© Digital Building Technologies | Gramazio Kohler Research | ETH Zurich
Master of Advanced Studies ETH in Architecture and Digital Fabrication
MAS 2020/21 Thesis Projects
Robotic assembly of self-interlocking spatial structures
This thesis project investigates the automated assembly of lightweight modules facilitated by the use of additive manufacturing in architecture. The research focuses on computationally designed 3D printed connections for space frame modules suitable for robotic and drone assembly. Spanning large distances with little material, these characteristics, space frame structures are being further explored and developed as a highly industrialized system. For the investigation of this thesis, the goal is to design an ultralight structure, whose rods are connected with 3D printed interlocking joints with no additional fasteners. This research aims to define a new system where the assembly of space frame structures is automated with the help of robotic arms and thus reduces human labor on site. 3D printed joints allow a high degree of flexibility in the design and guarantee a time-efficient fabrication. These processes generate a prefabricated system, in which single modules are first assembled manually and secondly those are directly mounted on-site with drones and robotic arms.
Students : Pascal Bach & Ilaria Giacomini
Thesis Supervisor : Marirena Kladeftira
In collaboration with : Marco Hutter, Robotic System Lab
Ground Truth: data-driven robotic sand excavation
Digital fabrication has the potential to enable more sustainable architectural construction, by allowing the use of abundantly available natural granular materials. However, these materials and the processes by which they are manipulated lead to increased levels of complexity. An emerging trend in the field of architecture is the use of machine learning models to predict material behaviour and fabrication parameters. This has the potential to short-cut expensive simulations while also being more adaptable to the current (i.e. actual) situation through 3D scanning. In this research, an image-to-image translation architecture that uses a generative adversarial network (GAN) is proposed to incorporate material behaviour in the generation of fabrication data. After training the GAN, (1) complex sand behaviour is predicted prior to robotic excavation; (2) the model is used to search for trajectories that produce a desired terrain geometry. Finally, this research aims to explore the possibilities to feed data-driven-models back into design exploration.
Ko Tsuruta & Simon Griffioen
Thesis Supervisors: Ryan Luke Johns, Jesús Medina Ibáñez
Robotic Sand Dropping
Robotic Sand Dropping (RSD) explores a robotic additive manufacturing technique for depositing sand droplets to fabricate ephemeral sand structures. With no strong architectural precedents it starts with the development of an incipient fabrication method to understand the material behavior of the robotic sand dropped structure. The thesis adopts an investigative, experimental and material-driven approach to seek possible geometric performances and applications of RSD. This requires the development of a custom-made end-effector based on a syringe system for dropping, as well as a computational process that translates the digital model into correct arrangement of points for the robot to carry out the precise positioning. Furthermore the robot setup is used to systematically test different mixtures of sand and binders to be able to predict various temporariness of structure accurately, and explore the design potential according to it.
Yen Fen Chan
Thesis Supervisors : Jesús Medina Ibáñez, David Jenny
Material Consultant : Anna Szabó
3D Printing for Kinetic Architecture: Rigidity, flexibility, and Actuation
Kinetic architecture is a concept through which buildings are designed to allow parts to move, which has been an attraction to architects and engineers. A number of kinetic buildings have been created from the late 20th century, using motors; however, the primary challenges of this method are mechanical units’ maintenance and their high energy demand. To address these shortcomings, this thesis explores 3D printing (3DP) in combination with shape memory alloys (SMAs) that can change their shape when a temperature stimulus is applied. A significant advantage of 3DP is the precise control of geometries, resulting in the effective control of stiffness that allows gaining high control over the shape-morphing behavior of kinetic elements. SMA-actuated 3D prints offer several advantages, including no noise, lower energy consumption, and longer working life.
In this context, this thesis develops an experimental program with a series of tests, enabling higher controllability of shape morphing behavior through complex geometric features and investigating possible approaches to the material- and energy-efficient actuation. For material-efficient shape-morphing, a minimum amount of SMA-wires is embedded as a muscle into flexible and motion-optimized 3D prints.
Thesis Supervisors : Kwon Hyunchul, Prof. Dr. Moslem Shahverdi (Empa)
Skin and Bone : Eggshell 3D Printed Formwork with Mesh Mould Reinforcement
This thesis topic aims to investigate the integration of two on-going and applied research projects at Gramazio Kohler Research: Eggshell and Mesh Mould, which operate within the digital concrete fabrication domain. Eggshell is a fabrication process that combines large-scale robotic fused deposition modeling 3D printing with simultaneous casting of a fast-hardening, set-on-demand concrete. Mesh Mould is a robotically fabricated stay-in-place formwork and reinforcement for waste-free non-standard concrete construction. In order to tackle the challenge of successfully integrating these technologies into a combined constructive system, this thesis focuses on physical prototyping experiments that incorporate robotic fused deposition modeling, 3D printing, digital material processing and robotic assembly to develop approaches that combine metal meshes with 3D printed formworks. The investigation of a coherent ‘skin and bones’ system is established as the main objective for this research topic and the resulting prototypical wall model showcases the integration of the two systems through a custom segmentation and connection approach developed and informed by the research process.
Thesis Supervisor : Joris Burger, Dr. Ena Lloret-Fritschi, Dr. Ammar Mirjan
4D Printed Formworks for Concrete Spraying
Concrete is the most used material after water and one of the most significant contributors to CO2 emission worldwide. Therefore, it is necessary to investigate how we can use less of this material in the near future. A project doing so is AeroCrete, a robotic spraying system that uses permeable meshes as the formgiving part of the concrete, with the objective of producing structurally optimized shapes. However, shaping these meshes remains complex and requires extensive manual labor. This thesis aims to explore the technology of 4D printing for the making of permeable and sprayable formworks using a predefined digital design pipeline. 4D printing is a process by which a flat printed geometry reshapes into a 3D geometry through the influence of applied environmental stimuli such as heat, moisture, pre-stress, pre-tension, and more. So far, this technology has mainly been developed for small-scale applications such as product or fashion design and medical devices. However, its potential application to architecture i s entirely new. Thus, the goal of this thesis is to explore the design potential of 4D printing when applied to stay-in-place formworks for concrete spraying, specifically within the project AeroCrete.
Wei Chengyuan & Guillaume Jami
Thesis Supervisors : Ena Lloret-Fritschi, Selen Ercan Jenny, Nicolas Feihl, Hamilton Forsythe
Imprinting Concrete with Glass Inlays
Conventional methods of construction lack management of material usage, time and are highly labour intensive hence automation in the construction industry is rapidly growing. Currently, most robotic systems cater to only individual functionalities required for building architecture. Combining multiple materials into one system could further innovations for the field. The research aims to develop a fabrication method for combining two materials with 3D printing and pick-and-place robotic processes to construct a transparent concrete wall. The core challenge with this research is the integration of elements being placed while printing. To be able to tackle this, the research will develop a computational tool which allows to design different strategies for placing glass elements while simultaneously printing concrete which is then translated through a series of prototypes. The research would also focus on bridging the communication between both the robots and to design an end-effector for this system. To illustrate this process we use recycled laboratory glass with the pick-and-place robot and white concrete is the material for 3D robotic printing. To conclude the research a final 1:1 scale prototype allows us to reflect and evaluate the method.
Liya Sunny Anthraper & Wei-Ting Chen
Thesis Supervisors : Ana Anton, Eleni Skevaki, Lex Reiter
Ice Formwork System for Concrete Shell Structures
Structural design is often centered on material efficiency (material quantity and strength) rather than material effectiveness (right application). Concrete shells are efficient structural forms that achieve strength through geometry and can significantly reduce resource consumption by placing material only where needed. However, the fabrication process of the concrete shells’ discretised components or the formwork system required to construct them, are often wasteful and remain challenging.
This thesis presents a material-efficient design and fabrication framework for casting of non-standard concrete elements for shell structures. The framework integrates computational design tools for shell structures with the ice formwork as fabrication system for concrete casting. The presented methodology demonstrates how the ice can be used to control the density and porosity of concrete while increasing the structural depth. Beyond material efficiency, the ornamental patterns of void created by ice, result in unique visual and lighting architectural qualities that are difficult to achieve with standard formwork systems for complex concrete shell structures.
Thesis Supervisor : Vasily Sitnikov
Thesis Consultant : Juney Lee
Graded Ice Matrix Structures
Concrete has historically been one of the most popular building materials thanks to its availability, durability and versatility. Thanks to its ability of taking virtually any form, a thoughtful design can allow us to build more with less. There are some important drawbacks in non-standard concrete casting. The production of non-standard concrete is labor intensive, time consuming, expensive and producing large amounts of formwork waste. Several researchers have focused on developing no-waste, recyclable and self-demolding concrete formwork material systems. Graded Ice Matrix Structures is a research for a controlled, lean-production-model of spatially graded concrete elements using ice formwork. It taps on Vasily Sitnikov’s Ice Formwork research and aims to contribute to the design and production of controlled and precise spatial grading of concrete components.
Ice aggregate is investigated in terms of geometry, size and packing densities within two main groups: regular and autonomous. The resulting concrete elements, negative cast of the graded ice matrices, are then evaluated for their structural integrity, their performance on a larger scale and their potential applications and affordances in architectural design.
Thesis Supervisor : Vasily Sitnikov
Advisor: Ena Lloret Fritschi
Augmented Wood Construction
Augmented wood construction explores the design, fabrication and assembly of intricate woodwork using computational design methods and visual augmentation, and applies and validates the concept in a real scale prototype of a balustrade for the historic Schatzalp hotel in Davos, Switzerland. As each balustrades don’t align with contemporary construction norms they will need to be replaced. Taken into consideration the changing circumstances of traditional craft, the question arises if new technologies such as augmented reality (AR) can be used in design and fabrication in order to identify an appropriate while efficient solution for the replacement and adaption of elements in historic buildings and preservation.
Thesis Supervisors : Petrus Aejmalaeus-Lindström ,Lauren Vasey, Matthias Helmreich and Fabian Kastner, Chair of Construction History and Preservation
In collaboration with Tim Sandy, Fadri Furrer INCON
Immersive Collaboration : Virtual and Augmented Reality for onsite fabrication
The existing use cases of extended reality (XR) technologies in the Architecture, Engineering and Construction (AEC) industry, highlight innovative augmented fabrication techniques and immersive virtual reality (VR) design processes. However, communication challenges during on-site/off-site collaborations that lead to ‘as design’-’as built’ conflicts and delays are not yet overcomed. The proposed method is based on the creation of an XR system incorporating the development of two applications that connect users with augmented reality (AR) and virtual reality (VR) devices. This method enables an online bidirectional data flow among collaborators on-site and off-site and offers new interactive workflows and communication patterns in design and AR-aided fabrication processes. To demonstrate our system, we selected two case studies that allow two users (AR, VR) to collaborate efficiently in two locations. The outcome is two transient installations out of sticky notes, computationally designed and dynamically adjusted to site-specific data. By testing our system, we discovered the potentials of merging XR technologies in design and fabrication processes. These new interactive interfaces facilitate design on the fly, online data exchange, and paperless construction site and could prevent delays, communication difficulties, and deviations from the original design.
Evgenia Angelaki & Foteini Salveridou
Thesis Supervisors : Daniela Mitterberger, Romana Rust and Lauren Vasey