Modeling Elongated Grains in Rolled Microstructures Using DREAM.3D

80 views
Skip to first unread message

李宜蓁

unread,
May 19, 2025, 12:58:44 PM5/19/25
to dream3d-users

Hi DREAM.3D team,

I'm currently trying to construct a model for fully rolled grains. I have a question regarding grain size input. Since the EBSD maps of rolled grains often show significant elongation along the TD and RD directions, it's not feasible to use an averaged grain size across multiple directions—as we commonly do for fully recrystallized structures—when performing grain size analysis via the intercept method.

However, DREAM.3D only allows the input of a single grain size value when constructing the grain structure. This seems to make it difficult to accurately model the elongated grain morphology typical of rolled microstructures.

Is there a way to assign different grain size values for different directions, such as specifying distinct sizes along TD, RD, and ND? This would allow the model to more closely represent the actual material condition.

Thank you very much!

Michael Jackson

unread,
May 23, 2025, 12:44:03 PM5/23/25
to 李宜蓁, dream3d-users
Hello,
  1: To be clear you are defining the Axis RATIOS between A, B and C in that dialog that opens up. When thinking about those ratios, the Feature ESD should be considered to be the A Axis length and the B and C Axis lengths are derived from there. 

2: The Insert Transformation Phase filter should be used if needed. If the material you are modeling is known to have twin relationships then using the filter would be useful.

--
Mike Jackson                    mike.j...@bluequartz.net
BlueQuartz Software         www.bluequartz.net
President/Owner               Dayton, Ohio
Principal DREAM.3D Developer


On May 23, 2025 at 09:10:35, 李宜蓁 <leeyc...@nycu.edu.tw> wrote:

Thank you very much for your reply — it was very helpful to me. I have two questions:

1.In the StatsGenerator filter, although it's possible to define the a/b/c axis lengths to simulate the rolled grain morphology, I'm unsure about the value to input for Mean Feature ESD. Should it correspond to the grain size in the RD or ND direction, or should it be the average across two or even all three directions?

2.When I select (03) Single Cubic Phase Rolled, the Insert Transformation Phase filter does not appear in the left-hand pipeline. Does this mean that for rolled grains without any phase transformation, this filter is not required? I previously included it, but it seemed to cause uneven compression in my subsequent mechanical simulation.

Thank you again for your reply!


Michael Jackson <mike.j...@bluequartz.net> 於 2025年5月23日 週五 下午9:00寫道:
Hello,
    If you go to the bookmark in the image below, we have a preset for rolled microstructures. 
    

What is not shown in that pipeline though is how we got to the values. If you click on the “Preset Statistic Models” and select the “Primary Rolled” a new dialog window will appear. This is where you set the aspect ratios for the rolled grains. The assumption here is that the grain size from the main part of the user interface of the StatsGenerator filter is the “A Axis Length”. From this information you should be able to model a rolled microstructure fairly closely.




Do let us know if this information helps.
--
Mike Jackson                    mike.j...@bluequartz.net
BlueQuartz Software         www.bluequartz.net
President/Owner               Dayton, Ohio
Principal DREAM.3D Developer


--
You received this message because you are subscribed to the Google Groups "dream3d-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dream3d-user...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dream3d-users/5b866608-0c9d-4a11-8df3-6a7117189a84n%40googlegroups.com.

Jack

unread,
May 23, 2025, 1:16:17 PM5/23/25
to Michael Jackson, 李宜蓁, dream3d-users
Hi LYC,

I never built a 3D microstructure from scratch, I have seen people doing that with DAMASK, and vorvonoi model. 

I think that if you somehow fulfill the population of the interceps lenght using the cycloids with the major radius orientedbin the 3 dimensions ( sonyou would have three distributions) and you distribuite the distance of the germs so that the distribution of the germs distances resamble it you can do it. But it is quite a challenging job. 

李宜蓁

unread,
May 26, 2025, 5:11:40 AM5/26/25
to dream3d-users

Dear Mr. Jackson,

Thank you for your helpful response regarding the axis ratios and Feature ESD in DREAM.3D.

I would like to confirm whether the following procedure is appropriate for modeling rolled grains:

  1. I calculated the average grain size along the Rolling Direction (RD) from my EBSD data and entered this value as the Feature ESD.

  2. I measured the length ratios (A:B:C) of several individual grains from EBSD (e.g., based on RD:TD:ND dimensions) and used the average ratio as the input for the Axis Ratios field in DREAM.3D.

In other words, I treated the RD grain size as the A-axis (major axis), and set the B and C axes according to the observed shape ratios. Does this approach correctly follow how DREAM.3D interprets the Feature ESD and axis ratios?

I would greatly appreciate your confirmation or any suggestions for improving this setup.

Best regards,

LYC

Michael Jackson 在 2025年5月24日 星期六凌晨12:44:03 [UTC+8] 的信中寫道:

李宜蓁

unread,
May 26, 2025, 5:13:56 AM5/26/25
to dream3d-users

Dear Jack,

Thank you very much for your kind reply and insights.

I appreciate you sharing your thoughts on the challenges of building a 3D microstructure from scratch. It's helpful to know that some people approach this with DAMASK and Voronoi models. Your suggestions regarding intercept length distributions and germ spacing give me a better sense of how complex this task can be.

Thanks again for taking the time to respond!


Jack 在 2025年5月24日 星期六凌晨1:16:17 [UTC+8] 的信中寫道:

Michael Jackson

unread,
May 26, 2025, 2:44:59 PM5/26/25
to 李宜蓁, dream3d-users
The procedure that you state in your email should work correctly. If it does not give expected results please report back to the group.

--
Mike Jackson                    mike.j...@bluequartz.net
BlueQuartz Software         www.bluequartz.net
President/Owner               Dayton, Ohio
Principal DREAM.3D Developer


Reply all
Reply to author
Forward
0 new messages