Optimization and Finite Element Modelling of Powder Segregation during Metal Injection Moulding

摘要

Segregation is frequently encountered in injection moulding. In order to eliminate this an optimization method is developed. Firstly a mathematical function is used to model the segregation process. Design of experiments is then used to select the parameters which have most influence on segregation. Four parameters are found to be significant in this respect: Powder volume fraction, Interaction coefficient, Powder density, Binder density. In order to minimize the effect of these parameters on segregation, a generic algorithm is proposed. Coupling simulation results with injection moulding software "FEAPIM" is found to lead to excessively long computer processing times. To shorten this, therefore, the surface response of the four parameters is constructed using the Moving Least Square Approximation method (MLSA). This is found to give a good approximation. The optimization strategy proposed in this paper has been applied to tensile test specimens. Good correlations have been achieved between predictions and experimentally produced components.