Hi all!
The 'automatic' option for the mobility in the 'Data for phase interaction' section of the driving file is a remnant from the times when there was no automatic time-stepping available in MICRESS. It will most probably vanish in the next MICRESS release!
The automatic mobility is based on a numerical segregation criterion, which prevents the interface from getting instable due to too large increments. For a given fixed time step size the maximum value for the mobility is calculated for each cell of the given interface, which is still numerically stable. The lower the value of the time step the higer the calculated mobility value. A variable 'prefactor' can be used to adjust the criterion to the individual necessities, a value lower than 1 would reduce the calculated mobility and therfore increase the numerical stability. 1.0 should be used by default.
The automatic mobility should be replacd by the automatic time stepping option (Time input data), which is clearly the better choice, because it calculates the global time step value using all mobility values of the phase interactions. The mobility of the interface is either a physical value or an important numerical parameter to adjust the correct interface kinetics ( look here ). Using the automatic time stepping the mobilities can be freely chosen for each interface while the purely numerical parameter 'time step' is adjusted according to the segregtion criterion. In addition to this conceptional argument the automatic time stepping shows better performance and includes extra stability checks for stoichiometric phases.
So the recommmendation is: don't use it anymore!
I would like to get feedback, if there is anyone still using the automatic mobility and does not want to be without!
Bernd
---
original message from Bernd