Reconfigurable Mould

Problem and objectives

The main objective was to develop a reconfigurable mould for producing volumes by casting crushed container glass. The system needed to support geometric variation without rebuilding the entire device and to use materials and technologies available locally.

The specific objectives were to:

• incorporate parametric modelling and 3D printing into the development of the matrix;
• manufacture and refine a prototype through laboratory testing;
• replace the lost-mould process with demountable, potentially reusable components.

System and design decisions

The device consists of three elements:

1. Matrix. A demountable base, walls, lid and separator printed in PLA 850. It defines the outer volume of the ceramic mould and allows its configuration to be adjusted.
2. Pattern. An interchangeable geometry placed within the matrix during forming. Its division into quadrants allows different variants to be combined.
3. Mould. A refractory ceramic component formed from the matrix and pattern. After firing, it receives the crushed glass for kiln casting.

The first proposal used six hexagonal components and an articulated metal frame. Digital analysis identified demoulding problems and fixed component positions that conflicted with the intended interchangeability. The second version reduced the number of joints and organised the system around demountable quadrants.

Every printed component has at least one flat face to avoid support material during FDM production. Perforations reduce material and printing time; joints, ties and guides control the assembly before the paste is introduced.

Development and validation

The matrix was modelled parametrically in Fusion 360 and printed in PLA 850. The editable feature history supported revisions to dimensions, joints and draft angles before manufacture.

The mould material used a refractory paste containing 30% Tinkar plastic clay, 30% kaolin and 40% fine grog, with 16–18% water. Its pseudoplastic behaviour required manual compaction to reproduce the details of the matrix.

Eight tests were documented. The initial tests refined the formulation, compaction and demoulding procedure. The non-porous PLA walls retained moisture and caused adhesion; adding a nylon film and partially removing the matrix after two days improved drying while preserving the form.

The process uses two thermal cycles: firing the ceramic mould at 1190°C and casting the glass at a peak of 875°C, followed by controlled annealing. The complete sequence covers matrix assembly, paste placement and drying, mould firing, glass charging, kiln casting and release of the finished piece.

Results and limitations

The prototype verified the fit of the printed components, the forming of the refractory mould and the separation of its quadrants. The eighth test completed one casting with crushed glass: it produced a solid volume, and the mould remained intact after the thermal cycle.

This result validates the system's basic operation in a laboratory test. The documentation does not yet establish the number of reuse cycles each mould can withstand, dimensional stability across repeated firings, or a quantified reduction in waste or cost.

Proposed next steps include adding a glass reservoir to compensate for volume reduction during fusion, validating guided stacking within the kiln, and investigating patterns produced through manual modelling, 3D scanning and parametric segmentation.

Publication and documentation

• Academic title: Diseño de molde reconfigurable para el desarrollo de volúmenes con molienda de vidrio hueco.

• Author: Santiago Abásolo Montes de Oca.

• Supervisor: Beatriz Amorín.

• Institution: Universidad de la República, FADU, Escuela Universitaria Centro de Diseño.

• Year: 2025.

• Official record and full thesis — Colibrí Repository, Universidad de la República