The use of COMSOL Multiphysics for drop test simulations has come a long way since we last tried it for this kind of simulation. Recently we conducted a small comparison of doing a drop test simulation of a cell phone in COMSOL with its implicit solver to solving the same problem with an external explicit solver. Especially for drop test simulations it is often favorable to use explicit solvers, as shown in the table below.
| Feature | Explicit | Implicit |
| Time Step Size | Small | Large |
| Computational Cost Per Step | Low | High |
| Total Computational Cost | High (many steps) | Lower (fewer steps) |
| Best for Large Deformations | ✅ Yes | ⚠ Limited |
| Best for Long Simulations | ❌ No | ✅ Yes |
| Handles Contact Well? | ✅ Yes | ⚠ Can be tricky |
The option to interface directly from COMSOL to an explicit solver can now be done with a new supplementary solver interface we are developing at Resolvent. With this interface it is possible to send the generated mesh and model definitions to the explicit solver and import the solutions back into COSMOL.
For this specific problem the external explicit solver was approximately 4 times faster than COMSOL with default settings, and 100 times faster when making use of mass scaling. We are very encouraged by these first attempts to use the COMSOL interface for solving problems also with a supplementary solver – do you have a good use case or example where this new kind of solver could be of benefit for you?
