Partners and International Organizations
(English)
|
Alusuisse (CH), Calcom (CH), Ecole Polytechnique Federale de Lausanne (CH), Elkem (N), Hoogovens (NL), Hydro (N), Institut National Polytechnique de Lorraine (F), Pechiney (F), Technische Universiteit Delft (NL), VAW (D)
|
Abstract
(English)
|
Cast aluminium of large format slabs often suffer from severe inhomegeneity in chemical compositions, due to so-called macrosegregation effects. This segregation leads to non-uniform product performance resulting in product rejection and energy losses of about 850'000'000 MJoule/year. Existing casting simulation models only give insight in thermal, fluid flow and stress distributions during the casting process. No reliable predictive tool is available to quantify compositional segregation. Therefore the European Community and 10 European Industries and Research Institutes have agreed to a Brite-Euram project (Proposal N0 BE95-1112 Contract N0 BRPR-CT95-0112) aimed at developing a numerical model describing and predicting segregation.
The following goals were set up at the beginning of the project: 1) Develop a microsegregation module that would handle multicomponent alloys, back-diffusion in the primary phase and precipitation of eutectic phases; 2) Couple the microsegregation module with phase diagram data; 3) Implement shrinkage into the macrosegregation calculation; 4) Couple the micro- and macrosegregation modules in order to account for both static and continuous casting situations; 5) Implement the movement of the grains into the code; 6) Validate the code with small casting experiments and benchmark calculations; and finally
7) Validate the code by comparison with real size DC casting macrosegregation measurements.
The main contributions to these goals can be summarised as follows: 1) EPFL and EMN, in collaboration with Sintef, defined a systematic approach to multicomponent alloy macrosegregation calculations; 2) EMN and Sintef developed a microsegregation module that was applicable to binary alloys, non-linear phase diagrams and eutectic precipitation, based on an approach developed jointly with EPFL; 3) EPFL focused on the development of a microsegregation module coupled with the software ThermoCalc for the ternary Al-Mg-Si alloy; 4) Considering the difficulty encountered with the Al-Mg-Si alloy, EPFL and Calcom developed a simpler approach for multicomponent alloys, based on a pre-existing simple model for binary alloys; 5) Calcom developed a new version of the FEM fluid flow solver in the software calcoMOS, based on a Galerkin Least Squares (OLS) method. This was done in order to handle more easily the shrinkage contribution; 6) Calcom and EPFL coupled the various microsegregation modules with the macroscopic code, including the shrinkage contribution and the transport of the solid phase, which required a mixed Eulerian-Lagrangian formulation; 7) EMN developed a module to calculate grain movement and this module was integrated in calcoMOS by Calcom; 8) EPFL defined some benchmark tests for static and small DC castings; 9) EPFL performed small scale axisymmetric casting experiments on Al-Mg and Al-Cu in order to validate the macrosegregation calculations; 10) EPFL and Calcom performed a large number of tests in order to assess the accuracy, stability and features of the new developments and to see the influence of physical phenomena (shrinkage, back-diffusion, linear/non-linear phase diagram, etc.) and calculation parameters (mesh size, time step, etc.) on the calculated results.
The goal of Calcom was also to finalise a version of the software calcoMOS and of the various modules that have been delivered to the project partners.
For more scientific details, the reader is referred to the various scientific reports and articles produced in the duration of the project, which are quoted below, and in particular to the two PhD theses written at EPFL by T. Jalanti and X. Doré on the topics of macro- and microsegregation, respectively.
|
