Welcome to the forums. Please post in English or French.

You are not logged in.

#1 2018-06-12 15:02:09

Anirudh
Member
Registered: 2017-09-11
Posts: 154

GROT_GDEP with springs

Hello all,
I have been recently doing some simulations with springs and I think springs don't work correctly under large rotations. As can be seen in the attached image, the only spring element expands unrealistically under a torque load with (DYNA_NON_LINE). I have kept the other triangular element inertia to minimum to avoid too much tension build up in spring.

I believe the issue is because only PETIT deformation behavior is allowed for spring elements.
Therefore, I need to create new spring element compatible with GROT_GDEP that could allow rotations many times over. (I could also use CABLE element but that would introduce additional mass in the system).
Might anyone suggest how to start off?

Thanks a lot.
Regards
Anirudh

Offline

#2 2018-06-13 13:16:59

MrSeb2000
Member
From: Canada
Registered: 2017-10-10
Posts: 16

Re: GROT_GDEP with springs

Hello Anirudh,

You are correct, GROT_GDEP cannot be used in the case of DIS_T and DIS_TR elements (see U3.11.02). The only options available are PETIT and PETIT_REAC. If you use relatively small time steps, you could use PETIT_REAC (stiffness matrix is calculated based on the actualised geometry).

If you really need to use GROT_GDEP, the CABLE element (with EC_SUR_E = 1 in DEFI_MATERIAU) would be the simplest approach. Regarding the additional mass, you just need to set the density of the material to zero. You need to choose Young's modulus E and area A, such as k_spring = EA/L, where L is the length of the cable element. I should mention that this approach cannot model rotational springs (with DIS_TR) because the CABLE element only has degrees of freedom in translation.

If you still need to develop a new spring element, I cannot help you, but I would recommend to have a look a the Module 5 (Development training) of the Training section of Code_Aster website.

Regards,

Sébastien

Offline

#3 2018-06-13 15:24:53

Anirudh
Member
Registered: 2017-09-11
Posts: 154

Re: GROT_GDEP with springs

Hi,
Thanks for reply.
CABLE element is a viable option but I suspect there would be a singular matrix error at times of compression during load cycle, as they have zero stiffness in compression. I will try to read the doc for implementation of new finite element. Also attaching the error of spring elongations. The attachment failed for some reason before.
Could anyone tell me the max allowable size of attachment? I am trying to upload a <100 kb gif file but the system says I exceed the size limit. I could attach files as big as 2MB in the past.
Regards
Anirudh

Last edited by Anirudh (2018-06-13 15:27:29)


Attachments:
elongation_spring.zip, Size: 91.2 KiB, Downloads: 37

Offline

#4 2018-06-13 16:15:55

MrSeb2000
Member
From: Canada
Registered: 2017-10-10
Posts: 16

Re: GROT_GDEP with springs

Anirudh,

Regarding the stiffness in compression for the CABLE element, you can adjust it in DEFI_MATERIAU with EC_SUR_E (ratio between E in compression and E in tension). If you use a value of 1 for EC_SUR_E, the CABLE element can also resist compression. Code_Aster uses a default value of 1e-4 for EC_SUR_E.

mat = DEFI_MATERIAU(
    ELAS=_F(E=...,
                    NU=...,
                    RHO=...,
                    ALPHA=...,
    CABLE=_F(EC_SUR_E=1),
)

Regards,

Sébastien

Offline

#5 2018-06-13 18:37:12

Anirudh
Member
Registered: 2017-09-11
Posts: 154

Re: GROT_GDEP with springs

Hi,
thanks for reply.
While we are at it, let me bring forward the original issue with springs under large rotations. I have tried two cases where I am applying torque Mz on the center node and recording the movement of the flag. The direction of moment is incorrect. Its opposite actually.
The simple motive is to spin the triangular flag around center point about Z axis many times over.
>Case1:
I have double nodes at the junction where spring meets the triangle. Also, a 0D element is applied on the node connected to spring. Both overlapping nodes are joined by LIAISON_DDL along X and Y. Spring is having very large stiffness along X, Y and Z axes. The result is an expanding spring as rotation angle gets larger.
>CASE 2:
I directly join the spring to the node of triangle element. Solver gives no error, but there are no displacements for any value of torque or spring stiffness.
Need to know why this is happening.
I also attach the corresponding files for both testcases.

Regards
Anirudh

Last edited by Anirudh (2018-06-14 03:59:25)


Attachments:
two_cases.png, Size: 79.83 KiB, Downloads: 43

Offline

#6 2018-06-13 18:42:08

Anirudh
Member
Registered: 2017-09-11
Posts: 154

Re: GROT_GDEP with springs

Case1


Attachments:
case1.zip, Size: 45.62 KiB, Downloads: 33

Offline

#7 2018-06-13 18:44:37

Anirudh
Member
Registered: 2017-09-11
Posts: 154

Re: GROT_GDEP with springs

case2


Attachments:
case2.zip, Size: 48.59 KiB, Downloads: 35

Offline

#8 2018-06-13 23:58:24

Ioannis
Member
Registered: 2016-04-01
Posts: 32

Re: GROT_GDEP with springs

Hi Anirudh,

For the second case if I understood well, the reason you get this result is because the shell elements implemented in code_aster do not possess a rotational in-plane DOF. This is typical in standard shell elements, you can look in the literature for "drilling dof" of shell elements. So, you cannot apply a torque directly in the node of the triangle.

Regards,
Ioannis

Last edited by Ioannis (2018-06-14 00:05:11)


Aether Engineering
- aethereng.com -

Offline

#9 2018-06-14 02:56:13

MrSeb2000
Member
From: Canada
Registered: 2017-10-10
Posts: 16

Re: GROT_GDEP with springs

Anirudh,

Since you apply a moment at one end of the spring. This means you cannot use the CABLE element, which has degrees of freedom in translation only. In that case, you could use beam elements (POU_D_* with GROT_GDEP).

I did some trials with case 1. When the GROT_GDEP option is used for the K_TR_D_L element, a warning is printed in the mess file saying only DEFORMATION='PETIT' is possible. I compared the results for PETIT and GROT_GDEP. They are the same, indicating that GROT_GDEP is ignored for DIS_TR elements. I also tried PETIT_REAC, and in that case, the displacements are much less than for PETIT (warning message is still printed though).

Regards,

Sébastien

Offline

#10 2018-06-14 03:04:22

Anirudh
Member
Registered: 2017-09-11
Posts: 154

Re: GROT_GDEP with springs

Hello,
Thats true Ioannis and that is applicable for solids too.
Sébastien, POU* for beam should work but I think it would be difficult to emulate a spring with constant stiffness K through beam.
What should be done in order to turn the whole thing with a torque on center node? Maybe a new element.
As an example, I need to be able to simulate the attached scenerio. (A tyre which rolls over an obstacle is rotated by a torque at the center, in a moving co-ordinate system.The center node is connected to edges with rigid links. Also, here we don't explicitly control the center of rotation of tyre, rather let the tyre dynamics determine it. However, a local constant torque is always applied at the center. Apologies for bad image quality ). Its taken from here: http:  // www. adina.com/newsgrp.shtml
Thanks.

Regards
Anirudh

Last edited by Anirudh (2018-06-14 04:15:17)


Attachments:
example_tyre.zip, Size: 77.53 KiB, Downloads: 37

Offline

#11 2018-06-14 04:18:27

Anirudh
Member
Registered: 2017-09-11
Posts: 154

Re: GROT_GDEP with springs

Hello,
The results of moment on beam. displacements are very low.


Attachments:
s5.png, Size: 20 KiB, Downloads: 34

Offline

#12 2018-06-14 04:23:52

Anirudh
Member
Registered: 2017-09-11
Posts: 154

Re: GROT_GDEP with springs

Hello,
Also attaching the beam  testcase.

PS: the attachment does not work correctly , If a post is edited to include code tags, the attachment is lost. Also, the size of attachments is restricted to 100Kb.

Thanks
Anirudh


Attachments:
beam_rotate.zip, Size: 3.25 KiB, Downloads: 37

Offline

Board footer