Projects / Programmes
Modelling rheological properties of ceramic-paraffine suspension for low-pressure injection- moulding
| Code |
Science |
Field |
Subfield |
| 2.07.01 |
Engineering sciences and technologies |
Computer science and informatics |
Computer structures, systems and networks - software |
| Code |
Science |
Field |
| T120 |
Technological sciences |
Systems engineering, computer technology |
artificial neural networks, modelling, simulations, injection-moulding, rheology
Researchers (4)
| no. |
Code |
Name and surname |
Research area |
Role |
Period |
No. of publicationsNo. of publications |
| 1. |
15654 |
PhD Matej Andrej Komelj |
Materials science and technology |
Researcher |
2003 - 2004 |
176 |
| 2. |
05601 |
PhD Franc Novak |
Computer science and informatics |
Head |
2003 - 2004 |
316 |
| 3. |
04292 |
PhD Saša Novak Krmpotič |
Materials science and technology |
Researcher |
2003 - 2004 |
668 |
| 4. |
11972 |
PhD Drago Torkar |
Computer science and informatics |
Researcher |
2003 - 2004 |
91 |
Organisations (1)
| no. |
Code |
Research organisation |
City |
Registration number |
No. of publicationsNo. of publications |
| 1. |
0106 |
Jožef Stefan Institute |
Ljubljana |
5051606000 |
90,724 |
Abstract
With the proposed project we intend to solve the complex problem of modelling the rheological properties of ceramic-paraffin suspensions with regard to their composition, a key parameter in low-pressure injection moulding in the production of ceramic parts at AET Tolmin.
In the first stage we will simulate the moulding process and establish the influence of process parameters on the flowability. In particular, we will pay special attention to the dynamic viscosity. For a specific mould we will study both the shear-rate distribution in the mould and its dynamic range. The results will be verified using real systems from production.
In the next stage, based on viscosity measurements in the determined shear-rate range, we will construct a model of the visco-elastic behaviour. The model parameters will be predicted by using a constructed artificial neural network (ANN) that is trained with experimental data. We will test various topologies and learning algorithms and the best performing one will be selected. The model and the ANN's behaviour will be thoroughly tested with new experimental data.
Finally, another ANN will be constructed to model the relationship between the ceramic composition and the desired rheological properties. This represents a complicated inverse multi-value problem; therefore, we will take special care with the design of the learning algorithm. The test set will comprise data from an experimental database and from artificially generated examples using the ANN from the previous stage.
The result will be a methodological approach to the preparation of ceramic suspensions for low-pressure injection moulding. This will enable us to select the appropriate suspension composition for a specific mould.
