Dear all,
If latent heat is used in a MICRESS simulation, there is an additional source of numerical instabilities, which is not covered by the automatic time stepping criteria. If the interface mobility of a growing phase is too high and the time step is too long, the latent heat release in one time step can be so big, that the driving force created from the corresponding temperature change is so big, that this causes an even bigger (opposite) release of heat in the next time step. This behaviour is difficult to predict, because it depends not only on the thermodynamics of the system, but also on the amout of interface in the calculation domain and also on the diffusion coefficients.
So, there is no systematic way to prevent that, other than reducing the time step manually if necessary. This can be done by properly specifying the optional parameters to the automatic option in the time input data, e.g.
automatic 0.95 0.95 1.E-5 1.E-7
The first two parameter are prefactors to the phase-field and segregation criterion, where default is 0.99.
The third value is the maximal allowed time step value which we use here to manually decrease the time step to a numerically stable value.
The last parameter is a minimal time step value which is used for other things (e.g. if in a single cell numerical problems would cause extremely low time step values, which would practically stop the simulation).
Before limiting the time step value, of course you should check whether you are using correct values for the resolution and interface mobilities, and whether the latent heat parameter are reasonable.
Bernd
---
original message from Bernd