Is it possible to simulate bonding (adhesion) during simulation of explosive welding process in Autodyn? I need it to simulate the generation of residual stresses in explosively welded plates.
High-velocity welding includes several bonding mechanisms that result in a wavey bonding shape at the interface. The formation of melting pockets at the tides of the waves and atomic bonding are among these phenomena. Autodyn is a simple and user-friendly FEM software but according to my experience, its material model and solvers need to be developed in order to capture the mentioned mechanisms, especially when your aim is to calculate the residual stresses. As the HAZ is rather small and the whole process is considered adiabatic, the importance of thermal residual stresses can be negligible. But, severe plastic deformation at the interface should be considered carefully. I suggest using meshless techniques such as SPH as the distortions are high. Also, coding in the autodyn environment will give better results. The paper below is a good reference for your work:
Article Towards better understanding of explosive welding by combina...
Dear Hamed Bakhtiari,
thank you for your answer, it is as I expected. I know this paper, it is a very good article, but according to me, Authors did not simulate the bonding mechanism only jet and wave formation. The very nice pictures of waves are done according to me during the unfinished process of collision. I did a similar simulation in Autodyn applying SPH but the waves are destroyed if the process has proceeded, the flyer plate bounced back. See the attached simulation.
Dear Professor Karolczuk,
As far as I know, FE simulations can not simulate atomic bonding mechanisms. Such mechanisms are simulated in molecular dynamic simulations. If you need to study the atomic bonding, MD might be better. Here is a sample paper:
Article Molecular Dynamics Simulations and Experimental Investigatio...
But mechanical locking and mechanical bondings can be studied in FEM and I think in the case of your study (evaluating the residual stresses), this would be enough. I saw the attached file. I don't know why your flyer plate moved back immediately after the collision and didn't make bonding. A possible solution that works in most cases is to use initial velocity condition instead of igniting explosives. By that I mean, you can calculate the collision velocity of the flyer plate (at the standoff distance of zero) and then apply this velocity to your flyer plate. In this case, explosive materials are not needed anymore and instead, their effect on the flyer plate velocity is considered. The simulation time is reduced considerably and the wavey interface is formed. For simplification, you can apply a constant velocity to your flyer plate at the S.D of zero. If it worked, then you can apply a variable field of velocity vectors across your flyer plate.
Regards
The MD simulation is conducted for very limited volume, very small. So it is not possible to simulate the residual stresses across flyer plate with thickness 1 mm or more. Yes, the application of initial velocity will reduce computation time but the final result will be similar. The flyer plate must move back (spring effect) since the bonding is not managed and the elastic energy is stored in the collided plates. If I modify the material parameters in the way that only plastic strains are generated there will be no elastic energy and no spring effect. Or another way is to force the flyer plate to stay at the base plate by application of additional, artificial pressure. But, such manipulations will not lead to the true distribution of residual stresses. In Autodyn we can simulate the joins but the nodes of two parts must have the same location which is not possible since the gap size must be positive.
I get stuck at these conclusions. May in the Abaqus it would be possible?
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators