AM Single Layer Simulation

  • 1.2K Views
  • Last Post 14 August 2018
  • Topic Is Solved
Kellen.traxel posted this 18 July 2018

Overall Single Track Geometry

What is the most efficient way to model the laser heat source moving across my symmetric, 3D, single-layer track? I have read about using commands in APDL, however, it is not clear how I can manipulate a nodal (or volume) heat source as a function of the simulation time and position. In the image above, I was able to model a nodal heat source at the starting location (t=0) using APDL commands in the "Transient Thermal" section of my model tree. Ultimately, my question is about the relation between the APDL commands in workbench (and their location in the model tree) and the actual simulation time, and how they could be combined using a DOWHILE loop of some sort so that the local laser coordinates can translate with respect to the overall model domain. I currently do not have access to the ACT toolbox or AdditivePrint in my ANSYS Mechanical license.

Thanks in advance for any information or resources.

Regards,

Kellen

 

Order By: Standard | Newest | Votes
Kellen.traxel posted this 14 August 2018

Hi Sandeep, i wanted to point you to my question posted elsewhere: https://studentcommunity.ansys.com/thread/stef-error/

Kellen.traxel posted this 14 August 2018

Hi Sandeep, it worked!

Thanks and I will post on a new discussion when I have further questions.

Regards,

Kellen

SandeepMedikonda posted this 14 August 2018

Kellen,

Here's what you do to get it to work:

1. Add the isparameter to the variable of interest in the xml file (recommend you to do this in the original download). In this case, I selected Radius and the Source Power Intensity:

      <property name="R1" caption="Radius of the Beam" control="float" default=" 5[mm]" unit="Length" isparameter="true" />
      <property name="LD" caption="Source Power Intensity" control="float" default=" 100[W mm^-2]" unit="Heat Flux" isparameter="true" />

2. Next, copy that file MovingHeat.xml and the folder MovingHeat found in the src folder to 

%appdata%\ansys\v191\ACT\extensions

Replace anything there although I think you would need just the xml file placed there.

3. Once you re-launch Workbench, select the Scripted MovingHeat extension instead of the Binary one.

4. Now once you re-launch Workbench, you will observe the parameter buttons appear:

Hope this helps.

Regards,

Sandeep

SandeepMedikonda posted this 14 August 2018

Kellen,

I will have to debug a little w.r.t parametrizing the model. I will keep you posted as I find out.

For Annealing, please try EKILL/EALIVE. There are numerous discussions in the forum on this. Try and include a blank load step after EKILL and use EALIVE. Please see this article and this discussion on XANSYS.

P.S: If you have found a solution to your initial query, please mark it as a solution (even if it is your own post) so that it would make things easier for someone going through it at a later time. Also, please post new questions as a new discussion and provide a link to the old discussion.

Regards,

Sandeep

 

Kellen.traxel posted this 14 August 2018

Hi Sandeep, another question I have about residual stress analysis is there a simple way to incorporate annealing of Ti6Al4V (i.e. when the annealing temperature is reached, displacement is equal to 0)?

Thanks as always,

Kellen

Kellen.traxel posted this 13 August 2018

I did, it did not work. One thing that might not be right is that I have stored the extension in the same folder as my ansys results files. Should the extension (.webex) be saved in the same folder as the ANSYS program is saved?

 

Kellen

SandeepMedikonda posted this 13 August 2018

Kellen, can you try uninstalling from the ACT start Page and re-install that extension again?

Kellen.traxel posted this 13 August 2018

This is what my xml and input screen show up as:

 

Kellen.traxel posted this 13 August 2018

Hi Sandeep,

Do I need to reinstall ANSYS or do anything additional for the parameterization to work? I input that line of code into the correct folder and am not seeing the parameterization checkbox.

Any info helps, as always.

 

SandeepMedikonda posted this 13 August 2018

Also,

With regards to parametrizing a variable just open the xml file (in the folder ~\CT_MovingHeat_R170_v4.1\MovingHeat\src) and for the input you want to parametrize just add isparameter="true".

      <property name="vel"  caption= "Velocity" control="float" default="5 [mm s^-1]"  unit="Velocity" isparameter="true"/>

Regards,

Sandeep

 

 

SandeepMedikonda posted this 13 August 2018

Kellen, I misquoted earlier here today, please see my edited response below:

[EDITED]:

I would recommend you to make sure that you are capturing half the area? So yes, it would be Power/(0.5*pi*radius^2). When you make it symmetric on the exact edge, you are modeling 2 lasers with a smaller radius, but we still want to capture the total area captured by the whole laser?

I would recommend you to try 2 test cases to confirm if you are getting similar results, One with the full body and one with the symmetric condition defined on the edge? Is this possible?

Regards,

Sandeep

Kellen.traxel posted this 13 August 2018

Sandeep, sorry for any confusion I will try to re-explain. I think my question is more on the concept of the symmetric boundary condition, and how to define a total surface flux if only a certain portion is physically modeled.

My model is simulating half of a single-track deposit width and is symmetric along the center of the laser spot, as it is traversing the surface of the material. Because of this, my model is really only being affected by half of the laser, as the other half is not physically modeled. So when I define the laser intensity as Power/LaserArea, does this mean I need to input Power/(pi*laserradius^2) or Power/(0.5*pi*laserradius^2), i.e. including the entire laser area or only half of the area.

Any information helps.

Thanks

Kellen

 

SandeepMedikonda posted this 13 August 2018

Yes, I believe C2 is defined by Power/LaserArea.

I don't quite understand your question with respect to symmetric B.C. Is your laser path being affected by the symmetric b.c?

If so, wouldn't the overall path covered by the laser be smaller? Why would the laser intensity have to change for this?

~Sandeep

Kellen.traxel posted this 13 August 2018

Hi Sandeep, yes I have this documentation. Thank you it helps. How should I define C2? Power/laserarea, as I mention above? Also, for a symmetric BC should I use half of the laser area in the flux definition vs. the whole area?

Thanks Kellen

SandeepMedikonda posted this 13 August 2018

Kellen,

Do you have the documentation? It should be available in the 'doc' folder of the downloaded extension. He is defining heat flux based on the following relation:

Let me know if this helps?

Regards,

Sandeep

Kellen.traxel posted this 12 August 2018

Sandeep, the better questions is: what is the best way to define the flux based on input parameters such as laser radius, power, absorption, etc.? I am defining mine as Power*absorption/(pi*laser_radius^2), but I am not sure what the developer had in mind for this definition. Is there a pdf that the developer might be able to provide on specifics of how to define this parameter based on the user's desire to adjust different parameters? 

Thanks in advance for any information you might be able to provide.

Warm regards,

Kellen

Kellen.traxel posted this 12 August 2018

Hi Sandeep, in my model I am dividing my domain in 2 and making it symmetric, in this case should my input heat flux be divided by half of the laser spot size or the entire spot size?

 

Thanks

Kellen

Kellen.traxel posted this 12 August 2018

Sandeep, I believe it is called 4.1, but it is definitely the one that pops up in the ACT store. Hope that helps.

 

Thanks

Kellen

SandeepMedikonda posted this 12 August 2018

Kellen, what version of the ACT extension are you using?

SandeepMedikonda posted this 11 August 2018

Kellen, I don't think this is currently possible, at least not without modifying the ACT code. Anyway, I've reached out to the ACT developer, will let you know once I hear back.

Regards,

Sandeep

Kellen.traxel posted this 11 August 2018

Hi Sandeep, I've had success thusfar with programming the laser heat flux. Is there any way to parameterize the input flux in order to do a design study, without altering the ACT code? If not, is there a more straightforward way of running parametric solutions that doesn't include making a new project for each set of paraemters?

 

Thanks

Kellen

SandeepMedikonda posted this 24 July 2018

Hi Kellen,

  Yes, it needs to be close to the time you are specifying in the Moving Heat Flux. I saw some EKILL commands being used in the code, so it's probably the birth-death technique. If you are interested, you should be able to access all of the APDL (+ python) code in the act extension:

~\ACT_MovingHeat_R170_v4.1\MovingHeat\src\MovingHeat

~Sandeep

Kellen.traxel posted this 24 July 2018

Got it to run. So I basically need to make sure that my end time is close to the actual travel time? Also, does the moving heat source utilize a "quiet" or birth and death technique so that there isn't  conduction through "unprinted" material?

Thanks

Kellen

SandeepMedikonda posted this 23 July 2018

Additionally, the reason it ran is that your End Time was way off 5 sec. Now, the path you specified is 0.015 m long and the velocity is 0.005, so it would take 3 sec. Now, in cases where it would take 2.7 sec then round it off and make it 3 seconds.

SandeepMedikonda posted this 23 July 2018

Ok Awesome, The speed could be because you do have a fine mesh or it could be the machine too. To speed it up you can change the 'number of segments' but it would reduce accuracy.

Kellen.traxel posted this 23 July 2018

Program running now, pretty slow so I will let you know when it finishes.

 

Thanks

Kellen

SandeepMedikonda posted this 23 July 2018

Kellen, Can you put 3sec as end time for moving heat source and try to solve it?

Kellen.traxel posted this 23 July 2018

Length of path is 0.015m

SandeepMedikonda posted this 23 July 2018

Kellen, can you post a snapshot of the length of the path (edge) and also the mesh plot?

Kellen.traxel posted this 23 July 2018

Still getting same issue as before. From log file:

Show More Posts
Close