How to attach temperature and velocity to rigid body

Dear Lammps Users,
I want to do an impact experiment with rigid copper spheres, and I wish to give the spheres an additional constant temperature, but I found that using
“fix 5 mobile rigid/nvt single temp 1000 1000 0.1”
“velocity mobile set 0 0 -20 sum yes units box”
is not feasible to give him the initial velocity and temperature at the same time, after that I attached the temperature alone without setting the velocity and found that it moved on its own, I looked through the previous maillist and found that I can’t seem to attach the temperature to the rigid body like this, is there any other way to attach the temperature and velocity to the copper ball at the same time?

Generally speaking, it makes no sense to thermostat a single rigid object in as much as it makes no sense to thermalize a single atom. Temperature as a concept is only meaningful for large numbers of degrees of freedom. It is only technically possible to compute the temperature from a small number of degrees of freedom because of the assumption that all degrees of freedom have on average the same kinetic energy, but that assumption becomes less meaningful the degrees of freedom you can average over.

In addition in your particular case, if you apply a thermostat to a particle that you also want to add a velocity to, you need to compute the temperature with the bias from the added velocity removed. Since the added velocity is translational kinetic energy, you now can only compute and control the rotational temperature, i.e. you reduce the number of degrees of freedom by another 3 from 6 to 3. Otherwise, the thermostat will - of course as it is its purpose - “thermostat away” the added velocity.

You can assign velocities to atoms of a rigid body, but you have to do that before you define the rigid body, since the rigid fixes will maintain the position and velocities of each rigid body (for translation and rotation) internally and then compute the positions and velocities of atoms from that. Thus after the rigid bodies are defined, any velocity command has no effect.

In summary, you should use “fix rigid” and not “fix rigid/nvt” and not try to maintain the temperature of your rigid object, and also put the “velocity” command before the “fix rigid” command for the net velocitiy.

Thank you for your recommendation.
I canceled nvt after using it for relaxation, and then I set nve for the substrate alone in the subsequent impact simulation, do I need to set the ensemble for the rigid copper sphere in the impact simulation ?

I have nothing to add to my recommendation. Your question sounds like you have not fully understood what I have explained.