Overview

While working for Atmos I was Project Lead on the Mobile Orchard, a sprawling, intricate and inhabitable sculpture of a full-scale English fruit tree, that was commissioned for the 2013 City of London Festival.

This ambitious project, which was created as the centrepiece of the 2013 Festival, travelled between six prominent locations within the City, pausing at each venue for one week where it acted as a spectacular backdrop for a series of musical and theatrical performances. The main sculpture was accompanied by dozens of young fruit tree saplings, together helping to create a dense grove that celebrates and promotes British orchards and produce, a passionate cause for many of our sponsors and partner organizations on this project.

The frequent changes of location necessitated an overall design that was quickly and easily demountable. In order to allow the entire Tree to be disassembled and moved from one venue to the next in a single day, we focused on creating a series of individual components (Roots, Branches, and Secondary Branches) that could be transported as discrete elements, before being lifted, positioned and secured in place by only two or three people. In contrast the core elements (Trunk Base and Trunk) are heavier pieces that can only be moved by hiab crane, but which act as ballast against the considerable loads created by self weight, wind, and people climbing and sitting within the Tree’s network of branches.

Early Design Investigations

Our initial sketches and studies presented two Trees – both a series of sinuous lines flowing through space – meandering all the way from the tip of the Roots, up the Trunk, and out to the end of each Branch. Visitors would sit under their canopies and be able to follow the line of the root they sat on, as it wound up the core and out into bifurcating branches.

Initial Concept Models:

• 

• 

• 

• 

These concepts encouraged us to investigate lasercut sheet steel as our primary material – one that could act as both underlying structure and external cladding, twisted into sinuous bands and perforated with a curved fold line, creating gently curved ‘V’ shaped sections that are far stronger than the original flat sheets.

[ Each of these pieces thus conformed to the rules of Curved Folding that I have discussed in great detail on the Project Page about Curved Folding. Much of my work at this stage was only possible because of the experience with developable geometry and fold simulation that I gained during my previous work with RoboFold. ]

These ‘V’ forms were then arrayed radially in a star formation around the Trunk’s central vertical axis. The result was a series of strands that were all unique, but together acted as a composite structure. Each one widened at their base to form generous shallow-folded Roots, before coming together and bunching into sharply defined striations.

Metal Option – Parametric Model:

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

We wished to use Curved Folds for a number of reasons. They made the material stronger — able to span greater distances in the Branches, and stiffer in the Trunk. They also halved the number of elements, in turn halving the amount of open joints and number of fixings necessary – which made the Trunks less ‘cluttered’ visually. Fewer open joints also reduced the likelihood of people hurting themselves on sharp edges (unlikely to happen anyway, as the edges of sheet steel are softly filleted by the heat from lasercutting). Lastly, because Curved Folds inherently stay developable throughout their formation — i.e. they never need to be stretched or squashed to conform — they can be hand folded and positioned. This was a significant factor given the short project schedule and limited budget.

We spent the initial phase of the project exploring the possibilities (and difficulties) of creating such an innovative structure in curved folds. At the end of this period we accepted that budgetary and time limitations necessitated that we only produce one Tree, and that we change to a more established material and assembly process.

For this reason we moved on to a design formed of CNC-routed birch plywood sheets.

CNC-Plywood Structure

We prototyped different sheet thicknesses and load tested different Branch lengths to ensure such an unusual assembly could withstand the forces of people moving around and atop them.

Prototype Tests:

• 

• 

• 

The core design centred on a trunk that leaned and cranked and bent and spiralled and twisted as far as engineering constraints would allow a structure with a deeply eccentric centre of gravity. It bundled 12 trunk segments each composed of thin timber sheets laminated together. Each Segment sat at a 30 degree angle to its neighbours, which would allow them to be simply machined by a 15-degree CNC bit and form a planar edge against those adjacent pieces. Internally we positioned a series of diminishing ellipsoidal compression ‘tree rings’, set along an eccentric spiral, which these Segments were wound and glued against, to provide bracing and structural stability.

Trunk:

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

• 

Trunk Base:

This eccentricity meant each Segment of the Trunk exhibited a different size on its outside face, and this governed its capacity to grow branches, as each extension jostled for limited territory within the trunk that they must socket into and tension against, while at their tips the clustering and proximity of neighbouring branches uniquely affected the degree of bifurcation and deviation in plan and section for each one.

• 

• 

• 

• 

• 

• 

Branches:

• 

• 

• 

• 

• 

• 

Branch Sockets and Junctions:

These drivers, along with a host of other structural constraints, generated a deeply rigorous relationship between Trunk and Branches – creating a parametric definition that incorporated many factors such as depth-to-cantilever ratios, variable bolt spacing and recessing, spot and strip lighting locations, and so on – as well as the programmatic desire to optimise geometrical paths of stairs and head-height enclosures.

• 

• 

• 

• 

Master Parametric Model

The design process was developed almost entirely within Rhino, Grasshopper and RhinoscriptVB, where a set of interdependant tools were built that both simulated the artificial growth of trees at a range of scales, as well as managing the automation, rationalisation, subdivision, contouring, slicing, separation, numbering and preparation of parts that was necessary in order to convert the complex 3D form into a geometry that could be CNC-routed, assembled and physically realised.

• 

• 

These different scripting modules were networked together to create a master parametric model that was able to steadily aggregate and resolve a highly complex set of varied constraints (engineering, ergonomic, economic, fabricational and logistical inputs) into a single system. The end result was a flexible model from which we could derive a wide range of different design solutions, while being certain that all the options it generated conformed to these vital criteria.

• 

• 

• 

• 

• 

This model allowed us to adopt a design process where every single variable of the sculpture could be adjusted and refined individually, and the effect on the whole assembly was instantly visible.

The variables controlled by the Master Parametric Model included, but were not limited to:

· Trunk Base :: Position of Connection Points between Base and Trunk above.

· Trunk Dimensions :: Height; Width and Tapering; Twist and localised eccentricity from this pure spiral; number of Segments.

· Trunk Structure :: Thickness; Number, Position and Depth of Tension Rings; Socket Depth, Width and Taper; Corresponding position and size of Tension Bolt holes and washer recesses on other side of Trunk.

· Branch Structure :: Length; Location of inflection points and size of Deviation in Plan and Elevation (which control each Branch’s ‘waviness’); Cross Section Changes; Plug Depth, Width and Taper.

· Programmatic Constraints :: Lowest Branch Heights; Step Heights and Positions; Crown Seat Position.

· Fabrication Constraints :: Slice Thickness; Part Labelling, Size and Layering; Arraying and Positioning of Orientated Pieces.

• 

• 

• 

Many of these variables were interlinked in some way — for example, changing an individual Branch’s length or deviation automatically adjusted its Plug Depth and Width, as well as the corresponding Trunk Socket dimensions. Other variables affected the entire set of components, so changing the material thickness (‘slice spacing’) would change the width of both the Branch Plug and Trunk Socket (and many other variables), as these were programmed to always remain a multiple of the plywood thickness, for obvious reasons.

Fabrication Process

The manufacturing hybridised a range of subtractive technologies (3-axis CNC routing with waterjetting and lasercutting), working with one of the only CNC operators in London that also operate as fabricator and contractor. The initial design specified 18mm plywood sheet – ubiquitous and by far the most economic per weight – monolithically CNC-contoured in 3mm steps, but prototyping revealed that the necessary flipping and under-cutting required for rounded 3-dimensionality dangerously undermined the vacuum suction of the machine bed, prompting the use of a simpler and more multiplied system of laminated slices of 4mm Birch – a new grid that the design’s inherent parametricism could instantly absorb.

Assembly:

• 

• 

• 

• 

• 

Automated assembly of the slices was impossible, so fabrication of the sculpture still necessitated close collaboration and iteration between designer and fabrication team – initially in the refinement of CNC cutting protocols, and later in the manual crafting of each element and their assembly into a unified structure.

This task was undertaken by the experts at fabrication company Nicholas Alexander and a huge number of volunteers, many of whom were students in Architecture and related design courses, who wanted to gain experience in fabrication and on-site work. This team produced the spectacular final sculpture, and ensured that even on a limited budget and with a tight deadline, this hugely ambitious project was a success.

Text above by Jeg Dudley + Alex Haw

Finished Sculpture:

• 

• 

• 

Project Team:

Project Status: Built in 2013

Client: City of London Festival (Emma McGovern)
Designer: Alex Haw / Atmos Studio (Jeg Dudley, Natalie Chelliah, Xiaolin Gu, Miriam Fernandez)

Project Role: Project Lead for Atmos

Structural Engineering: Blue Engineering (James Nevin)
Lighting Design: Arup Lighting (Arfon Davies, Dwayene Shillingford)
Fabrication: Nicholas Alexander (Jak Drinnan, Nicholas Runeckles, Anna Baker)…

…and a huge number of enthusiastic and talented volunteers (see Atmos’ project page for details)… :

Further Information:

· Atmos website project page.
· The Mobile Orchard micro-Website.

· The 12-Segmented Trunk Base and Trunk, before assembly.
· The Assembled Plywood primary structure.

· More images of the intricate parametric model I created to control the Tree’s form.

Publications:

· The Orchard was presented by Atmos/ Alex Haw at the Fabricate 2014 conference, in Zurich. Also read the accompanying proceedings online here, or published in print here.

· A very positive write-up of the Orchard, published in the Architect’s Journal, August 2013 — here, and here.