Foster + Partners | Graduate Show 2019 by Alejandro Guerrero Neira

Graduate Show 2019

Contents

4 Participants 5 Foreword Architecture 9 49 77 105 151 179

Studio Studio Studio Studio Studio Studio

1 2 3 4 5 6

(S1) (S2) (S3) (S4) (S5) (S6)

Specialist Teams 215 231 239 249 261

Environmental Engineering (EE) Structural Engineering (SE) Communications (COM) Modelshop (MS) Urban Design (UD)

268

Organisers and Volunteers

Participants

Aikaterini Lasithiotaki Alessandro Bonfiglio AndrĂŠs GonzĂĄlez Molino Christos Chatzakis Cristina Manta Larisa Tsydenova Megan Makinson Ting Yi Chan Yiki Liong

Aanisah Chowdhury Alejandro Guerrero Neira Chen Shen Josh Corfield Michael Silvestri Wioletta Sarara

Alex Bramhill Ignacio Diaz-Maroto Rivas Konstantinos Chatzimanolis Natalie Cheung Stefan Necula Beatrice Melli

Adrian Siu Andrew Yuen Callum Rowland Camille Dunlop Eugene Jahng Gabriel Pavlides Kasia Gryszkiewicz Kevin Herhusky Pablo Valero Escolano Ruoming Song Will Chen

Jingyi Chen Joey Aoun Margaux Coquelin

Stavros Tseranidis

COM

Celia Shiels Hannah Gasparutti

Alessandro Magliani Chen-Yong Tan Chris Cheung Dominic Oliver Elizabeth Morgan Lang Jin

Felix Burke Paulina Grenda

Emilie Li

Christian Kennedy Daniela Andrade Salgado Silva Duncan Catterall Fadhil Fadhil Heidi Han Ola Urbanska Sean Bailey Yanni Hajigeorgis

Foreword

As a young student, I had a total disregard for rituals, which I thought were outdated and old-fashioned practices. Perhaps there is some truth in that, however, there is one ritual at Foster + Partners that I am increasingly attached to, and that is the annual Graduate Show. Everyone – new architects, visual artists, urban and industrial designers, structural and environmental engineers – all are invited to present their recent student work to peers and senior colleagues in a spirit of sharing, enquiry and learning. As a practice, we are extremely lucky to welcome such talented individuals who want to join our team here. Last year we had 31,000 applications, of which 282 joined the practice, from 22 institutes spread over seven different countries. And interestingly for an architectural practice, less then half of them were architects! This year, the quality of the work was as high as ever. From masterplans to dwellings, the project explored a variety of scales. Materiality was a particularly popular theme, as was sustainability – with investigations into biomaterials, water scarcity, coastline erosion and rising sea levels. Inspiration was sought in the mundane – fire safety codes – as well as the poetic – cybernetic theory. The projects spanned the length and breadth of the globe, reflecting the multicultural nature of our team.

As always, the show was organised for the graduates and by the graduates – so everyone is to be congratulated for an inspiring event. Particular praise for helping with the organisation goes to Alejandro Guerrero Neira, Alessandro Magliani, Dominic Oliver, Gabriel Pavlides, Jingyi Chen, Joey Aoun, Josh Corfield, Margaux Coquelin, Stefan Necula and Yanni Hajigeorgis. I must also pay tribute to a number of my colleagues who work incredibly hard behind the scenes to make this event happen every year – Narinder Sagoo as sponsor and overall Project Director; Charlotte Sword as Project Manager; Charles Sabberton, Chloe Bingham as Project Co-ordinator; Mike Bass & his team in Document Control who worked tirelessly printing everything with a special mention to Tony Wenban; Alicia Cox & her team for doing a great job with the food, drink and hospitality; David Gilliard who set up the IT; The Warehouse Team for moving everything around, with a special mention to Craig Edwards; then Rupert Goddard for organising the graphics; Sam Strudwick for putting together this great catalogue; and Steve Teare, Aaron Hargreaves and Nigel Young for capturing it all on camera. Congratulations to all – I look forward to next year’s event with great anticipation!

Studio 1

Aikaterini Lasithiotaki Alessandro Bonfiglio AndrĂŠs GonzĂĄlez Molino Christos Chatzakis Cristina Manta Larisa Tsydenova Megan Makinson Ting-Yi Chan Yiki Liong

The Analogical Journey within the Layered City Aikaterini Lasithiotaki

Course: Master in Architecture University: De Montfort University Country: United Kingdom Current Project: Bloomberg Year Start: 2018

This project suggests an allegorical and literal journey within the layers of Heraklion city. It reintroduces the cityâ&#x20AC;&#x2122;s urban fragments as sustainable spaces that respond to its environmental needs and makes them active elements of the future city. The proposal emphasizes on the cityâ&#x20AC;&#x2122;s lost authenticity of space and its problematic public areas, while it stresses the boundaries between the different uses of the public spaces by generating a network of spaces that respond to the current social and cultural needs of the city. The theatricality which is found both in everyday life and in cultural and spiritual processions that currently happen in the city, is reflected in a sequence of cross programming public spaces, such as inhabitable walls, theatrical stages, urban vineyards and small interventions on existing buildings, which stimulates a flow of people who escape from the qualities of the busy city and allow themselves to experience the fluidity of space and use. An attempt to orient the individuals in time and space through alternations of collective moments and isolation, whilst they engage to the eternal journey of life and death.

1 12

2 Graduate Show 2019

3 14

1. The main public square. 2. Fragments of the city, concept explorations. 3. Cross programming spaces in section and 3D. 4. Carnival procession, entrance to the public courtyard. 5. Physical model studies.

5 Graduate Show 2019

Kaunas Concert Centre Alessandro Bonfiglio

Course: Laurea Magistrale in Architettura University: Politecnico di Genova Country: Italy Current Project: Haymarket Edinburgh Year Start: 2018

The project is intended to re-orientate the city towards the riverside, the new building acts as a catalyst to revitalise the cityâ&#x20AC;&#x2122;s south bank, creating the impetus for further riverside development. The Hall was therefore imagined as a lantern that states its presence and its vitality to the historical centre with the purpose of engaging people to explore the other side of the river. The complex is composed by four different volumes linked together underneath by an infrastructural platform where all the back of house activities take place.. On top of the platform it is provided a public square that offers the public an outdoor performance space, protected on the edges by the volumes, two of them host the main and secondary hall, and are provided with a hatched glass faĂ§ade, the other two containing services related to these spaces are conceived as carved out monolithic blocks of concrete. The contrast between the sculptural shape of the main hall and the austere, rigid and semitransparent cube containing it, creates an interesting in-between space that distributes the hall on all sides. The development also includes a new public park, a riverside pedestrian and cycle route, and a dock for tourist vessels arriving by river. 16

1. Contex model picture. 2. Building model picture. 3. Section model picture. 4. Axonometric exploded. 5. V iew from the other side of the river. 6. View of the public square.

6 Graduate Show 2019

PRODUCED BY AN AUTODESK EDUCATIONAL PRODUCT

32 16 44

20 45

PRODUCED BY AN AUTODESK EDUCATIONAL PRODUCT

8 18

7. Detail section of a tribune. 8. Concert tall interior view. 9. General plan. 10. Section of the main hall.

PRODUCED BY AN AUTODESK EDUCATIONAL PRODUCT

Graduate Show 2019

10 19

Infinite Voxels AndrĂŠs GonzĂĄlez Molino

Course: MArch Architectural Design University: University College London Country: United Kingdom Current Project: China Merchants HQ Year Start: 2018 1

Intile is a project that uses additive manufacturing, such as plastic extrusion, to materialise designs based on discrete design principles. This might seem antithetical, but even though the design space is constrained to a discrete set, it is still large enough that fabrication methods used for serialisation are less efficient and cost-effective. The spatial extrusion fabrication method allows architect to create lighter and more efficient forms without any material waste. This process by itself does not allowed for the creation of inhabitable spaces as it cannot fulfil all the necessary conditions, such as insolation, waterproofing, acoustic protection and so on. Infinite Voxels uses a hybrid approach that combines the advantages of spatial extrusion with other materials and methods, such as machines polysteryne. We can create a catalogue of blocks that avoids the habitability problem commonly found in digital architecture. The rise of 3D printing has delivered the concept of mass customisation instead of mass production. We propose an alternative way to achieve mass/customization by allowing a large, but not infinite set of possible design variations.

4 20

1. Architectural approach. 2. Fabrication _ printing process. 3. Geometry test. 4. Fabrication _ prototypes. 5. Architectural speculation _ domino approach.

5 Graduate Show 2019

6 22

6. Architectural speculation. 7. Computation_ infinite voxels aggregation.

7 Graduate Show 2019

Rawbot: Discrete Raw Materials and AR Assembly Christos Chatzakis

Project Rawbot is rethinking the construction process and questions the role of automation in architecture. In a day when automated construction technologies are focusing on robots, Rawbot develops an Augmented Reality smartphone app that utilizes people and brings automation to rural places in the most affordable way, without the use of any specialized equipement or infrastucture. The system uses raw, untreated bamboo poles to deliver cheap, sustainable and quickly assembled housing solutions that entirely skip the material standardization process. This becomes possible through the development of an adjustable and reversable joint for bamboo, designed to be used by untrained people. Course: MArch Architectural Design University: University College London Country: United Kingdom Current Project: Battesrsea Power Station Year Start: 2018

1 24

Cone

Clip Core

2 Graduate Show 2019

3 26

1. Architectural speculation perspective. 2. Computational logic, joint fabrication and data collection to architectural piece. 3. Architectural speculation view. 4. Machine learning for the detection of bamboo points of interest and Augmented Reality (AR) smartphone app for the assembly. 5. Proof of concept â&#x20AC;&#x201C; assembled model.

5 Graduate Show 2019

‘Industrial Metamorphosis’: An Automobile Factory Typology Cristina Manta

Course: MArch Architecture University: The Bartlett School of Architecture Country: United Kingdom Current Project: Principal Place Residential Year Start: 2018

‘Industrial Metamorphosis’ proposes the development of a new automobile factory typology in Zil, Moscow, Russia. The programme expands on Moscow’s industrial heritage and speculates on the potential rebirth of the Zil Automobile Manufacturing Site, while responding the the larger issues of masterplanning and industrial reuse currently present in the capital. The project proposes a model of changeable architecture based on the dynamic flow of self-organising or emergent systems, from micro, internal scenarios to macro, urban ones. This model is used to generate ambiguous ‘soft’ spaces, constantly reorganising themselves according to the users’ participation and their interaction with the environment, according to local rules and macro behaviours. Given their capacity to create novel structures with dynamic experiential implications, the principles of emergent systems are applied throughout the design process- on conceptual, formal and programmatic levels.

1 28

2 Graduate Show 2019

3 30

1. Zil Industrial Park Aerial View.

1918 Housing and Industrial Diagram

2. Flow Field Exploration and Structural Intergration. Compression

3. Zil Interior View and Section.

Expansion

4. Industrial Pods. 5. Physical Model.

2018 Housing and Industrial Diagram

Compression

Expansion

5 Graduate Show 2019

Hudson Yards Delirium Larisa Tsydenova

ASSEMBLY OF THE URBAN STACK

Course: MArch RIBA ARB Part II University: University of Westminster Country: United Kingdom Current Project: China Merchants Bank HQ Year Start: 2018 1

Manhattan is a felicity of skyscrapers, large building blocks, the Grid with endless avenues, with a boundary ring made out of highways. It decorates the city skyline, conferring it’s such an iconic look that is familiar to everyone, also represents hegemony of power and money. What is the American Dream like and how such Delirium can be referred not only to the city, but also the rest of the global world? Despite the secrecy of capitalism, there is a strong exhibitionist nature. The ‘Delirium’ is a socio-economic entrepreneurial Expo Multiverse at Hudson Yards accompanied with a transportation terminal and acts as a hub for cultural escape in the city. The Buyer’s Club - conference halls with lounge areas and in-between activities - is juxtaposed with social outputs (recreation grounds and public terraces) spread across the Stackers; accompanied by the Oculus, separating the overall mass vertically in addition to horizontal terrace domain, where curves act up for the human interaction. As a Social Condenser, the Delirium celebrates an open exposition of cultural events where business emporium and public domains coevolve, colour and deployment of patterns and apertures transforms the city block in New York into a new public metropolitan domain.

4 32

1. Stackersâ&#x20AC;&#x2122; assembly. 2. Poche model (1:250, 2m*0.6m*0.75m) representing the Groundscraper. 3. Highly articulated Shell. 4. Within context. 5. Architectural cake stack - Social Condenser conceptual representation.

5 Graduate Show 2019

6. ‘The Buyer’s club’ Graduate Show 2019

Inst. Architectural Relics Megan Makinson

Course: BSc Architecture University: University College London Country: United Kingdom Current Project: Miami Beach House Year Start: 2018

Situated in the Chicago Loop, The Institute is designed as both an Archive and Construction Trade School. Chicago has a capitalist view on preservation, where a building can only be saved from demolishment if both historical significance is proved and a new economic ‘creative re-use’ can be found. This created a tradition of Architectural salvaging, starting in the 60s by Richard Nikkel, where large fragments of demolished architecture were saved. The Institute, following this tradition, ‘creatively re-uses’ fragments of lost Chicago architecture, moved and placed in the Institute. These fragments are curated to be didactic models for the students whilst being an imaginative way to inspire the importance of new preservation in still standing architecture in Chicago.

2 36

3 Graduate Show 2019

4 38

1. Initial Facade Test.

5. Evening Street view render.

2. The Facade Library.

6. Project Two â&#x20AC;&#x201C; A Boarding House for the Conditioning of Method Actors. Case study records by The Matron.

3. Section. 4. Internal Render â&#x20AC;&#x201C; the Workers cottage wood workshop.

6 Graduate Show 2019

At Your Own Risk Ting-Yi Chan

Course: Architecture University: Architectural Association Country: United Kingdom Current Project: Dachong Tower 7 Year Start: 2018

The project looks social problems that revolves around the upbringing of Hong Kong children. With endless after school tutorial classes (even in the weekends) Hong Kong children in average gets less free time than the city’s prisoners. Another contributing factor would be the lack of rule-free public spaces in Hong Kong. A series of “Playeatres” inhabit the city, hijacking and activating the everyday architecture that Hong Kong children walk pass in their tight daily schedules. The project encourages engagement and interaction, utilising Augusto Boal’s Forum Theatre where the audience (aka. The children) participates in the acting of the play with the actors (aka. The Playeatres). Researching on psychologist James Gibson’s work on the “affordances” of objects, the designs explore the idea on how one thing as banal as a steel frame could be “afforded” in the imaginations of children as infinite usages. Through creating a list of parasitic elements that created these “Playeatres”, there is no need to design for playgrounds anymore but public spaces.

Graduate Show 2019

1. The Big Softie. 2. The Competition. 3. The Drop. 4. The Enclave. 5. The Portals.

The project is a world building process that wants to remind curators and architects to be aware of the fact that the city is not a well-controlled white cube, but a place dominated by unchecked, multifarious differences that encourages usage, activity and staging, giving the place of Hong Kong a specific conception of space.

Graduate Show 2019

Dacha Micro-Cities: An Alternative Mode of Russian Living Yiki Liong

Course: Architecture MArch (ARB/RIBA Part 2) University: The Bartlett School of Architecture, UCL Country: United Kingdom Current Project: Principal Place Residential Year Start: 2018

The project explores the thresholds between what is man-made and what is nature. The immediate appearance of fractured landscapes are constructed from micro-voxels; the medium used to explore the interplay between nature and synthetics. The exurbs of Russian cities are characterised by Dachas, a housing typology in which residents have a close relationship with nature. In contrast, the city centre of Moscow is plagued with low quality Kruschevka towers. In light of the different characteristics of each typology, the concept looks to create an alternative micro-city which balances the necessities of protection from harsh climates, while maintaining fluid public spaces. By implementing passive energy systems at a master-plan scale, the effect of extreme weather conditions may be mitigated within the city to provide thermally stable, habitable environments. Thermal comfort specifications and affiliated buffer-zoning has been employed to derive suitable divisions between public and private spaces. The arrangements determined from investigations provides the basis for the development of highly textured, habitable micro-city networks.

1 44

2 Graduate Show 2019

3 46

1. Masterplan fragment: cluster aerial perspective. 2. The final masterplan. 3. Masterplan fragment: final model. 4. Micro-city cellular clusters: aerial perspective. 5. Conceptual micro-voxel digital exploration.

5 Graduate Show 2019

Studio 2

Aanisah Chowdhury Alejandro Guerrero Neira Chen Shen Josh Corfield Michael Silvestri Wioletta Sarara

Healing + Renewal With Clay Aanisah Chowdhury

In my thesis projects I explored the themes of memory, loss, rebirth and renewal through the medium of natural clay. The first project responds to the villages lost during severe flooding in Lincolnshire, specifically focusing on an individualâ&#x20AC;&#x2122;s mourning process of losing his wife. Using clay excavated on the site, a programme for ceramics workshops and sculpture celebrates lives lost for public use. When exploring rebirth and renewal in the second project, I designed 3D clay printing factory that could help revamp Grimsbyâ&#x20AC;&#x2122;s neglected appearance. The factory collects and prepares clay from the back of the site to make 3D clay printed work. The ceramics created are sold or displayed after being fired and glazed in several different ways throughout the building. In the central theatre, the visitors can watch the large 3D printer make large scale ceramics that will be used as public artwork.

Course: BArch Architecture University: University of Nottingham Country: United Kingdom Current Project: Beverly Hills Masterplan Year Start: 2018

1 52

6 Graduate Show 2019

PIT FIRING PREP

3D LARGE PRITNING LARGE GLAZING

CLAY STORAGE

DRY ROOM

FREE FORM STRUNG CLAY PRINTING

CLAY EXTRUDER

3D ROBOTIC MACHINES

SORTING

WASHROOM

WOOD FIRING KILN

DIP GLAZING + GLAZING LABS CLAY EXTRUDER + 3D ROBOTIC MACHINES

AIR/ SMOKE DY RIN

FREE FORM STRUNG CLAY PRINTING

STORAGE

DRY ROOM

KILN

WASHROOM

ROOF EXHIBITON SPACE GALLERY CLAY STORAGE

CAFE

RETAIL + THEATRE

CAFE

R TO LA CA ES

GLAZING CLAY MAKING

ELECTIRC

CLAY PICK UP

ROOF EXHIBITON SPACE GALLERY

DRY ROOM

PIT FIRING

Workers Movement

Public Movement

Clay Movement + Process

Pottery Firing Methods 7

1. Abstract of collage of husband losing his wife floods that happened on the site. 2. Clay modelling the building to understand the organic nature of clay. 3. Connecting 5 balls represent the five villages that have been lost in the floods, how the balls connect and create space. 4. Exploring how to inhabit clay. 5. Modelling the building and drawing on top to visualise and develop the form. 6. Village of Lost Sculptures – Section A-A - Healing loss with clay.

7. 3D Clay Printing Factory – Section A-A - Renewal of Grimsby’s appearance through 3D Clay Printing. 8. Conceptual collage of interactive theatrical central atrium. 9. Clay feeding into the building from the back of the site (orange) and public movement (blue). 10. Sectional collage of materiality and activity of the 3D printing factory.

10 Graduate Show 2019

LA INDUSTRIAL Alejandro Guerrero Neira

Course: Master in Architecture University: ETSAM - Universidad Politécnica de Madrid Country: Spain Current Project: Amaravati Secretariat Year Start: 2018

- WHAT - LA INDUSTRIAL is a citizens’ laboratory that seeks to activate a local and open culture in the city of Madrid, throughout a program based on cross-disciplinary and digitally mediated practices that enhance new processes of cultural management, digital innovation and social transformation. It is as a place of encounter that reflects the current ways of thinking and producing culture. - WHERE - LA INDUSTRIAL is in Entrevías, a district located south of Madrid. A territory with a strong identity, which has its origin in the migration phenomena that occurred in the urban transition of Spain in the mid-20th century. Since then, its population have fought for the recognition of their social, educational and cultural rights against other areas of the city. - WHY - LA INDUSTRIAL aims to decentralize the cultural activity of Madrid and strengthen the concept of Culture of Proximity. The Culture, understood as an active movement, associated with the involvement of the citizens in the production of the neighborhoods’ activity. The Proximity understood as a citizenship right, measures the quality of our democracy, because the greater the proximity of the territory, the more rigorous and satisfactory the participation. 56

1. Peripheral neighbourhoods in Madrid. 2. Culture of proximity. 3. T he southern neighbourhood of EntrevĂas. 4. Preexisting dwellings by architect F.S. de Oiza - redrawn in egyptian perspective. 5. Models during the design process.

5 Graduate Show 2019

7 58

6. Urban regeneration. 7. Activity mapping and anylisis. 8. From voids to plazas. 9. LA INDUSTRIAL – section and plan. 10. Media facade - detail drawings. 11. LA INDUSTRIAL – interior spaces.

10 Graduate Show 2019

11 59

MYCO-MENSA: Bartlett B-pro Architectural Design Program Chen Shen

Course: MArch Architectural Design University: The Bartlett School of Architecture Country: China Current Project: Beverly Hills Master Plan Year Start: 2018 1

MYCO: MENSA from RC7 is a group focusing on create grow architecture with bio-receptive geometry and bio-material that is a self-growing, filamentous, natural composite. The mycelium of Pleurotus ostreatus (white oyster mushroom), was grown on a 3D-printing material, wood-fill filament to produce the bio-material then used to bio-fabricate a digital designed architectural geometry such as a table. The geometry needs to provide suitable conditions for mycelium growth and fantastic visual effect. Furthermore, we developed an easy printed logic in digital design, which can help our prototype to be printed by robot arm. Mycelium growth experiments, at the same time, are processed to explore the positive relationship between properties of mycelium growth and details of geometry, such as different density gradient of structure system, growing region of mycelium colonization and thickness of the material. This project aims to design a scaffold in geometry which in biocomposites called Woodfill and collocated by mycelium. Mycelium could reinforce structure strength and then become a growth material. We anticipate that this biodesign prototype could achieve diverse practical functions and sustainable meanings.

4 60

1. Digital language: differential line growth. 2. Growth on 3D printed woodfill faliment sample. 3. Fabrication methods: robotic arm printing. 4. Myco-mensa design: first prototype. 5. Exploration of digital language.

5 Graduate Show 2019

2 62

1. Myco-mensa design: final prototype. 2. Robotic arm printing prototype Photograph: final design. 3. Architectural scale design: cabin design. 4. Urban scale design.

4 Graduate Show 2019

The City Archive Josh Corfield

Course: MArch RIBA Part II University: The Bartlett School of Architecture Country: United Kingdom Current Project: Amaravati Secretariat Year Start: 2018

The Indian culture is manifested through myths, fables and fantasies of gods and other worlds. Storytelling represents a notion inextricably linked with knowledge and learning, a type of knowledge that is told and passed on, interpreted and exaggerated. This project employs storytelling as a way of preserving and interpreting the changing city of Varanasi, exploiting an age old custom to protect the past, present and future city. I understand Varanasi as an open book, a nucleus of generational stories that are inherently linked to context and place. Through the layers of religion, tourism and time the city itself is telling stories. The design itself is intended to evoke stories through creating atmospheric conditions imagined from traditional Indian temple architecture. The idea of relative identity between architectural elements and an architectural language, in combination with the creation of centres and sub-centres aim to create evocative, sensorial spaces that invoke stories within the users. Through spontaneous, everyday storytelling and planned performance, stories are recorded, analysed and preserved, cataloguing the changing city.

1 64

2 Graduate Show 2019

3 65

4 66

1. Second floor plan. 2. Axonometric view of archive. 3. Axonometric view of public space. 4. Interior archive view. 5. S un courtyard. 6. Longitudinal section.

6 Graduate Show 2019

Double Architecture: The Copy of the Ducal Palace of Urbino Michael Silvestri

It’s been more than one hundred years since Otto Rank published Der Dopperlgänger: a pioneering psychoanalytical study regarding doubles. Since then most disciplines developed researches and theories on the same subject, expanding on the work of the Austrian psychoanalyst. Architecture, on the other hand, faced an ever-increasing number of doubles without bothering to set up a proper theory. The aim of this thesis is thus to fill the gap between the theory and practice of double architecture. The research is based on the correlation between pairs of buildings and twins because the latters have been subject to countless studies that shed a light on the complex dynamics of every relationship and can therefore be applied to other disciplines. Seen as duplicity is related to identity matters I’ve developed a project for Urbino, a city split between residents and students. Studies prove that the two entities can no longer co-exist: the students need another city. Following Aldo Rossi’s celebration of the ability of monuments to structure a city, the Ducal Palace is edited and copied in the area where most of the students currently live, it becomes a symbol for the new city.

Course: MSc Architectural Design University: Università Iuav di Venezia Country: Italy Current Project: Beverly Hills Masterplan Year Start: 2019

Norra Tornen.

Torre Garisenda e Torre degli Asinelli.

Grove at Grand Bay.

018

019

030

Johanna ed Eva Gill.

Cabañas en el Río.

Maschere al Carnevale di Venezia.

Murale delle Absolute Towers.

055

072

Marilyn Monroe.

054

Frøsilo.

Gemelli siamesi.

The Cloud.

031

Attentati al World Trade Center.

Torres Siamesas.

Chang ed Eng Bunker.

073

People’s Building.

Ideogramma cinese rén.

104

110

105

117

116

111

122

123

Copenhagen Gate.

Salk Institute for Biological Studies.

Partenone.

190

191

Stretta di mano.

Ginger Rogers e Fred Astaire.

258

259

Elbphilharmonie Hamburg.

268

Pilotis di Villa Savoye.

Philharmonie Berlin.

253

252

Costruzione del The Silo.

Edificio per uffici della Nationale-Nederlanden.

Pilotis di Villa Dall’Ava.

Disegno preoperatorio.

FRAC.

Dorian Gray e il suo ritratto nel film Dorian Gray.

National Stadium.

269

Nido d’uccello.

298

303

302

315

320

321

one onevisits visitsaamonument monumentof of this thistype, type,for forexample examplethe thePalazzo Palazzodella dellaRagione RagioneininPadua, Padua,one oneisisalways alwayssurprised surprisedby byaaseries seriesof of questions questionsintimately intimatelyassociated associatedwith withit.it.In Inparticular, particular,one oneisisstruck struck by bythe themultiplicity multiplicityof of functions functionsthat thataabuilding buildingof of this thistype typecan cancontain containover overtime timeand andhow howthese thesefunctions functionsare areentirely entirelyindependent independentof of the theform. form.At Atthe thesame sametime, time,ititisisprecisely preciselythe theform formthat that impresses impressesus; us;we welive liveititand andexperience experienceit,it,and andininturn turnititstructures structuresthe thecity». city». Aldo AldoRossi, Rossi,L’architettura L’architetturadella dellacittà, città,english englishtranslation translationDiane DianeGhirardo Ghirardoand andJoan JoanOckman, Ockman,The TheArchitecture Architectureofof the theCity, City,The TheMIT MITPress, Press,Cambridge-London Cambridge-London1982 1982

ORIGINAL ORIGINAL

DOUBLE DOUBLE

«Koolhaas’s «Koolhaas’stheoretical theoreticalprogram programdoesn’t doesn’tassume assumethat thatthe thefunction functionisisable ableto todetermine determinethe theform; form;there’s there’saaclear clearwill willto toexceed exceedaacategorical categoricalrule ruleof of the themodern modernmovement, movement,proving provingthat thatthe thefuncfunc2 tion tioncan canadapt adaptitself itself to toaapreconceived preconceivedform formasaswell wellasasbeing beingassociated associatedwith withdifferent differentforms. forms.The Thefinal finalgoal goalisisto toprove provethat thatparticular particularneeds needscan canbe befulfilled fulfilledininspaces spacesthat thatare areundifferentiated undifferentiated or orrestrictive, restrictive,they theycan cantake takeadvantage advantageof of complex complexspaces spaceswhich whicharen’t aren’tnecessarily necessarilylogical». logical». Graduate Show 2019 69 Lamberto LambertoIppolito, Ippolito,La Lavilla villadel delNovecento, Novecento,Firenze FirenzeUniversity UniversityPress, Press,Firenze Firenze2009 2009(personally (personallytranslated) translated)

Double architecture

Il doppio in architettura

Thesis supervisor: Professor Sara Marini | A.Y. 2017-2018 | Exterior view

NEW DUCAL PALACE

3 70

VIII.

XV.

«The hanging garden faces the closeby hill like the terrace of a japanese garden and looks at this spectacular scene from the rectangles of its false windows». Leonardo Benevolo e Paolo Boninsegna, Urbino, Laterza, Roma-Bari 1986 (personally translated)

NUOVO PALAZZO DUCALE

«[...] il muro e la porta del Palazzo Ducale di Urbino sono ripensate come un diaframma [...] uno spazio abitabile nel quale [...] il muro diventa intervallo». Federico Bilò (a cura di), A partire da Giancarlo De Carlo, Gangemi, Roma 2007

1. Double architecture case studies. 2. The Ducal Palace of Urbino: original (left), copy (right). 3. Site plan of the Colle dei Cappuccini. 4. Exterior view of the New Ducal Palace. 5. Exterior view of the New Ducal Palace. 6. Section of the New Ducal Palace.

6 Graduate Show 2019

Into the Climate: Improving Living Conditions in a Bioclimatic Manner Wioletta Sarara

shaping the settlement in harmony with climate

living in the symbiosis with nature - a healthy circular community

north wind

houses can be shift every 1m

2 er s

ols t ic e

w inte

r so

ls ti

north wind

south wind

2. tilting the facades to the degree of summer and winter solstice - shadow all year round

excavated earth stormwater canals

3. creating butterﬂy roof for better vertical ventilation and rain harvesting 3

houses made from natural materials

hills preventing ﬂooding

rainwater for solar energy potable use

fertilizer biogas from black water and compost

edible crops

water ﬂows clean water to extra water to the river storm canals for irrigation

water treatment plants as natural building materials different water usage

constructed wetlands stage two

sceptic tank surpluses for non-potable for gray outer people water water storage

water treatment banana plantation plants as natural materials

constructed wetlands stage one

sceptic tank water taken from the river

water circulation

3. rammed earth service walls enabling the integration of essential utilities

organic matter circulation

Indonesia struggles with informal settlements. In Cigondewah, Bandung, people live on an offthe-grid area in self-built houses made from poorquality materials unsuitable for the hot-humid climate. ‘Into the climate’ aims to improve those conditions by creating a bioclimatic and selfsustaining in water and energy development. Dwellings are shaped accordingly to the tropical climate by applying bioclimatic principles learned from local vernacular architecture. Additional, they are made from natural, breathable materials available on site, using circular building methods and creating no waste aside. The whole community is supported by the closed circulation of water and organic matter, additionally obtaining food and building materials from this cycle.

a low-tech co-housing design from natural building materials

community based construction

Course: Arch. Engineering Graduation Studio University: Delft University of Technology Country: Poland Current Project: Meyongdong Apple Store Year Start: 2018

1. The interconnected 3 ideas of contemporary vernacular settlement. 2. Current situation in Cigondewah, Bandung. 3. Shaping architecture accordingly to the local hot-humid climate. 4. Living in the symbiosis with nature – a healthy circular community. 5. Hand-made model of co-housing. 72

5 Graduate Show 2019

meeting points stormwater canals banana plantation hills

the heart - main meeting point where surpluses of water are available for external communities

constructed wetlands - stage 1

during dry season outdoor theatre constructed wetlands - stage 2 vegetable crops biogas

water used for watering the gardens 7

6. Section through co-housing - ground floor workshops and commmon spaces, upper floors single or double floor appartments. 7. Urban plan of the holisitic community. 8. A low-tech, co-housing designed from natural materials available on site.

private inside spaces

extended living space/ corridor

bamboo conctruction prefabricated on site natural infil pannels rammed earth brick service wall

public outside spaces and inner courtyard

public inside spaces/ workshop

entrance zone

8 Graduate Show 2019

Studio 3 Alex Bramhill Konstantinos Chatzimanolis Natalie Cheung Ai Wen Beatrice Melli Stefan Necula Ignacio Diaz-Matroto Rivas

The Modern Hansestadt of the New Silk Road Alex Bramhill

Course: MArch Architecture University: University College London Country: United Kingdom Current Project: Lusail Towers Year Start: 2018

“The Modern Hansestadt” explores how a new Chinese-led Asian-European trade route, called the New Silk Road, will affect the planned European terminus of Duisburg. The project explores how the influx of trade will create a bustling metropolis around a new intermodal transport hub. A new long-spanning megastructure (2km²) over Duisburg’s existing docks will be the foundation of an ascendant, ever-in-flux city – a melting pot of Eastern and Western cultures. Within the structural framework, there are districts containing city elements such as marketplaces, trading, and accommodation. Market stalls, shops, houses, and containers are transported through the scheme via transportation routes. These are inserted into the districts and attached to the megastructure or, in the case of containers, transported to their onward journey down the Rhine. Utilising genetic algorithms for structural and spatial analysis, the Modern Hansestadt considers the design optimisation and successful inhabitation of a physical European framework for the New Silk Road.

1 80

1. The HauptstraĂ&#x;e â&#x20AC;&#x201C; The main street of the scheme, lined by Hanseaticinspired houses.

3. Section demonstrating how the transportation network runs through, and is integrated into, the new city.

2. The main structure of the scheme allow goods transportation between the New Silk Road and the Rhine, as well as being occupied by traders, creating a new city.

3 Graduate Show 2019

4 82

4. Detailed fragment illustrating how buildings attach to the structure of the Modern Hansestadt. 5. The Modern Hansestadt is a threedimensional trading city, providing a bridge between continental trading powers.

5 Graduate Show 2019

6. Below the HauptstraĂ&#x;e, the workings of the city are revealed. Goods and buildings enter the scheme to be transported onward or placed within the scheme respectively.

6 83

Traces of Disappearance Beatrice Melli

Course: AA Undergraduate University: Architectural Association Country: Italy Current Project: Lusail Towers Year Start: 2018

The project explores the concept of disappearing as a way of disrupting the understanding of the architectural space. Questioning the limits of perception. In a heavily digitized world, gestures and actions are becoming automated and as a result immaterial, intangible. Bringing us to a kind of virtualization of the world. The more we upload aspects of ourselves on the internet the more the role of architecture should engage us in other ways, leaving a trace. I believe in designing a correlation between a vision and tactile reading of the space, where the role of physical architecture should re-engage us with the tangibility of the world. For we are not dealing with disappearance as the natural phenomena, but rather an â&#x20AC;&#x2DC;inventionâ&#x20AC;&#x2122; of human beings, it is disappearance as a type of art dealing with time and memory of both the collective and individual body. Disappearance of the social act, which according to Maurizio Ferraris, an Italian modern philosopher: the social act, which requires at least two people to initiate, is the bases of our society. The natural faculty that allows to construct culture. I consider the role of architecture to be a fundamental player in leaving traces as spatial experiences to be explored and remembered by the individual and collective body. 1 84

2 Graduate Show 2019

1. Disappearance Its definitions vary in this intangible ambivalences, there are no defined boundaries or limits. It deals with time, change and evolution. One should engage in order to understand the space. 2. The Crease A language of recognizable forms that dissolve the limits of the vanishing point. The crease can be read both as a ceiling and/or ground condition, keeping a clear shift and disconnection between the surfaces. A transition that questions the boundaries of the oblique, where the sloped surface acts as an extension of both the vertical and horizontal field. 3. Blurring The Line of Perception The space creates a correlation between a vision and tactile reading of architecture. A diffuse atmosphere of vanishing traces, where the user begins to question disappearance. Where the light plays the role of the intangible material, hiding and showing different spaces. No single point of focus. This way one breaks out of their habitual figure ground orientation and is prompt to engage with the continuity of the space.

3 Graduate Show 2019

A.I. Architect: Automating Concept Design in the Residential Sector By Utilising Genetic Algorithms Ignacio Diaz-Maroto Rivas

Course: MSc, Building Information Management University: University of Westminster Country: United Kingdom Current Project: Sydney Metro Year Start: 2018

This research paper explores the capacity to automate the architectural design process, hoping that this can lead to improved construction productivity, and a greater reach of good design principles. The paper proposes that AI has the potential to generate architectural design. A basic pilot system is developed to tackle concept design in affordable housing in the UK. The system uses Genetic Algorithms to generate a volumetric design in Grasshopper and a 3D model is generated in Revit using Dynamo. The design principles are determined by a human Architect, but the system designs the actual buildings. In essence, the Architect designs a Virtual Architect. The pilot system is tested on three sites with existing buildings. The premise is that if the system can output layouts similar to the existing buildings, then it is successful.

1 88

2 Graduate Show 2019

3 90

1. Structure of pilot sytem. 2. Scenarios used for testing the pilot system, showing existing buildings on the left and showing generated layouts on the right. 3. Pilot System improved and developed to tackle paired towers in Hong Kong. 4. Description of Stage 1 Grasshopper script. 5. Descrption of Stage 2 Dynamo script

5 Graduate Show 2019

Water Cycle Seaskyscraper – Highly Integrated Synthesis Konstantinos Chatzimanolis

Course: MSc Architectural Computation University: The Bartlett School of Architecture, UCL Country: Greece Current Project: Lusail Towers Year Start: 2018

The “Water Cycle Seaskyscraper” examines the water scarcity issue through an innovative multiparametric, highly integrated synthesis of an architectural prototype. It can be characterized as a sustainable solution for water production. The procedure uses seawater, which under a specific treatment is collected, distilled, filtered, metalized, and finally it produces clean and tasty water. Within this framework, the current project suggests the development of such high-rise mechanism, which is positioned in the sea. The application of parametric principles that take place in this investigation aims to optimize the water production performance and at the same time the functional composition of living spaces.

Dimpola Thesis - University of Cyprus

2 92

Water Cycle Seaskyscraper Highly Integrated Synthesis

1. Section diagram – integrated water generation unit. 2. Architectural model – scale 1:500. 3. Bottom-up view – architectural visualisation.

Case Study: Aegean Water Cycle Seaskyscraper

3 Graduate Show 2019

Case Study:

Aegean Water Cycle Seaskyscraper

4 94

4. Night scene long distance view â&#x20AC;&#x201C; architectural visualisation. 5. Bird view â&#x20AC;&#x201C; architectural visualisation.

Konstantinos Chatzimanolis

5 Graduate Show 2019

Authoring Nature, A Fable of Singapore Natalie Cheung Ai Wen

Portrait

Course: MArch Architecture University: National University of Singapore Country: Singapore Current Project: Bonifacio Residential Tower Year Start: 2018

Authoring Nature is a commentary on Singapore’s over-control of nature within the City. Known as “Garden City” and “City in a Garden”, the country constantly uses nature, both flaura and fauna, as an asset towards nation-branding. While cities continue to expand into natural territories, an increasing number of wild animals have started to encroach into our domestic territories, causing an increase in tension between city-goers and the wild. “Get rid of these birds and butterflies, they’re dirtying my green wall”, this quote overheard was telling of how we saw nature within the city. We often idealised nature for it’s aesthetic value, however reject its natural effects. This three-act thesis narrates the state of a highly-raved endangered animal - the river otter, within Singapore’s skyline and premium site - the Marina Bay Sands. Begetting the question - What exactly does Nature mean to us in the City?

1 96

2 Graduate Show 2019

3 98

1. A Natureâ&#x20AC;&#x2122;s Fable. 2. The Picturesque / A City Sanctuary. 3. Inside-Out / Outside In. 4. The Architecture Beneath. 5. N ature Beyond Us.

5 Graduate Show 2019

RE_Scent: The Institute of Smells Stefan Necula

Course: MArch RIBA Part II University: The Bartlett School of Architecture Country: Romania Current Project: Wing Tai Year Start: 2018

The architectural proposition aims to speculate a unique typology for cultural preservation, generated by the relationship odours have to both mundane activities and the sacred customs in Varanasi. Within the olfaction dimension concepts of space, memory and odours are investigated to renegotiate the importance smells have in influencing both sacred and profane realms. Therefore, the project aim is to identify the intangible qualities smells have in relation to culture and heritage as well as their potential to preserve such qualities. The hidden qualities odours have in relation to time and memory represent the core of the project and impart reason to the investigation. To allow for a clearer understanding of the subject, which the proposal is exploring, an olfactory journey through the ancient city of Varanasi is the stage for the narrative.

1 100

3 Graduate Show 2019

101

4 102

The experiments are challenging the conventional manner in which spaces are experienced by renegotiating the importance odours have in creating unique spatial experiences; addressing the question: â&#x20AC;&#x153;How to make visible the invisible?â&#x20AC;?

1. Riverfront. 2. Plan. 3. Views of the experimental spaces. 4. RE_Scent Overview. 5. E xperimental odours: Physical interpretation.

Graduate Show 2019

103

104

Studio 4

Adrian Yankee Siu Andrew Yuen Callum Rowland Camille Dunlop Eugene Jahng Gabriel Pavlides Kasia Gryszkiewicz Kevin Herhusky Pablo Valero Escolano Ruoming Song Yinghua (Will) Chen

The Diggers Festival of Peace Adrian Yankee Siu

Course: MArch Architecture (RIBA/ARB Part 2) University: The Bartlett School of Architecture, UCL Country: United Kingdom Current Project: South Sabah Masterplan Year Start: 2018

Formerly inhabited in Runnymede, an eco-activist community called “The Diggers” crafted an off-grid woodland village to provoke a land reform. By highlighting this disused land in London, they advocated an ecologically sustainable living model with cultivation and dwellings built with local materials, establishing a sense of cult around mother nature. In 2015, these squatters were eventually evicted from their self-constructed homeland, after a lawsuit over the land ownership against property developers. A new speculative settlement is proposed to relocate these eco-activists in London, where they collectively realise this alternative Arcadia. This provides an opportunity to imagine the utopian visions that squatting communities attempted to create over the years. The experimental masterplan in Benfleet extends the legacies of the Diggers and establishes a living ecosystem governed by the changes in nature, such as sun movements and seasonal transitions, which echoes with the ecologically sustainable lifestyle of the activists. Through this annual pilgrimage of nature, “Festival of Peace” is hosted to celebrate their reform, as well as to promote social autonomy and equality to the outsiders and newcomers... 1 108

1. Deconstructing the Diggers nomadic dwelling. 2. Orrey - Sun movement studying instrument.

2 Graduate Show 2019

109

3 110

3. Diorama drawing â&#x20AC;&#x201C; elements for a sustainable settlement. 4. Production sector â&#x20AC;&#x201C; the Diggers settlement. 5. Diorama model.

5 Graduate Show 2019

111

Expanded Domesticity: Beyond Confined Domestic Spaces In Hong Kong Andrew Yuen

Course: AA Undergraduate University: AA School of Architecture Country: Hong Kong Current Project: South Sabah Masterplan Year Start: 2018 1

The project reacts against the constrained living conditions of migrant workers in Hong Kong, where confinement and practicality shape the domesticity in the small city. A new series of housing typologies is designed to expand domestic life through literal and phenomenal expansion, without increasing the average square meterage. They come as the form of new build housing blocks and parasites on the existing building facades, with the rooftops contributing to communal programs. Each new housing typology incorporates mirror to achieve deep and infinite view by reflection, translucent material to introduce natural light to confront with the rigidity of conventional solid wall.

4 112

1. New public housing typologies. 2. Material - mirror, timber & translucency. 3. Walking up to the apartment. 4. New rooftop - communal space. 5. Plan - the new Sham Shui Po.

5 Graduate Show 2019

113

6 114

6. City view - the new Sham Shui Po. 7. The new public housing block.

7 Graduate Show 2019

115

Architecture Against Conflagration Callum Rowland

1. The Pagoda ‘Snake-Dance’.

with Chris McCallum and Rishabh Shah

4. Exhibition, vertical evacuation (3/3).

2. Exhibition, the wall (1/3). 3. Exhibition. Refuge (2/3).

Course: MA Architecture (ARB/RIBA Part I) University: University of Edinburgh Country: United Kingdom Current Project: Bloomberg Year Start: 2018

Informed by the traditional Japanese Pagoda, a contemporary structural language of stepped loads has emerged. Applied to the residential tower block typology of Tokyo, this system has responded to the unpredictable and devastating threat of conflagration which has arisen as a consequence of coalescing geographic, historical and architectural features. Operating across a variety of scales and scenarios – from day-to-day life to the event of widespread urban fire, and from individual intervention to urban strategy – the Kitasenju Fire Wall safeguards local communities whilst contributing to a city of ever-increasing density.

1 116

2 Graduate Show 2019

4 117

High Risk Low Risk Evacuation Spaces

5 118

5. Tokyo, map of risk and evacuation. 6. Model photograph, service modules. 7. Model photograph, a forest of tension cables.

6 Graduate Show 2019

119

Pipeline Hijacking Camille Dunlop

Course: BSc Architecture (RIBA Part 1) University: University College London Country: United Kingdom Current Project: Phnom Penh International Airport Year Start: 2018 1

â&#x20AC;&#x2DC;Pipeline Hijackingâ&#x20AC;&#x2122; is a speculative glimpse into alternative means of dwelling and building construction. The project imagines how a community could hijack, grow along, and be informed by, an existing infrastructure of an Icelandic Geothermal Pipeline Network which supplies a constant stream of hot water. The resulting architecture becomes an ecosystem of hot water pipes with a variety of habitable climates, unique spatial experiences and an ethos of sustainable living.

3 120

1. Elevation of one dwelling. 2. Site map highlighting Geothermal Power Plant and the Pipeline within the landscape.

3. Short section through one dwelling and the surrounding community baths. 4. Plan of one cluster of community including private dwelling and community spaces.

4 Graduate Show 2019

121

5 122

5. Plan of a single isolated dwelling. 6. A series of wall details from a variety of domestic rooms, including a bedroom, a bathroom and a living area. These details creating a variety of specific habitable climates and individual spatial experiences.

6 Graduate Show 2019

123

Willet’s Point Estuary Eugene Jahng

Course: M.S Advanced Architectural Design University: Columbia University GSAPP Country: USA Current Project: Intuitive Year Start: 2018

The Willet’s Point Estuary connects the isolated ‘Iron Triangle’ of Queens, NY to the growing industrial neighborhood of the borough, utilizing the nearby Flushing Bay and Creek to generate hydroelectric power for the entire community district. The polluted river and greywater from surrounding districts is filtered into the Estuary’s wastewater treatment facility and wetland. The resulting purified water is collected into a final reservoir to be released into the bay, transforming the former Valley of Ash into a living, self-sustaining Battery for the community. A racing venue for rowing boats would be ideal given the creek and bay, with public boathouses for storage and maintenance. Two boathouses mark the beginning and end of the race, while providing a public boat storage space for the rowing community. Races start from the Secondary Boathouse on the Creek around the site into the estuary and end in front of the Primary Boathouse. Public access points into the bleachers are on street level from elevators or walkways down the slope. Open roads trailing down the public wetland provide further circulation. Private access points to the wastewater facility are connected from the highway to the docks. Trailers pulling boats can also access the docks from these points.

2 124

1. Site Conditions with overlaying flood hazard zones.

3. Panoramic Render of site seen from Citi Field.

2. Site Approach diagrams from existing site, sea level rises, extended creek, Cut and fill, and private/public program.

4. Section Perspective through the estuary.

4 Graduate Show 2019

125

5 126

5. Axonometric Diagram showing the final hydroelectric power plant and wastewater treatment facility extending into wetlands and public boathouses. 6. Section through entire site from flushing creek and primary reservoir to the final reservoir and estuary.

6 Graduate Show 2019

127

House of Bardud Gabriel Pavlides

Portrait

Course: Bsc Architecture University: University College London, Country: Cyprus Current Project: Phnom Penh International Airport Year Start: 2018

With fishing villages on the coast of Iceland disappearing along with their century old values and traditions, the project aims to offer a new model for the countryâ&#x20AC;&#x2122;s declining rural fishing industry and its population. Addressing local environmental issues such as eroding coast line, rising sea levels and volcanic activity, the structure rises out of the sea bed with a platform element housing the communal living spaces, surrounded by floating islands serving as sustainable fish farms. Eroding cement mixed with seeds enhances the biodiversity in the area while new layers of floating concrete allow the structure to grow parallel to the new community it houses off the coast of Iceland.

1 128

1. Axonometric section through time. 2. Under water view. 3. B irdâ&#x20AC;&#x2122;s eye view.

3 Graduate Show 2019

129

Microfactory: The Fabrication Club for Pro and Amateurs Kasia Gryszkiewicz

Haggerston Bath is the abandoned Grade II Listed Edwardian building in Hackney, London. The purpose of the project is to bring back life into the current Urban Ghost and turn this magnificent building into a community focused hub that provides an opportunity for students and new start-ups to invent, create and fabricate. The goal is to preserve Hackneyâ&#x20AC;&#x2122;s heritage but also introduce the new infrastructure above the historic swimming pool thus creating the symbolic meaning of old reflecting new; the old space is preserved but the new infrastructure enables the contemporary inventions.

Course: Architecture BA (Hons) University: Leicester School of Architecture Country: Poland Current Project: WingTai Redevelopment Year Start: 2018

1 130

1. Long section through the building. 2. Inside view. 3. Program explanation.

2 Graduate Show 2019

3 131

4 132

4. 4. Detailed section. 5. Inside view. 6. Program explanation.

5 Graduate Show 2019

6 133

Infrastructures of Memory Kevin Herhusky

Course: BArch University: California Polytechnic, San Luis Obispo Country: United States Current Project: Intuitive Surgical Year Start: 2018

Architecture was once the primary vessel of memory preservation, but eventually passed its responsibility to new inventions, such as the printing press, books, photos, etc… leading up to perhaps the most prominent current vessel of memory: the internet and its infinite cloud of data. Following this realization, this thesis proposes a revival of an architecture saturated in memory, at the anchor point connecting the current vessel of the internet to architecture: the data center. For all the data center contains and represents, it remains an incredible lost spatial opportunity – the data center of today is essentially a memory prison, a massive warehouse full of servers and cables, and not much else… but it should be the opposite, a memory palace, a grand spectacle of preservation! Through a predictive fiction, the data center is embedded within the infrastructural landscape of the vacant and deteriorating Folsom State Prison. An intricate layering of program partitions and augments the site, retrofitting the prison into memory palace - a grand spectacle of preservation – where an expanding cloud condenses data into a hybrid of digital and physical space - establishing a new phygital vernacular. Imperfect, low-res techniques of preservation and highly-articulated systems of infrastructure weave together into an extra-dimensional tapestry of memory. 134

1-3

1-4. E xperiment – New techniques of phygital preservation. 5. Precedent – Google maps imperfect texturing. 6. Exterior axon – An intricate tapestry of memory saturates the infrastructural landscape.

6 Graduate Show 2019

135

8 136

7. Exterior Axon-elevation.

8. Interior – The grand spectacle of memory preservation. 9 -11. Physical model – overlayed with digital projection.

11 Graduate Show 2019

137

Rewriting the Limits. Centre of the Culture of the Sea Pablo Valero Escolano

Portrait

Course: Master in Architecture University: ETSAM (Universidad PolitĂŠcnica de Madrid) Country: Spain Current Project: Project Floyd Year Start: 2018

The Pasaia Bay, a place marked by the greatness of its nature and not far away on time was a site with high cultural and social interest that nowadays seem to be hidden under concrete and cranes. The out of control evolution of the Industry at the Bay was the trigger for the chaotic situation that today this place is living. Different reallities seek to coexist but none of them reach to get it. The Industry tends to dissapear at some points of the bay, looking for the balance, leaving a indelible print; inmense concrete platfomrs that gained ground to the sea. The proposal seek to go back to the origins, recovering the human scale and the close relationship wich San Pedro and its inhabitants had uppon a time with the sea. To fullfil this target, it is proposed to work on the ground that once uppon a time was introduced to generate those concrete platforms. A place for joy and meeting with the bay and its multipe landscape as backdrop.

1 138

Propossal Concept

OPTION 2

2 Graduate Show 2019

139

Centre of the Culture of the Sea Longitudinal Section

3 140

1. Horizon view. 2. Sketches and views. 3. Longitudinal section + general axo. 4. Axo program. 5. Auditorium interior view.

5 Graduate Show 2019

141

Writing / Automata Ruoming Song

Course: AA Diploma University: AA School of Architecture Country: United Kingdom Current Project: Phnom Penh International Airport Year Start: 2018

Reyner Banham proposed to treat architecture as a “black box”, recognised by its output though unknown in its content. The project unfolds the “black box” of an architect’s mind as a series of conceptual machines. By combining semiotic readings of architecture with functionalist references to the scientific logic, the machines bridge the gap between abstract thinking and concretised forms, between the creative subject and the automata. It reconstructs diagrammatic flows and operation sequences that engage with dissimilar conceptual platforms, that build upon divergent sub-strands of disciplinary thinking - autonomy vs contingency, object vs system, history vs imagination. Overall, the production process turns these tensions into complementary forces that sustain the evolution of architecture. As the architectural form remains in transition, the project also suggests ways to animate static elements of European historical cities such as Rome, subject to misreadings, clashes and adaptations.

1 142

3 Graduate Show 2019

4 143

t tex

on fc to

ing

s Lo

tail

tio ca

pli

atis

atio

m ste Sy

tio ga

gre Ag

i yp et

ch Ar ng

pi ty

o ot Pr

olo

rph

Pro t

oty pe

ha ta

re rea

rali

sed

ag a

TYPE RE-GENERATOR 144

Through the generifying process, an architectural object loses its specificity and becomes an archetype. The machine then allows the archetype to be multiplied and eventually generate new prototypes that regain specificity.

1. Meta-factory of conceptual machines. 2. Language of Danteum. 3. Disciplinary maze. 4. Machine of co-existence. 5. Type re-generator. 6. Diagrammatic flow of automata.

6 Graduate Show 2019

145

Russian Cultural Centre Yinghua (Will) Chen

Portrait

Course: Architecture MArch (ARB/RIBA Part 2) University: The Bartlett School of Architecture, UCL Country: United Kingdom Current Project: Phnom Penh International Airport Year Start: 2018

A project that observes the nature of the urban systeam, the chemistry of the city, infrastructure and architecture. Inspired by the theory of space syntax, a humanguided â&#x20AC;&#x2DC;attractor systemâ&#x20AC;&#x2122; manipulates geometry influencing the relationships between and the deformation of experienced architectural space. The urban proposal is for the representation of a new American city, multi-vertical layers recomposing the city context within one system. Politically, this cultural centre introduces Russian art and culture to the context of New York and serves as a concertration of multi-cultural exchange, immigration and education. This is an integration of the city of St Petersburg within one architecutral intervention on a vibrant corner of New York City.

1 146

2 Graduate Show 2019

147

3 148

1. Perspective drawing. 2. Conceptual drawings and process model render. 3. Material performance and Prototype exploration. 4. Sectional model of the inserted junction.

4 Graduate Show 2019

149

150

Studio 5

Alessandro Magliani Chen-Yong Tan Chris Cheung Dominic Oliver Elizabeth Morgan Lang Jin

Objects Against the Canon Entropy and the New Barbarians Alessandro Magliani

Course: Architecture University: Architectural Association Country: Italy Current Project: The Red Sea Project Year Start: 2018

The project seeks to stir up the historical context of Italian cities by tapping into their undiscovered, ‘barbaric’ potential. It is focused on the transformation of architectural form extracted from several urban sites into a series of itinerant objects, targeting the static and restrictive idea of the architectural ‘canon’. Buildings that are part of the historical tradition of several Italian cities are selected and gradually taken apart, to form a series of new ensembles, which are independent from their original sites, and are free to roam in and out of the city to interact with radically different contexts. Once they reach a different setting, the objects are able to deploy and expand into temporary exhibitions that show the original buildings samples in alternative, anti-hierarchical layouts. The project is ultimately an invitation to look at the existing form of our cities as a rich mining ground, that can be used freely to sample existing pieces of architecture in order to create new forms and assemblages within fluid and adaptable disciplinary frameworks.

154

1. Crash Scene. 2. Venetian Float. 3. Milanese Cart. 4. Venice unbuilds itself. 5. Meta-Italian Landscape (following page).

3 Graduate Show 2019

4 155

156

5 Graduate Show 2019

157

Cocina Comunitaria: Centro Habana Foodhall Chen-Yong Tan

Course: MArch Architecture University: Universityof Bath Country: United Kingdom Current Project: Huaxing Longsheng Year Start: 2015

The Cocina Comunitaria is a Community Foodhall in Centro Havana with the aim of providing a public platform for the preservation, dissemination and exchange of knowledge and skills related to nutrition, food security and gastronomy. The project adopts the analogy of the ajiaco, a national dish of Cuba used to describe the island nation as a melting pot. The cooking of the ajiaco symbolises the complex formation of Cuban culture, where the stew is constantly simmering and is never ready. As Cuba is at the cusp of change, new cultural ingredients and foreign culinary influences would keep entering the pot of Cuba. Here, the lesson of ajiaco would emphasise the importance of assimilation, as opposed to the replacement of the old with the new, for the future transformation and continuous enrichment of Cuban food culture.

3 158

Landscapes of Cultivation: a. Shade Garden b. Cultivation Garden Cloister c. Scent Garden d. Kitchen Garden e. Herb & Spice Garden

1. Calle Blanco Elevation 2. Section through Calle Blanco Refectory 3. Section through Garden Cloister

b e

4 Graduate Show 2019

159

160

4. 5. The Refectory. 6. Cultivation Garden Cloister. 7. Teaching Kitchen. 8. Calle Colon Morning Market Street. 9. Calle Blanco Evening Street Food Market. 10. Central Market Hall.

10 Graduate Show 2019

161

Promote Walkability: Reclaiming Pontcannaâ&#x20AC;&#x2122;s Backlane Chris Cheung

The idea of promoting walkability in the community is not only about improving the existing pathway, but also to redefine underutilized pocket spaces in-between urban fabric. To meet the increasing demand on spaces for local start-up creative businesses, and connect them with the community, a Creative Hub is proposed along the backlane. It provides a venue for creative people to meet, share creative ideas and collaborate with others. It will become a pilot scheme for the development of a walkable creative community in Pontcanna, and a catalyst to encourage local residence opening up their underutilized backyard towards the backlane with their own creative businesses. Course: Master of Architecture University: Cardiff University Country: Hong Kong Current Project: Shanghai Lingang Art Centre Year Start: 2018

Centralized Utility System

Restaurant (1/F): restaurant space connects to coworking space via the bridge extension to G/F when the event space is not in use Event space exhibition area and Bar (G/F): support space for the major event space, with bar as well for informal meeting and gathering

0.5

1.5

Sheltered Backlane Space: newly paved backlane as shared surface, which connect closely to the event space on both side it will become pedestrianized area during event and conference, acting as foyer and informal meeting space, linking the exhibition area, bar and the major event space

Event & Conference Space: double height space that can open up to the sheltered backlane visual connection from upper floor connected to the coworking space multifunctional space for event, talk, conference, award ceremony, mini concert, creative industry related performance, movie display

1 162

2 Graduate Show 2019

3 163

Legend: 01 Reception 02 Waiting Area 03 Workshop 04 Gallery 05 Storage 06 Workshop (Large) Pontcanna Creative Hub

Sharing Block: for people to explore 0

Pontcanna Creative Hub

Legend: 11 Reception 12 Coworking Studio 13 Presentation space 14 Quiet Workplace 15 Private Room 16 Meeting Space 17 Temporary Exhibition Area 18 Event Conference Space 19 Centralized Utility Service 20 Storage

Collaboration Block: for people meet and create

5 164

1. Perspective section of Collaboration Block. 2. Backlane view to Sharing Block (Entrance Building). 3. Backlane view to Collaboration Block. 4. Section and perspective of Sharing Block. Arrival Area / Gallery Space / Corridor to Workshop Space / Main Workshop Space. 5. Section and perspective of Collaboration Block. Co-working Space / Connection Space / Main Hall for Casual Activities / Main Hall for Formal Events. Existing backlane with solid wall on both sides. It is mainly used by vehicles to access to individual backyard or carparking. There are potential to redefine backlane where provides a secured yet open atmosphere.

6. Phase change to site with the proposal of Creative Hub as pilot scheme to Pontcanna Backlane. 7. Axonometric view to the Creative Hub Backlane Pilot Scheme.

Pilot scheme provides a creative hub as catalyst that encourage people to use the underutilised backlane differently. With the provision of different programmes, the scheme start to be populated by people.

Future domestric development, which individual open up their own backyard towards the backlane, with development of the own creative businesses, contributing to the walkable creative community.

6 Graduate Show 2019

7 165

Neural Interface Dominic Oliver

Course: BSc Architecture University: University of Bath Country: United Kingdom Current Project: Shanghai Lingang Art Centre Year Start: 2018

An exploration of Posthumanism and Cybernetic theory, Neural Interface aims to dissolve the barriers between corporeal and mental space. Through an experimental approach to rehabilitation, injured veterans use neuroprosthetics to aid researchers map neurons, ultimately hoping to link brain and computer. The building becomes a prosthesis of the user, a collective consciousness. Responsive, empathetic environments interact with the veterans and create diaphanous thresholds. Soft robotics grow from the infrastructural frame transforming space through feedback loops. This creates a fluid architecture, physicalising the mass hallucinations of digital realities. A soaring Posthuman temple is clad in electronic waste, growing as technology consumption increases. When computers shrink to brain implants, the tower starts to decay, becoming a monument to the tangible manifestation of digital life. As the cobalt crisis worsens, the ruin will transform into a material salvage.

1 166

2 Graduate Show 2019

167

3 168

1. Tentacular Thinking. 2. Fluid Architecture. 3. Cloud Mind. 4. Responsive Environments. 5. Ethereal Therapy. 6. â&#x20AC;&#x2DC;Dataspaceâ&#x20AC;&#x2122;. 7. Autopoeisis. 8. Political Machines. 9. Seismic Flux.

5 Graduate Show 2019

169

The Vertical Urban Sports Center Elizabeth Morgan

Course: Masters of Architecture University: Northumbria University Country: United Kingdom Current Project: Techcombank II HQ Year Start: 2016

This design project seeks to explore the concept of a vertical urban sports centre encouraging better health and well-being in the youth of Newcastle. The proposed structural system is a framed tube constructed of a hybrid system Route of Old George’s Stairs of steel and timber. Through active design,Historic theJourney scheme will hope to improve health and wellLocation of Cale Cross Market being encouraging the use of the community Castle stairs to journey up through the site and into the tower. The design of the tower is to include aspects supporting health and well-being in every opportunity, fully integrating nature into Key Historic Route from the Castle to the Quayside the design of the building fabric.

GEORGE’S STAIRS FROM SIDE STREET TO TYNE BRIDGE

Historic Route from the Castle to the Quayside

New Stairs Details: External Public Route

Connectivity - All Major Boundaries

Accessability - EntrancesHistoric to Building Route fromRoute the Castle to the Quayside Historic from the Castle to the Quayside

Stairs Design: Direct and Clear, Inspiration from Painter’s Heugh Stairs. SITE DIAGRAMS Height of Level Change: 18m

Historic George’s Stairs Route: Shortcut from Side street, Cale Cross Market to the end of Pilgrim Street Existing Georges Stairs Route: Connects ‘Side’ Street to walkways on Site (Public Right of Way removed in 2013) New George’s Stairs Route: Conecting ‘Side’ Street to Tyne Bridge Site Design Components: Transparency and Connectivity

Connectivity - All Major Boundaries

Historic Route from the Castle to the Quayside Route of New George’s Stairs

Route of Painter’s Heugh Stairs

Tyne Bridge Route

Access Points to Building

Railway Viaduct Route ‘Side’ Street Route

Route of Old

Line of Viaduct

Route of Old George’s Stairs

View out to cityscape

View back to Structural Line Tower perpendicular to ‘ Side’ Street Line of ‘Side’ Street Connectivity - All Major Boundaries

Route of New George’s Stairs

Route of Painter’s Heugh Stairs

Tyne Bridge Route

Access Points to Building

Railway Viaduct Route ‘Side’ Street Route

‘Side’ Street Ro

Castle

Location of Services

Railway Viaduct Route

Location of Cale Cross Market

‘Side’ Street Route

Key

perpend

Key Structural Line perpendicular to Viaduct ISOMETRIC OF EXISTING INFRASTRUCTURE

Structural Line perpendicular to ‘ Side’ Street Line of ‘Side’ Street

Location of Services

View out to cityscape View back to Tower Boundary Relationships to Structure

Pass through Viaduct

Castle Structural Line perpendicular to Viaduct Line of Viaduct

Key

Structural Line perpendicular to ‘ Side’ Street Line of ‘Side’ Street

Location of Services View out to cityscape View back to Tower Boundary Relationships to Structure

Visability out to City Scape - Location of Services

4 170

Structur perpend

Line of Viaduct Structura

Pass through Viaduct

Accessability - Entrances to Building Visability out to City Scape - Location ofKey Services

RailwayRoute Viaduc Route ‘Side’ S

Structural Line to Viadu perpendicular Line of V to Viaduct

Access Points to Building

Tyne Bridge Route

Historic Journey

Boundary Relationships to Structure

Route of Painter’s Heugh Stairs

Route of New George’s Stairs

Line of Viaduct

Location of Cale Cross Market

Tyne Bridge Ro Railway

Location ofCross CaleMarket Cross Market Castle

Structural Line perpendicular to Viaduct

Building Name: Cale Cross House and Tyne House Current Use: Offices (Not Fully Occupied) Existing Infrastructure: Series of Stairs and Walkways Stairs from ‘Side’ Street: George’s Stairs Current Stairs Network: Public Right of Way removed in 2013 Issues with Existing Infrastructure: Loss of Direction, No Transparency and No Obvious Direction

Route of Old George’s Stairs

Historic Journey

Location of Cale

Castle

Historic Journey

Route ofGeorge New George’s Stair Tyne Br

Historic Journey

Pass through Viaduct

Route o

Route of Old George’s Stairs George’s Stairs

Side’ Str Structural Line Line of ‘S perpendicular to Side’ Street Line of ‘Side’ St

1. George’s Stairs – urban stitching. 2. Site diagram – connectivity, all major site boundaries. 3. Isometric of existing infrastructure. 4. Internal studio stairs. 5. Team urban garden.

5 Graduate Show 2019

171

6 172

6. Active design: creating a walkable vertical community. 7. Sectional perspective through tower looking west.

SECTIONAL PERSPECTIVE THROUGH TOWER LOOKING WEST

Graduate Show 2019

7 173

Casa De La Musica: Havana Music Centre Lang Jin

Course: Master in Architecture University: University of Bath Country: China Current Project: Jing An Zhangyuan Tower Year Start: 2018

This project takes the context of Cuba as the island of music. It responds to the rising issue of a professional venue shortage and the lack of recreational space in the neighbourhood area, by embedding a new music centre in the Centro Havana. This project aims to support the value chain of the local music industry as well as providing public recreational space for the local people. The scheme is inspired by the unique Cuban streetlife and traditional Cuban Casino Salsa dance. It aims to extend urban space into the building, by having a brick vaulted open theatre as the central space. The rest of space follows a hierarchy of vaults, which is acoustically differentiated between public space and private music rooms.

1 174

Graduate Show 2019

9 175

176

1. Central open theatre. 2. Dancing Patterns of Cuban Salsa Casino. 3. Hierarchy of vaults. 4. V ertical circulation interacts with the vaults in different ways. 5. Central open theatre and secondary restaurant area. 6. Night view of central open theatre. 7. Top view of central open theatre. 8. Secondary restaurant area. 9. Teaching zone for musicians and dancers are located on the second floor. 10. Dance studio. 11. Bar area. 12. Sectional perspective.

12 Graduate Show 2019

177

178

Studio 6

Christian Kennedy Daniela A. Salgado Silva Duncan Catteral Fadhil Fadhil Heidi Han Ola Urbanska Sean Bailey Yannis Hajigeorgis

The Crown Collective Christian Kennedy

The Crown Collective places community as its spiritual centrepiece. It groups vertical communities in three story hubs. These hubs open onto a shared sky garden and community courtyard. The communities are built around shell and core housing which provides â&#x20AC;&#x2DC;ready to camp inâ&#x20AC;&#x2122; housing: a structural, weather-proof, thermally and acoustically insulated shell which can be partitioned as the inhabitant desires. The Crown Collective provides a commercial and civic heart combining a forum, market and public baths. It weaves these programs together to provide a venue of participation. Which in turn is connected to the vein of a rapid transport system onward to the rest of London. Course: BArch Architecture University: University Of Nottingham Country: United Kingdom Current Project: Whiteleys Year Start: 2018

182

Graduate Show 2019

183

184

Graduate Show 2019

185

The Ecstatic Void of Colour Daniela Andrade Salgado Silva

Course: MArch Architecture University: Cardiff University Country: United Kingdom Current Project: 425 Park Avenue, Citadel fit-out Year Start: 2018

186

Johann Wolfgang von Goethe believed that ‘light and darkness, light and its absence, are all necessary to the production of colour... Colour itself is a degree of darkness’ and thus lives within shadow. Colour provides an experience of indefinable pleasure and produces definite, specific states on the human mind. In the proposed architectural intervention we re-inhabit East London’s infrastructural artefact, The Bishopsgate Goods Yard, as a place of reflection and utter amelioration; a refuge characterised as a Bathhouse for Spitalfields. Inspired by the empirical works of James Turrell and Olafur Eliasson, its architecture will serve as a place for healing as a journey through the architectural sublime. Users will filter through the proposal, encountering the metaphysical power of colour’s influence onto the human condition.

GROUND FLOOR PLAN. @ 1.200

CROSS SECTION THROUGH EXISTING VAULTS + THE GOODS WALK.

3 Graduate Show 2019

187

BAFFLES OF ACCESSION

Vertical connection to the Goods Walk and Pool Hall.

INDOOR PURITY POOL

Submersion after the culmination of steps to amelioration.

4 POOL HALL

Repetition of coloured shadow.

188

ENERGY CHAMBER

Consumption of coloured steam.

3. 2.

10.

6. 8.

RADIANCE CHAMBER TECTONIC SECTIONAL PERSPECTIVE. 1. Wall Build up: 200mm reinforced concrete; DPM layer; 100 mm rigid insulation; vapour barrier; 100mm reinforced concrete; 30mm salt brick tiles 2. 350mm Concrete Parapet with Reflective pool on roof

1. Compression and Release Sectional perspective through the scheme.

8. Treated Larch timber plank walkway with steel support

2. Progression of Amelioration Ground Floor Plan of the Bathhouse. 3. Journey Through the Vaults - Long section through the Bishopsgate Goods Yard vaults. 4. Progression of Colour - Spaces filled with colour for healing. 5. Radiance Chamber - Technical investigation into one of the ‘floating’ chambers in the gardens. 6. Purity - External walkways over reflective pools and swimming pools.

6 Graduate Show 2019

189

Timber City: The revival of Hullâ&#x20AC;&#x2122;s industrial past Duncan Catterall

Course: MArch Architecture University: University of Westminster Country: United Kingdom Current Project: 6W 77th St Penthouse Year Start: 2018

Once the UKâ&#x20AC;&#x2122;s largest importer of timber, Hulls industry has seen extensive decline, further exacerbated by exceptionally high flood risks. The Timber City masterplan repurposes abandoned industrial land in the heart of Hull, converted into a series of urban wetlands, attenuating flood waters while the rich industrial heritage is preserved, celebrated and enhanced, starting a high-tec industrial revival, pioneering glulam and CLT technologies. At its centre, the Assembly Hall. The worldâ&#x20AC;&#x2122;s largest fully timber structure, encapsulating production lines for prefabricated timber buildings, pushing the boundaries to establish innovative, sustainable, building technologies, founded in Hull and exported to the world.

1. View from across one of the new public squares towards the assembly hall and timber research offcies. 2. Masterplan axo of proposed timber factory and new city centre. 3. Plan of the assembly hall. 4. View across urban wetlands and the completed masterplan proposal of a pioneering timber city. 5. 5. Looking down one of the new canals into the assembly hall where prefabricated structures are being assembled. 4 190

5 Graduate Show 2019

191

6 192

6. Dual Use of the assembly hall. As a hall for the assembly of prefabricated structures and when not in use, as a hall of assembly for the people of Hull. 7. Exploded structural diagram. Made from an entirely timber structure utilising glulam, clt and japanese joinery to include no metal connections.

Double curved timber window panel

Timber window frames follow double curved structural layout

Offset curved 125 degree squewed glulam beam

The lateral bracing beam is extended using pole tenon joints with splices

Dovetail splice joint used for ďŹ xing rail into base of mudsill for clean and strong join.

Offset curved 55 degree squewed glulam beam

Double intersecting glulam curved beams

Outer curved central glulam beam

Inner curved central glulam beam Timber lean to window framing

Collumn retention plates

Lean to window timber frame

Triple plug connection, comprising of central duck nest vertical member accepting short male, long male and female beams.

Quadruple goose neck frame connection

Tile support frame with integrated cross bracing

Secondary timber moment frame and working platforms

Tapered timber piles

Timber piles connecting to mudsill and mudsill bracing

Triple corner tenon joint to mudsill

Twisted timber tiles

Mini sliding tenon joint connects ďŹ xing rail. Bracing beam secured to column with wedged overlap in counter-sunk groove.

Graduate Show 2019

193

The Living Cell Fadhil Fadhil

The research focuses on the evolution of non-relational architectural objects. It is a process that contextualizes the fabrication process and decontextualizes the resultant architectural object ‘product’ This project develops the concept of manufacturing ‘the living cell’ a division of architectural objects into systematically prefabricated parts. Significantly, architectural objects can exist at a multitude of different scales. The first scale is the living cell scale, full building assemblies that are constructed through a series of parts. The second scale is the micro-object that is replicated, mass produced and assembled, generating tangible interior environments Course: Architecture MArch (ARB/RIBA Part 2) University: The Bartlett School of Architectures, UCL Country: United Kingdom Current Project: Greenville County Office Building Year Start: 2018

1 194

3 Graduate Show 2019

4 195

5 196

1. Drawing showing the accumulation of the micro object generating spatial interior environments. 2. Exterior perspective showing full building assembly. 3. Experiential interior perspective showing the characteristics of the interior environment 4. Perspective drawing showing an X-Ray of the exterior shell encapsulating the Interior assembly. 5. Exterior perspective showing full building assembly. 6. Exterior perspective showing full building assembly. 7. Drawing showing the fabrication method of generating the living cell.

7 Graduate Show 2019

197

SORT: Building as a Cultural Machine Heidi Han

Course: AADRL MArch in Architecture + Urbanism University: Architectural Association Country: United States Current Project: Hankook Headquarters, Aswan Hospital Year Start: 2018

SORT is a research project that speculates the future of culture through the investigation of a ‘building-as-machine’ – a ‘plug-and-play’ model. It explores a future where our environments could allow for participation of different multi-influential agents. It questions if a building, its contents, and occupants could all be active participants – having the ability to navigate through a dynamic infrastructure of decision making and organization. It challenges architecture as we know today: one that is mostly a passive entity, conceived, materialized, and occupied. It speculates if architecture could be a sorting machine - a dynamic warehouse that is constantly changing and responding. The project investigates the agency of three main elements: the agency of a building, the agency of the contents of a building, and the agency of the occupants or visitors of the building. Utilising the Centre Pompidou as a case study, the project proposes a new way of experiencing the future of culture through a cultural machine. The machine organizes our world where intelligent systems of the agents make decisions, and influence the outcome of others. The building transforms based on the properties and logistical rules of a mobile transformative unit prototype. The objects sort through computational curation and visitors can engage in a dynamic environment where architecture becomes infrastructure.

1 198

Graduate Show 2019

7 199

8 200

1. Interior View. 2. Iterative Arrangement on Pompidou Site. 3. Iterative Arrangement on a Hypothetical Urban Park Site. 4. Genetic Algorithm Sample Study. 5. Catalogue of Iterations of Building Arrangement Based on Genetic Algorithm Code. 6. Exploded Axo of Unit Prototype Parts. 7. Unit Prototype Transformation. 8. Interior Views. 9. Mobility Prototype Study Model. 10. Surface Simulation.

10 Graduate Show 2019

201

Collective Exchange: Inspiring Dialogue Through Social and Morphological Palimpsest Ola Urbanska

“We do not only exist in a spatial and material reality, we also inhabit cultural, mental and temporal realities. (…) Architecture is essentially an art form of reconciliation and mediation and in addition to settling us in space and place, landscapes and buildings articulate our experiences of duration and time between the polarities of past and future.” Juhani Pallasmaa, ‘Space, Place, Memory and Imagination’

Course: MArch Architecture University: Welsh School of Architecture, Cardiff University Country: United Kingdom Current Project: Bel Air Residence Year Start: 2018

Collective Exchange focuses on social inclusion of Palermitans, migrants and refugees, looking to establish itself as a platform for dialogue. It promotes collaboration and community empowerment, drawing connections between past, present and future through the programmatic agenda and its architectural translation. New moments of porosity re-embed the dilapidated cluster of buildings into the site and city, fusing the proposal with existing structure. Layers of social and morphological palimpsest are emphasised through an unorthodox heritage strategy and tectonic approach. They create opportunities for exchange whilst paying tribute to Palermo’s history of syncretism, trade and resilience.

2 202

EXPE R I E NCES

KNOWLE DG E

VALU ES & TRADITIONS

FU NCTIONS

PRODUCTS

SKILLS

Exchange of experiences as a process of bonding and coping with difficult events from the past.

Exchange of knowledge and intangible skills, eg. languages, literature, creative writing, etc.

Exchange of cultural heritage through performances, exhibitions and wider-public conversations.

Exchange and fluidity between the functions, supported by the layout and connections between them.

Exchange of fresh produce or items created in the workshops, and an ignition point for micro-businesses.

Exchange of practical skills, including making, crafting and building.

4 Graduate Show 2019

203

5 204

1. Section through the scheme showing the sequence of spaces: the exchanges between mass and void, light and shadow. 2. South elevation of the scheme illustrating integration with the surrounding buildings. 3. Heritage strategy axonometric defining retained parts of existing structure. 4. Types of exchange between the users. 5. Plan of the scheme highlighting daily activities. 6. Illustration of the south entrance to the Collective Exchange. 7. Illustration of the north entrance to the Collective Exchange.

7 Graduate Show 2019

205

Material and Place Sean Bailey

Course: MArch Architecture University: Welsh School of Architecture Country: United Kingdom Current Project: Innovation Pavilion Year Start: 2018

To many, medieval market towns in the UK conjure images of narrow alleys, small street markets, local pubs, parish churches and more. These places make up the social environments in many of these towns and associate with the familiar motifs found in the historic fabric that have occupied these towns for centuries. Perhaps, no motif encompasses many medieval towns more-so than the visually dominant exposed timber framed building. This is certainly true for the small rural town of Leominster. In Leominster, the world of traditional fabrication has been substituted for the globalized world of computerized, mass produced and standardized components and this has consequently resulted in vastly contrasting urban landscapes. It has been debated that it is within this process of making that the qualities of intricacy and the translation of beauty have disappeared from design. This project ultimately looks to achieve a continuity between the two worlds of tranditional and contemporary fabrication in the context of Leominsterâ&#x20AC;&#x2122;s central social environment.

1 206

OPTION 2 2

4 Graduate Show 2019

207

5 208

1. A Grammar of Sheet Steel. 2. Exploded Axonometric: Core Strucure. 3. Conceptual Spaces. 4. Sectional Perspective, Central gallery. 5. 1.2mm Galvanised Sheet Steel Structure 1:1 model. 6. Spatial Qualities Conceptual Model. A Structural Sheet Steel System. 7. 7. 2m x 1.8m Wall Mounted Topographical. Model of Leominster in Mild Steel.

7 Graduate Show 2019

209

‘Pilot Park’ Yannis Hajigeorgis

The project develops a hybrid typology of parkland, for the modern ex-industrial city. Characterised by innovation and manufacture, its littered with historic relics of its industrial past, some revitalised others preserved in honour connecting the future to the past. The proposal for the former Reckitt owned Morely Street takes inspiration from the city’s famous Ultramarine pigment industry, with a century of roots that stretch from industry and manufacture to philanthropy and public service. The site, that was once prosperous, offers a beacon...a Bastion of industrial pride within its northern setting. for the people of Hull. Course: MArch Architecture University: University of Westminster Country: United Kingdom Current Project: Hellenikon and Aswan Heart Foundation Year Start: 2018

1 210 P I L O T P A R K - c H Imn E y v I Ew “ c EL EBR A T I n G H U L L ’ S cO L O U R FU L P A S T ”

2 Graduate Show 2019

3 211

P ILOT P A RK - EA ST En TRA n cE GA TE “c ELEBRA TInG HULL’S c OLOURFUL P A ST”

4 212

1. Pilot Park Tower top view. 2. Pilot Park projection. 3. A celebratory venue. 4. A welcome path â&#x20AC;&#x201C; history re-birthed. 5. Meandering paths. 6. Pilot Park â&#x20AC;&#x201C; architectural birds eye.

6 Graduate Show 2019

213

214

Environnemental Engineering

Jingyi Chen Joey Aoun Margaux Coquelin

Double Skin Facade Structure Impact on Indoor PM2.5 and Thermal Performance of High-rise Residential Tower â&#x20AC;&#x201D; Case study in London Jingyi Chen

Course: MSc Environmental Design and Engineering University: University College London Country: United Kingdom Current Project: China Merchant Bank Year Start: 2018

This study explores different structures of Double skin facade (DSF) impact on IEQ and energy saving on a high-rise residential tower in London â&#x20AC;&#x201D; Norland House. Different structures of DSF include Window-box, Corridor, Shaft-box, and Multi-storey DSF. The result shows that Shaft-box and Multi-storey DSF are more efficient by reducing 40% heating demand and 30% overcooling risks, however, leading to 20% higher overheating risk compared to base case (no DSF). Furthermore, units on South facade always have lower heating load due to more solar radiation, while North facing units have lower PM2.5 concentration by the help of leeward negative pressure.

Environment Overview Summer Solstice Jun 21st

Norland House Equinox March/Sept 21st

Winter Solstice Dec 21st Dominate Wind: WWS, 4w.19m/s

Outdoor PM2.5 concentration: 17ug/m 3 Noise Level: 65-69dB

c 218

Dominate Wind: WWS, 4.19m/s

Norland House Details 22F

Roof U-value: 2.19 W/m2•K

18F

External window U-value: 1.67 W/m2•K G-value: 0.40

14F

10F

External Wall U-value: 2.15 W/m2•K

Floor U-value: 1.20 W/m2•K

Double Skin Facade Structures

There are four kinds of Double Skin Facade structures generally: Window-Box DSF Corridor DSF Shaft - box DSF Multi - storey DSF This study simulated the naturally ventilated DSF, in which the air movement is forced by vertical air density difference and wind pressure.

Window - Box (WB)

Shaft - Box (SB)

Graduate Show 2019

Corridor (CR)

Air inside DSF get heated by solar radiation, hot air with lower density floating to the top, and cold air with higher density dropping down to the bottom. The air movement can enhance indoor ventilation. When the DSF openings are closed, it can become a solar glazing zone, which can havest solar gain and heat indoor places.

Multi-storey (MS)

219

Study Workflow

.idf file

.csv file

• Sketchup is used to building Norland House and DSF geometry • Indoor openings are set as multizone connection through Energyplus plug-in • Completed multizone geometry is exported as .idf file.

• As per research, indoor PM2.5 source of residential house is mainly from cooking, which is 1.84e10 5 µg/hr (Hu et al., 2014), the deposition rate is 0.4h-1 (Howard-Reed, Wallace, & Emmerich, 2003). • Energyplus is used to set up thermal properties for building envelope, Indoor and outdoor PM2.5 level, etc. • London weather file is assigned • Run the simulation, results will be exported as a .csv file.

• Microsoft Excel is used to compare indoor PM2.5 concentration, thermal comfort and annual heating load of each options • A simple SASS Analysis is done to research analyse the correlation between orientation/storeys and variables.

Results and Discussion

Indoor PM2.5, ug/m3

Indoor PM2.5 Concentration

+13%

+13% +8%

+8%

WHO limitation

Base case

• As per WHO requirement, indoor PM2.5 concentration should be less than 25ug/m3. • Base case has lowest PM2.5 with 21ug/m3 • DSF options will increase indoor PM2.5 by 8% ~13%, which are still below WHO limitation.

Base model

Thermal Comfort

61%

Base model WB

68% Overheating

77% Comfort

78% Overcooling

75% 25

100 Percentage of time, %

220

• Free running season is from May to September, according to Adaptative model that the thermal comfort range should be 20~25 ˚C • Shaft-box and Multi-storey DSF can extremely improve thermal comfort percentage up to 75%, compared to 61% in base case • However, DSF also increases unexpected overheating risks

0 -0.2

Window-box

Corridor

Column

Multi-storey

Annual heating load-0.4 -0.6

Annual Heating Load, GJ

2000

SOUTHWEST

NORTHEAST

Figure 30: Average wind pressure on Southwest and Northeast facade -37% -39% -36% -40%

1500 even though annual mean ventilation is encouraged by SB and MS, annual mean However, PM2.5 concentration still increases. To penetrate the reason behind it, same typical design days January 1st and Jun 21st are analysed below with sample Unit 6 in 6th floor.

• London heating season starts from October to April, according to CIBSE guide A, heating setpoint for residential units should be 17-22˚C • Base case has highest annual heating load with 1900 GJ • DSF could help reduce annual heating load by more than 35%, especially Shaft-box and Multi-storey DSF could reduce up to 40%.

1000

4.6.2. Ventilation rate and PM2.5 concentration in typical design days 4.6.2.1 Heating design day 0

Figure 31 plots ventilation rate and PM2.5 councentration on Janurary 1stMS on hourly basis . Peak Base model WB CR SB values happen during cooking time bacause of the same window opening time set with cooking during heating season. It’s noticable that the base case has higher ventilation rate 0.7m3/s than DSF cases 0.6m3/s on this day , therefore leading to lower peak PM2.5 concentration Indoor air flow vs cases PM2.5(0.15ug/m3.) concentration: Typical winter day Jan 1st (0.12ug/m3) thanrate other

0.8

0.6

0.4

4.5.2.1 Heating design day

A typical winter day Jan 1st is presented to show relationship between the ventilation rate with PM2.5 concentration, unit 6 on 6th Floor is tested:

PM2.5 concentration in 6F U6_Jan 1st, ug/m3 2018 September

Ventilation rate in 6F U6_Jan 1st, Jingyi Chen m3/s

0.2

23:00

21:00

19:00

17:00

15:00

13:00

11:00

09:00

07:00

05:00

03:00

01:00

0 As 0Figure 25 shown, peak heating load happens during cooking time 09:00-10:00, 12:00-14:00 and 18:00-20:00 with 75, 42 and 40MJ separately because of kitchen window opening (set in Methodology chapter). DSF’s max-benefit (50% reduction) happens in the meantime since hot air from DSF transferring from base cavity BW CR inside CL through MS open kitchen base window BW duringCRdaytime. CL WhileMS 18:00 to 20:00 DSF cavity can be a thermal buffer reducing heat loss after sunset. Real-time solar heat gain chart on the left shows the same trend. DSF can increase solar gain from 0.5 vs heating load: Typical winter Jan 1st toSolar 3 MJgain at noon, contributing to heating loadday reduction.

Figure 31: Ventilation rate and PM2.5 concentration in 6F U6 on Jan. 1st

Hourly heating load in 6F U6_Jan. 1st, MJ

Solar heat gain in 6F U6_Jan 1st, MJ 80

41 40

0 Base

Box-window

Corridor

Column

01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00

Multi-storey

Base case

Box-window

Collumn

Multi-storey

Figure 26 and 27 plots thePM hourly heating load and indoor operative of 6F Annual heating load temperature Thermalprofile comfort 2.5 concentration Parameter unit 6 on June 21st. It’s noticeable from solar gain chart that SB and MS collect a great number R P R P R P of solar gain up to 18GJ, five times more than that of WB and CR, 10 times of that in base case. This large amount of solar gain leads to -0.32 continuous overheating starting from 12:00 P1_orientation 0.85precisely 0.00 0.00 -0.08 0.03 at noon to midnight as shown in Figure 27.

P3_DSF structure 20

0.92

-0.03

0.72

Solar heat gain in 6F U6_Jun 21st, MJ

-0.26

0.00

-0.37

0.00

• Double skin facade case has extremely higher solar gain than base case scenario, almost 5 times more. • Base case scenario has 80 MJ peak heating load, double of that in double skin facade. • Heating load reduction according to solar gain, Double skin facade leading to lower peak heating load because it help collect more solar gain.

A SPSS analysis is operated to test the correlation between unit orientation, floor and Double skin Facade structure and PM2.5 concentration, heating load and thermal comfort:

SPSS Analysis 4.5.2.2 Free-running design day

-0.01

A typical winter day Jan 1st is presented to show the solar gain with heating load, unit 6 on 6th Floor is tested:

Corridor

Figure 25: Solar heat gain and hourly heating load in 6F U6 on Jan. 1st

P2_floor

• Base case scenario has higher indoor airflow rate than the Double skin facade cases. • Base case scenario has peak PM2.5 concentration as 30 ug/m3, which is lower than Double skin facade cases as 50 ug/m3. • PM2.5 concentration reduces according to indoor air flow rate, Double skin facade leading to more PM2.5 concentration because it reduces natural ventilation.

-0.13

0.08

-0.16

0.03

• P1_orientation correlates with PM2.5 concentration most, Northeast façade units have lower PM2.5 concentration than those on Southwest façade. • Orientation and structure both have significant medium strong association with annual heating load (R=-0.319, P=0<0.05 and R=-0.366, P=0<0.05), as Southwest façade has lower heating load, meanwhile SB and MS help reducing more. • DSF structure is the main parameter with slightly negative statistical association with R=-0.155 (P=0.032<0.05), corresponding to the finding that WB and CR better improve thermal comfort.

10 5 0

Graduate Show 2019

221

Cellular Approach to Energy Efficient Cities: A Bespoke Modelling Framework Joey Aoun

Course: MSc Urban Energy Technology and Policy University: Newcastle University Country: United Kingdom Current Project: New Phnom Penh International Airport Year Start: 2018

Data Preparation Model Occupancy at OA Level

The research approaches the city as a living organism. An organism defined by an assembly of urban cells. A cell scale which is large enough to identify patterns of energy consumption and supply beyond the boundaries of a single building, but small enough to address tangible solutions. The research output is a comprehensive energy modelling framework to empower planning teams with methods and tools necessary to improve the efficiency of cities in early design phase. The framework was then applied on the Ridgeway (new) area, a low voltage substation, in Newcastle Upon Tyne, the United Kingdom. The recorded output was then validated against open-sourced empirical data and peer energy models. The framework spans not only different disciplines, but also the levels and system boundaries of analysis, i.e. Individual to aggregated consumption patterns and behaviour, the element of time to dynamically link demand and supply, and the element of space to connect the building to its urban context. This study is one step closer to integrating engineering simulation and planning methods in a multi-scale urban energy workflow.

Pre-simulation Model

Main Simulation Model Estimated Number of Occupants per Property

Archetypes Characterization

OS Address Base Premium

OS Mastermap Building

Address Matching

Height

OS Mastermap Building Polygon

Polygons Simpliﬁ cation

2.5D Model Generation

Masses to thermal Zones

Applied Thermal Zones Characteristics

Simulation Settings Setup

Model Simulation

Analysis Model

OS Mastermap Complementary Layers

UPRNs Modelled Heating Domestic EPC Data

EPC Variables Simpliﬁ cation

Reﬁ ned Domestic EPC Database

Non-Domestic EPC &

EPC & DEC Variables

DEC Data

Simpliﬁ cation

Reﬁ ned EPC & DEC Database

& Electricity Consumptions

Match Collected Data to UPRNs Model Validation

Compiled Rating List & SMV Data

Local weather Data

Reﬁ ned Dataset

Climatic & MicroClimatic Conditions

Empirical Data & Peer Energy Models

1 222

1. Energy modelling framework. 2. State of the art facts and figures associated with the EU energy market.

3. Research philosophy: Local area energy planning. 4. Energy modelling approaches identified from an in-depth literature review & analysis of peer urban building energy models.

$ Country Performance (%) Top 25

25-50

50-75

<25

N/A

*World Energy Trilemma Index 2018 Regional Overviews

The EU Energy ‘Trilemma’ • Efﬁ ciency First • Energy Security • Energy Affordability • Environmental Impact

Demand-response

Envelope retrofi t

Renewable potential

HVAC systems retrofi t

2 The city seen as ...a living organism

...an assembly of urban cells.

A cell scale large enough to identify patterns of energy

..but small enough to address tangible solutions.

Bottom-Up vs Top-Down

Individual Building data vs Archetypes

Power Network & Urban Buildings Integrated vs in isolation

Occupant Behaviour & Energy Simulation Integrated vs in isolation

Domestic & Non-domestic Mixed-use vs in isolation

Urban Building & its Urban Context Integrated vs in isolation

4 Graduate Show 2019

223

Urban fabric layers

Building geometric data

Power Network & Infrastructure

Weather Data

Urban Fabric Layers

Investigate Site Potentials

Number of UPRNs

HVAC System

Building Fabric Characteristics

Window to Wall Ratio

Energy Schedules

Hot Water Use

Shape fi le

Building nongeometric data

Occupancy data and schedules

XML

nomis Ofﬁ cial labour market statistics CAMBRIDGE ARCHITECTURAL RESEARCH LIMITED

Empirical Data & Peer Energy Models

Number of Occupants

Identify patterns of use

XML

Buildings Height

SHP

Buildings polygons

Energy modelling interface

Geometry modelling interface

1-2-3

Energy simulation engine

CSV

2.5D Model

energy plus

Results

1- Identifying non-energy consuming UPRNs. XML

2- Buildings’ polygons simpliﬁ cation.

Urban Building Data

3- Matching UPRNs with buildings’ polygons.

Model Calibration

h2 h1 + h2 2

5- Floor/ glazing generation: • Identifi cation of thermal zones per fl oor; • Assigning window to wall ratio identifi ed per building archetype.

3 224

Weather Data

4- 2.5D roof height extrusion.

Non-geometric data*

Name

Roof

Facade

Slab

Partition

Ground

Adiabatic

Default

Construction layers+ Uvalue

True - False

Priority Zone

Surface Refl ection

Comment

Data Source

True - False

0 - 0.5 - 1

Construction layers+ Uvalue

Thermal mass True - False

* A sample from the XML input ﬁ le: Zones construction data input for a default unit

12000

Results

4000

base 8000 7000

4000

6000

2000

5000

2000

8000

8000 7000

7000

4000 3000

4000

2000 400.00

3000

2000

2000 1000 350.00

4000

3000

2000

1000

1000 0

1000

0 0 0 DHW Electric Equipment Heating Load Zone Lights ElectricDWH Equipment Heating Load DWH Results Electric EPC CHM data (14944) Equipment Heating Zone Load Lights Zone DWH Electric LightsEquipment Electric Equipment Heating Load Heating Load DWH Heating Load Results EPC CHM data (14944) Electric Equipment Results EPC CHM data (14944) Results EPC Results CHM data EPC (14944) CHM data (14944)

Zone Lights

DWH

Zone Lights 4510048537) Grand Total

Validation of the recorded output for one of the units (UPRN in Ridgeway (new) neighbourhood in comparison with Cambridge Housing Model (CHM) & Energy Performance Certifi cates (EPCs) data

200.00

150.00

100.00

% of error when comparing the simulated output against ...

50.00

Mar

Avr

1st Qtr

0.00

2nd Qtr

Aug

-15% Sep

Oct

... EPC Data

Nov

Dec

+16% 1st Qtr

2nd Qtr

... CHM Data (14944)

500.00

3.00

450.00 2.50

500.00 500.00

300.00

500.00 500.00

250.00

450.00

500.00

450.00 450.00

200.00

400.00

350.00

Grand Total

DHW Heating Load Electric Equipment Zone Lights Grand Total

1.50

350.00 0.50

300.00 300.00 DHW 350.00 300.00 DHW Heating Load 250.00 Energy consumption Breakdown 300.00 DHW Heating Load Electric Equipment 0.00 for the month of May for the 250.00 Nov Jan Feb300.00 Mar Avr May Heating Jun Load Jul Aug Sep Equipment Oct Dec DHW Electric Feb Mar Avr May Jun Jul Sep OctZone Lights Nov Dec Ridgeway New neighbourhood 250.00 Aug Electric Heating Load FebDHW Mar Jun Equipment Jul Aug Zone Lights Sep Oct Grand Total Nov 200.00Dec Dec Mar Avr May Feb Avr May Jun Jul Aug Sep Oct Nov 250.00 Zone Electric Equipment Total HeatingMar Load FebMonthly Avr May recorded Jun Lights Jul Ridgeway Aug Grand Oct Nov Dec 200.00 energy consumption for the (new)Sep neighbourhood 250.00 Grand Total 200.00 Electric EquipmentZone Lights 150.00 Grand Total 200.00 Zone Lights 150.00 Grand Total 200.00 150.00 100.00 150.00 100.00 150.00 Graduate Show 2019 100.00 50.00 100.00 50.00 100.00 50.00

50.00

350.00

2.00

1.00

400.00 400.00

100.00

Zone Lights 400.00 Grand Total

400.00

450.00 150.00

Jan Jan Jan Jan

DHW DHW DHW Heating Load Heating Load DHW Heating Load Electric Equipment Electric Equipment Heating Load Electric Equipment Zone Lights DHW 450.00 Zone Lights Electric Equipment Zone GrandLights Total Heating Load Electric Equipment Grand Total Zone Lights Grand Total

350.00

ENERGY CONSUMPTION (KWH)

400.00

100.00 50.00 50.00 50.00

50.00 0.00 0.00 0.00

Jul

ENERGY CONSUMPTION (MWH)

200.00 150.00 150.00 150.00 150.00 100.00 100.00 100.00

Jun

ENERGY CONSUMPTION (MWH)

250.00 200.00 200.00 200.00

May

... NEED Data

450.00 400.00 400.00 400.00 400.00 350.00 350.00 350.00 350.00 300.00 300.00 300.00 300.00 250.00 250.00 250.00

+11%

ENERGY CONSUMPTION (MWH)

500.00 Feb 500.00 500.00 500.00 450.00 450.00 450.00

ENERGY CONSUMPTION (MWH)

Jan

ENERGY CONSUMPTION (MWH)

0.00

ENERGY ENERGY CONSUMPTION CONSUMPTION (MWH) (MWH) ENERGY CONSUMPTION (MWH) ENERGY CONSUMPTION (MWH)

ENERGY CONSUMPTION (MWH)

4000 5000

6000

250.00

5000 Electricity Consumption

6000

5000 450.00

5000

Gas Consumption

8000

300.00

Electricity Consumption

7000 Validation of the recorded output for the whole Ridgeway (new) neighbourhood in comparison with the National Energy Effi ciency Data (NEED) data 6000 6000

500.00

3000

+16%

8000

Gas Consumption

7000

ENERGY CONSUMPTION (MWH)

4000

... Decoding the energy language of today’s urban building stock, to investigate potential energy scenarios for the city of tomorrow.

6000

8000 6000

‘

Energy Results Zone Lights NEED Electric Equipment Heating Load Use Intensity DWH 15 to 250 KWh/sqm shown on this scale Results EPC CHM data (14944)

8000

10000

NEED

+11%1000 base

10000

-55% base

2000 12000

+106%

7. A sample of the simulated model output emphasising its spatiotemporal resolution.

+3%

4000

-4% base

5. Results Validation against opensourced empirical data & peer energy models.

+40% unknown base

6. Recorded energy output, at a specific hour during the day, visualized for the Ridgeway (new) neighbourhood.

3. Model generation: A step by step integrative modelling process. 3000

-32%

4. Model pre-simulation: Energy modelling workflow.

2. Model characterization: 5000 Data input aggregated at different scales.

-35%

7000 1. Data processing: Open-sourced urban building and power network 6000 data.

base

8000

Electric Equipment

Zone Lights

Heating Load

DHW

DHW DHW Heating Load DHW Heating Load Electric Equipment Heating Load DHW Electric Equipment Zone Lights Electric Equipment Heating Load Zone Lights DHW Grand Total Zone Lights Equipmentrecorded Grand Total 4510048537) in Heating HourlyLoad energyElectric consumption for one of the units (UPRN Grand Total neighbourhood Electric EquipmentZone Lights Ridgeway (new) Grand Total Zone Lights

0.00

Grand Total

225

Understanding the City of Tomorrow through a Holistic Urban Approach Margaux Coquelin

Course: MSc Urban Engineering University: EIVP - Paris School of Engineering Country: France Current Project: King Salman Park Year Start: 2018 1

‘By 2020, every Parisian will be within a seven minute walk from a green space’ states the Paris Climate Plan. The Urban Cool Island project works towards achieving this target by developing a study and a state of art to provide guidance principles for the City of Paris. The second project ‘Place de l’Eglise Charonne’ links both urban planning and sociology sciences to understand what makes a public realm successful by understanding the population needs and expectations through observations and interviews. 2

4 226

1. Effect of Urban Heat Island. 2. Daytime and night-time effect of Urban Heat Island. 3. 2003 European heat wave.

5. A combination of solutions for Urban Cool Island. 6. Paris Urban Cool Island Map and App.

4. Paris smog air pollution.

6 Graduate Show 2019

227

7 228

±±

1:500 1:500

e n nnaro aroCh Chde de rc rc Pa Padu u in dm in e mh he C

7. Place Charonne context and observations.

Diagnosis and analysis of Etatdes deslieux lieux Zone d'étude Etat Zone d'étude the plazaCharonne Charonne

8. Diagnosis and analysis of the plaza. 9. Planning intentions.

Cimetière de Charonne Cimetière de Charonne

Caractérisation l'espace Context Caractérisation dede l'espace

Green space Espace Végétalisé Espace Végétalisé

Church garden Jardin de l'Eglise Jardin de l'Eglise

! ^

Wall

Zebra crossing Passage Piéton Passage Piéton Banc Bench Banc

Eglise Saint-Germain de Charonne Eglise Saint-Germain de Charonne

Tree Arbre Arbre

Flower pot de Fleurs PotPot de Fleurs MurMur Escaliers Stairs Escaliers

Key factors affecting the uses

Des facteurs influencants usages Des facteurs influencants lesles usages Ru

Ru St e

Zone fortement ombragée Shaded area Zone fortement ombragée

St en dh nd al ha l

Car park Zone de Stationnement Zone de Stationnement Retail Linéaire Commerces Linéaire Commerces Point de point vue View Point de vue

de Rencontre Common meeting point de Rencontre ÿ ÿLieuLieu o

Disparity of uses Une disparité des flux des usages Une disparité des flux et et des usages

e lais t-B e ain lais eS t-B ain eS

t ole gn g n Ba Bade de e e Ru

let

Médiathèque Marguerite Duras Médiathèque Marguerite Duras

Piétons Pedestrian FluxFlux Piétons Automobile Car flow FluxFlux Automobile

flow

Zone de conflit Conflict ofd'usage use Zone de conflit d'usage

10 5

10 Mètres 10 Mètres

Margaux COQUELIN Margaux COQUELIN Victor DEBELHOIR Victor DEBELHOIR Hannah GOLDBERG Hannah GOLDBERG Clara LE COCQ Clara LE COCQ Yang LI Yang LI

±± 1:500

1:500

min

du C

e e nn nn ar o ar o Ch Ch de de rc rc Pa Pa du in m he

Planning intentions Plan Plan d'objectifs d'objectifs Place Place de de l'Eglise l'Eglise Charonne Charonne Cimetière de Charonne Cimetière de Charonne

Caractérisation Caractérisation de l'espace de l'espace Context

Vegetation Espace Espace Végétalisé Végétalisé

Church garden JardinJardin de l'Eglise de l'Eglise

Tree ! Arbre Arbre Wall Mur Mur

Shaded area Zone fortement Zone fortement ombragée ombragée

View Point de Point vue de vue

Retail Linéaire Linéaire Commerces Commerces

Passage Passage PiétonPiéton Zebra crossing

Eglise Saint-Germain Eglise Saint-Germain de Charonne de Charonne

Legibility, accessibility diversification Une Une modification modification de l'espace... deand l'espace... Road De la De Voirie la Voirie

30 30km/h km/h Zone Zone 30km/h Ru

St e

Ru nd

area

Zone de Zone ralentissement de ralentissement Speed bump St e

Ralentisseur Ralentisseur Trapézoidal Trapézoidal Zebra crossing

Park Du Parc Du Parc

Public spaces De l'Espace PublicPublic De l'Espace

Escaliers Escaliers Stairs

TerrainTerrain de Pétanque de Pétanque

ole

t-B

lais

Flower pot ^ Pot dePot Fleurs de Fleurs

Planters Bac à Bac fleurs à fleurs Bench Banc Banc

( !

Médiathèque Médiathèque Marguerite Marguerite Duras Duras

( !

ain eS

ole

Petanque strip Zone de Zone calme de calme Quite area Mur Végétalisé Mur Végétalisé Green wall Arbre Arbre élagué élagué Pruned tree Elevateur Elevateur Disabled access

Redefining the uses des des ... pour ... pour une une redéfinition redéfinition usages usages

Car flow Flux Automobile Flux Automobile Flux Piétons Flux Piétons Pedestrian

5 10

10 Mètres 10 Mètres

flow

Rencontre de Rencontre Common meeting point ÿ Lieu ÿ deLieu MargauxMargaux COQUELIN COQUELIN Victor DEBELHOIR Victor DEBELHOIR Marché Marché aux Fleurs aux market Fleurs Flower

Hannah Hannah GOLDBERG GOLDBERG Clara LEClara COCQ LE COCQ Yang LI Yang LI

9 Graduate Show 2019

229

230

Structural Engineering

Stavros Tseranidis

283

S. Tseranidis et al. / Automation in Construction 72 (2016) 279–293 Summary of models, nuisance parameters and MATLAB functions considered.

1 2 3 4 5 6 7 MATLAB

NN Table 2 Summary of models,Number nuisanceof parameters neurons and MATLAB functions considered. 6:6:30

Machine Learning for Structural Design Stavros Tseranidis

1 2 3 4 5 6 7 MATLAB

1 2 3 MATLAB

1 2 3 MATLAB MARS

1 2 MATLAB

Number of layers NN Training function Number ofchecks neurons Maximum validation Number of layers Internal ratio of training data Training function Internal ratio of validation data Maximum validation checks Internal ratio of test data Internal ratio of training data feedforwardnet Internal ratio of validation[34] data RBFN

1:2 trainlm 6:6:30 10 1:2 0.8 trainlm 0.2 10 0

Internal ratio of test data feedforwardnet [34]

Mean squared error goal RBFN Spread of radial basis functions Mean squared error goal Maximum number of neurons Spread of radial basis functions newrb [36]

{cubic, linear} 10:10:40

Maximum number of functions ARES Lab [38]

Course: MSc Computation for Design and Optimization University: Massachusetts Institute of Technology (MIT) Country: USA Current Project: Lusail Towers / Santa Clara Year Start: 2018

This research explores the use of machine learning and data-driven approximation algorithms in the early-stage design of structures. The use of machine learning models to rapidly evaluate design performance can lead to a more in-depth exploration of a design space and reduce computational time of optimization algorithms. Original contributions include novel metrics and visualization techniques for understanding model error and a new robustness framework that accounts for variability in model comparison. These concepts are applied to several case studies including a multi-objective airport terminal design case study which, importantly, considers both structural material volume (embodied energy) and operational energy consumption.

10:10:300 6 0.8 10:10:300 True 6 5

Bootstrap sample size ratio Sample with replacement Minimum leaf observations

0.8 True 5

TreeBagger [35]

0.5:0.5:4

600

newrbe [37]

Machine{exponential, Gaussian, linear, spherical, spline} Learning DACE Toolbox [39] generation of accurate DATA new data performance is the main objective of

Correlation function

they studied, while Jinalong et al. [27] R ,forRAAE, and generation of accurate new data performance the main objective of RMAE. All ofmodel thesecomparisons error metrics are summarized withuses their the construction the surrogate model in the firstisplace. RMAE. All ofstands these for error metrics deviation). are summarized along with their for- Several the construction the to surrogate model in the first place. mulas in Table 1 (STD standard techniquesof exist estimate the generalization error and are mulas inthat Table 1 (STDastands for standard deviation). techniques to[15] estimate the generalization error and are Error measures provide more direct and comprehensive quanexplainedSeveral in detail in Queipoexist et al. and Viana et al. [30]. The simplest measures metric that provide a more direct and comprehensive explained in detail in Queipo al. [15] and Viana al. [30]. The simplest titative model Error performance do not exist; some alternative ap- quanmethod is to split a given datasetetinto train and testetdata, construct the titative model performance metric do not exist; some alternative apmethod is to split a given dataset into train and test data, construct the proaches to address this limitation are presented later in this paper. model with the train data, and then compute the error in the test data proaches to address this limitation are presented later in this paper. model with the train data, and then compute the error in the test data Special focus is also given to the visualization of the results, as normaland take this as an estimate of the generalization error. Another techSpecial focus is also given to the visualization of the results, as normaland take this as an estimate of the generalization error. Another techization and visualization are problem specific and can have a significant nique is called cross-validation (CV), in which the original dataset is ization and visualization are problem specific and can have a significant nique is called cross-validation (CV), in which the original dataset is impact on how error is understood. split and into parts and the model is(image trained all all thethe parts except Fig. 2. Seven-bar truss (a) variables (b)kanalytically computed space from with [14]).with impact on how error is understood. split into k parts anddesign the model is trained parts except one, which is used as the set of previous case. The procedure one, which is used astest the test setthe of the previous case. The procedure is thenisrepeated until until eacheach one one of the k sets hashas served as as the test set. then repeated of the k sets served the test set.models problems. One of the first attempts of this kind in the field of structural different models is shown in Fig. 3. Details on the approximation 2.4. Robustness surrogate modeling By taking the mean of theoftest set a more robust generalization 2.4. in Robustness inengineering surrogate modeling By taking mean testerrors, set in errors, a more was by Schmit and Miura [7] in a NASA report in 1976. A the used later arethe presented Section 2.2. robust generalization error estimate is produced. Another advantage to CV is that a measure error estimate is produced. Another advantage to CV is that a measure review of the application of approximation methods in structural engiIt is also noteworthy that surrogate modeling is used extensively in As mentioned, previously mentioned, it for is crucial for a surrogate modeling the variability the error be obtained taking, example, As previously it is crucial a surrogate modeling appli- appliof the of variability of theoferror can can be obtained by by taking, forfor example, neering was published by Barthelemy and Haftka in 1993 [8]. The the aerospace industry. The basic principles remain the same across discation to have an acceptable generalization error. Thistorefers to an error the standard deviation of computed the computed k test errors. If this procecation to have an acceptable generalization error. This refers an error the standard deviation of the k test setset errors. If this procemethods inIn this review paper aredata response methodolciplinesthe since thenumber methodology relies on data.ofQueipo et al. [15] of on thenew model onexplored new data. this context, new meansurface data is repeated of times as solely the number samples estimate of estimate the model data. In this context, new data mean data dure isdure repeated the samesame number of times as the number of samples ogy (RSM) and neural networks (NN). The paper predicted that addihave made a thorough overview of the common practices of surrogate samples thatbeen haveused not been used at any point in the construction the original dataset, which means only a single sample used samples that have not at any point in the construction of the of the in the in original dataset, which means thatthat only a single sample is isused methods would emerge and the the practice would expand. Today, modeling. Theythen alsoerror applied those techniques inthe an MOO problem is tional the most important error required since the rapid every time theset, testthen set, the error is called PRESS, and method model. Thismodel. is theThis most important error required since rapid every time as theas test the is called PRESS, and the method although the computational power has increased exponentially from from the aerospace industry. Another comprehensive survey of blackLarge integer optimization problems are NP-complete, therefore Machine Learning is required 1 20 years ago, the engineering problems that modern designers face box approximation method for use in high-dimensional design probhave also increased dramatically in scale, and therefore surrogate lems is included in [16]. modeling has been studied and applied extensively. Other research has attempted to integrate performance evaluation In 1992,1Hajela and1 Berke [9] contributed an overview of the use of into parametric design in architectural and civil structures. Mueller neural networks in structural engineering problems. They noted that and Ochsendorf [17] considered an evolutionary design space explora1 1.13 1.18 1.11 1.02 1 this approximation technique evaluation tion,1.22 Shi and Wang [18] examined performance-driven design from an 100 kNcould be useful in the rapid 4.0 m 1.13 1.22 1.18 1.11 1.02 100 kN 4.0 m 3.0 m of simulations such as nonlinear structural analysis. Neural networks architect's perspective, while Granadeiro et al. [19] studied the integra2.0 m 2.0 m 3.0 m potential for use in 2.0 m approximation models 2.0 m and still have significant tion of energy simulations into early-stage building design. Each of X3 this field today,Xsince nonlinear structural analysis is frequently perthese approaches could benefit from the use of surrogate modeling, 3 formed. Researchers have also been using approximation algorithms which is the main contribution of this paper. 1.0 m in structural engineering for the dynamic properties and response eval1.0 m 4.0 m uation and optimization of structures [10], seismic risk assessment [11], 2.2. Model 1.01 1.29 types 0.98 0.91 1.07 X1 0 1 and energy MOO 4.0 m are extensively X2 simulations [12]. Energy simulations 1.01 1.29 0.98 0.91 1.07 1.0 m X 4.0 m paper, since they can be extremely computationally Several additional methods have been developed over the years to 0 1 1 examined in this X2 0 1.0 m expensive and have become more important and widely used in the deapproximate data and have been used in surrogate modeling applica4.0 m 0 sign of contemporary buildings and infrastructure. tions. Common methods include polynomial regression (PRG) and re0 The use of approximation algorithms in conceptual structural and sponse surface methodology (RSM) [20], in which a polynomial 0 architectural design was also examined by Swift and Batill in 1991 function is fitted to a dataset using least squares regression. These 1.12 1.29 1.06 0.96 [13]. This research focused on truss problems, with the nodal positions methods have been used in 1.17 many engineering problems, including 2 as variables and structural weight as the objective, while sampling a those given in [15]. (a) (b) 1.17 1.12 1.29 1.06 0.96 design space and later approximating it using neural networks. One of the most widely used surrogate modeling method in engiA similar approach was followed by Mueller [14], with a seven-bar neering problems Kriging (KRIG). Since it was formally established (a) (b) iswith Fig. 5. Seven-bar truss topology and design variables (a) (Image from [14]) sampled designs score (b). truss problem examined as shown in Fig. 2a. The variables were again in its current form [21], it has been applied extensively ([11,15,22]). In the positions of the truss nodes, specifically the vertical positions as sampled relateddesigns work, Gano et al. (b). [23] compared Kriging with 2nd-order polynoFig. 5. Seven-bar topology and design variablesnodal (a) (Image from [14]) with score shown along with their ranges in Fig. 2a. The design space (with the mial regression. Chung and Alonso [22] also compared 2nd-order RSM structural weight as the objective score) computed analytically, without and Kriging for an aerospace case study and concluded that both models approximation, is shown in Fig. 2b. The approximated design space for performed well and are a realistic methodology for engineering design.

1 variable 50 discrete values 50 design combinations 2 variables 50 discrete values 1,225 design combinations 3 variables 50 discrete values 19,600 design combinations 4 variables 50 discrete values 230,300 design combinations 5 variables 50 discrete values 2,118,760 design combinations 6 variables 50 discrete values 15,890,700 design combinations 7 variables 50 discrete values 99,884,400 design combinations 8 variables 50 discrete values 536,878,650 design combinations 9 variables 50 discrete values 2,505,433,700 design combinations 10 variables 50 discrete values 10,272,278,170 design combinations

S. Tseranidis et al. / Automation in Construction 72 (2016) 279–293

281

Surfaces generated from machine learning algorithms

Theoretical surface

Source: C.T. Mueller, “Computational Exploration of the Structural Design Space”, MIT, 2014

Fig. 3. Seven-bar truss approximated design space for different parameters (image from [14]).

4 234

281

0.5:0.5:4

Degree ofKRIG polynomial regression function {0,1,2} Correlation function {exponential, Gaussian, linear, spherical, spline} {cubic, linear} Degree of polynomial regression function {0,1,2} DACE Toolbox [39]

Traditional Scientific Method10:10:40 THEORY model comparisons they studied, while Jin et al. [27] uses R , RAAE, and 2 MATLAB

0.8 0.2 0

Number of trees Number RF of variables to sample Bootstrap sample size ratio Number of trees Sample with replacement Number of variables to sample Minimum leaf observations

RBFNE TreeBagger [35] S. Tseranidis et al. / Automation in Construction 72 (2016) 279–293 0.005 Spread of radial basis functions RBFNE 0.5:0.5:4 0.005 Spread of radial basis functions 600 0.5:0.5:4 newrbe [37]

Maximum number of neurons newrb [36] KRIG

Piecewise function MARS type Maximum number of functions 1 Piecewise function type ARES Lab [38]

Fig. 2. Seven-bar truss (a) variables and (b) analytically computed design space (image from [14]).

problems. One of the ﬁrst attempts of this kind in the ﬁeld of structural engineering was by Schmit and Miura [7] in a NASA report in 1976. A review of the application of approximation methods in structural engi-

different models is shown in Fig. 3. Details on the approximation models used later are presented in Section 2.2. 18:extensively Design samples of Seven Bar truss GH tool It is also noteworthy that surrogate modelingFigure is used in Figuregenerated 18: Design by samples of Seven Bar truss generated by GH tool

RF RF

best best

RBFN RBFN

best best

performance performance

RBFNE RBFNE

other engineering disciplines but have not been studied and applied extensively in the fields of civil and architectural engineering. This research investigates surrogate modeling techniques and evaluates Machine learning enables 4. Theoretical design space of seven6. Many different curves can fit the them optimization over large number of bar truss (first from left) and same data. case studies. is given to the development of a variables/parameters. on various related approximations of the sameFocus using 7. For multiple variables, many different machine learning (rest). holistic framework that is generalizable. multi-dimensional Fig. 1 displays core Seven-bar truss problem formulation surfacesacan fit concept SingleRun Run (left) and design combination 5. Seven-bar truss design variables the same data. Single in surrogate modeling. The circles represent the available data, with performance scores (right). performance scatter plots (top) and machine learningbeing model comparison many different models able to fit them. The art in surrogate Design space exploration. (down). modeling is to choose the one that will also fit new data well.

best best

performance performance

Seven-bar truss parametric performance plots S. Tseranidisscatter et al. /S.Automation in al. Construction 72 in (2016) 279–293 Tseranidis et / Automation Construction 72 (2016) 279–293

1.15

1.1 1.1 1.05

Structural score Structural score

1.15

1.1 1.1

1.05 1.05

Name EXP GAUSS LIN SPHERICAL CUBIC SPLINE

1 1 -0.5 -0.5

1 1 0 -0.5 0 -0.5 x1

0.5 0 0 0.5 x1

train models

1.15

1.1 1.1 1.05 1.05

select

1.15 1.15

1.1 1.1

1.05 1.05

performance

1 1 0 -0.5 0 -0.5 x2 assess assess performance performance of1.15 chosenmodel model of chosen

0.5 0 0 0.5 x2

1 1 -0.5 -0.5

0.5 0.5

1 1 0 -0.5

0.5 0 0.5 0 x3

0.5 0.5

Different fi t curves 1.15 1.15

1.1 1.1 1.1 1.1 9. Single-run Single-run workﬂow. workﬂow. Fig. 9. 1.05 1.05 1.05 1.05

1.1 1.1

1.05 1.05

1.15

285 285

0 -0.5 x3 x3

1.15

parameters

1.1 1.1

1.15 1.15

compute

1.05 1.05

performance performance

1.05 1.05 performance

best 1 best 1 -0.5 0.5 -0.50.5

train modelsscore compute select Predicted Structural Predicted score Structural for prediction best model model Predicted score for different different Structural prediction best 1.15 score 1.15 Predicted Predicted Structural score Structural nuisance error nuisance 1.15 error

Structural Structuralscore score

Structural score Structural score

Actual Structural score Structural score best Actual Structural score Actual best 1.15 1.15score Actual Structural score Structural 1.15 Actual

1.05 KRIG KRIG

Structural score Structural score

1 2 3 4 5 6

285 285

S. Tseranidis et al. /S.Automation Construction 72in(2016) 279–293 Tseranidis etinal. / Automation Construction 72 (2016) 279–293

MARS MARS

1.15 1.15

1000 10

f(x)

order to to make make predictions.Therefore, Therefore,the themain main evaluationmetric metricofofaa order predictions. evaluation 1 1 1 1 1 1 1 well 1 1 1 model 1 at model is is how how well itit0.5 performs atmaking makingpredictions predictions onnew newdata, data,or or performs on 0 0.5 -0.5 0 -0.5 0.5 0 -0.5 0.5 0 -0.5 -0.5 0.5 0 -0.5 0.5 0 -0.5 is to0.5 0 the 0.5other -0.5 how 0 itit generalizes. 0.5 -0.5test 0 the -0.5 0 -0.5 0 0.5 -0.5 0 0.5 in other words, how well generalizes. The test set has role0.5 When the the in words, well The set has ofof Whenan anapproximation approximation model model is developed, goal make x1 x2 x3 therole x1 x2 x3 x1 x2 x3 x1 x2 in x3 data “new” data data in aa way waythat thatititcomprises comprisespristine pristine datathat thatwere werenot notused used model in “new” model “learn” “learn” aa speciﬁc speciﬁc physical physical process or numerical simulation in Seven-bar truss machine learning model comparison Fig. 7.RBFN, Seven-bar truss, variablescatter vs. performance plots—test set. Fig. 7. Seven-bar truss, variable vs. RBFN, performance plots—test scatter set. 4.1. 4.1.Understanding Understanding error error and and visualization visualization 1

Fig. 7.RBFN, Seven-bar truss, variablescatter vs. performance Seven-barFig. truss7. | Structure Seven-bar truss, variable vs.RBFN, performance plots—testscatter set. plots—test set. 10 1022 9.9e-01 9.9e-01

AverageAverage training time [sec] Average time [sec] time [sec]

common sampling schemes, included in the GH sampling tool,3.4. are Ran- example 3.4. Design example common sampling schemes, included in theincluded GH sampling are Ran- tool, common sampling schemes, in thetool, GH sampling areDesign Ran- example 3.4. Design example common sampling schemes, included in the GH sampling are Ran3.4. Design dom, Grid, and Latin Hypercube [40].scheme Thetool, Random refers to dom, Grid, anddom, LatinGrid, Hypercube [40]. The Random refersscheme to -1and Latin Hypercube [40]. The Random scheme refers to dom, Grid, andsampling Latin Hypercube [40]. The Random scheme refers to 10-1 10 random values (commonly uniformly distributed) for each A design example of a parametric seven-bar truss is introduced 11 sampling random valuesrandom (commonly uniformly each10 design example of a example parametric truss is introduced 1.8e-01 distributed) 1.7e-01 (commonly sampling values uniformlyfor distributed) forA A design of aseven-bar parametric seven-bar truss is introduced sampling random values (commonly uniformly distributed) for each10 all combinaAeach design example ofisaused parametric seven-bar truss istointroduced 14.53 14.53 variable within a specified range. The Grid scheme samples here and in the following sections illustrate the presented variable withinvariable a specified range. The Gridrange. scheme all combinahere and is used theis following sections to illustrate presented within a specified Thesamples Grid scheme samples all combinaherein used in the following sectionsthe to illustrate the presented variable withintions a specified range. The Grid scheme samples all combinahere and is used inand the following sections to illustrate the presented of variable values with equidistant spacing across each variable, methodology. The “initial” geometry of the problem is shown in Fig. tions of variable values with equidistant spacing acrossspacing each variable, The “initial” geometry of the problem is shown in Fig. 6.32 6.32 -2 tions of variable values with equidistant across eachmethodology. variable, methodology. The “initial” geometry of the problem is shown in Fig. -2 tions of variable values with equidistant spacing across each variable, 10 methodology. TheThe “initial” geometry of the problem is shown in as Fig. 10 again within a specified range. The Latin Hypercube scheme includes , x , and x shown (along 5a. variables of the problem are x 3.90 again within aagain specified range. The Latin Hypercube scheme includes (along The variables of the 1 shown 2 3 3.90 problem are x 1 , x 2 , and x 3 as within a specified range. The Latin Hypercube scheme5a. includes 5a. The variables of the problem are xas 1, x 2 , and x 3 as shown (along 1.2e-02 again within a sampling specified range. The Latin Hypercube scheme includes x2,[14]). and x3This shown (along 5a. variables the problem areFig. x1,5a 1.2e-02 random values for each variable, but inthat such a with wayThe that 2.33 withoftheir bounds in case study is for research 2.33 sampling random valuesrandom for each variable, but in such abut wayin their bounds in Fig. 5a [14]). This case study iscase for research 00 a way sampling values for each variable, such that with their bounds in Fig. 5a [14]). This study is for research 10 10 sampling random values for each variable, but 5.2e-03 in such athe way thatspace. with bounds in Fig. [14]). This case is study is for for research 1.56 1.56 purposes 5.2e-03 ensures even distribution of samples across design The their ap- and and5athe methodology intended more complicated ensures even distribution ofdistribution samples across the design space. The appurposes the methodology is intended for more complicated 3.3e-03 ensures even of samples across purposes and the methodology is intended for more complicated -3 3.3e-03the design space. The ap10-3 ensures even distribution of samples across the design space. Theisappurposes the methodology is structural intended for complicated 10 propriate number of samples that should be used problem specific andproblems where the andmore energy analysis has significant propriate number of samples that bethat used is problem specific the structural and energy analysis has analysis significant 1.6e-03 propriate number ofshould samples should be used is problemproblems specific where problems where the structural and energy has significant 1.6e-03 propriate number of samples that should be used is problem specific problems where the structural and energy analysis has significant 0.50 0.50 and commonly determinedA empirically. A higher number of dimencomputational needs. and commonlyand determined higher number of dimenneeds. commonlyempirically. determinedAempirically. A higher number ofcomputational dimencomputational needs. and commonlysions/variables determined empirically. higher number of dimenneeds. lead to a higher number of samples needed incomputational order to The variables examined were the horizontal – x1 ––and vertical – sions/variablessions/variables lead to a higher number of samples needed in order to – and vertical The variables examined were the horizontal – x 1 lead to a higher number of samples needed in order The variables examined were the – x1 – –and vertical – -1 -4 -4 sions/variablesachieve lead to a10 number ofuniformity samples needed in order to –asand vertical Theto variables were horizontal – xhorizontal 10 10higher 10 -1 the same level of orKRIG distance between samples. –examined position of onethe node of shown the truss, shown inalso Fig. 5a and also xRBFNE 2 one achieve the same levelthe of same uniformity or distance between samples. – position nodeKRIG of in1 as Fig. 5a andin RF MARS NN level RF of RBFN RBFNE 2NN NN RF RBFN RBFN MARS KRIG NN RF RBFN RBFNE MARS MARS achieve uniformity orKRIG distance between xxsamples. –one position of the onetruss, node as of the truss, Fig. 5a and also xof 2RBFNE achieve the same level of uniformity or distance between samples. of node of the truss, as inthat Fig.shown 5a and also 2 – position Model type Model type type This increase inofthe number of samples isinexponential in the the Grid Model – position of shown the node lies in the axis of symmevertical x3the Model type Model type Model type This increase This in the numberin samples is of exponential the Grid in position––of node that lies in the axis of symmevertical – the xthe 3 –vertical increase the number samples is exponential the Grid – position of the node that lies in the axis of symmex 3 the node that lies in the axis of symmeThis increase in the number of samples is exponential in the Grid the vertical – x – position of 3 scheme. This is the mainofdisadvantage of thefor Grid scheme for try problems try of truss. The structure is constrained to be symmetric. The design scheme. This isscheme. the main disadvantage the Grid scheme problems of truss. The is constrained be symmetric. design The This is the main disadvantage of (a) theMSE Grid scheme for problems trystructure of truss. The structure isto constrained to beThe symmetric. design Fig. 10. Seven-bar truss and time comparison bar chart. scheme. This iswith the main disadvantage of the GridThe scheme problems try of truss. structure is constrained to be symmetric. The design Fig. 10. Seven-bar truss (a) for MSE and (b) (b) build build time comparison barThe chart. a large number of variables. Latin Hypercube is the scheme space was sampled based on the three (3) nodal position variables, with a large number variables. Latin Hypercube is the schemeis thespace was sampled based on the based three on (3) nodal position variables, with a of large numberThe of variables. The Latin Hypercube scheme space was sampled three (3) nodal position variables, 5 the with a large number ofthe variables. The Latin Hypercube is the scheme space was sampled based on the three (3) nodal position variables, used for case studies of the current paper. The specific number of and each design was analyzed for gravity loads (accounting for used for the case studies the current paper. The specific of number and each was analyzed foranalyzed gravity loads (accounting for used for theof studies of the current paper.number The specific of design and each design was for gravity loads (accounting for used for the case studies ofcase the current specific number of and each design was analyzed for gravity the loads (accounting samples used for each casepaper. studyThe is given later. steel buckling, which enhances nonlinear nature for of the problem). samples used for each case study is given later. steel buckling,steel which enhances theenhances nonlinear nature of the nature problem). samples used for each case study is given later. buckling, which the nonlinear of the problem). samples used for each case study istogiven later. steel buckling,Awhich enhances theof nonlinear of the problem). All datasets used train and assess the models have been normalized. A iscrossthemay required crossperformance score the“Many formnature Fig. surrogates becrossconsistent i (where All datasets used train and assess the models have been normalized. Lthe AΣA isiLiLthe required A performance score of the form ΣA i (where Allto used to train and assess the models have been normalized. A is the required A performance score ofi1. form ΣA i (where All datasets used todatasets train and assess theused models have been normalized. A is the required crossA performance score of area the form iLi (where The normalization scheme in this study adjusts each variable so that areasectional andofΣA L member the length of member i) wasas determined as the The normalization used inscheme this study adjusts variable thatvariable sectional and L thearea length i)of was determined the Thescheme normalization used in thiseach study adjustsso each so that sectional and L the length member i) was determined as the The normalization scheme used in this study adjusts each variable so that and L the length of member i) was determined as the it0has a amean of 0 and a standard deviation of 1. Once a datasetsectional has been areaobjective. it has a mean ofit and standard deviation of 1. deviation Once a dataset has been a mean of 0 and a standard of 1. Once a datasetobjective. has been objective. it has a mean ofcreated 0has and athrough standard deviation of 1. Once a dataset has train/validation/test been objective. simulation, before thetrain/validation/test partition into Using the developed sampling Grasshopper sampling tool described above, created throughcreated simulation, before the partition into Using the developed Grasshopper tool described above, through simulation, before the partition into train/validation/test Using the developed Grasshopper sampling tool described above, created throughset, simulation, before the column partitionand into train/validation/test Using Grasshopper sampling tool described above, each input variable the output vector was reduced by its the developed data were collected to later fit approximation models. Fig. 5b gives 15 set, each input variable and thecolumn outputand vector reduced its reduced databy were to later fit approximation models. Fig. 5b gives 15 set, eachcolumn input variable thewas output vectorby was its collected data were collected to later fit approximation models. Fig. 5b gives 15 set, each input variable column and thebyoutput vectordeviation. was reduced by its data were collected to produced later fit approximation models. Fig. 5btool. gives 15 design has a mean and then divided its standard Normalization is impordesigns by the developed sampling Each mean and thenmean divided bythen its standard deviation. Normalization is impordesigns by the developed sampling tool. Each design and divided by its standard deviation. Normalization is impor-produced designs produced by the developed sampling tool. has Eacha design has a mean and then tant divided by its standard deviation. Normalization is impordesigns produced by the developed sampling tool. Each design has a an approxto bring all the variables in the same range and prevent assigning unperformance score associated with it, which is used to train tant to bring alltant the variables inthe thevariables same range andsame prevent assigning un- assigning performance score associated with it, whichwith is used to train approxto bring all in the range and prevent unperformance score associated it, which is an used to train an approxtant to bring all realistic the variables in the same range and prevent assigning unperformance score associated importance and bias toward some variables. imation model. with it, which is used to train an approxrealistic importance andimportance bias toward some imation model. realistic and biasvariables. toward some variables. imation model. realistic importance and bias toward someseparated variables.test set was used to assess imation model. Finally, the completely the perFinally, the completely separated test separated set was used assess the perFinally, the completely testtoset was used to assess the perFinally, the completely testInset wasto used to assess the per-the results and formance ofseparated the models. order better comprehend formance of the models. of In order to better comprehend results andthe formance the models. In order to betterthe comprehend results and formance of thethe models. In aorder to better comprehend the values results(the and output for the effect of model, the performance score 4. Error and visualization the effect of a model, theof performance values (the output for(the theoutput 4. Error the effect a model, thescore performance score values for theand visualization 4. Error and visualization the effect of a model, thewere performance score values (the output for the design 4. Error models) normalized so that that the the best-performing in the theand visualization models) were models) normalized sonormalized that the best-performing design in the design in were so best-performing models) were sampled normalized soreceives that the best-performing the accordingly. data score of 11 and thedesign rest areinscaled scaled Inthe thiscrucial section, the crucial crucial issues of of visualizing model's perforperforsampled data receives score of 1 and the rest are scaled accordingly. In this section, issues of visualizing a visualizing model's perforsampledadata receives aa score of and the rest are accordingly. In this section, the issues aa model's sampled data receives a score of 1 of and performs the rest are scaled accordingly. In this section, the crucial issuesitsoferror visualizing a model's perforFor example, example, score two times worse thanthe thebest best design mance and measuring are discussed. Several visualizationtechtechFor example, a For score of 2 performs two worse than theworse best design mance and measuring its measuring error are discussed. Several visualization techaa score of 22times performs two times than design mance and its error are discussed. Several visualization For example, a score of 2 performs two times worse than the best design mance and measuring its described error are discussed. Several visualization in the the test test set. set. niques are are and used inthe thecase case studies. techin the test set. in niques are described and used in the case studies. niques described and used in studies. in the test set. niques are described and used in the case studies. MSE MSE MSE

Table 9: KRIG correlation functions considered

Data point

In Figure 13 an example dataset from [38] is shown with different configurations of the parameters to showcase the difference they can make. The black dots are the sampled points used for training the Kriging model and the surfaces are the result of applying the trained model on axfine grid.

(a) Graduate Show 2019

(b)

with the data ” (Figure inspired from

(c)

7 235

[15] ).

8. The thesis-proposed machine learning methodology. 9. Single run workflow. 10. M ethodology implemented on different machine learning model types.

287

error error

in Construction 72 (2016) 279–293

nuisanceparameters parameters nuisance

Neural Networks

S. Tseranidis et al. / Automation in Construction 72 (2016) 279–293

114

Published with MATLAB® R2014b

Random Forests

Published with MATLAB® R2014b

Radial-Basis Function Networks

117

119

Fig. 11. Robust model comparison framework sketch for a single model. Training Set

Validation Set

Test Set

To address this, another scalar value for the error is needed for comtrained models parison, which is somehow internal to the data itself without any effect from the model. One approach to obtain such a number would be to performance NN best its error on the validamake a random prediction model and calculate tion set. It would be necessary to repeat this process a number of times to obtain an interval to which a t-test could be performed to determine whether or if performance RF a trained model has an effect on the predictions, best it can be considered random. A more simple approach, proposed here as the “flat” model approach, assists in a more rapid and qualitative evaluation of a model's performance a model error. This approach, instead of random predictions, considers RBFN best that predicts a single value for each data point in a set. The proposedSingle Run value is the arithmetic mean of the dataset. As a result, the model used to compute the error is a “flat” model that always outputs the performance RBFNE same value as shown in Fig. 8. It serves as a good benchmark (like the best random prediction) since it does not reflect in any way the structure of the data, and a surrogate model must surely have a considerably lower error than the “flat” model's in order to be considered adequate. performance MARS best comparison on the test For the seven-bar truss example, the model set is shown in Fig. 10a, where the “flat” model error is the black dashed Published with MATLAB® R2014b line at the top of the plot. The y-axis scale is logarithmic and it can be ob121 Radial-Basis Function Networks Exact served that some than the “flat” performance KRIG models performed substantially better model, which is the desired effect. To furtherbestexplain the “flat” model, the scatter plot produced for the test set is included in Fig. 8. The predicted value is always train the same (mean of Actual values), models compute select assess and the “flat” for different prediction best model performance model error is the MSEnuisance of the model plot. error represented inofthis chosen scatter model parameters Finally, a different way to obtain an intuitive prediction error value is Fig. 9. Single-run workflow. to take the root of the mean square normalized error, with the formula 9 shown as RMSNE in Table 1. This would be an order errortoestimate of the permake predictions. Therefore, the main evaluation metric of a Understanding error and visualization model is how well it performs at making predictions on new data, or centage of the error on each sample entry.

When an approximation 236 model is developed, the goal is to make the el “learn” a speciﬁc physical process or numerical simulation in Seven-bar truss | Structure 0

122 role of in other words, how well it generalizes. The test set has the “new” data in a way that it comprises pristine data that were not used

10 2

Published with MATLAB® R2014b

Multivariate Adaptive 124 Regression Splines

Kriging Models 128

was carried out in Rhino and Grasshopper. For the structural perforergy evaluation, it was decided that 1000 samples should be used, remance, the objective of structural weight was calculated using the finite quiring approximately 6 h to generate for the current case study. The element modeler Karamba, which is a plugin for Grasshopper [41]. For number of samples used for the Training/Validation/Test datasets all energy simulations, the plugin ARCHSIM was used [42]. This plugin were 600/200/200, respectively. connects the parametric model in Rhino to EnergyPlus [43], an energy analysis program available from the U.S. Department of Energy. More 6.3. Approximation results and discussion 11. Design case study: Model information about this problem, the setup, and all assumptions used performance in the simulations can be found in [44]. Each designscatter evaluationplots. required The presented approximation framework was applied to the case approximately 0.2 s for the structure and 25 s for the energy, making study design problem for several different climates. The model compari12. Published paper on thesis-proposed surrogate modeling an ideal solution. son bar charts are shown in Fig. 14 for the energy simulations in Abu Because this case study involved energy simulations, data were colDhabi (cooling dominated), Sydney (mostly cooling), and Boston methodology. lected by simulating the same structure in different climates and orien(heating and cooling). For all these climates, the surrogate models perD esign case study inspiration. tations. The case study for13. NS orientation (the longitudinal axis of Fig. formed at an order of magnitude better than the “flat” model benchmark. 12) is examined here. The NN model performed the best, followed by the KRIG. The approxima14. Design case study problem There is a trade-off between how many samples can be generated tion performance was significantly better for the structure score (Fig. 14d) and the accuracy that is desiredformulation. in the approximation model. If more than for the energy score across all model types. NN, RBFN, and KRIG all Automation in Construction 72 (2016) 279–293

(a)

Predicted Value

1.6

1.4

1.2

Automation in Construction

1.4 -10%

journal homepage: www.elsevier.com/locate/autcon

1.2

Data-driven approximation algorithms for rapid performance evaluation Proceedings of the International Association for Shell and optimization of civil structuresand Spatial Structures (IASS) S Amsterdam Future Visions 1.2

Predicted Value

1.6

1.4

1.6

1.2

1.6

+10%

1.4

1.6

-10%

1.6

1.2

1.6

+10%

Building Technology Program, Department of Architecture, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

-10%

1.2

1.6

Stavros Tseranidis ⁎, Nathan C. Brown, Caitlin T. Mueller

-10%

1.2

1.4

+10%

1.4

1.2

1.6

+10%

1.4

a r t i c l e

1.6

1.2

1.6

+10%

1.4

i n f o

a b s t r a c t

Article history: Received 1 September 2015 Received in revised form 20 January 2016 Accepted 6 February 2016 Available online 24 March 2016

1.6

+10%

This paper explores the use of data-driven approximation algorithms, often called surrogate modeling, in the early-stage design of structures. The use of surrogate models to rapidly evaluate design performance can lead to a more in-depth exploration of a design space and reduce computational time of optimization algorithms. While this approach has been widely developed and used in related disciplines such as aerospace engineering, there are few examples of its application in civil engineering. This paper focuses on the general use of surrogate modeling in the design of civil structures and examines six model types that span a wide range of characteristics. Original contributions include novel metrics and visualization techniques for understanding model error and a new robustness framework that accounts for variability in model comparison. These concepts are applied to a multi-objective case study of an airport terminal design that considers both structural material volume and operational energy consumption. © 2016 Elsevier B.V. All rights reserved.

Keywords: Surrogate modeling Machine learning Approximation Structural design

Predicted Value

3.2 Design space formulation A parametric model of the airport design was generated in Rhinoceros and Grasshop following geometric parameters: overhang length, starting edge location, starting hin of the left half-arch, angle of the right half-arch, and glazing ratio (see Figure 3). evaluation of both structural and energy performance, the model included linear eleme the structural pieces of steel as well as surfaces representing the exterior envelope The ranges for each parameter were determined based on general design problem con a minimum requirement for structure clearance on both the interior and exterior of the 1.4

1.4

-10%

1.2

Predicted Value

1.4

1.2

1.4

1.6

+10%

3.5

(d)

Contents lists available at ScienceDirect

+10%

-10% 1.2

(c)

KRIG

1.6

+10%

-10%

(b)

RBFNE

1.6

+10%

1.2

1.4

-10%

1.5

2.5

Actual Value

3.5

1.4

The engineering design process for civil structures often requires computationally expensive analysis and simulation runs within a limited timeframe. When the assessment of a design's performance takes hours or days, the potential for exploring many solutions and signiﬁcantly improving design quality is limited. This paper addresses these limitations by investigating the application of surrogate modeling, a data-driven approximation technique, to civil engineering in order to empower designers to achieve more efﬁcient and innovative solutions through rapid performance evaluation.

1.6

+10%

2.5

1.5

2.5

1.2

3.5

2.5

S. Tseranidis et al. / Automation in Construction 72 (2016) 279–293

1. Introduction

1.6

+10%

3.5

288

-10%

1.1. Design optimization for civil structures

1.5

2.5

3.5

1.5

Actual Value

2.5

Optimization is an established and widely used method for highperformance engineering design. However, unlike in other engineering disciplines, the optimization objectives and constraints in civil and architectural structures such as buildings and bridges are not always easily quantified and expressed in equations. Instead, they often require human intuition and initiative to materialize. For this reason, design optimization in civil engineering has yet to reach its full potential. There are a few examples where it has been successfully applied for to buildings, Fig. 13. Inspiration the design case study (image from [44]). such as in braced frame systems for tall buildings by Skidmore, Owings, 1.2. Need for computational speed and Merrill (SOM) [1], but these cases remain exceptional. In contemporary architectural design, a wide variety of braced frame systems have The exploration of a design problem and various optimal solutions on the specific data sets used for training and validation. The met should ideally happen in real time, so that the designer is most produc⁎ Corresponding author at: 77 Massachusetts Avenue, Room 5-418, Cambridge, MA tive. Researchhere has shown thatto rapid response time can in signifiproposed seeks characterize theresult effects of variability due to 02139, USA. Tel.: +1 617 852 1737. E-mail address: stser@mit.edu (S. Tseranidis). cant productivity and economic gains [3]. The upper threshold for

3.5

Actual Value

Fig. 15. (a) Abu Dhabi, (b) Sydney, (c) Boston—Energy Overall, (d) Structure.

1 X1

67.7 m

5. Robustness

A robust model comparison methodology has been developed and is 1 case study probdescribed in detail in this section. It is applied in several http://dx.doi.org/10.1016/j.autcon.2016.02.002 lems,0with the results showcased theB.V.next section. 0926-5805/© 2016 in Elsevier All rights reserved. The motivation for this methodology is the need for a way to quantitatively compare the performance of different families of models in approximating the same dataset. An important goal of the methodology is the extraction value. 2 of an interval of a model's error in addition to the average error 0 run of3 training The way to obtain this interval is not to make a single models on a given4 train/validation set configuration, but to use many configurations and make several runs. This section first describes the structure of a single run. Next, the process of generating and aggregating 6 the models is explained. Overall, additional runs in order to compare this section introduces the main framework that has generated the results of the architectural case studies.

X 0

been employed recently, including both the example in [1], as well as Qingdaobei Station(High Line 23, the more irregular design by Neil M. Denari Architects New York City [2]). While High Line 23 is Shandong, not structurally optimized, Qingdao, China its architectural success is closely linked to the visual expression of its AREP (architect) structural system and geometry. As demonstrated by these buildings, MaP3 (structural engineer) there often needs to be a balance between quantitative and qualitative objectives in design optimization. Both quantitative and qualitative design goals pose challenges for the use of optimization in terms of computational speed. First, simulations such as structural analysis and predictions of building energy consumption often require signiﬁcant computational power, increasing with the complexity and size of the project. Thus, optimization algorithms can require substantial execution time, slowing down or impeding the design process. On the other hand, the qualitative nature of civil engineering design, with its hard-to-quantify considerations as described above, can require many iterations, and the ﬁnal design is unlikely to be selected from a single optimization run. In practice, the combination of problem formulation challenges and long simulation times means that optimization is rarely used in the design of architectural and civil structures. Even quantitatively comparing several design alternatives can be too time-consuming, resulting in limited exploration 13 of the design space and likely a poorly performing design.

data set so that the model's general robustness can be underst One method to obtain a variability measure for the error is cr validation, which was described in Section 3. In that section, it also argued that when an abundance of data is available, it is prefer to choose a completely separate validation set and avoid cross-val tion. When considering both 1) abundance of data, and 2) a need an error variability measure, this paper proposes a methodolog training the models for several different separated training and val tion sets. 5 In detail, the aforementioned procedure for the single run is rep ed for many different random partitions of the training and valida data pool. This means that the training and validation data are po and a certain number of random data partitions are generated in tr ing and validation sets with a constant number of samples in each The validation set error (MSE) is stored for each different nuisance rameter and every model type. Afterwards, the mean of the MSE each parameter across all partitions is calculated. Then, the nuisance rameter with the minimum arithmetic mean MSE is chosen and the rors for this speciﬁc parameter for each partition14are studied as desired measure of the variability. The standard deviation is a me that can be extracted from this information, or alternatively, the ra 237 and the scatter can be examined. This process is applied for e model and the output is an error measure with the accompanying iability for each of the six models. For one model, the process is s

15.0 m

X = Glazing Ratio (0.3-0.5)

5.1. Single run

45.7 m

In the ﬁrst place, it is assumed that there is a single training set and a separate single validation set. A framework was developed in MATLAB to train all six different types of models examined in this paper for different nuisance parameters and within each model type. The nuisance parameters considered for each model have been summarized in Table 2. The training set (the same for all the models) is ﬁrst used to train the models. Then, the mean squared error of those models is com-

Figure 3: Design parameters for the case study Graduate Show 2019

238

Communications

Cecilia Shiels Hannah Gasparutti

The Connection Project Cecilia Shiels

Course: BA Interior Architecture and Design University: Nottingham Trent University Country: United Kingdom Year Start: 2018

My final project was influenced by the findings of my dissertation, based on the negative effects of the growth of palm oil. Palm oil was introduced into the UK as a machinery lubricant during the industrial revolution and has since become one of the leading contributors to deforestation today. Through the reintroduction of nature to a former industrial complex in Leeds city center I was able to create a green oasis, allowing nature to reclaim land lost to the industrial era. Research and development of ideas led me to focus on three user groups, which I aimed to connect with nature in different ways for different reasons. A horticultural research facility, connected to Leeds Universities school of earth and the environment offers researchers and students with a base from which to conduct their work. A seed bank runs through all floors of the facility, this not only provides a method of preserving nature for future generations, it also links to the sites past. A cooling system maintains the temperature of the seeds, and though historically smoke could be seen billowing from the three chimneys on the site, now the air is clear with steam being produced through the cooling system of the seed bank.

YOU’RE PROBABLY UNAWARE OF THE DESTRUCTION YOU’RE CAUSING.

Palm oil is in 50% of the items you consume daily. You brush your teeth with it, you moisturise your face with it, you eat it. Growth of the ingredient directly contributes to the destruction of rainforests in Indonesia and Malaysia. Destroying habitats of endangered animals. Many large brands say they are waiting for consumer demand for alternatives. They rely on our ignorance. Don’t let them. Palm oil is the in 50% ofQR the code. Scan items you consume daily, you brush your teeth with it, you moisturise your face with it, you eat it. Growth of the ingredient directly contributes to the deforestation of rainforests in Indonesia and Malaysia. Destroying habitats of many now endangered animals such as the orangutan. Many large brands say they are waiting for consumer demand for a sustainable alternative. They know what they’re doing, and rely on our ignorance. Don’t let them. Find out more. Scan the QR code.

1 242

A childrenâ&#x20AC;&#x2122;s nursery lies to the east of the main building, enabling children to form deep, meaningful connections with nature from a young age, encouraging them to grow into more environmentally conscious adults. Research led to the creation of a park like space for the residents and workers of the city, as a reflection of the parks and heaths that were opened to the public during the industrial revolution and of the appreciation of nature that grew from the era.

1. Poster highlighting the lack of awareness surrounding the palm oil issue. Smoke also represents wild fires lit to clear ground for palm oil plantation. 2. South Elevation | Showing existing building before redesign. 3. North Elevation.

3 Graduate Show 2019

243

6 244

4. Visual | Showing seed bank within internal study space. 5. Entrance to horticultural research facility. 6. Section. 7. Axonometric visual of horticultural research facility and childrens nursery. 8. Childrenâ&#x20AC;&#x2122;s nursery.

7 Graduate Show 2019

8 245

Behind my portfolio Hannah Gasparutti

Course: BA (Hons) Photography University: Norwich University of the Arts Country: United Kingdom Year Start: 2018

My photographs are not directly based on a particular topic or are a part of projects, they are all individually stylised in a way to represent myself, my upbringing and personal interests. Having Italian family I was introduced to the Literature /Film genre Giallo which later has gone on to be influence in the way that I create and view images. Giallo explores crime and psychological horror in a very stylish manor. Key characteristics of Giallo is very bold colour and very stylish architecture and clothing – Every aspect that makes up the genre is very rich feeling. Idolising the way the Giallo filmmakers would stylise their films I have gone on to create images that explore the richness in colour, fashion and architecture. The photograph on the right hand side is one of my images that has been loosely inspired by this Genre, in particular by Dario Argento’s film Suspiria, this through expoloring the meaning of ‘Suspiria de Profundis’ which is a Latin translation meaning sighs from the depths.

246

Graduate Show 2019

247

248

Modelshop

Felix Burke Paulina Grenda

Wind Turbine Installation Vessel 1:200 & other Projects Felix Burke

Combining my passions of model making and sustainability, I decided to design and construct a wind turbine installation vessel for my university final year project. The real vessel has the complex task of transporting and installing 200m tall wind turbines at sea and while my model is 1:200 scale, it was a daunting but exciting challenge. I drew the model from scratch as no plans were available and used a wide range of processes and materials including CNC routing the Hull, 3D printing the turbine nacelles and blades and fully rigging the crane to make it functional.

Course: Model Design: Model Effects BA (Hons) University: University of Hertfordshire Country: United Kingdom Current Project: 6W 77th Street Year Start: 2018

The next page shows two other projects made during my course. The first is an OskKosh Striker 3000 airport fire engine at 1:24 scale. Built as part of my 2nd year vehicle project, I learnt many new processes on this model including designing and printing 3D components for the first time. Featuring a fully populated cab, detailed suspension and markings which I printed myself. The second was created for a new learning Hub within Cassiobury Park, the aim of the model was to show clearly what the original Lodge building consisted of as it has since been demolished, built from laser engraved chemiwood from original drawings and sprayed using acrylics.

1 252

2 Graduate Show 2019

3 253

4 254

1. View of the bow section of the vessel showing the asymmetrical design. 2. All round view of the OskKosh Striker 3000, with interior and suspension detail shown. 3. Cassiobury Lodge all round view and top plan showing the layout of the building. 4. View of Model as a whole, showing the vessel fully jacked up. 5. Starboard quarter, with three 3D printed Azimuth thrusters. 6. Deck layout with fully working and rigged crane on the port side.

6 Graduate Show 2019

255

Bodie Island Lighthouse, NC, USA (1:20, Cut away model) Paulina Grenda

Course: Model Design BA (Hons) University: University of Hertfordshire Country: United Kingdom Current Project: Rotation Sketch Modelshop Year Start: 2018

The Bodie Island Ligthouse is situated on the coast of North Caroline, USA, in front of the visitor centre. Built in 1872, the structure stands at 52 meters tall and is a great example of Victorian architecture. The idea for this model was sparked whilst visiting MusĂŠe des Arts et MĂŠtiers in Paris where I saw a model of a lighthouse. Upon further research, I found the Bodie Island lighthouse and I believed that the visitor centre would benefit from a cut away model that would showcase the details of the upper part of the structure, that otherwise is hard to notice for visitors admiring from below. The model was hand fabricated in 12 weeks. This time was split into 4 phases: 1) Research 2) Measured Drawing 3) Rhino Model 4) Test pieces & model making 1

256

1. Bodie Island Lighthouse. Image available from: http:// lighthousefriends.com/light. asp?ID=357

4. Close up of stair detail in the Rhino model. 5. Close up of the external detail in the Rhino model.

2. Scaled drawing.

Bodie Island Lighthouse, North Carolina, USA

Graduate Show 2019

3. Overview of Rhino model.

5 257

I wanted to make the most of my last year at University by challenging myself to use all the available machinery in the workshop. Happy to say, I was successful! Some of the machinery and materials used included: -Lathe -Milling machine -Vacuum former

-Acrylic/PET -MDF/ Veneer -Chemiwood

-Laser cutter -Polyurethane resin -3D Printers

This also involved multiple handmade jigs that were mostly used for heatbending and gluing, as shown in the images below.

258

1. Turning chemiwood to make 8th floor of the model

6. Priming components

12. Assembly of roof: flashing

7. Silicone mould of bricks

13. Assembly of the roof: details

2. Turning chemiwood to make the form to vacuum form the lens

8. 8th floor covered in Polyurethane resin bricks which were heat bent using jig in Figure 4

14. Painted roof

3. Milling out space for windows in previously turned 8th floor (Figure 1)

10. 3D printed components

5. Vacuum forming the lens

11. Assembly of the roof: tiles

15 Graduate Show 2019

9. Painting test

4. Heat bending jig

15. Front of the finished model 16. Back of the finished model

16 259

260

Urban Design

Emilie Jinyi Li

Autonomous Campus Emilie Jinyi Li

Course: MArch Urban Design University: The Bartlett, UCL Country: United Kingdom Current Project: Mawan Qianhai Masterplan Year Start: 2018

Autonomous mobility will be the new urban paradigm. Speculating on the trajectory of autonomous technology - computer vision and machine learning will enable a sensed environment beyond human comprehension to understand the intricate relationship between us and nature, allowing us to adapt our mobility and urban spaces established on the digitally sensed-scape and algorithms. Emancipating functions from fixed geographical locations and our pre-existing knowledge of space, mobility and form, an alternative design approach is offered where our future urban communities can be dynamically and adaptively co-created with its nature capital as priorities, redefining the relationship into an integrated, symbiotic system. In the mass customisation of a unit, preliminary analysis of the projected space usage is undertaken to ascertain true needs over and beyond preferences. And the resultant adaptation of the optimal form of unit is a negotiation between the inside and outside. The autonomous campus city, will be a cyborg-city comprising of many mass-customised dynamic individual urban unit elements, with the city scale smart control algorithm. The form of the co-created city will be mobile and dynamic. 1 264

1. Mass-customised Unit - Unit form determine by needs and preferences. 2. Unit adaptability - Responsive to its surrounding sensed environment. 3. Multimaterial 3D printing - Technology of printing parts with a range of mechanical, optical or thermal properties.

Tango black 95A

35o

25o

70A Vero clear 4

25o

45o

72o 95o

2 Graduate Show 2019

3 265

4 266

4. Computer vision - Sensing urban spaces. 5. Dynamic autonomous campus â&#x20AC;&#x201C; Adaptive collective community

5 Graduate Show 2019

267

Organisers and Volunteers

Organisers

Mike Bass Chloe Bingham Alicia Cox Craig Edwards David Gilliard Rupert Goddard Aaron Hargreaves Charles Sabberton Narinder Sagoo Sam Strudwick Charlotte Sword Stephen Teare Tony Wenban Nigel Young

Floor Planning

Stefan Necula Gabriel Pavlides Alejandro Guerrero Neira

Wall Plaque

Josh Corfield Yanni Hajigeorgis

Poster

Dominic Oliver Alessandro Magliani

Book

Joey Aoun Jingyi Chen Margaux Coquelin

Riverside, 22 Hester Road London SW11 4AN T +44 20 7738 0455 fosterandpartners.com