Probabilistische numerische Untersuchung zum Einfluss variierender Gießparameter und Geometrie auf die Gefügequalität von Feingussprodukten
David Pusch, Robin Schmidt, Matthias Voigt, Konrad Vogeler
abstract:
In the present paper the application of probabilistic methods to the numerical simulation of an investment casting process is studied, using the example of a step wedge. The objective of the combination of probabilistic methods and numerical casting is the optimization of the manufacturing process with respect to the metallurgical quality of the casting or the targeted adjustment of the process by identification and appropriate modification of the determining process values. In order to achieve this, the influence of scattering process parameters on the change in microstructure of the cast component is analyzed in terms of a sensitivity analysis.
The paper focuses on the investigation of the actual pouring process under vacuum and the adjacent cooling as well as solidification. The process parameters as the input quantity are distinguishing the process; these are e.g. casting temperature, shell temperature or geometry (wall thickness). Characteristics of the microstructure as the output quantity are the porosity and the secondary dendrite arms spacing in the whole casted part. The deterministic calculation is done with the commercial software ProCAST of the ESI GmbH.
The Monte Carlo method paired with an optimized Latin Hypercube sampling and a statistical evaluation of the results is used to identify the scatter of the microstructure of the cast part and its sensitivity towards the input parameters of the casting process. Measures for the sensitivity are the correlation coefficient or rank correlation coefficient as well as the coefficient of importance along with a coefficient of determination of the used response surface. A better understanding of the discovered relations is shown by plotting the statistical data on the finite elements mesh.
Keywords:
monte carlo method, investment casting, probabilistic, numerical casting simulation, ProCAST, IN738LC
reference:
Pusch, D., Schmidt, R., Voigt, M., Vogeler, K., 2013
"Probabilistische numerische Untersuchung zum Einfluss variierender Gießparameter und Geometrie auf die Gefügequalität von Feingussprodukten",
Proceedings of International Colloquium of the Cluster of Excellence ECEMP 2013.