A note on numerical simulations

The most common model for turbulence that used in the die casting industry for simulating the flow in cavity is k-k- . This model is based on several assumptions

1. isentropic homogeneous turbulence,
2. constant material properties (or a mild change of the properties),
3. continuous medium (only liquid (or gas), no mixing of the gas, liquid and solid what so ever), and
4. the dissipation does not play a significant role (transition to laminar flow).

The k-k- model is considered reasonable for the cases where these assumptions are not far from reality. It has been shown, and should be expected, that in cases where assumptions are far from reality, the k-k- model produces erroneous results. Clearly, if we cannot determine whether the flow is turbulent and in what zone, the assumption of isentropic homogeneous turbulence is very questionable. Furthermore, if the change to turbulence just occurred, one cannot expect the turbulence to have sufficient time to become isentropic homogeneous. As if this is not enough complication, we have to consider the effects of change of properties as results temperature change. Large changes in the properties such as viscosity have between observed in many alloys especially in the mushy zone.

While the assumption of the continuous medium is a semi reasonable in the shot sleeve and runner, it is far from reality in the die cavity. As discussed previously, the flow is atomized and it is expected to have a large fraction of the air in the liquid metal and conversely some liquid metal drops in the air/gas phase. In such cases, the isentropic homogeneous assumption is very dubious.

For these reasons the assumption of k-k- model seems unreasonable unless good experiments can show that the choice of the turbulence model does not matter for the calculation.