Jump to content
Tiago

quasi static analysis

Recommended Posts

Hello,

 

I have completed a dynamic crash analysis on hypercrash and wanted to convert it to quasi-static. The dynamic analysis lasts no longer than 100 ms so I can use a low time-step.

 

How should I proceed to be able to use radioss for a 2  minute simulation with a higher time step? Is is possible with explicit analysis in radioss or should I use another software?

 

I changed the velocity to an imposed velocity and I am using no mass in order to make it quasi static. The minimum time step that I am introducing is being ignored when the calculation software starts. 

 

Any advice?

 

Thanks in advance.

 

Share this post


Link to post
Share on other sites
Guest

Hi Tiago,

You can do a quasi static analysis in RADIOSS. You can try running with a higher time step, but always check for the mass error in the engine out file.

Can you elaborate more on the quasi static analysis you are trying to perform, that is the analysis objective.

Share this post


Link to post
Share on other sites

Hi George,

 

The quasi static analysis must consist in a simulation of a compression test of a structure so I will impose a velocity in the end nodes of the structure and crush it onto a rigid wall. 

 

I increased the time step to 0.1 s but since the beginning of the calculation, radioss is using a time step of 9e-4 what makes the expected time raise to close to 300 hours.

 

How can I impose the time step? I am using DT1 and DTIX control cards to impose minimum maximum and initial time step

Share this post


Link to post
Share on other sites

I have seen many times explained that the time step must be in the order of the time that a sound wave takes to cross an element. In explicit dynamic analysis that works because the crash tests takes about 100 ms to occur. 

 

A quasi static test with a speed of 2mm per second takes more than 1 minute so the same time step traduces on a computation time of 14 days. 

 

If I try to impose a minimum time step of around 0.1s it is ignored and a time step of 9e-4 is used by radioss. I don't know why.

 

And even if i can impose it, will it be accurate? There's any way to make it implicit so we don't have to worry with the time step? or can radioss perform it as explicit with a high time step and still have accuracy?

 

Adding a control card DT_ELTYPE_KEYWORD_IFLAG it runs for a bigger time step but the error is getting to -95% since the start

Share this post


Link to post
Share on other sites

I spotted an error in the material model.

 

Corrected it and radioss runs now, the problem is that with a time step of 1ms, the mass error is getting to 5e+6. 

With this time step the calculation requires 2000 seconds to run. If i use an even smaller time step will raise the time to many hours or even more than a day

 

Using a time step of 0.1 ms and reducing the time of the test i could get the calculation process down to 6000s but the mass error is still 1e+4

Share this post


Link to post
Share on other sites
Guest

Hi Tiago,

Explicit method is commonly used in crashworthiness analysis due to its capability to solve highly non-linear problems without numerous iterations and convergence problems. However, the time step for explicit methods is limited by the time that the physical wave crosses the element. Therefore, to avoid large amount of CPU time, the explicit method is usually used for non-linear dynamic problems with a short period of simulation duration, that lasts for milliseconds normally.

In your case the run time provided is 125 seconds which is very much high. You can give a try with implicit options available in RADIOSS, but there are chances of convergence issues.

There are many examples and documents available on implicit in RADIOSS. Please go through it and try implementing the same.

Share this post


Link to post
Share on other sites
Guest

Hi Tiago,
There are documents and examples in Help Menu on implicit analysis in RADIOSS. Please search for implicit analysis in the Help Menu.

There will be difference in results (deformation, stresses developed...etc)when velocity is increased.

impli.PNG

Share this post


Link to post
Share on other sites
Guest

Hi,

Normally for a crash analysis the run time will be more especially with 4gb RAM.

But its difficult to predict the simulation time as it directly depends on the time step of the model. If the time step is low the run time can drastically increase.

You can try with /DT/NODA/CST keyword which can reduce the run time.

Share this post


Link to post
Share on other sites
Guest

Hi,

This is bit tougher to answer, as we cannot provide  a general answer. Node and element counts in a model also affect the run time along with the above statements.

I suggest you to download our full car crash model and run the same. The model file can be downloaded from the link below:

This will give you an idea on the estimated run time.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
You are posting as a guest. If you have an account, please sign in.
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...