LS Dyna

LS Dyna

Axial Collapse of Column in Explicit Dynamics

    • Shajee45610
      Subscriber

      Hey there!


      I'm working on my final year project at my university where I am assessing the structural crashworthiness of an unconventional vehicle and looking to improve it. A part of my project involves simulating the behavior under impact which I did under explicit dynamics since the complete software was available. The model itself is complex and this discussion is of two parts which I would greatly appreciate any input and help so that I can complete my project successfully and graduate.


      Part 1


      The first concern I have is on the generation of cracks in my model. In explicit dynamics I have set erosion controls for removal of elements under material failure, however since the entire element is removed (instead of a corner of the element or the point where it fails), the behavior of the impact is questionable. That is, instead of the structure undergoing plastic deformation and cracks, it completely breaks apart due to multiple element erosion. Initially the mesh size was very low and there wasn't any failure at this point. However, since I knew the results wouldn't converge I increased the mesh and that's when the structure completely fractured. Is there any way I can generate cracks or element separation at a point rather than the entire element being removed? I am limited in time as well as computing power hence I can't greatly increase my mesh size as well nor reduce my time inc.


      I wish I could share the file with you but the nature of the project is such that I am not in a position to do so due to confidentiality. However I will share the video file instead of what I am talking about through this google drive link of the behavior of the structure in low mesh and higher mesh conditions.  The material used is Steel 1006 with a Max Equivalent Plastic Strain Failure of 0.4


      https://drive.google.com/drive/folders/1XKuB68f7ZnBMGJQ1JMXbyGJIovZmMLGK?usp=sharing


       


      Part 2


       


      The second concern I have is in studying axially deformable columns in order to use them as energy absorbing members (like in a crumple-zone). I have attached the workbench file and have run the simulation with the erosion turned on and off. With erosion turned on I run into a similar problem as in part 1. The entire structure does not undergo deformation, producing the characteristic folds as seen in the image included in the Google Drive Link. Instead, since the elements are removed, the behaviour looks different. When I keep erosion off, I run into the problem of energy error being too large. I used explicit dynamics because with the necessary improvements I plan to make I will run a full scale model to once again study behaviour under impact. 


       


      TLR for part 2 - I expected to see the characteristic folds throughout the structure and not just to the two corners and also for the elements to not be removed as it did. The material used was arbitrarily selected as Aluminium and I just played around with its failure criterion. I played a lot with mesh density to see how it is but couldn't get what I was looking for.


       


      I would greatly appreciate some help in this, it would mean a lot and I'll heavily be in depth as it would help me complete my degree.


       


      Best regards!


      Shajee

    • peteroznewman
      Subscriber

      Hello Shajee,


      You might get something out of reading this discussion.


      Here is a Static Structural model that is showing the classic folding you mentioned above. 


      Right click on the video after it starts and select Loop to make the video play over and over.



       


      Regards,


      Peter

    • Sandeep Medikonda
      Ansys Employee

      Shajee,


        For the first part of your question. Did you try having a Bonded type contact along with a Breakable property specified.?


        If you haven't looked at this give this a try.


      Regards,


      Sandeep

    • Shajee45610
      Subscriber

      Hi Peter,


      Thank you for your quick response and apologies for the delay on my end, I went to sleep right after I posted the discussion and had a busy Monday Morning. I am yet to fully watch your edit videos on that discussion but what caught my attention was towards the end of the discussion when you created a joint at the intersection of three tubes and constrained it to Y=0. While it looks a lot more realistic, My concern is that if the behavior can be considered more accurate with your edit or without since we are artificially constraining a joint to Y=0. I use Ansys 18.2 so I am doubtful if joints are supported.


      For clarification, I am running my model under EuroNCAP Full Width Frontal Collision protocols, where the model is at 13.89 m/s (50 km/h). I have included a rigid wall as well as a floor and block masses instead of point masses. I ran simple models initially and found that with 'Erosion Under Material Failure Turned Off', I run into Energy error, which is why it is included in the final model.


      Is it possible for elements to simply break off from one another instead of being completely removed from the equation? Because this behavior is causing a massive fracture to generate as seen in the videos. If it is not possible, is it justifiable to state that 'Based on the simulation, the probability of high plastic deformation and failure is within said region'?


      The element removal behaviour is of concern in part 2 as well because I run into energy error when I turn erosion off. However, as in your video and a few others I just found, the classic folds seem to be generated only under static structural analysis. So I am guessing this behaviour is due to the explicit nature of the collision? I attached some video files of part 2 as well.


      Edit: For some reason, when I try to upload video files, it says file extension not allowed. So please check my drive link again.


      Best Regards


      Shajee

    • Shajee45610
      Subscriber

      Hi Sandeep,


      No I didn't include it in my model but I will see where I can include it. I'm guessing some of the energy will be absorbed by breakage correct?


      Thank you


      Shajee

    • Rob
      Ansys Employee

      I've just unstuck some posts from the Approval system: I think this has caused a duplicate, but I'll leave them for now. 

    • Sandeep Medikonda
      Ansys Employee

      Not only that but I this is a more accurate way to represent crack and fracture compared to element erosion. Check out this section on Fracture and Crack here.


      Regards,


      Sandeep

    • Shajee45610
      Subscriber
      Hey all
      I managed to get the results I sought after for part 2. Turns out that mesh type also plays a big role when stimulating axial crush columns in explicit dynamics. While the manner of deformations differed between hexahedral and tetrahedral type meshes, excessive element degradation wasn't witnessed in tetrahedral mesh. I'll have to compare results to real world tests to see what's ideal for simulation for my complete model.

      Thanks for your inputs!
      Regards
      Shajee
    • Sandeep Medikonda
      Ansys Employee

      Hi Shajee,


      If a post answers your original question, please mark it as a solution (‘Is Solution') so that it might help someone in the future (Even if it is your own post). This will help reduce repetitive questions and help provide better support on this Forum.


      Check out this post on the Guidelines/Best Practices for the Student Community.


      Regards,


      Sandeep

Viewing 8 reply threads
  • You must be logged in to reply to this topic.