Stats Generator

[Sihwa Sung]

Hello,

I want to make a dream3D file with Stats Generator.

I have some questions for this programm.

Q1. How to determine mu and sigma? I have numerical data for grain size distribution versus area fraction. Are mu and sigma arbitrary value to fit the graph or could the numerical data(.xlsx) be imported?

Q2. After creating data, in the ODF section, I can select the 'Bulk Load Weight and Spreads' with the current virsion. In this case, what kind of .txt file is needed? I mean is it ok with the .txt file which has only three Euler angles? or other imformation is also needed?

Q3. I heard that with the current virsion of Dream 3D, dream 3d output file could be changed as input file into the Abaqus directly. Please explain the process to conversion.

I'll wait for your answer. Thank you.

Sincerely

Sihwa Sung

[Sean Donegan]

Hi Sihwa,

I'll try and answer your questions individually:

Q1. A good place to start learning about the parameters in StatsGenerator is by checking out the DREAM.3D user manual. You can access the user manual by launching DREAM.3D and then going to the Help menu > DREAM.3D Help. This opens up the user manual. In the user manual, head to the Auxiliary Tools > StatsGenerator section. Here we explain what the parameters in StatsGenerator mean and the process one can use to select them, including mu and sigma. Check the manual for an in-depth explanation, but the short version is that mu represents the average of the lognormal grain size distribution, which is actually the average value of log(size of grains), where "size of grains" is computed as an equivalent sphere diameter. Similarly, sigma is the standard deviation of the lognormal grain size distribution, which is actually the standard deviation of the values of log(size of grains). The actual plot that is drawn is in "real space" units (microns, etc.), so the best approach is the modify the mu and sigma values until the graph matches your actual data. Alternatively, if you are able to fit your data to a lognormal distribution to determine values for mu and sigma, you could use these as the inputs in StatsGenerator. 

Q2. In the current version of DREAM.3D (v6.1.77), the bulk load option uses a file format that can be generated by DREAM.3D using a filter called "Write StatsGenerator ODF Angles File". If you have orientation data in a format like .ang or .ctf, you can read it into DREAM.3D as usual, and then feed your Euler angles into this filter to generate an "angles" file.  The filter will generate a separate file automatically for each phase in the system.  You can then use this file as the input to the bulk load option in StatsGenerator.  If you have orientations in a format that DREAM.3D cannot read in manually, you could create the file yourself. The format of this file is currently:

Number of Euler Angles

phi1 PHI phi2 weight sigma

....... 

where the Euler angles are in the Bunge convention. So, if you had three orientations of (90, 0, 0) all with weights and sigmas of 1, the file would look like this:

3

90 0 0 1 1

90 0 0 1 1

90 0 0 1 1

The Euler angles can be in degrees or radians, just make sure to select the correct option for your data when importing into StatsGenerator.

Q3. There are two filters in DREAM.3D that can export Abaqus files: the "Abaqus Hexahedron Writer" filter and the "Write Abaqus Surface Mesh" filter. Check out these filters' documentation for in-depth details. The filter documentation is located in the DREAM.3D user manual. The hexahedron writer takes a voxelized geometry as input, and will simply write the voxels as hexahedra. This results in a very "blocky" volume mesh, but serves as a quick and easy way to get a structure into Abaqus. For example, you could create a synthetic structure using stats from StatsGenerator and then export this voxelized synthetic structure using the hexahedron writer. The surface mesh writer writes out a surface mesh of a 3D volume as "shell" elements. This filter requires a surface mesh geometry as input, so in the example of a synthetic microstructure, you would have to first surface mesh the structure to use this filter. Note that surface mesh writer does not result in a volume mesh.  Both the Abaqus writers produce INP files.

hope that helps,

-Sean

[Sihwa Sung]

Dear Sean Donegan

Thank you for your answer. I was really helpful to me.

Following your explanation and tutorial of Dream 3d, I got 5 kinds of .inp files.(_elset, _nodes, _elems, _sects.inp)

But there is no information of Euler angles for each element sets. I need this information for Abaqus.

Is it possible to get this information by using the pipeline "Abaqus Hexahedron Writer"?

If  it is impossible, how to connect between euler angles and element sets? Is there another manner?

During making Abaqus input file with Dream 3d programm, the .csv file which has the information about axisEulerAngle_0, axisEulerAngle_1, and axisEulerAngle_3 was made.

What's the meaning of these three angles? Thier maximum value is 6.xxx, so I think it's not Euler angles in degree.

I'll expect your answer.

Thank you.

Sincerely

Sihwa Sung 

2015년 7월 22일 수요일 오후 4시 6분 43초 UTC+2, Sean Donegan 님의 말:

[Michael Groeber]

Sihwa

The Euler angles in that file are in radians (6.xxx is 2*pi).  I believe that .csv file should have one Euler angle set (phi1, PHI, phi2) for each element set (or grain).  If the number of Euler angle sets in the file does not match the number of element sets, please let us know.

Best

Mike Groeber 

[Sihwa Sung]

Thank you for your answer.

The number of feature ID and Material (Grain) number are the same.
I want to confirm that the Feature_ID is exactly match to the Material number?
For example, does the Grain_Mat10 has the same Euler angles with Feature_ID 10?

Friendly
Sihwa

2015년 7월 29일 수요일 오후 8시 54분 47초 UTC+2, Michael Groeber 님의 말:

[Sean Donegan]

HI Sihwa,

Yes, the Feature IDs/Grain IDs/Material numbers should all represent the same index, so Grain_Mat10 should have the same Euler angles as Feature_ID 10.  

-Sean

On Wednesday, July 22, 2015 at 8:44:34 AM UTC-4, Sihwa Sung wrote:

[Quan Jiao]

Hi Sean,

I just started to use Stat Generator, I think there's a typo in the GUI, the plus sign (red circled in the fig) in my understanding should be a minus sign.

Best,

Quan

On Wednesday, July 22, 2015 at 8:44:34 AM UTC-4, Sihwa Sung wrote:

[Sean Donegan]

Hi Quan,

Yes, this seems to be a typo, nice catch!  We'll fix it for the next release.

thanks,

-Sean

On Wednesday, July 22, 2015 at 8:44:34 AM UTC-4, Sihwa Sung wrote: