Averaging the driving force across the interface

solid-solid phase transformations, influence of stresses and strains
Post Reply
kamalnath
Posts: 13
Joined: Mon Dec 11, 2017 9:17 pm
anti_bot: 333
Location: Columbus, Ohio

Averaging the driving force across the interface

Post by kamalnath » Sat Oct 31, 2020 8:45 pm

Dear all,

There was a discussion about driving force averaging in the MICRESS manual.

<<< ------- From Manual (Vol 2, Section 3.2.1)----- >>>
3.2.1
Driving force options
The driving force options in the phase-interaction input allow increasing numerical stability and (together with the
interface mobility and the antitrapping model) calibrating the correct interface kinetics. They consist of parameters
on averaging, limiting and (in future) noise.
<<< ------------- >>>

I want to know more about this ? Is there a MICRESS publication (or) some other literature where they discuss about this ? For solid-solid transformation (eg:precipitation problem) is this still necessary ?

Regards,
Kamal

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Averaging the driving force across the interface

Post by Bernd » Sun Nov 01, 2020 9:03 pm

Hi Kamal,

The "dG_options" include some functions which all are about the driving force, but have very different background:

# 'DeltaG' options: default
# avg ...[] max ...[J/cm^3] smooth ...[Deg] noise ...[J/cm^3] offset ...[J/cm^3]
avg 0.50 max 100 smooth 45.0

The option which you address is the averaging of the driving force. It means that inside the interface region, the driving force is not calculated independently in each grid cell but averaged over a certain length along the gradient direction through the interface. The 'avg' parameter defines the averaging length and must be between 0 (no averaging) to 1 (not limited averaging length). We generally propose 0.5 as default, which corresponds to an averaging length which is equal to the interface thickness. The exact way the average is calculated has been described in an old thread but did not change since then. Unfortunately this is not published elsewhere.

The other options are:
max: smooth cutting of the driving force at a given value (positive or negative).
smooth: random noise on the gradient direction. Reduces grid anisotropy in many cases
noise: adds random noise on each grid cell independently
offset: Adds a constant value to the driving force, e.g. to compensate for curvature in a 1d-simulation, or for correction of thermodynamic databases

Bernd

kamalnath
Posts: 13
Joined: Mon Dec 11, 2017 9:17 pm
anti_bot: 333
Location: Columbus, Ohio

Re: Averaging the driving force across the interface

Post by kamalnath » Sun Nov 01, 2020 9:15 pm

Thank you Brend. One more question, for solid-solid transformation, is this averaging still necessary ? We don't expect solute trapping in solid-phase transformations right ?

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Averaging the driving force across the interface

Post by Bernd » Mon Nov 02, 2020 1:47 pm

Hi Kamal,

It makes no difference whether it is solidification or solid-solid transformation: If the diffusion length of the elements is not large compared to the numerical interface thickness, there will be a gradient in driving force over the interface, and there is the possibility of the interface to become unstable. Please note that even when there is no instability, the mobility correction ("mob_corr") function is calibrated to give the best results when using an averaging value of 0.5.

Bernd

Post Reply