Hello fellow Abaqus users,
I've been working with Abaqus for the past two years and have encountered a significant issue in my simulations related to the plastic deformation of sheet metals. Specifically, I am trying to simulate a two-stage process: the first stage involves the conventional deep drawing of a metal blank, and the second stage is referred to as "flattening," where another tool deforms the drawn blank in an attempt to make it flat by pushing it backward.
In my simulations, I'm using Abaqus Explicit, where the tools are modeled as rigid surfaces, and the blank is represented as an isotropic material using shell elements with a thickness of 1 mm.
The problem I'm facing is that when I impose a movement of 14 mm for the punch in the simulations, I obtain results indicating that the punch is moving only 13.52 mm. Similarly, in the flattening stage, I set the flattening dies to arrive at a distance of 1 mm, but they actually arrive at 0.8 mm. This discrepancy of just 0.2 mm in the final stage of flattening could significantly impacts the simulated force.
Here are some details about my simulation setup:
- Blank is modeled with shell elements (squares with 2 mm dimensions).
- I'm using the penalty method for contact (not kinematic contact).
- Tangential behavior is defined with a penalty of 0.15, representing the friction coefficient.
- Normal behavior is modeled as "hard contact" to prevent the blank from penetrating the tools, and detachment is allowed after contact.
I haven't made any significant modifications to other parameters, and I've used default configurations for shell elements in the mesh module. I also haven't implemented any contact control algorithms. Interestingly, I've noticed that this strange behavior of the rigid tools not reaching their intended positions doesn't occur when using 3D elements.
I did some mesh sensitivity and this seems not to be the problem.
My questions are:
1) how can I have a precise control on the movement of the tools in Abaqus? Form me having 0,5mm of difference can be problematic.
2) Do you think this issue is related to the mesh or some aspect of the contact algorithm? Or maybe related to the mesh?
I'd appreciate any insights or suggestions to help me resolve this problem and achieve more accurate simulation results.
Thank you in advance for your assistance!
Best regards
Daniele Farioli The issue you are experiencing in your Abaqus Explicit simulations, where the tools do not reach their intended positions with high precision, can be attributed to a complex interplay of factors. Firstly, to achieve precise control over tool movement, attention to various simulation parameters is crucial. You should consider reducing the time step size to enhance the accuracy of the dynamic simulations and define boundary conditions to ensure that the tools are accurately constrained and loaded. Furthermore, adjust the contact settings, especially the penalty factor, contact, and damping, to better capture the interaction between the tools and the blank. Second, the issue is likely related to both the mesh and contact algorithm. Despite performing mesh sensitivity analysis, it's essential to ensure that the mesh resolution around the contact regions is appropriate. Distorted or poorly conditioned elements in these areas can significantly affect contact behavior. Additionally, consider utilizing more advanced contact algorithms such as General Contact or Contact Pair for improved accuracy in handling complex contact interactions. This shift may demand more computational resources but can yield more accurate results. Ultimately, this challenge necessitates an iterative refinement process, taking into account the aspects of time step size, contact settings, mesh quality, and solver parameters, all with a keen eye on the physical behavior of the system. You can connect with me on my WhatsApp for further guidance; https://wa.me/+923440907874
Hi Engr. Tufail, actually your suggestion of reducing the time step size seems working and giving a proper control on the tool movement.
With more regards to my specific problem of forming and flattening, I faced two issues:
1) During forming the punch was not moving as I imposed. But I checked that the punch was sort of bounced away of a small quantity by the blank during the holding phase (in the holding phase the blank holder is pushing the blank against the die, thus inducing small oscillations in the center of the blank which caused the punch -because of contact activate- to move up a bit giving a small mismatch with punch stroke in the end of the forming step). I solved this issue by blocking the vertical motion of the punch in the BC during the "holding step", which was free in the first model.
2) During flattening I observed that all movements/rotations of the upper and lower flattening plates were imposed and/or locked so the mismatch observed was not explainable as in case 1. To solve the issue, I simply reduced the time step of the simulation of 2 orders of magnitude.
Thanks for your help and time, I will still try to refine more the model and check if it is really representative of reality even with such short time step for flattening (from 1 s to 0.01 s).
:D
Daniele Farioli Thank you for considering the suggestions I provided earlier. Good luck :)
Hi Daniele,
I have faced similar issues. Remeshing and using *Fixed Mass Scaling seem to work
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders