MAS 2017 / 18

Numerical Sculpting for Fabrication Informed Spatial Printing

This thesis presents a method that is based on Volumetric Modelling (VM) for the generation of fabrication-informed geometry. VM offers modelling techniques for designing with great geometrical flexibility, what can be metaphorically called ‘Numerical Sculpting’, as well as for mapping and operating with fabrication constraints. These properties have not been widely explored in relation to spatial printing and its constraints. In this thesis the VM techniques are investigated through the design of wire arc additive manufacturing (WAAM) connections of standard elements that are robotically fabricated in place. A novel toolset of VM techniques is proposed that creates three dimensional printing paths meant for WAAM fabrication. A design language for WAAM connections is developed and results are fabricated as demonstrators.

Tutors: Inés Ariza, Mathias Bernhard

Student: Ioanna Mitroupolou


Additive nodes for non-regular space frames

Building non-regular structures of any kind, which deviate from the efficiency of industrialized production methods, is a difficult and expensive endeavor. With the development and rise in availability of consumer- and industrial-level additive manufacturing technologies it seems that, for the first time after an era of rapid industrialization, the production of unique building elements might become feasable again.

In order to enrich the toolbox for mass-customization and to open up the architectural design space a computational tool for creating 3d-printable nodes for non-regular timber space frames was developed. The tool handles the geometric and logistic complexity and provides the user directly with the necessary data for production. This enables the designer to spend more time on exploring the wide range of structural possibilities and focus on developing unique and aesthetically interesting compositions.

Tutors: Marirena Kladeftira, Matthias Leschok

Student: Alexander Enz


Design for Concrete Extrusion Printing

Concrete Extrusion Printing is a 3d printing technique using a fluid, extrudable mix of concrete that hardens upon deposition. Sharing the advantages of other digital fabrication techniques, Concrete Extrusion Printingis ideal for the creation of bespoke, freeform constructions that would otherwise require elaborate and expensive formworks that go to waste. Unique to the method is the flexibility to precisely deposit material for geometric complexity, material efficiency and structural needs, whilst also having, amongst other additive manufacturing technologies, an incredibly fast vertical build rate allowing for large scale implementation in architecture.

Despite increasing interest as a topic of research, Concrete Extrusion Printinghas yet to produce a novel aesthetic for such architectural applications. Enabled by the research context of ETH Zurich, the RFL setup for large prefab parts, and the versatility of a 6-axis robotic arm, this thesis will propose a design language that emerges from these unique advantages and constraints, exploring an expanded design space that is inextricably tied to the specific materiality, geometric opportunities and fabrication constraints of the method. Employing digital technologies, a new expression for concrete can emerge with the potential to imbue architecture with novel roles and form.

Tutors: Ana Anton, Lex Reiter

Students: Angela Yoo, Jun Su

Augmented OnSite Brickwork.jpg

Augmented On-Site Brickwork

The thesis "Augmented On-Site Brickwork" aims to expand the design space of manual brickwork assemblies to complex geometries using Augmented Reality (AR) technology. The firm materiality of brick and mortar as a building material, combined with computational design tools and fabrication methods, can form a novel convergence to produce complex spatial assemblies, integrating acoustic, daylight, or structural performance. The goal of this research is to assess the potentials and the feasibility of combining both human dexterity and digital precision realizing such brickwork. By that, the proposed augmented assembly method aims to provide a viable alternative to robotic fabrication in efficiency, robustness, and design possibilities. This thesis is conducted in collaboration the Robotics Systems Lab (RSL), ETH Zürich.

Tutors: Kathrin Dörfler, Selen Ercan, Tobias Bonwetsch, Timothy Sandy

Student: Fernando Ceña Martínez


Flexible Formwork for Concrete Curves

The thesis explores a flexible formwork system for materializing concrete curves in space. Due to the resistance and elasticity of the mould consisting of 3D printed TPU (thermoplastic polyurethane) joints and off-the-shelf PE (Polyethylene) pipes it is possible to build thin concrete structures with a minimal amount of material. The flexibility of the mould allows for self-calibration of a catenary network when the concrete is poured. Hereby, material behavior and computation are collectively used as tools for form-finding, calibration and fabrication. 


The research demonstrates how to embed material behavior and structural requirements into the computational model for a flexible system. Employing material properties and physics as tools for form-finding, in both, digital and physical models. The research explores the potentials and challenges of the system through physical prototyping and computational simulation continuously informing the design iterations and possible architectural formulations for concrete curves in space resulting from the language of the system. 

Tutor: Rena Giesecke

Student: Francisco Regalado


FDM Formwork for Discrete Concrete Stairs

Concrete has always been in the forefront of construction materials thanks to its inherent potential of being cast into any conceivable shape. Even though computational tools in architecture, through digital fabrication, have improved customized production and enabled a new architectural design freedom, the construction industry has no sustainable solutions for the production of free form complex concrete structures yet. In this context, the present research investigates FDM formwork for (ultra)-high-performance fiber-reinforced concrete to produce large-scale building components from discrete elements. This fabrication method provides designers with an unprecedented freedom in regard to mass customization and facilitates formwork fabrication off-site within a low-cost sustainable technique. The potential of the synergy between concrete and 3D Printing is showcased through the design and fabrication of a staircase, deepening its figurative, sculptural and functional aspects with its extensive capacity to give character to the architectural space.

Tutor: Andrei Jipa

Students: Georgia Chousou, Matteo Lomaglio


Performative Digital Composites for Lightweight Furniture

Towards lightweight architecture, the adversities of new fabrication technologies and materials have been continuously challenged. Meanwhile, 3D-printing (3DP) technology opens up the potential of mass-customization, which is a demand of the times. This research aims to develop a novel computational design in aids of a fabrication system utilizing lightweight materials, PLA (Polylactic Acid) and CFRP (Carbon Fiber Reinforced Polymer). The robotic tool-path generator, which is the outcome of the system based on the fabrication-aware design enables automated manufacturing, suitable for plastic fused deposition modeling (FDM) 3DP. The primary design driver is the visualization of the invisible structural flow that results from the physical interaction of human and object. Nowadays 3D-printed objects tend to be heavy, which is caused by the fundamental adhesion problem of FDM 3DP. As an alternative, adaptive topological design was applied to the usage of the structural data which enables strategic and flexible deployment of multi-material. The system demonstrates enhanced material efficiency and performance over other cases such as injection mold chair, tested in the one to one final demonstrator of the chaise lounge.

Tutor: Hyunchul Kwon

Students: Moon Young Jeong, Frank Cheng-Huang Lin


Robotic Concrete Spraying and Surface Articulation

Robotic AeroCrete is a method developed for producing geometrically complex thin-shell textile-reinforced concrete structures with articulated surface textures using a mobile robotic set-up. It is achieved by using air pressure to spray fiber-reinforced concrete with a controlled layer-thickness onto a permeable formwork such as a carbon-fiber mesh, which acts as reinforcement after the concrete has cured.

The robotic process involves the following steps: The permeable formwork is scanned to register its location and to reconstruct a surface geometry, from which the trajectories for the robotic glass-fibre reinforced concrete spraying procedures are derived. Firstly forming a structural frame to stabilize the mesh, followed by a full structural cover. Lastly, micro-fiber reinforced concrete is sprayed for an aesthetic surface articulation. 

To validate the full process, a final demonstrator for a bus stop is produced. This demonstrator highlights the high structural efficiency of such a thin-shell concrete structure, and the architecturally high-resolutionsurface texture inherent to its own process. The thesis research is conducted as a collaboration with the Industrial partner Bürgin Creations.

Tutors: Alexander Nikolas Walzer, Kathrin Dörfler

Students: Jetana Ruangjun, Nizar Taha


New Timber-Timber connections for robotic joinery

The traditional interlocking connections elaborated in close relationship to the equipment available to manufacture the joint geometry and the assembly process. They mainly rely on form closure through a tight-fitting and/or interlocking connection. The multi-robotic collaborative assembly allows rethinking how interlocking connections are formed and assembled. One of our main idea in this research is about how digital fabrication can translate and improve the traditional craft. By using CNC milling machine to design wood joinery this thesis investigates on interlocking wood-wood connections for linear elements in one spatial node, assembled by the slide, rotate and lock. At the same time investigation on assembly sequence of each element.

Tutor: Aleksandra Anna Apolinarska

Student: Sahar Barzani

spatial nodes oriented dowels.jpg

Spatial Nodes, Oriented Dowels

The thesis proposes to further investigate the joint typology that was developed within the MAS project “Gradual Assemblies, a summer pavilion for the Istituto Svizzero in Rome”, linear timber elements connected through spatially inclined dowels. Gradual Assemblies focused on the exploration of a simple connection detail applied on a large-scale prototype with complex geometry. Two oriented dowels, locking both rotation and movement of all elements formed the connection between all timber elements. Here, the potential of this connection detail to form an adaptive and performative joint typology will be explored with the goal to increase the overall structural stiffness. In addition to this, a computational method based on a network data structure allows to distribute and generate the joints in a larger aggregated system to test the adaptation capability in different situations.

Tutors: David Jenny, Andreas Thoma, Dr. Mario Rinke

Student: Yao Wang


Augmented Metal

Neuronal Stool is a series of stool created through generative design engine. It pushes the limit of the geometrical complexity produced by combining 3d-printed sand moulds and aluminum casting.


The design engine incorporates aluminum behavior in its geometrical rules to grow the curves for the stool and generates numbers of design alternatives. Then similar design alternatives are clustered into groups by machine learning. On the engine's user interface, users could choose design cluster and adjust performance parameters (castability, stability, and cost), based on which the engine generates a fitness landscape of designs within the cluster for users to choose for fabrication. It allows users to explore a variety of designs with options of their desired properties.


In this project, we fabricated two stools generated from the design engine. Every design thereby produced is guaranteed to be feasible for casting as the design engine considers material behavior of casting in the design process. Neuronal Stool demonstrates the potential of human-machine collaboration in exploring design space.

Tutors: Aghaei Meibodi Mania, Benjamin Dillenburger

Students: Haruna Okawa, ZongRu Wu


Rebar Assemblies

Despite innovations in computational design and digital fabrication, little attention has been paid so far to investigate on the structural and manufacturing implications of novel reinforcement distributions in concrete building components. In this context, a computational framework for the generation and structural validation of reinforcement layouts in reinforced concrete surfaces is presented. Featuring custom reinforced concrete design and non-linear finite analysis modules, the framework retakes on the concept of principal stresses and proposes a distance-constrained algorithm for generative tracing of stress-aligned reinforcement lines. 

Tutors: Zhao Ma

Students: Rafael Pastrana