problema crash su honeycomb con patran/nastran (solutore marc)

[Claude87]

Hello everyone!...I am an engineering student and I have a serious problem in simulating the impact of a steel ball on the honeycomb (aluminum, modeled as elastoplasty) of a sandwich panel (subline that we are simulating only the impact on the honeycomb or there are no skins in the model). I am using patran as preprocessor and nastran as solutor (only 600 implied non-linear, i.e. marc). the ball was attributed a speed of 5 m/s. the model seems ok however when I analyze with nastran I can not have some hint of crash. we have reduced the time steps for the analysis (we fixed the simulation time on 0.1-10 ms) but in the file of marc always results a exit number of 3015, which indicates a convergence problem if not mistaken. I just can't figure out where the problem is, even though I have to stress that I use this software recently. I would be very grateful if you give me some help or advice. Thank you in advance!

 

[MacGG]

Hello claude87,

what it means "I can't get some kind of crash.".

However if you use the sol600, let's talk about medium/low speed impact (not properly crashing), in any case, I suggest you read the manual md nastran application examples, which you can also find in http://www.scribd.com/ search nastran, impact, and you will find a number of examples that should help you.

Hi.

♪

 

[Claude87]

hi macgg!...they are actually impacted at low speed...and it is actually not very clear that phrase. ..however I meant that as soon as I arrive at the moment I should have the impact (I made two quick accounts by hand) nothing really happens: the ball stops and the panel does not deform as it should (its movements are almost null). I thought there might be a problem with elastoplasty but I don't think so. In fact I have also read some examples of application on the manuals of nastran and md nastran that use the sol 600 and is prorpio that the problem: I imposed the analysis by following the instructions of the manuals and yet I arrive at 3015 or the analysis converges but gives me null movements both for the sphere and for the honeycomb (as if everything remained still!!). So I can't understand what's wrong, even because we did a simple static test analysis (without ball, applying a force to the panel) and that works....in any case I thank you for the advice and I will continue to try.

 

[MacGG]

Hello claude87,

Okay, excuse the banal question, does the sphere have inertial properties?
I guess it's a rigid contact body, you control it at speed?
how does the energy of the body vary according to time?

Hi.

♪

 

[Claude87]

Hello, macgg!

then, initially we shaped the sphere as a rigid contact body, however analysis did not converge (always the usual 3015) and then we meshato also the sphere. I explain: we have created a spherical surface with shell elements, that is, we have passed from a simple rigid body (solid) to a deformable spherical surface. then we also tried to create a sphere with solid elements (hex and wedge). in these last two cases we created contact as a deformable body and then we attributed to the knots of the ball an initial speed. In fact, I think it would be more appropriate to use a rigid sphere without mesh because in fact we are interested in what happens to Honeycomb not the sphere. In short, we are proceeding for attempts and although it seems to be a relatively simple teaching exercise we cannot come out of it..... as far as energy in theory is what we have to find, is the goal of our work: we have to evaluate the energy change of the system or better how much energy is dissipated by the panel (the ball is attributed only the speed, we do not even consider gravity). ..however now that you make me think when we left the road of the rigid solid sphere we had not reduced much time steps for analysis... could that be? ? ...in any case I thank you for the availability and for the help macgg... in case you have other suggestions thanks in advance!

Bye!

 

[MacGG]

Hello claude87,

I apologize if I insist, but the sphere, whether rigid or flexible, has an inertia? or does it have mass properties (both rigid, and elements that make it flexible)?
said this, how does the overall energy of the same vary during at least the simulation stretch you do?

Hi.

♪

 

[Claude87]

Hello, macgg!

then the ball realized with shell elements has a mass because to the elements I assigned a material (steel) and a thickness... to the rigid sphere instead I have not assigned any properties. ..only the contact of rigid type and the initial speed... regarding the energy are a bit in difficulty because we didn't think of this aspect...that is, we didn't set it in the output request section because we wanted to first verify that the analysis worked and then in a later phase we would set more accurately all the outputs we need to complete the work.. .

 

[MacGG]

Hello claude87,

In any case, I believe that this output is of fundamental importance and so I recommend that you calculate it.

If you do not define inertial properties, there is no energy to deform/break anything.
The calculation, compared to the rigid sphere, may not converge because the parameters of the solutor are not correct or however the orientation of the elements is not correct. the external faces of the elements, are those that can go in contact with the rest (default). verify it both on the ball and on the panel.

Hi.

♪

 

[Claude87]

Hello, macgg!

after days and days we managed to come to the end of it...in reality we abandoned marc and we passed to the sol700 also because I noticed that usually the problems of impact are treated with explicit solutor...with the rigid sphere continues to have problems (probably because we did not allocate an inertia), however using a ball realized with shell elements the simulation works...the crushed panel and gives the expected deformations...I thank you for the advice that it treated us relatively simple

to the next magg and thanks again.. .

Bye!

 

[dinoMsc]

hi also I am using the sol7100 for such an example my ball though is def (matd024), while the slain dv is rigid.

I richiest the glstat and my sliding interdace energy is negative... Can you tell me if it's normal? ? ?

ps: sn in case of null friction.