluated results of the various developed systems for every fitness criteria are tested separately, which are compared and studied to find the efficiency of each system within that criteria. The system having the maximum range for a fitness criteria, which would in turn have the maximum flexibility, is chosen, as the system is required to optimize over extreme ranges of climatic conditions over the year. Maximum range Solar Exposure - 3 Strut Tensegrity System Option 3 Maximum range Angle of Incidence(AOI) - 4 Strut Tensegrity System Option 1
Photograph showing the interconnections between the struts and springs of the physical prototype (3 strut tensegrity System)
1
2
3
4
6
5
1 - 9: Fabrication process 10 - 15: Details of the prototype Above : Photograph of FINAL PASSIVE SYSTEM PROTOTYPE
7
8
9
10
11
12
Development Process of Passive System Prototype
Project Information:
Name: Form-Found Membrane Activated Gridshells Project type: Academic Research Location: Architectural Association, London
PROJECT ABSTRACT
This project explores the feasibility of using a jointed mesh system composed of cruciform-base units, which may rotate at their centres and about each other, in different configurations, to form gridshells of varying spatial arrangements. The curved surfaces of the gridshells are developed by attaching a tensile membrane to the mesh configuration, at a given eccentricity, thus inducing bending within the system. The mesh is configured into a particular pattern, before being actuated by the membrane. Once the pattern is set, the membrane is attached. The final forms have particular spatial qualities – area, curvature, height and internal lighting effects – which depend on the mesh pattern used.It was envisaged that to achieve a particular, desired spatial arrangement (in response to particular site parameters such as the flow of people or the increased need for shelter), the mesh could be configured as necessary while flat on the ground, and the membrane applied to generate the curved surface. If, at a later point in time, a different spatial arrangement were required, the gridshell could be dismantled, the mesh re-arranged and the membrane applied to generate the new form.
INTRODUCTION
This project deals with a material system, with very specific rules which govern it, that is used to create gridshell structures which enclose (or partially enclose) space. As shall become evident, the proposed system is responsive in that it can be manually dismantled, re-arranged and reerected, depending on the user’s spatial requirements of it. The response does not occur in real-time and is not actuated actively by any particular environmental factor. The research was focussed on developing the material system itself, and how, by discovering and manipulating the system’s particular properties, the spatial or environmental condition/s it could control, would become evident. The brief specifies that the project should involve “responsive surfaces capable of adapting to both their environment and their users”. It states that the material system should be able to produce “three distinct spatial effects throughout a 24-hour cycle”. It also requires that the “system is structurally effective and its differentiation will... intensify at least one environmental condition.
Form-Found Membrane Activated GridShells
Opening and Closing Mechanism
THE BASE UNIT
Early experimentation led to the development of a cruciform base unit, with a central hinge as shown, allowing the cruciform to open and close. It was originally intended that the four ends of the cruciform would be used to control the boundaries of a membrane (to allow for some form of dynamic control of porosity).
30mm
5mm MDF
120mm
DEVELOPING CURVATURE
By applying the tensile force representative of a membrane (the rubber bands in the prototype models shown) at an eccentricity, it was noted that bending is easily caused in the cruciform base. This is because the applied force at an eccentricity causes a moment to be induced, which in turn causes bending. The value of the moment is given by: M = F*e The radius of curvature R due to the induced bending is given by: R = (E*I) / M (This is the general theory of bending in elastic structures) .The terms E and I are properties of the cruciform-base section in bending and indicate the section’s intrinsic resistance to the bending. It is thus evident that curvature in the base unit may be controlled and is dependent on: Applied Tensile Force - F Eccentricity - e Cruciform Stiffness - EI
GLOBAL ARRANGEMENTS: PROTOTYPE 1
First attempts at developing a global arrangement yielded a prototype where base units were attached at opposite ends as shown. Applying a tensile force (the rubber bands) on one side formed a doubly-curved assembly. Varying the central angle of any cruciform in the assembly, prior to applying the eccentric tensile force, causes a change in the global curved geometry. It was therefore evident that further studies were necessary to identify possible mesh rotations and the resulting curved geometries. Digital models also indicate that the system is capable of reversing curvature, by changing the side where the eccentric tensile force is applied.
Cruciform-Base for Prototype Models
F
M
T
e C Inducing curvature by applying the tensile force at an eccentricity.
Increasing the eccentricity causing further curvature to be induced. The tensile force also causes a compressive force to be developed in cruciform base. 90O
FLAT PANEL
CURVED PANEL
RESULTING SURFACE 30O
PROTOTYPE 2
The development of the next prototype was aimed at dealing with the following issues (discussed earlier): - Membrane Enclosure - Joint Locking and Structural Stability - Offset between Gridshell Units It was also noted that the tensile force required to stretch the membrane could perhaps be used to induce curvature in the grid. Further work was also necessary on the mesh patterns and their respective, resulting geometries when curved.
RESISTANCE TO ELONGATION
CONTRACTION
MEMBRANE ENCLOSURE
Studies were carried out where separate membranes were joined to four central joints of separate units within a grid configuration. However, this setup was only stable while the joints were locked (tightened bolts). Once unlocked, the grid becomes a mechanism and the tensile force in the membrane causes rotations and the grid to close in the direction on the tensile force. A further experiment was conducted, where a larger piece of membrane was attached to all four corners of every unit arranged in a grid set up. It was noted that while the joints were unlocked, contrary to expectations, the membrane did not cause the system to close but in fact blocked the mechanism. In structural terms, the two-way tensile action of the membrane could perhaps provide the triangulation required to make the structure stable.
RESISTANCE TO ELONGATION
CONTRACTION
CURVATURE Membrane study where membrane is attached to grid at all unit corners. The contraction of the system in one direction requires an elongation in the traverse direction (as the system contracts). This is limited by the membrane’s pull.
Grid closing due to membrane tensile force, when joints unlocked.
Prototype model using membrane to actuate bending.
Previously, in the first prototype, curvature was achieved through the use of tensile cables (rubber bands). However, it was noted that it could also (probably) be achieved by using the tensile force due to the stretching of the membrane, itself. Prototype models revealed that this was the case. As can be seen when the membrane is attached, very close to the grid elements (therefore, at 0 eccentricity), no curvature is observed. As the membrane is moved away from the grid, the curvatuce induced increases.
Mesh Initial Position
DIGITAL MODELLING OF PROPOSED SYSTEM
For Prototype 2’s behaviour and qualities to be explored in depth, it was necessary to develop digital experiments quickly. Since the final form depends on bending against gravity, and is thus form-found, this meant that the digital experiments could not simply be geometric models, but needed to take into consideration the actual physical behaviour. A work-method needed to be developed to be able to quickly produce different digital mesh patterns which would subsequently be imported into a structural modeller. This work-method is represented in the flow chart opposite. The process is fairly long and complex, and to reach a single solution (let alone a range of solutions) requires many options to be considered. These include the unit size, footprint area and proportions (when flat), mesh pattern, plate thickness and material and tensile force in the membrane. In the larger digital meshes tested, this became especially difficult, as the effect of the self-weight increases drastically and enormous tensile forces in the membrane are required to induce the necessary counteracting bending. This perhaps suggests a material limit to the spans over which the system may be used, which would be scope for further investigation. The complexity of this process was a severe limit in producing feasible outcomes. Larger solutions with many nodes, covering larger spans, also need considerable amounts of computation time to produce a result.
Mesh Actuated Position
Mesh Angle: 90O - 90O
Family 1
40O - 40O
Family 2
STRAND TEST FAMILIES
A series of small scale (consisting only of 50 units) formfound families were developed to prove that the proposed working method could produce results for different mesh patterns, and to explore the resulting issues. At this stage, gravity was not introduced into the models as it was assumed that bending would not be affected at this small scale (1m when mesh is expanded completely). Patterns vary but the following used throughout: 180x180mm 50 Unit (Square Footprint) Mesh Arrangement Plates 10mm Mahogany Tensile Force 100N Eccentricity 75mm
90O - 20O
Family 3
20O - 50O
Family 4
APPROACHING VALID SOLUTIONS
Solution 1 Mesh Angles: 40O - 40O
Solution 2 Mesh Angles: 90O - 90O
Solution 1
Solution 3 Mesh Angles: 30O - 70O
Solution 2
Dis - Actuation
Re-Configuration
Since it was proved that large-scale digital prototypes could be modelled and produce satisfactory results on a particular scale (span of around 10m), the next step was to try to develop different solutions for the same mesh, in different configurations. A square mesh of 450 units was set into three distinct configurations (using the pattern mechanisms described earlier). These were then actuated by tensile forces of the same magnitude (representative of the different membranes of the same stiffness. It was thus possible to achieve three distinct solutions. The three final results occupy different footprints, are of different heights and produce different lighting effects (observed from within). 300x300x30mm Struc. Steel Units Eccentricity e=260mm Tensile Force F=10KN Rod Diameter = 30mm Bending stresses induced in the units of the mesh vary from one solution to another but remain within a reasonable range.
DYNAMIC STRUCTURE
Actuation
Solution 3
It was originally intended that the structure be actively dynamic in its responsiveness - responding directly to a stimulus, kinematically. The structural logic of Prototype 2, however, does not permit this. It allows, instead, for a different sort of “responsiveness� where the structure is capable of a range of different configurations (each one with particular spatial qualities) and the user must disassemble and reassemble the structure according to his requirements. The proposed assembly logic is as follows. The mesh pattern is configured and the structure is actuated. For a different configuration, the mesh is dis-actuated (by reducing to 0 the membrane eccentricity), the membrane removed and the mesh is reconfigured as necessary. Membrane is then reapplied to actuate the structure and obtain the new resulting form, with different spatial configurations. This process, admittedly is labour-intensive, and future work would be sure to explore how the proposed system could be made more responsive, while retaining the idea of forming the gridshell with a membrane (in tension) at an eccentricity. Active responsiveness could also (or alternatively) be explored in how the eccentricity is controlled.
TENSILE FORCE
TENSILE FORCE IN MEMBRANE CRUCIFORM BASE
MEMBRANE
HINGES
SPATIAL QUALITIES
The families were also used to evaluate spatial conditions for lighting, inside any space they may enclose, by using simple shadow renders, for different hours of the day. The different patterns of the external gridshell produce varying shadows on the underlying membrane, which are perceived from the inside. The footprint area of the families is also noted, since while the number, position and scale of the units in the grid remains the same, different patterns cause the same mesh to occupy different footprints.
TENSILE MEMBERS ADDED TO ATTAIN STRUCTURAL TRIANGULATION
Several individual frameworks are attached to each other to form the desired structure.
Assembly of frameworks
Triangulation of Structure using Tensile Membrane
PROTOTYPE ASSESSMENT AND CONCLUSIONS
FORM-FINDING PROCESS COMPUTE SOLUTION INPUT PARAMETERS: AREA+MESH SIZE HEIGHT OF CURVATURE REQ. SPATIAL QUALITIES
COMPUTATIONAL MODEL (GRASSHOPPER) TO GENERATE FLAT MESH TRY: NO. OF UNITS PLATE DIMENSIONS ECCENTRICITY ANGLE PATTERN
RESULTANT FORM
ASSESS CURVATURE: IS IT HIGH ENOUGH?
NO
INCREASE FORCE OR INCREASE ECCENTRICITY OR LOWER PLATE THICKNESS (INCREASES PLATE STRESS)
RE-COMPUTE
YES
INCREASE PLATE THICKNESS (DECREASE IN CURVATURE)
RE-COMPUTE
NO
CHANGE FORCE ECCENTRICITY ON DIFFERENT UNITS
YES IMPORT GEOMETRY INTO STRUCTURAL MODELLER (STRAND7)
APPLICATION OF FINITE ELEMENT MESH
ASSESS PLATE STRESSES: ARE THEY TOO HIGH?
NO
IS REQUIRED FORM DEVELOPED (PROGRAMMATIC REQUIREMENTS)?
ASSIGN: PLATE THICKNESS PLATE MATERIAL MEMBRANE TENSILE FORCE
YES
NO
ARE REQUIRED SPATIAL QUALITIES DEVELOPED?
YES
REPEAT PROCESS FOR DIFFERENT CONFIGURATION
CHANGE PATTERN
VALID SOLUTION
EXPORT MEMBER DATA
It appears feasible to generate gridshell structures by using the tensile forces intrinsic to stretched membranes, to induce the curved form. This is done by replacing the cables which apply an eccentric force on the cruciform bases, to generate a moment, in Prototype 1, with a membrane which serves exactly the same function. The forms that have been achieved so far are limited to sinclastic surfaces. The mesh, consisting of a grid-arrangement of cruciform units, can be manipulated into different patterns (which depend on the angle between the elements composing the cruciform bases) before the membrane is applied. Once the membrane is applied, different mesh patterns lead to different final outcomes (with their own particular spatial qualities). The offset between adjacent gridshell units (in Prototype 1) has been removed to ensure that force is transmitted directly between units. The membrane has been applied globally over all units in the mesh, rather than locally, to achieve complete enclosure. The digital work-method proposed allows for a range of mesh patterns to be modelled and then imported into a structural modeller, where actual force and behaviour is used to obtain final forms. This formfinding process has been used to develop a number of test prototypes at different scales. Larger spans imply greater required tensile forces in the membrane to counter gravity. The system may possess an intrinsic material limit to the spans it is capable of achieving without the tensile forces in the membrane becoming unrealistically large. Further tests, with base units of different dimensions would be required to explore this further. T Their spatial qualities are distinct. The exploration of further configurations was limited by the computationally-heavy nature of the work-method. The process of assembly has been described and is labourintensive. More work is necessary to explore how it can be improved. The tensile force within the membrane introduces a degree of triangulation to the otherwise flexible mesh mechanism. This triangulation, however is not a complete one of the whole structure. The introduction of further cables would be required to bring about the full triangulation of the structure. As it is, the structure relies on friction to ensure structural stability. The triangulation introduced by the membrane is sufficient to block the mechanism described as Mesh Mechnism 1. Mesh Mechanism 2 is not blocked but for it to become active, movement would be required across many joints (where it is resisted by friction). Given the large number of joints over which friction is developed, relying on friction does not seem out of the question. Further work is required to verify this.
Computational Design of Architecture and Urban Systems
This section of the portfolio illustrates the exploration of Spatial Qualities in architecture and urban systems. The research focuses on development of design through real time optimization and analysis using genetic and evolutionary algorithms.
Project Information:
Name: Supercity, Shenzhen Project type: Urban Development International Competition Location: Archi-Union, Shanghai
Project Description:
To design a Urban Patch with super economic functions in the Shenzhen Bay, Super Headquarters Base which will support industries, supplementing it with functions such as venues for international conferences, exhibitions and cultural and art programs. The main content for the competition is urban and architectural design which plans for the central area of the Super Headquarters. It is planned that 35.2 hectares of land will be used with a building area of 1.5 - 1.7 million square meters. The overall structure for the super headquarters will be one Cloud City Center, two featured streets and a few three-dimensional city groups. The Cloud City Center, as the core of the headquarters’ functions crucial to the area’s ecology and the intelligence and energy it aims to embody, is the priority and forefront of the construction. This notice solicits globally creative design plans with an international perspective based on its existing planning that includes three landmark towers, two land plots for cultural facilities and a central park.
SuperCity , Shenzhen
Urban Development International Competition
Solar Insolation in OPTIONS Summer (Wh/m2 ) day
Digital Tools Used: Rhino + Grasshopper(Geco Plugin) + Ecotect Multiple Configurations Based on Re-Arrangement of Towers TO DETERMINE THE OVERALL FORM
Strategy 1
One of the main approach to design this supercity was to create an energy efficient development. The first evaluation parameter chosen to this approach was to provide better solar insolation through appropriate arrangement of towers. To test the efficiency of arrangement of towers, Multiple Configurations of the arrangement of towers were created. Solar Insolation analysis were performed on various options to understand how each arrangement had their own influence. Rhino + Grasshopper + Geco + Ecotect were used evaluate the Solar insolation analysis. The analysis were considered for both Winters and Summers. Large variations were found in the performance of different options with respect to summers and winters. Therefore an optimum option for summers and winter were chosen.
Option 1
Option 2
Option 3
Option 4
OPTION 6 OPTION 8 OPTION 5 OPTION 7 OPTION 3 OPTION 4 OPTION 1 OPTION 2
651.06 690.7 694.89 724.06 762.32 844.1 861.02 862.87
Solar Insolation in OPTIONS Winter (Wh/m2 ) day OPTION 2 OPTION 1 OPTION 4 OPTION 3 OPTION 5 OPTION 6 OPTION 7 OPTION 8
625.17 622.95 611.8 591.38 533.3 529.81 501.09 482.5
Optimum configuration for solar insolation for summer & winter
Option 5
Option 6
Option 7
Option 8
OPTION 3 Spread-Out & Taller Towers Towards the Center
Solar Insolation Analysis summer & winter Multiple configurations based onfor re-arrangement of towers Multiple configurations based on re-arrangement of towers To determine the overall form
BUILDING 2
BUILDING 1
BUILDING 3
Basic Master plan with Boundary Limits of 3 towers(Left) Projection of rays of visual sight from the South Facade towards the Sea(Right)
Strategy 2
As the site was located on the shores of Shenzhen Bay, another criteria was chosen to develop the design was to provide maximum view for the building occupants towards the South facing the sea by reducing visual obstruction by other towers. To maximize the view appropriate orientation of the three towers, a genetic algorithm was created to reduce the cumbersome process of creating multiple options through various permutations and combinations of the location of three towers within the 3 boundary limits.
Digital Tools Used: Rhino + Grasshopper(Sonic Plugin) + Galapagos
Position of Towers: B1- 2, B2 - 8, B3 - 8 Least Visual Obstruction: 9%
Multiple Options displaying the projection of rays of visual sight from the South Facade towards the Sea
Form: Top: -30deg Bottom: 30deg Visual Obstruction Percentage 16.5%
Strategy 3
Another strategy chosen to provide maximum view for the building occupants of the rear tower on the South facade by changing the form of the towers. The form of the tower was manipulated by controlling the angles of facade.
Digital Tools Used: Rhino + Grasshopper(Sonic Plugin) + Galapagos
Form: Top: -30deg Bottom: -20deg Percentage 10%
Form: Top: -30deg Bottom: -20deg Percentage 10%
Form: Top: 15deg Bottom: 0deg Percentage 26.5%
Form: Top: 20deg Bottom: 30deg Percentage 22.5%
Project Information:
Name: DaChan Industrial Development, Shekou Project type: Architectural Development International Competition Location: Archi-Union, Shanghai
Project Description:
To provide upgrading and redevelopment design for the existing DaChan buildings, from the perspective of sustainable development. Develop the area into a Creative Cultural Park based on the existing spatial carrier of DaChan. Support the industrial development of the area and the creation of an industry-university-research base by activatin the site for low-intensity and low-density development, and provide featured places in consideration of the redevelopment/utilization of the existing buildings. The most challenging part of the project was to redesign the silo by retaining the existing building spaces and characteristics, create dynamic spaces, and guide the subsequent implementation process.
DaChan Industrial Development, Shekou Architectural Development International Competition
Exploring the possibilties of Combined Spaces by Destruction of Interior Walls Combination of Silos
Destruction of interior wall
Length
Old/New Area
Area Achieved
Silo Area - 77m2
0m2
20.49m
Silo Area - 154m2 Total Area - 190m2
36m2
32.60m
Silo Area - 154m2 Total Area - 207m2
53m2
32.60m
Silo Area - 154m2 Total Area - 225m2
71m2
15.55m
Silo Area - 154m2 Total Area - 190m2
36m2
31.10m
Silo Area - 308m2 Total Area - 345m2
37m2
0m
Strategy 1
The first approach in restoring the building was to understand the spaces through various analysis as the silos have huge limitation due to their shape and size, some amount of reconstruction were found to be inevitable. Hence Area analysis with combination of multiple silos spaces through minimum destruction of interior walls was initial strategy chosen in restoration of the building. Most efficient choice of spaces were found to be optimum selection between options with minimum destruction of walls and maximum reclamation of single larger space.
Final Spaces
Exploring the possibilities Exhibition Spaces
Interior spaces are found to be narrow and obstructed by supporting walls which make it inefficient for exhibition or gallery spaces
Silos provide the most unobstructed space and are located towards the periphery allowing possibilities for natural lighting and ventilation
Strategy 2
Interesting flow in the movement of visitors was an important requirement to be considered for a cultural activity with exhibition or gallery spaces. The movement between the silos were found Option2 - L2 to be difficult as the supporting structural walls which could not be altered, hence the movement between the silos connected by bridges were found to be better alternate solution.
Option1 - L1
Option1 - L2
Option2 - L1
Option3 - L1
Option3 - L2
Option4 - L1
Option4 - L2
Option5 - L1
Option5 - L2
Option6 - L1
Option6 - L2
Importance of Bridges
1. To separate the exhibits between each other 2. To experience of transition space 3. To connect silos on either side using the dead spaces with minimum floor area
Possibilities of connection between the silos
Selected Pattern Option for Exhibition Spaces B
B’ B
B’
A
A’ A
A’
Option7 - L1
Connections possible with minimum damage to structure of the silos with the distance being considerable
Option7 - L2
Selected Option allows considerable possibility of connection between silos by using spaces alternate opposite silos of the building.
Type of Spaces in Museums Spaces
Connections possible but the central wall between the silos causes split in the width of the ramp
Gallery
Permanent Exhibits
Exhibition
Temporary Exhibits
Exploration of vertical access through Levels Option 7a - Gallery(Permanent exhibits) Connections possible with minimum damage to structure of the silos but with no possibility of transition space
Section AA’
Spaces are developed on 2 different levels(+0.0m & 2.5m) which facilitate for gallery(where visitors enjoy more time to view the exhibits)
Section BB’
Option 7b - Exhibition(Temporary Exhibits) Spaces are developed on Multiple levels(+0.0m & 2.5m). This pattern facilitate visitors to forcibly move continuously on an opening night for exhibitions.
Connections not possible due to the presence of the supporting wall Elevation AA’
Typical Floor: Lower & Upper Level Combined
Exhibition Spaces, Bridges & Ramps
Upper Level @ +2.5m
Lower Level @ +0.0m
Both Levels - Plan (Overlapping)
Both Levels - Perspective View Both Levels - Exploded Perspective View
Cafe/ Observatory space
Exhibition space Museum reception
Ramps Exhibition Spaces Bridges
Master Sectional elevation
5m 5m
Museum
Auditorium
Library Lecture Halls Office
Rock Climbing
5m Lift Core
Lift Core
Lift Core
10m Height Atrium
5m 40m
Lift lobby
5m
Water Diving
5m
curator room ClUb reception
Gaming rooms
Museum Entrance/ reception
5m 5m
Exhibition Club Service
The spaces inside the silos are interconnected through bridges to reduce damage to the structure and supporting walls. Alternate crisscross movement with mezzanine floor with double height reduced the damage and created a smooth flow withing the structure.
Type 1 Cultural Museums Galleries Exhibition Radial Arrangement Grid Arrangement
Radial & Grid
Radial Arrangement Grid Arrangement
Radial & Grid
Type 2 Educational Libraries
Strategy 3
To understand the potential of silo shape and spaces numerous functions through multiple configurations were tested. The types of functions were mainly divided into cultural commercial educational and recreational. Each function was tested with various criteria which best suited the silo which would preserve the architectural and retain the maximum characteristics of the building.
Type 3 Educational Classroom, Lecture halls Studios, Workshops Type 4 Commercial Offices Conference / Meeting Rooms Type 5 Commercial Shopping Supermarkets
Type 6 Recreational Auditorium Theatre Arenas(Concert halls)
Lecture Hall
Classroom
Live Art Studios
Office - Workstations Office - Meeting Office - Conference
Shopping - Radial
Shopping - Grid
Supermarket
Audi - Radial
Audi - Grid
Concert Hall
Radial
Grid
Martial Arts
Boxing
Type 7 Recreational Cafeterias Restaurants Food courts Type 8 Recreational Martial Arts Boxing SPA/ GYM Gym
Programme
Architectural Value based on Circulation
Cultural
Intergration
Modification Requirements
Design
Possibility
Cultural Museums Galleries Exhibition Centers Educational Libraries Classroom, Lecture halls Workshops Commercial Offices Conference / Meeting Rooms Shops(Arcade) Shopping Mall Supermarkets Recreational Auditorium Theatre Arenas(Concert halls) Bars Cafeterias Restaurants Food courts Casino SPA/ GYM Water Diving Indoor Wall Climbing Martial Arts, Boxing
NOTE
LEGEND
-
-
-
-
-
-
-
-
Some programme have strict standards which make the design difficult to experience the existing space LOW MED HIGH
Possible - May Be Not possible
LOW HIGH
Possible - May Be Not possible
Design Parameters
Typology
Display & Viewing Area Display & Viewing Area Display & Viewing Area
Type 1 Type 1 Type 1
Display & Access Area Presentation & Seating Area Presentation & Seating Area
Type 2 Type 3 Type 3
Storage, Selling Area Display & Access Area
Type 4 Type 5
Presentation & Seating Area Seating Area Seating Area Seating Area Seating Area Spaces Dependent on typology Spaces Dependent on typology Spaces Dependent on typology Spaces Dependent on typology Spaces Dependent on typology
Type 6 Type 3 Type 7 Type 7 Type 7
Unique Types
Project Information:
Name: Interconnected Urban Spaces Project type: Academic Research Location: Architectural Association, London
ABSTRACT
The brief of the project indicated a research experiment on a patch of the city of London, studying the designed system thought out all the levels of integration. The preferential location of the site indicated for our system a determined social group, that of floating people. Those parameters, connectivity and integration, orientated the two parallel formation strategies, aiming on reorganising and reconnecting the urban tissue and producing at the same time, multiple integration spaces for the new habitants. Morphological exploration of the system’s materialization led to an interlinked network of green spaces and urban blocks, combined with amenities distribution, producing differentiated spatial configurations.
Interconnected Urban Spaces
Urban Development Research
Concept
The brief of the project was to develop a patch of the city, of London, placed on the Isle of Dogs, through research on its multiple scales. Taking advantage of the site’s connectivity to the rest of the city, as analysed on previous chapter, we chose a specific social group to be placed as new residents. Investigating on the floating population combined with the topology of the chosen site, a methodology of rules was developed in order to facilitate their needs through programmatic uses and spatial organisations. A new connectivity network is organised, introducing public green spaces in relation to the water borders encountered within the patch (docs + Thames waterfront). The whole new network is based on the proximity to the transit places (indicated as primary nodes), organising at the same time routes between those and the green spaces (secondary nodes). Moving away from the conventional approach of converting those connecting routes to streets, the block structures constitute the connection lines, combined with pedestrian paths. The new structures of buildings represent the whole network in a diagrammatic, operational as well as morphological level and variable densities are dispersed among them. At the scale of the blocks, buildings and “cells”, an integration network is organised, aiming on encouraging if not imposing the integration between constantly variable residents, whose identity is multidimensional.
METHODOLOGY
Outlining the methodology that was taken, the research in combination with the design process, focuses on two, parallel routes. First are the strategy created for the connectivity network and the hierarchies fabricated within it. Having as a starting point the existing transit/transport network, a set of rules lies beneath the distribution of open green spaces as well as the connecting tissue between the two of them. The integration network is organised at the same time. The typologies of the “cells” and the buildings are defined and organised to clusters, aggregating a variety of shared spaces, varying from semi-private to open public spaces. The overall performance of the system is evaluated through the criteria of proximity and integration, primary needs for the specified users. Additionally, the system is calibrated through the use of environmentally defined criteria (wind and light performance), optimising the performance of its spatial organisations.
Connections
Placement of the nodes for the connectivity network is followed by the connections and how could the variable and numerous probabilities can be evaluated and straiten through a new set of rules. Proximity again plays a determinant role, removing those connections whose length is bigger that 800 meters, meaning more than 10 minutes’ walk which is considered acceptable walking distance for everyday use. The coherence of the system is ensured, and all the parts of the site are interconnected optimally not only towards the transit points, but also in-between the open public spaces. Then another set of rules was needed to calibrate the connections of the system with the existing morphology, firstly eliminating those which intersect with the dock’s outline and then if intersection occurs outside a node, randomly erase one of the intersecting connections. The randomness that was introduced at this stage, allowed a variety of acceptable solutions for the site, each of which meets the requirements that were set by the evaluation criteria. Each of the nodes has different weight and facilitates different functions; therefore, the range of influence, meaning the abstracted footprint of the space needed for these functions, varies between primary and secondary nodes. Influenced by those ranges, the length of each connection is redefined, practically reduced, according to its vicinities. Having in mind that each connection has differently defined starting and ending point, an inevitable hierarchy was created between them, classifying them to three categories: primary-to-primary connection, primaryto-secondary connection and secondary-to-secondary connection.
OPTION 1 LOW DENSITY
Density Distribution
The connectivity of the site as its main characteristic oriented us towards a transit or floating population. Therefore, a research was made towards the cities with most transient population in the world, in order to comprehend and define the aiming density for the given patch of London in comparison with its covering area. Surprising enough, the given area has one of the highest densities in London, although the whole city has relatively small number, compared to New York. The first experiment is based on reaching that specific density, within the 1km2 we are examining, which will not be homogeneous in its appearance, and evaluate the blocks’ heights achieved. Based on that, we could evaluate the system and how would it adapt/encounter a population growth that would change the starting density, in the future or in a second experiment. The overall density of 25.000 people / km2 is divided to three density ranges: low (10.000 people / km2) , medium (15.000 people / km2) and high (27.000 people / km2). The ranges are distributed within the connections, reaching the overall population of 25.000 people. The different densities have different morphological outcome on the system, as indicated on the diagrammatic studies that follow, since the additional residential demand pushes the system for more vertical growth. Connections with length smaller than 400 meters accommodate only low density, whereas longer ones, according to their peaks take medium and high densities. In that case, proximity to primary nodes implies high density, whether proximity to secondary node implies medium.
OPTION 2 MEDIUM DENSITY
OPTION 3 HIGH DENSITY
OPTION 4 MEDIUM-HIGH DENSITY
OPTION 5
HOUSING
The intention is to establish different hierarchies of interaction from the level of the housing to the level of the to the common spaces and ultimately public spaces. The proposed housing solution will be divided in 3 basic categories, Private Spaces, Semi-Private Spaces and Public spaces each with a range of squared meters required depending upon the standards. Additional to the standards, there are two main aspects to be considered respective to this particular issue. One is the duration of stay, established from the previous studies the categories: Short term until three months, medium term from three to twelve months, and Long stay from twelve to 48 months. The second factor is the number of people sharing the private Units, this is Single, Couple, and Families (from three to five members). In conclusion, the dimensions of the Private Units will be in a range depending upon whether it is for Singles, Couples or Families, and their respective Semi-Private Units will depend on the duration of the stay of each, smaller for short term and bigger for Long term. This Semi-Private Units can be shared between various categories of inhabitants, establishing a first level of interaction. This interaction is termed as intra-cluster interaction.
AMENITIES
Other levels of interaction will be established at different scales, from the cluster scale, block scale to the superblock scale. The amenities are strategically located considering easy accessibility and also to let them act as an additional level of interaction. The nodes will be spaces for the connection of all the network and act as open spaces to entertain commercial activities to justify their locations and easy accessibility.
TYPOLOGIES/CLUSTERS
The organization of the Private and Semi – private units forms a continuous feedback loop process. The basic stacking of the clusters was informed by the new layers of interactive spaces and cut-outs to facilitate better lighting conditions into the interiors with also achieving porosity. As illustrated in the diagrams, three rows of 6m are placed along a middle axis in order to fit the different typologies for both the Private and Semi-private Units. The Private Units are located at the rows on the outside, one at each side of the middle axis; meanwhile their respective Semi-private Units are placed in the row in the middle. In order to allow the light in and ventilation into the interiors, the Semi-private Units in the middle row are offset from the central reference axis. By doing this the exposed area can increase in a 50% if the cluster is isolated. Some rules for the offsetting are established to make sure that the Private Units are physically connected to their respective Semi-private Units. One is that an overlap of at least 1.50 meters between the Units is required to facilitate the transition from them. Also, as shown in the diagram, if the Semi-private Unit with less squared meters is located close to the central reference axis, this allows for more units to be connected and grow.
SECTIONAL EVALUATION
The sections are further evaluated towards its limitations. As the clusters are designed to form buildings with varying density, the exercise demanded stacking of such clusters over each other to achieve the respective densities. Starting with low density blocks, 7 floors with the intermediate floor for interaction on 4th floor is easily accessible from both the lower as well as the higher floors. The same concept is later applied to the medium and high rise density blocks, i.e. for 12 floors, 16 floors and 20 floors. It was impractical to go vertical than 20 floors as in that case, the intermediate floor for interaction will not be easily accessible from the extremes of the buildings, thus the entire idea of building such spaces will be lost. This helped in defining the height limitations for the buildings proposed and thus the ranges for densities are defined The buildings of height in range of 7th to 11th are low rise, from 12th to 15th are medium rise and from 16th to 20th storeys fall in high density. Using these heights as guidelines for calculating the population per building for respective densities, the ranges are distributed accordingly to meet the desired density.
CONCLUSION
Synopsizing the project, our aim for designing a well articulated and integrated network was achieved by transforming the connections to spaces and through them link the desired nodes - attractor points. At the same time the system is capable of adapting to density growths, using the proposed strategies, occupying more of the open spaceof the area. Nevertheless there could have been further morphological exploration of the produced system, in coherence with enviromental conditions and performances as well as reevaluation of produced spaces and their qualities, on the bigger scale of the open urban spaces.