A Journey to Digital Manufacture

Advances in computational design and digital fabrication are revolutionising the architecture and construction industries. Foster + Partners is embracing these technologies to improve the material efficiency of building elements as well as the methods with which they are constructed.

For Bucky, the goal of the architect-engineer was to challenge accepted conventions. He was a master of the art of ‘technology transfer’, harnessing new industries to produce pioneering solutions to old problems. Within the current context of computational design, in which computer systems now actively assist in the design process, and digital fabrication, where data drives the manufacturing process, he would likely be inspired by the cutting-edge tools and processes offering new opportunities in architecture and construction. Today, through digital analysis, individual components can be tailored to achieve the maximum strength while using the minimum material – to do ‘more with less’ – subsequently reducing the material waste and embodied energy of buildings. At the same time, the ability to directly fabricate parts from digital designs enables greater complexity of structure, customisation of parts and rapid manufacture.

These developments and technological opportunities are driving our research into material and structural optimisation and advanced construction methods. The Specialist Modelling Group (SMG) – a multidisciplinary team of designers at the intersection of computational design, building physics and research at Foster + Partners – has been leading the inquiry into efficient shapes and materials and their application at various levels of design, from modelling all the way to construction.

'The benefits for the wider office were immediately apparent.'

One of the first projects that took advantage of these technologies was the Yacht Plus boat (2005). It became apparent that the complex geometry of the hull and superstructure would take too long to model and would lack the necessary accuracy if traditional hand techniques were used. Three-dimensional printing proved to be a rewarding process – different individual models could be produced from one digital file, and the design team and client were able to see an exact model of the yacht for the first time, enabling quick and accurate reworking. The benefits for the wider office were immediately apparent – the 3D process was fast enough for the architectural design cycle, and could better handle the increasing number of complex, computer-assisted projects.

Further to its accurate and rapid representation of complex geometric structures, 3D printing also assisted analysis of designs on an operational level. Colour printing capabilities allowed for the visual representation of environmental data, such as solar exposure and airflow, directly onto the model. Architects and engineers could then more easily assess environmental and community impact, and tailor their designs accordingly. This method of analysis was deployed in designs for the Duisburg Harbour Masterplan in Germany (2007), which saw the successful urban renewal of a former industrial harbour into a well-connected, mixed-use quarter. The masterplan for this area had a strong ecological agenda, one which prioritised alternative methods of energy generation and which set the tone for later infrastructure projects. The in-depth analysis of sunlight paths on structures within the plan allowed for the optimum design and placement of solar panels across buildings, and thus the effective harnessing of available energy.

'To extend such potential from models to buildings in the real world, a change in scale was needed.'

These digital advances had immediate benefits at the level of design and modelling but, naturally, the ease and efficiency of rapid prototyping rose the question of how this technology might be used at the level of construction. To extend such potential from models to buildings in the real world, a change in scale was needed, as well as the application of more conventional construction materials beyond the sintered powder used for modelling purposes. The desire to investigate at this level led to affiliations with research partners that were already developing these new technologies themselves.

We then designed, simulated, fabricated and physically tested a range of prints for different construction applications. These included a curving wall made from experimental bricks that were designed for optimal thermal, acoustic and structural performance. Through such tests, we discovered that taking advantage of the characteristics of the fabrication method, such as the behaviour of the material and the printer’s movements, could lead to unique, customised building elements. Mastering materials, tool design and robotic control were all key to being able to fabricate the complex structures being generated digitally.

'The project offered a first taste of what it could be like to design, print and manufacture a large-scale final product.'

Another experiment undertaken with the consortium at Loughborough University directly demonstrated the real-world potential of rapid prototyping technologies. As part of Save the Children’s ‘Festival of Trees’ charity event, we helped produced a two-metre-high nylon tree. It comprised two main elements – a silver-plated cone containing an LED light within and an outer layer of ‘foliage’. Due to its size and complex geometry (the design of the leaves was generated by a customised computer programme), it was necessary to manufacture the structure in components. At this point, construction and fabrication parameters were still relatively strict – it was crucial that all components could be efficiently nested within the rapid prototyping machine’s build chamber. Nevertheless, the project offered a first taste of what it could be like to design, print and manufacture a large-scale final product.

We designed a structure inspired by the lightweight internal structure of bird bones. Following the principles of optimisation in nature, the structure worked on the premise that material is expensive but shape is cheap – it costs very little to devise a complex shape in nature, it’s data driven, but material takes energy to produce and mould. The efficient internal structure ensured the minimal use of regolith binder whilst maximising the structure’s strength – an important factor when considering the Moon’s volatile surface environment.

A full-scale prototype printed using a material approximating lunar regolith demonstrated the feasibility – in both production and structural terms – of the design, which has since been used to inform subsequent iterations of the lunar habitat. Not only has this research contributed to a coherent vision of permanent habitation on the Moon, and potential developments on Mars, but it has similarly contributed to the discovery and advancing of optimal structures that can be used here on Earth.

In this case, the research was geared towards developing a computational design tool that could drive the distribution and alignment of materials within a building part according to the stresses inside the required shape. The open-cell micro-structure of bones varies in density and alignment in response to stresses, so that regions of high stress in your bones are nearly solid and vice versa. We eventually developed this research into a software tool that facilitates the design of 3D-printed, structurally efficient building blocks of a non-standardised, customisable shape. And critically, at speed – the generation of multiple designs and subsequent tests undertaken to determine the highest-performing, most materially efficient structure for a design were no longer necessary. The best structure for a component could be determined at the click of a button.

With Maggie’s, the goal was for a cantilevering glulam beam that could be optimised to minimise weight, maximise strength, and demonstrate a unique design aesthetic. The practice’s research into optimised structures, efficient materials and the development of software that could best distribute this material according to specific structural requirements made this project possible. And interestingly, it revealed the aesthetic, as well as sustainable, potential behind structural optimisation.

The Maggie’s beam is a promising candidate for rapid manufacturing, although the graduation from rapid prototyping has yet to occur at the level of architecture. Yet advances in other industries are encouraging – many Formula 1 teams already take advantage of rapid manufacturing. Renault F1 cars use a cooling duct manufactured using 3D printing. Its complex geometry meant that it couldn’t be manufactured using traditional methods. Significantly, the duct has gone through several iterations as the car has changed and evolved, and the speed of the manufacturing process means it is able to keep up.

Our role in the consortium was to develop a new design approach incorporating all the manufacturing constraints – such as the resolution of the metal deposition, the reach of the robots, or the radii of machining tools – and produce a demonstrator part to show its potential application in the construction sector. Taking inspiration from Maggie’s, we sought to create a structurally and materially optimised steel cantilever beam.

'This is a truly advanced method of construction, and one that offers a flexibility and sustainability so far unseen in the construction industry.'

Over three years, we helped developed the tools and associated software that would make the printing of metal building elements possible. Incorporating both additive (3D printing) and subtractive (milling – a standard manufacturing process) properties, the robot combines new and existing technologies to produce a method of fabrication that can not only produce customised, optimised pieces, but also do so at speed.

This is a truly advanced method of construction, and one that offers a flexibility and sustainability so far unseen in the construction industry. At this early stage, 3D printing remains relatively costly as a method of production compared to conventional techniques, but the experimental designs developed during this project helped us understand where value can be added. That is, for optimised bespoke parts with complex geometries.