Die compaction simulation simplifying the application of a complex constitutive model using numerical and physical experiments /

by Wagle, Gautam Subhash.

The die compaction process is a rapid net-shape manufacturing process that yields low strength parts which are then sintered to create a functional part. The sintering stage induces shrinkage inversely proportional to the density of the part and distortion if density gradients are present. For an accurate description of the final shape and size, the amount of shrinkage and distortion must be anticipated and incorporated into the original tool design. Numerical modeling tools can be used to develop protocols to attain desired compact properties. The die compaction process is a complex process as the material undergoes particle rearrangement followed by plastic deformation. Modeling the process requires the use of a complex material model. The modified Drucker-Prager “cap” model can describe the loose powder response and the particle deformation under compaction loads. This research looks at simplifying the application of a numerical finite element model defined using this complex model. The significance of the parameters that define the material model on the results has been determined using a robust sensitivity analysis technique. The results from the analysis have been used to identify the critical parameters for density predictions. Since the recommended testing procedures used to characterize the material parameters are expensive and difficult to perform, alternative testing methods have been investigated for ease in industrial application. A testing protocol using a test method proposed by Coube and Riedel (2000) and a technique proposed in this research has been developed for characterizing the material parameters for the “cap” model. The protocol has been applied for characterizing two commonly used metal powders for die compaction applications: a water atomized A1000C iron powder and a water atomized 316L stainless steel powder. A numerical finite element model with the characterized parameters has been verified for application to predict density gradients in a die compacted part by comparing the numerically predicted density distribution to the density field obtained from physical measurements. The results of the verification establishes a good predictive capability of the numerical model and the testing protocol developed in this research. iii