Awards
Awards
Cast Glass Giants
Creating massive free-form cast glass structures with the aid of Topology Optimization and 3D-printed sand molds
To efficiently achieve large, monolithic cast glass structures, TU Delft is developing structural Topology Optimization algorithms that incorporate the characteristics of glass as a material and of casting as a manufacturing method. With the aid of Topology Optimization, intriguing designs of monolithic cast glass structures of decreased mass and thus, of shorter annealing times can be generated, resulting in considerably improved manufacturability in terms of time, energy and cost-efficiency.
To realize such, free-form customized cast glass structures of complex geometry, the team is investigating the use of 3D-printed sand molds as a high-accuracy, cost-effective fabrication method.
The challenge of glass recycling
Despite the common notion that glass is a 100% recyclable, currently, only the glass container industry implements its successful closed-loop recycling back to the same product. For most other glass products, including architectural and automotive glass, the closed-loop recycling rate is remarkably less and generally restrained to pre-consumer waste. Multiple technical and supply-chain barriers hinder glass recycling. These include recipe incompatibility, contamination by foreign matter or even color, labor-intensive disassembly of multi-material products, logistical challenges such as transportation costs and waste treatment, absence of recycling facilities for specialty glass, and strict quality standards. Of equal importance is the lack of high-value products designed from recycled glass. As a result, high-value glass products at their end-of-life are either landfilled or downcycled to low-value applications (e.g. aggregate). In essence, glass waste, remains a significant and unresolved problem.
The underexplored shaping potential of cast glass
Cast glass is a promising new structural material for architectural applications. It has a vast shaping potential; it can virtually take any shape and size, allowing us to envision storey-high glass columns and entire monolithic envelopes. So far, this shaping potential has been little explored in the built environment, mainly due to the perplex and lengthy annealing process involved for glass components of substantial mass/thickness, which in turn renders their production unrealistic. As a result, structural cast glass components are typically applied in the form of solid blocks comparable in size to standard bricks. Key factors that affect the annealing time are the thermal expansion coefficient of glass, the number of surfaces exposed to cooling, and the thickness and overall mass of the object. Essentially, the thinner the cross-section of the cast glass component, the exponentially less the annealing time.
Topology Optimization for generating mass-optimized, monolithic cast glass structures
With this in mind, TU Delft has been developing structural Topology Optimization (TO) algorithms for designing monolithic, 3-dimensional cast glass structures in architectural scale. To ensure the structural integrity and manufacturability of the cast glass structures, novel Topology Optimization algorithms are developed with constraints that refer both to the glass material/structural properties, such as principal stresses, deflection and strain energy, and to criteria that ensue from the annealing and fabrication processes. Although considerable research has been devoted to improving the manufacturability of TO structures, the constraints associated with casting have received little attention and the consideration of annealing has largely been ignored. In the case of cast glass, incorporating annealing constraints into the optimization problem is an essential feature that needs to be accommodated, whereas the tensile strength of glass, which is at least a magnitude lower that its compressive strength, invokes asymmetric stress failure criteria that cannot be captured by conventional ductile plasticity surfaces or uniform stress constraints. This asymmetric stress behavior can be achieved by either implementing the Drucker-Prager criterion or unified functions that can serve different failure criteria. The conjoined research by TU Delft and MIT has so far led in the development of three distinct topology optimization algorithms, tackling both 2D and 3D optimization, which take into consideration the aforementioned peculiarities of glass as a material and of casting as its fabrication method.
Realizing massive cast glass structures with the aid of 3D-printed sand molds
But how can we fabricate such complex, highly customized cast glass geometries in a cost-efficient way? Disposable, 3D-printed sand molds of high accuracy, already used for metal castings are a promising solution. Compared to the laborious and time-consuming process of standard investment cast molds, 3D-printed sand molds are quick and easy to make and allow for great complexity in shapes, including undercuts and voids. Such molds are particularly suitable for casting complex geometries, e.g. such as the ones resulting from a topological optimization design process. Currently, glass casting on disposable molds faces the major drawback of a resulting rough and opaque glass surface quality, requiring considerable post-processing to yield a glossy, smooth surface. This in turn results in a compromised dimensional accuracy and on increased time and production costs. If the surface remains unprocessed, it can greatly affect not only the visual but also the mechanical properties of the cast glass element. Thus, to investigate the potential of this mold technology in glass casting, several experiments are conducted at the Glass Lab of TU Delft focusing on identifying the best binder and sand combination for glass casting and on achieving a glossy surface quality directly upon demolding without any additional post-processing. Therefore, the research focuses on a series of kiln-cast laboratory experiments at various max. firing temperatures and annealing schedules, involving mold samples printed by ExOne using different sand types and binders and treated with various refractory coatings, coating combinations and protective layers for surface finishing. The results showcase the high potential of this technology for customized, highly complex structural cast glass elements.
Credits
Topology Optimization algorithms for cast glass
R&D : TU Delft (Faidra Oikonomopoulou, Charalampos Andriotis, Anna Maria Koniari, Eva Schoenmaker, Telesilla Bristogianni) in collaboration with
MIT (Josephine V. Carstensen, Jackson L. Jewett)
Glass casting on 3D printed sand moulds
R&D : TU Delft (Menandros Ioannidis, Faidra Oikonomopoulou, Telesilla Bristogianni)
Material sponsorship and consultation: ExOne (Andreas Muller, Leonhard Stöckle)
Hüttenes-Albertus - HA Group (David Hein)
