General Mechanical

General Mechanical

Transient analysis with modal superposition

    • glouvincs01
      Subscriber


      Hi, 


      I would to simulate a shock on this system ( I have to simplify for confidentiality purposes). 


      The system is composed of four parts :



      • The black one modeled as a rigid body: clamped 

      • The yellow one modeled as a flexible body 

      • The blue one modeled as a flexible body  (it is a screw)

      • The green one modeled as a rigid body with only the vertical direction motion allowed


      The contacts between all the parts (yellow//black: closed - yellow//green: closed and yellow/greenpen) are defined as frictional. 


      I would like to use a transient analysis with modal superposition to have a fast solution. 


      So i tried and used  three blocks: A structural static (for the preload involving bolt pretension) -> a modal analysis -> a transient analysis. 


      The study converged but the probleme is that i have a very high frequency for the first eigen value : about 100 000 Hz. I shoud have around a 1000 Hz.. 


      I read that the modal analysis does not take into account non linearity, and so the frictionnal contact is transformed into a bounded contact... 


      But why do i obtain such a high stiffness? I checked the mass properties... etc 


       



      • I think that the other alternative for this study is a full transient analysis. I applied a sinusoidal profile for the shock. How should i choose the time step ?  I read that the time step is about 1/(20*fmax). How do i estimate the frequency if the modal analysis went wrong ? 


      Thank you for your answers.  


       


       


       

    • Wenlong
      Ansys Employee

      Hi glouvincs01,


       


      It is not surprising if you have your first frequency very high because, as you mentioned, all the contact becomes linear bonded contact and all the bodies act as a whole. Imagine you do a modal analysis on the blue body only.


      For shock wave propagation problems, I would use an Explicit solver. You can use either Ansys Explicit or Ansys LS-DYNA.


       


      Regards,


      Wenlong


       


       


       


       

    • glouvincs01
      Subscriber

      Hi Wenlong, 


      Thank you for your answer... 


      I have a last question regarding the element type. I used a shell element for the yellow part. I'm doing a parametric study on the thickness of the beam. I was expecting less stress when we increase the thickness in bending condition. However, it is not the case...


      Do Shell elements present some restriction? 


      Best regards,


      David

    • Wenlong
      Ansys Employee

      Hi David,


       


      I don't think the shell element has any restrictions on predicting inaccurate stress. It's more likely about the model set up. You can confirm that by running a simple simulation like a cantilever beam. 


       


      Regards,


      Wenlong


       

Viewing 3 reply threads
  • You must be logged in to reply to this topic.