MAS 2018 / 19
Computational Clay Coral City
An investigation in robotic aggregation of clay for the design and fabrication of engineered coral reefs.
Robotic fabrication processes in architecture and construction
aim to bridge the complexity of computational design
methods with an efficient and coherent production technique.
For the robotic assembly of discrete elements, this often means
to fully control positioning and paths to physically match a digital
model, as well as using physically rigid building elements. However,
digital fabrication can also leverage new forms of control for
material processes with complex behaviors.
Unlike stiff materials where tight tolerances have most of the time
to be met for correct assembly, malleable ones allow for loose
connections to self level and interlock when enough force to deform
is applied. This thesis investigates such a process. It explores
and develops a common workflow for both the digital modelisation
and the physical production of robotic clay aggregation. The context
of engineered coral reefs, was used as a valuable design input
as well as a purposeful real world case to evaluate with.
Students: Eleni Skevaki, Nicolas Feihl
Tutors: Mathias Bernhard, Marie Griesmar, David Jenny, Jesus Medina
External Colaborators: Ulrike Pfreundt, Marius Neamtu
Digital fabrication technologies and additive manufacturing techniques opened new opportunities and challenges for the construction industry. Concrete Extrusion 3D Printing (CE3DP), introduced valuable opportunities for large-scale architectural elements design and fabrication. However, possible segmentation and assembly strategies havenít been exploited and remain as a limiting factor for the expansion of concrete 3d printing to industrial scale. In this context, the research focused on design and fabrication possibilities of assembly interfaces, an essential topic for scaled up 3D printed concrete components.
Students: Aya Shaker, Noor Khader
Tutors: Ana Anton, Dr. Lex Reiter
Controlled deformations as Design Drivers in Digital Casting
Conceptual simulation of Digital Concrete behaviour: Decades of research in concrete rheology lead to chemically/physically accurate viscous simulations at the cost of time and significant computational power. Contrasting to such, are conceptual simulations done for animation and/or film industry. These approaches often offer greater flexibility, tremendous art control and ‘realistic’ material behaviour. This thesis aims to develop a conceptual approach to describe and simulate the complex material behaviour during a novel digital fabrication process called Digital Casting and integrate the findings to a process-specific design tool. Through Digital control of fluidity and curing speed the concrete sets on demand, allowing for weakly supported formwork construction. Thesis simulation tries to encapsulate the entire process through various material behaviour (cloth; for weakly supported formwork, accelerated concrete behaviour; as viscous material) and various interactions that take place between them to form a design tool for further exploration. For sake of time, many areas that could stand to be a research on its own have been looked over to prioritise a completely procedural system that can later be broken into and further explored.
Student: Rahul Girish
Tutors: Dr. Ena Lloret-Fritschi, Anna Szabo
Generative Modelling with Design Constraints
Reinforcement Learning for Furniture Generation
Generative design has been explored in architecture to produce unprecedented geometries, however, in most cases, design constraints are limited during the design process. Limitations of existing generative design strategies include topological inconsistencies of the output geometries, dense design outputs, as the format used is often voxels or point clouds, and finally out-of-scope design constraints. In order to overcome such shortcomings, a novel reinforcement learning (RL) framework is explored in order to design a series of furniture with embedded design and fabrication constraints.
Experiments demonstrate that the proposed method of RL can generate a family of tables of unique aesthetics, satisfying topological consistency under design constraints. The model is successfully designed to fill a complicated geometry with a single stroke avoiding self-intersections as much as possible, meaning that the agent learns motion planning available for any geometry. In addition, a diverse range of visual expressions can be obtained by differentiating the geometric interpretations of the agent’s trail, e.g. piping along the trail or binary solidified voxels.
Student: Yuta Akizuki
Tutors: Prof. Dr. Benjamin Dillenburger, Dr. Mathias Bernhard, Reza Kakooee, Marirena Kladeftira
Digital Diffuser Design
This research focuses on the geometrical form, taking curved surface as a typology, as well as its sound diffusion effect. By means of digital design tool, 3D printed digital materiality, surfaces with com-plex formal conditions are computationally designed and fabricated. Robotic acoustics evaluation technique precisely detects the sound
diffusion manner of the surface and output data, which is visualized and fed back to the design optimization. The whole process generates an interactive design cycle in architectural acoustics.
Student: Chaoyu Du
Tutors: Achilleas Xydis
The research thesis presents the exploration to carving texture and pattern on 2.5D surface utilising the complex gestures of robotic movement, combining robotic fabrication together with customized tools to efficiently subtract clay material to fabricate the geometry in high resolution. The process is translating human carving skill to robotics and understand the fabrication data through a series of basic experiment to generating the toolpath from the curvature of geometry for carving clay material to get different carving effects on robotic movement. The workflow is unlike the pick-and-place style robotic control, it is considered about the carving effects generated by the whole tool path and tool orientation. The techniques of fabrication system is applied on the rapidly prototyping, and the further application shows the potential in subtractive robotic fabrication in clay material.
Student: Chen Ying-Shiuan
Tutor: Zhao Ma
Robotic timber assembly technologies present novel fabrication methods to achieve digital assembly of timber structures. The sequential roof and DFAB projects apply these technologies to realise full scale architectural structures. These digital assembly processes present the question of what the design process could be.
Parametric tools in architectural design present the possibility of extensive geometric freedom in the absence of performance. The results of such a process is often not fabricable as they are unaware of the fabrication constrains. Particularly, in the initial design stages, it is not only time intensive but also constrains exploring new design possibilities. Addressing the need for a fabrication aware design process, this thesis proposes a grammar based design methodology by creating shape transformation rules with fabrication constraints and capabilities integrated into an interactive computational design tool, Timber_Grammar. The tool provides functionality to apply and combine grammar transformations to create multiple design possibilities that are fabricable.
Student: Keerthana Udaykumar
Tutor: Victor Leung
Designing a Multi-Property Dispenser For Grading
Colored glass in architecture has a long history, ranging from stained colored glass, to high performance glass with foil inlays and functional grading. The potential of glass to perform as a structural material and the need for sustainable construction materials has opened the door for rethinking the role of glass within architecture. However, commercial production methods of glass do not allow for color grading glass. Yet one artisanal production method - namely the fusing of glass - has the potential to be pushed past its traditional scope.
The goal of this thesis was to develop a robotic tool that was able to grade colors with a certain accuracy in glass. The developed multi-property dispenser allows for the creation of volumetric designs that incorporate color and translucency gradients using a robotic arm as a 3D drawing machine.
Students: Jonas Van den Bulcke, Sofia Michopoulou
Tutors: Rena Giesecke, Pietro Odaglio
3D Printed Foam-Cored Sandwich Panel Façade
Foam cored sandwich panels have a high strength to weight ratio; and therefore, enable the construction of large structural spans and cantilevers, creating efficient and lightweight architecture. Conventional methods for fabricating sandwich. panels are efficient for standard geometries, but when used for complex geometries, they become time, cost and labor intensive due to the need of unique and expensive molds. Therefore, such applications can be afforded mainly in industries with high budgets like the aerospace industry, or in standardized manufacturing where molds have multiple uses like in the car industry.
Thus, more efficient fabrication solutions are needed for the production of bespoke sandwich panels. Additive Manufacturing (AM) technologies like Fused Deposition Modeling (FDM) 3D printing (3DP) offer freedom and flexibility in the design of non-standard geometries and are furthermore material efficient, cost-effective and require minimal labour allowing mold-less fabrication. Taking into account the advantages of 3DP the goal of this research is to produce a large-scale freeform composite sandwich element. This will be accomplished by initially investigating insulating foams and internal core geometries for integrated panel functionalities (insulation and translucency gradients), as well as the design of smart connection details for assembly.
Students: Nik Eftekhar Olivo, Yoana Taseva
Tutors: Matthias Leschok, Hyunchul Kwon
Design Freedom in Eggshell
The thesis researches a design space to an innovative technique developed by GKR in “Smart Dynamic Casting.” It focuses on testing the limitations of the fabrication method with ambitious design objectives. A clear methodology of testing with the already provided tools generates a framework of conclusions and interrelated data which is used to create one-to-one scale prototypes. The research starts with questions that arose from previous findings and continues unto further relevant questions. This helps achieve or draw closer the goals that the initial doctoral research strived to reach, such as: (a) cast-on-demand concrete (b) prototyping complex geometries (c) integrated reinforcement. And this way become a bigger player in “Digital Fabrication with Concrete".
Students: Antonio Barney, Wenqian Yang
Tutors: Joris Burger, Lukas Gebhard