Dear all,
I got several times the question why in simulations of the gamma-alpha transformation the phase 0 has to be included even if it is not needed, and sometimes even physical paramters have to be given for this phase.
The reason has to do with the past of MICRESS as a solidification tool. At those times it seemed reasonable to reserve phase number 0 for the melt and to give it some special status: The liquid phase exists at the beginning as grain number 0 automatically everywhere, no different grains of phase 0 can be defined and no anisotropy is allowed. This is still the case today, and it would take quite a lot of effort to change this consistently in all the code.
For the gamma-alpha case, the phases 1 and 2 have to be used for simulation because both consist of several grains. The inital grain structure of phase 1 is created by randomly setting grains into the phase 0 background, and if the radii were too small, free areas of phase 0 would be left over. MICRESS cannot know before whether this is the case or not.
The user knows it before, and therefore he or she does not need to define phase interactions for phase 0 with the other phases. In the case of using the TQ interface, if there is no interaction with phase 0 using the database, no identification of phase 0 in the database is necessary. But if latent heat is used, then phase 0, even if inert, needs to have thermophysical data like the enthalpy and the heat capacity if it is present in the simulation domain - and this cannot be ruled out!
On the other hand, nobody is forced to use only phase 0 for the liquid! Each phase number can be assigned to liquid by choice of the thermodynamic data, and even more than one (MICRESS-) phase can be assigned to the same thermodynamic phase. The only inconvenience of the special status of phase 0 is that for most solid-solid purposes is remains free, and in some cases a dummy value for the enthalpy and heat capacity has to be given.
Bernd