To the second question:
In the nucleation data, you specify the orientation relationship between the new grain and the parent grain, e.g. by using "parent_relation". However, this has only effect on the orientation of the new grain.
In the misorientation input, you specify the interface properties as function of the orientation relation. Here, you can e.g. introduce specific misorientation models for specific orientation relations in terms of interface energy and mobility. However, this has no influence on the actual orientation of the new grains.
Bernd
Heat Treatment simulation of Ni-alloy
Re: Heat Treatment simulation of Ni-alloy
Hej Bernd,
Thanks for the inputs.
BR
Chamara
Thanks for the inputs.
BR
Chamara
Re: Heat Treatment simulation of Ni-alloy
Hi Bernd,
Are there any recommendation for selecting maximum time step for phase-field solver for isothermal HT simulations
BR
Chamara
Are there any recommendation for selecting maximum time step for phase-field solver for isothermal HT simulations
BR
Chamara
Re: Heat Treatment simulation of Ni-alloy
Hi Chamara,
In heat treatment simulations automatic time stepping sometimes creates very large phase-field time steps, because diffusion is slow. Slow diffusion in turn leads to low values for the diffusion limited interface mobility if "mob_corr" is used. Normally, this is not a problem because the processes to be simulated are also slow.
However, there may be cases where there are faster processes which do not affect the phase-field time step, but nevertheless should be resolved properly. Examples for that are nucleation, the application of cooling or heating rates, if we start from a strongly off-equilibrium initial state, or if there is a fast diffusing species which is not yet in equilibrium.
In all these cases it is necessary to restrict the time step value with an upper limit. The value which you would chose depends on on the time scale of the mentioned fast processes. As a general rule, it makes sense to reduce the maximum time step at least at the beginning of the simulation, because you typically start off-equilibrium.
In case of stress coupling it is important to know that reducing the phase-field time step will increase the frequency of calls to the stress solver. This increases calculation time, but also may be helpful for convergence.
Bernd
In heat treatment simulations automatic time stepping sometimes creates very large phase-field time steps, because diffusion is slow. Slow diffusion in turn leads to low values for the diffusion limited interface mobility if "mob_corr" is used. Normally, this is not a problem because the processes to be simulated are also slow.
However, there may be cases where there are faster processes which do not affect the phase-field time step, but nevertheless should be resolved properly. Examples for that are nucleation, the application of cooling or heating rates, if we start from a strongly off-equilibrium initial state, or if there is a fast diffusing species which is not yet in equilibrium.
In all these cases it is necessary to restrict the time step value with an upper limit. The value which you would chose depends on on the time scale of the mentioned fast processes. As a general rule, it makes sense to reduce the maximum time step at least at the beginning of the simulation, because you typically start off-equilibrium.
In case of stress coupling it is important to know that reducing the phase-field time step will increase the frequency of calls to the stress solver. This increases calculation time, but also may be helpful for convergence.
Bernd