Defining symmetry planes

1,698 views
Skip to first unread message

Jacob Wilson

unread,
Jan 28, 2021, 6:32:10 PM1/28/21
to LS-DYNA2
I have been working on a problem in LS Dyna that has planar symmetry. The model has both SPH nodes and FEM components. 

At first, to impose symmetry conditions, I was creating nodes sets and imposing the symmetry constraints along the cut surfaces manually. i.e. for a symmetry plane oriented along the +x-axis, I would place restriction on x translation and y/z rotation. However I recently discovered both (i) *BOUNDARY_SPH_SYMMETRY_PLANE and (ii) *BOUNDARY_SPC_SYMMETRY_PLANE.

Is there a fundamental reason to use those keycards over a manual method other than just saving time? Also, I assume that (i) is used if and only if SPH particles are involved and (ii) is used if FEM nodes are to be constrained. Is there any issue in using both definitions in a single model?

As a final note, does anyone have any special tips or tricks they've found helpful/necessary in creating physically meaningful symmetric models? As of now, I pretty much only impose the symmetry constraints and toggle ISYM to 1 on the contact cards.

Sincerely,
Jacob

l...@schwer.net

unread,
Jan 29, 2021, 9:39:31 AM1/29/21
to Jacob Wilson, LS-DYNA2

Jacob -- Due to the nature of the SPH Kernel Function – typically Gaussian – it is difficult to prescribe essential boundary conditions, e.g. displacements, at the particle.

 

LS-DYNA, and other SPH methods, typical workaround this problem by the use of so called “ghost particles” generated internally that mirror the particles on one side of a SPH symmetry plane.

 

For traditional FEM nodes, *BOUNDARY_SPC_SYMMETRY_PLANE is a user convenience that provides nodal symmetry constraints identical with those users manually identify as nodes and constraints.

 

I know of no tips/tricks for using symmetry – but I am saddened by its apparent lack of use by too many modelers unnecessarily creating full models with obvious symmetry.

 

                --len

--
You received this message because you are subscribed to the Google Groups "LS-DYNA2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ls-dyna2+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/a974b518-ea8a-440c-97c0-785dba29caa7n%40googlegroups.com.

Jacob Wilson

unread,
Jan 29, 2021, 12:36:55 PM1/29/21
to LS-DYNA2
Thank you for your insight, Len. I tried to implement the SPH symmetry plane recently, and I must have done something incorrectly as it resulted in the generation of a massive amount of energy. 
Within the keycard, I specified the symmetry plane's direction vector with the vector tail point on the symmetry plane surface and the vector tip oriented so that the vector points toward the real SPH particles in the model. I do this for two different symmetry planes in the model. Is that the correct approach?

Sincerely,
Jacob

James M. Kennedy

unread,
Jan 29, 2021, 2:24:59 PM1/29/21
to Jacob Wilson, LS-DYNA2

Dear Jacob,

 

To build a symmetry model, you model only the necessary symmetric parts and

then apply any appropriate boundary conditions (to those nodes on the symmetry

plane). For example, if you have half-plane symmetry, you need to model only

one-half of the full model and then apply the nodal boundary conditions along

that plane of symmetry.

 

*node (translational, tc, and rotational constraints, rc)
*NODE
$# node x y z tc rc
1 210.0 0.0 100.0 4 6

*boundary_spc_node (translational and rotational constraints)
*BOUNDARY_SPC
$# nid cid dofx dofy dofz dofrx dofry dofrz
1 1 1 1 0 0 1 1
2 1 1 0 0 0 1 1

A boundary_spc_set option is also available
*BOUNDARY_SPC
$# nsid cid dofx dofy dofz dofrx dofry dofrz
1 1 1 1 0 0 1 1

*constrained_global (global boundary constraint plane)
*CONSTRAINED_GLOBAL
$# tc rc dir x y z
1 0 1 0.0 0.0 0.0

*constrained_local (local boundary constraint plane)
*CONSTRAINED_LOCAL
$# tc rc dir x y z cid
1 0 1 0.0 0.0 0.0 11

_____________________________________

 

When new nodes and elements are constructed after adaptive remeshing,

I believe that *boundary_spc_symmetry_plane is a popular option.

 

_____________________________________

 

Sincerely,

James M. Kennedy

KBS2 Inc.

January 29, 2021

Jacob Wilson

unread,
Jan 29, 2021, 2:49:09 PM1/29/21
to LS-DYNA2
Thank you for that information James. I have been able to implement the symmetry plane condition with FEM elements using *boundary_spc_symmetry_plane, but I have been unsuccessful at using the corresponding *boundary command in constraining SPH particles :/

James M. Kennedy

unread,
Jan 29, 2021, 3:11:38 PM1/29/21
to Jacob Wilson, LS-DYNA2

Dear Jacob,

 

See if this note is of any help?

 

Please check your input for your SPH symmetry planes. I think you may have the

direction incorrect for your initial symmetry plane.

 

*BOUNDARY_SPH_SYMMETRY_PLANE

$#     vtx       vty       vtz       vhx       vhy       vhz  

$       0.0       0.0       0.0      -1.0       0.0       0.0

       0.0       0.0       0.0       1.0       0.0       0.0

$ add

*BOUNDARY_SPH_SYMMETRY_PLANE

$#     vtx       vty       vtz       vhx       vhy       vhz  

       0.0       0.0       0.0       0.0       0.0      -1.0

 

--------------------

 

Another question previously posed to another user.

 

What version/release of LS-DYNA are you running?

I used the following Dev release

     |  Version : smp d Dev       Date: 10/02/2019     |
     |  Revision: 141173          Time: 17:43:57       |

And was able to successfully start your simulation. Please see the attached figure sent
to your personal email address.

Several suggestions in your modeling:

1. I would use 1/4 symmetry modeling instead of 1/2 symmetry modeling. This should
reduce your CPU by a factor of two. Simply add another symmetry plane to achieve
1/4 symmetry (for example, something like this):

*BOUNDARY_SPH_SYMMETRY_PLANE
$#     vtx       vty       vtz       vhx       vhy       vhz   
       0.0       0.0       0.0      -1.0       0.0       0.0
*BOUNDARY_SPH_SYMMETRY_PLANE
$#     vtx       vty       vtz       vhx       vhy       vhz   
       0.0       0.0       0.0       0.0       0.0      -1.0

--------------------

Jacob Wilson

unread,
Jan 29, 2021, 3:29:14 PM1/29/21
to LS-DYNA2
Interestingly, in either orientation of the normal (pointing into or out of the SPH model), the simulation shows the same issue: uncontrolled energy increase. 

I have attached a figure of my current model. Maybe I am not able to use the SPH constraint for my geometry? The yellow plane presumably represents the SPH symmetry plane. My goal is to model a planar cut of a sphere impacting a surface. 

                                                                                        PlaneCut.png

Jacob Wilson

unread,
Jan 29, 2021, 3:31:24 PM1/29/21
to LS-DYNA2
I will clarify that the part colored red is the SPH portion. I'd like to constrain the face parallel to the cut (in the plane of the screen).

l...@schwer.net

unread,
Jan 29, 2021, 5:34:03 PM1/29/21
to Jacob Wilson, LS-DYNA2

Jacob –

 

1\ Does your model run without the SPH symmetry plane(s)?

 

2\ Is your model completely in the positive x half space, i.e. all x-coordinates positive?

 

3\ Does your model run with one SPH symmetry plane?

With one symmetry plane, does it align with the Y-Z plane, i.e. SPH particles closest to the symmetry plane have an x-coordinate of +h/2?

 

4\I am not sure you can have two similarly oriented SPH symmetry planes.

I have used three orthogonal SPH symmetry planes in the past.

 

5\ What SPH FORM are you using?

 

                --len

 

From: ls-d...@googlegroups.com <ls-d...@googlegroups.com> On Behalf Of Jacob Wilson
Sent: Friday, January 29, 2021 12:29 PM
To: LS-DYNA2 <ls-d...@googlegroups.com>
Subject: Re: [LS-DYNA2] Defining symmetry planes

 

Interestingly, in either orientation of the normal (pointing into or out of the SPH model), the simulation shows the same issue: uncontrolled energy increase. 

 

I have attached a figure of my current model. Maybe I am not able to use the SPH constraint for my geometry? The yellow plane presumably represents the SPH symmetry plane. My goal is to model a planar cut of a sphere impacting a surface. 

 

                                                                                        

image001.png

Jacob Wilson

unread,
Jan 29, 2021, 6:30:38 PM1/29/21
to LS-DYNA2
Len and all,

According to the manual's definition of the SPH symmetry plane coordinate parameters, the symmetry plane normal should be pointing into the body being constrained. I just tried the opposite, orienting the normal away from the body, and things seem to be working well... very strange.

Len, to answer some of your questions:

(1) Yes, the model runs when *Boundary_SPH_Sym_Plane is replaced with regular *Boundary_SPC constraints. I'm just worried that the SPC constraints imposed on the SPH nodes will lead to bad results. 

(2) The model is entirely in the +z half of Cartesian space (including z=0). Does that affect the performance of symmetric constraints? 

(5) I'm using SPH form 1 with the default artificial viscosity formulation.




l...@schwer.net

unread,
Jan 30, 2021, 9:37:09 AM1/30/21
to Jacob Wilson, LS-DYNA2

Jacob –

 

Whatever works – always remember it is not what anyone or the manual says, it is what works that counts  😊   --len

                                                                                        Image removed by sender.

image002.jpg

Jacob Wilson

unread,
Jan 30, 2021, 10:33:41 AM1/30/21
to LS-DYNA2

Ha! Very true. 

James M. Kennedy

unread,
Jan 30, 2021, 1:08:36 PM1/30/21
to Ahmad Rahmati-Alaei, LS-DYNA2

Dear Ahmad,

 

I am not aware of any R11 restriction for this feature.

 

In the example I previously provided, ¼ symmetry was assumed with two planes of symmetry.

 

Also, please see the following presentation:

 

Fragassa, C, Topalovic, M., Pavlovic, A., and Vulovic, S., “Dealing with the Effect of Air in

Fluid Structure Interaction by Coupled SPH-FEM Methods”, Materials 2019, Vol,12, Issue 7,

Article 1162, April, 2019.

 

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6480372/

https://www.mdpi.com/1996-1944/12/7/1162

 

Sincerely,

James M. Kennedy

KBS2 Inc.

January 30, 2021

From: Ahmad Rahmati-Alaei [mailto:ahmadra...@gmail.com]
Sent: Friday, January 29, 2021 10:50 PM
To: James M. Kennedy <j...@kbs2.com>
Subject: Re: [LS-DYNA2] Defining symmetry planes

 

Dear Dr. Kennedy

As I understand it, only one SPH symmetry boundary condition (with BOUNDARY_SPC_SYMMETRY_PLANE) can be set for the model by LS-DYNA R11. It is true?
My 3D-SPH model has 1/4 symmetry and needs two symmetry boundary conditions.

 

Regards.

Ahmad

 

‫‪James M. Kennedy‏ <‪j...@kbs2.com‏> در تاریخ جمعه ۲۹ ژانویهٔ ۲۰۲۱ ساعت ۲۲:۵۵ نوشت:

Jacob Wilson

unread,
Jan 30, 2021, 2:02:46 PM1/30/21
to LS-DYNA2
Does anyone know if we are able to view the ghost SPH particles created by *Boundary_SPH_Symmetry_Plane either using LSPP or by some other means? 

-Jacob
Reply all
Reply to author
Forward
0 new messages