Problem with the calculation of dislocation density (GND) in HCP structure

[Matheus Brozovic Gariglio]

Hi, 

I'm writing a code in mtex-5.4.0 version (inspired by https://mtex-toolbox.github.io/GND.html) in order to calculate the dislocation density of both alpha (HCP) and beta (BCC) phases of the titanium alloys I'm studying.

I'm able to get good results, but, in the 2 cases that I tested, there is always a pixel of the alpha phase that I get a very huge value for the dislocation density, around 1e20 times higher than all others, affecting drastically the mean value that I'm calculating.

I would like to know if there's some kind of bug in the code, if I'm not adapting it correctly for the case of HCP structures, or it's something that could happen for these cases.

Please find attached the code for the calculation of GNDs in both phases and the link for 2 ebsd files that I've found this same problem (https://cloud.mines-paristech.fr/index.php/s/JCn3pw9N58Wefsl).

Kind regards,

Matheus

[Rüdiger]

Hi,

I'm wondering if this only occurs in those files which have a
non-uniform stepsize, such as the ones you shared.

Cheers,
Rüdiger

[Rüdiger]

Hi,
you define your dislocation systems just into one direction
symmetrise(...'antipodal'). This may lead to situations where no
solution can be found.

Using the following may work:
dS_A = [slipSystem.basal(cs_A) ...
        slipSystem.prismaticA(cs_A) ...
        slipSystem.pyramidalCA(cs_A)]
dS_A = dislocationSystem(dS_A.symmetrise)


Hope this helps.

Cheers,
Rüdiger

[Matheus Brozovic Gariglio]

Hi Rüdiger,

Yes, it worked perfectly. Thanks a lot again! :)

Cheers,

Matheus