The detailed description of the additive manufacturing process by experimental means alone could become totally unaffordable due to the high cost in both human and material resources. Such a complex process can be modelled in terms of partial differential equations which, due to their complexity as non-uniform geometries, heterogeneity and non-linearity of the materials, and complex boundary conditions, cannot be solved in exact form and it is for this reason that numerical approximations remain how an appropriate, if not the only, strategy to tackle them, which in turn requires the use of intensive computing power.

This is precisely the case that (i) numerical modelling and (ii) high performance computing are two of Cenaero’s core competences. Indeed, within the framework of this project, a sintering model has been implemented using the finite element method with a pressure-velocity formulation that allows the treatment of large deformations. This is a specifically developed and well refined implementation with, among other things, two noteworthy features (i) the material parameters depend on the local density and (ii) such density evolution is related to the permanent mechanical deformation. These features establish the link between the thermodynamics and mechanics of the physical phenomena in question

A numerical model has already been well tested and validated. In addition and during this validation, a thorough investigation of the appropriate material parameters to be used in the numerical calculations was carried out. It was found that, in general, the current implementation gives very satisfactory results. However, several niches of opportunity for improvement and refinement were identified.

In order to further identify and prioritise the above improvements and refinements, a detailed study of the influence of material parameters on the sintering model was prepared, conducted and analysed. A brief description of its general methodology, some remarkable results and the general conclusions of the study will be outlined in the following.

Since (a) the variables with the most important impact on the mechanical performance of the finished part are (i) the density distribution and (ii) the mechanical contraction during the sintering process, and (b) both can be accurately obtained as output variables from the already available numerical implementation, the material input parameters, i.e. eleven thermal or mechanical variables, are increased and decreased respectively in the order of 1%, 2%, 4%, 8%, 16%, 32% and 64%, The material input parameters, eleven thermal or mechanical variables, are increased and decreased respectively by 1%, 2%, 4%, 8%, 16%, 32% and 64% in order to detail their respective influence on the final part.

Figure 1 shows the typical responses of the output variables, i.e. density (left) and shrinkage (right), in these cases when varying the transition temperature (top) and grain temperature (bottom). Next to these responses, the base case (red line), which most closely resembles the reference case (magenta line), is also displayed for comparison. It is important to note that in all cases the base and reference cases do not match perfectly, even though the reference case is almost always within the spam covered by the various responses. This fact reinforces the validation of the available numerical implementation and fulfils one of the main objectives of the work described here.