Node deactivated from interface
Altair EmployeeHello,
I use Radioss for a dynamic impact problem, but also for the moment to validate my model through a quasi-static analysis.
A first component compresses a second component with an imposed velocity. The interface between them is of type 7. In dynamic explicit simulation the interface works fine but in the implicit simulation, I get a warning in interface : 'NODE ... DE-ACTIVATED FROM INTERFACE'. If I make an 'initial penetration' (initial gap < gapmin entered into type 7 interface) I do not get this problem (however I get initial constraints I want to avoid).
Would anyone know what the problem might be ?
Hereunder a part of my engine.out file:
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
CYCLE TIME TIME-STEP ELEMENT ERROR I-ENERGY K-ENERGY T K-ENERGY R EXT-WORK MAS.ERR
0 0.000 0.2000E+05 SOLID 39426 0.0% 0.000 0.000 0.000 0.000 0.000
ANIMATION FILE: pression_statique_07A001 WRITTEN
***** WARNING : ELEMENT FORMULATION ISOLID= 0
IS NOT AVAILABLE FOR STIFFNESS MATRIX BUILDING, USING GENERIC ONE INSTEAD
POSSIBLE CONVERGING ISSUE. *****
*--------- STIFFNESS MATRIX SETUP ---------*
SYMBOLIC DIM : ND = 155532 NZ = 2897175 NB_MAX = 26
FINAL DIM : ND = 47480 NZ = 761383 NB_MAX = 26
--STIFFNESS MATRIX WILL BE REFORMED AFTER EACH 3 ITERATIONS--
* CONVERGED WITH 0 ITERATIONS, |du|/|u|,|r|/|r0|,|dE|/|E|= 0.1000E+01 0.0000E+00 0.1000E+01
WARNING INTERFACE 701
NODE 13703 DE-ACTIVATED FROM INTERFACE
WARNING INTERFACE 701
NODE 13548 DE-ACTIVATED FROM INTERFACE
WARNING INTERFACE 701
NODE 13693 DE-ACTIVATED FROM INTERFACE
ANIMATION FILE: pression_statique_07A002 WRITTEN
* IMPLICIT COMPUTATION TERMINATED WITH TOTAL NONLINEAR ITERATIONS: 0
* TOTAL NUM.OF MATRIX FACTORIZATION AND PCG ITERATION: 0 0
RESTART FILES: pression_statique_07_0001_[0001-0001].rst WRITTEN
-------------
** CPU USER TIME **
#PROC [K]SETUP [M]SETUP IMPL.SOLV IMPL.TOTAL
1 .4862E+02 .0000E+00 .0000E+00 .4930E+02
#PROC CONT.SORT CONT. F ELEMENT KIN.COND. INTEGR. I/O TASK0 ASSEMB. RESOL
1 .5591E-01 .6813E-02 .2678E+00 .2133E-01 .4654E-03 .5012E+00 .9117E-03 .1199E-01 .5030E+02
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Thank you a lot,
Kinds regards,
Alice
Answers
-
Hi Alice,
As a workaround, can you set the Irem_gap as 1 in interface parameters and run the analysis. Please update
0 -
Hello George,
Thanks for your quick reply. I did not see it right away.
I solved my problem of slave nodes deactivation by initiating the implicit time step (/IMPL/DTINI). I tried setting Irem_gap to 1 but it did not change the fact that the interface is not taken into account.
Now, there isn't even an error. No slave node is deactivated but the animation shows a penetration. I constantly change the parameters in the interface as well as in the engine implicit options. But I don't manage to solve the issue of penetration. I have performed a checking of the implicit model, giving two warnings :
***** WARNING : ELEMENT FORMULATION ISOLID= 0
IS NOT AVAILABLE FOR STIFFNESS MATRIX BUILDING, USING GENERIC ONE INSTEAD
POSSIBLE CONVERGING ISSUE. *****and
** WARNING :IMPLICIT LOADING DATA **
Concerning the first warning, I first used the /IMPLICIT option ni the control cards but this gave me a lot of errors as Isolid= 14 in implicit default options and this is not compatible with tetra mesh. So I set it back to the generally default '0' value but it might give 'converging issues'. Would you know how I can manage this problem ? And if this might cause m y problem of penetration without errors ?
Concerning the second warning, I don't understand its meaning, nore did I find any help in the forum about this.
Please find attached my engine out file for the implicit checking computation.
Thank you in advance for helping me out.
Alice
0 -
Hi Alice,
Altair Forum User said:***** WARNING : ELEMENT FORMULATION ISOLID= 0
IS NOT AVAILABLE FOR STIFFNESS MATRIX BUILDING, USING GENERIC ONE INSTEAD
POSSIBLE CONVERGING ISSUE. *****Can you try with a different element formulation and see whether the warning is still there?. Normally for 4 node solid tetrahedron there is a drawback of low convergence. As a workaround, I can suggest you to use 10 node solid tetrahedron element and check.
Altair Forum User said:** WARNING :IMPLICIT LOADING DATA **
In some cases, RADIOSS does not detect any loading upon part and stops supposing that the input was false.
The suggested workarounds for this issue are:
1.Check loading in the model.
2.Increase the initial time step.
3.Creating a slight contact penetration in cases where loading is due to contact.
.
0 -
Hello George,
Thank you for your quick reply.
Altair Forum User said:Can you try with a different element formulation and see whether the warning is still there?. Normally for 4 node solid tetrahedron there is a drawback of low convergence. As a workaround, I can suggest you to use 10 node solid tetrahedron element and check.
I have already tried different element formulations but nothing works.
I indeed have 4-node tetrahedron elements. I tried yesterday to change them to 10-node tetrahedrons but did not find how to choose this. How can I tell Hypermesh I want 10-node tetrahedrons ? In a video I saw that Hypermesh gives the choice of first and second order tetra elements but I do not have this choice (see attached word document).
Altair Forum User said:In some cases, RADIOSS does not detect any loading upon part and stops supposing that the input was false.
The suggested workarounds for this issue are:
1.Check loading in the model.
3.Creating a slight contact penetration in cases where loading is due to contact.
Concerning the second warning: my loading consists of a first component compressing a second one; so indeed I have a loading created through contact. In my first simulations I had initial penetration of the contact gap without purpose. This made it work but I have one major problem with this: I have initial loading and this is clearly an outlying result (10^5 N at t = 0, when having 150 N after 0,2 mm compression during real testings).
Altair Forum User said:2.Increase the initial time step.
My initial timestep is of 10 for the moment (/IMPL/DTINI = 10), I will set it higher and come back to you.
Thanks again,
Alice
0 -
Hi Alice,In 3D panel, you will get Order Change option where this function can be used to change a group of elements from first order elements to second order elements.
<?xml version="1.0" encoding="UTF-8"?>
0 -
Hello George,
I managed to change tetra4 -> Brick (degenerated tetrahedron) by unchecking the 'match existing mesh box'. But would still be interested in knowing how to mesh in tetra10.
I changed Isolid to 17 (what to me seems best) and then to 14 (implicit default value) and both warnings disappear.
However my problem remains...
I will come back to you after having checked a higher inital time step and slight initial penetration.
0 -
Hi Alice,
Which HyperWorks version you are using?.In Help Menu>>Click on Updates and System Information and please attach a screen shot of it.
0 -
Hello Geroge,
Thank yo ufor the help on second order elements. I am currently running the simulation on tetra10 and will come back to you on this.
Find attached the screenshot of the updates information. I use version 14.0
<?xml version="1.0" encoding="UTF-8"?>
0 -
Hi Alice,
Please update to the latest version HyperWorks 14.0.130.
For the time being, you can use Order Change option to change from first order elements to second order elements.
0
