As noted by @antonellasuccar some of the lethe-fluid-particles test rely on arrangment of particles that are generated through DEM. The issue with this is that the arrangement is not perfectly deterministic and that regenerating the arrangement of particles tends to generate a new results for their velocity during the simulation.
We should refactor the CFD-DEM and the VANS tests that rely on particles to have them use particles with a clearly deterministic position (for example always the same arrangement drawn from a file) instead of using a packing made in-situ through DEM. This would make them a lot more robust and they would be better tests this way.
As noted by @antonellasuccar some of the lethe-fluid-particles test rely on arrangment of particles that are generated through DEM. The issue with this is that the arrangement is not perfectly deterministic and that regenerating the arrangement of particles tends to generate a new results for their velocity during the simulation.
We should refactor the CFD-DEM and the VANS tests that rely on particles to have them use particles with a clearly deterministic position (for example always the same arrangement drawn from a file) instead of using a packing made in-situ through DEM. This would make them a lot more robust and they would be better tests this way.