Submitted by NSalloum on Tue, 10/16/2018 - 09:42


I was wondering what is the exact physical meaning of the characteristic impact velocity, (i.e. characteristicVelocity), and how does it affect my simulation in LIGGGHTS.

Especially that I read in the documentations that it is basically introduced in the Hookean contact model, but it is rather involved in many scripts not dealing at all with Hook's contact model.

best wishes,

Daniel Queteschiner's picture

Daniel Queteschiner | Tue, 10/16/2018 - 15:26

To express the normal stiffness k_n in the Hooke model as a function of material parameters and relate it to the Hertz model, k_n_Hooke is chosen such that in the case of an undamped, perfectly elastic collision the maximum overlap delta_n_Hooke = v_0 * sqrt(m*/k_n_Hooke) is equal to the maximum overlap in the Hertz model. Here, v_0 is the relative particle velocity at the beginning of the binary contact. (k_n_Hooke = v_0^2 m* / delta_n_Hooke^2 = v_0^2 m* / delta_n_Hertz^2 )
And that is how v_0 gets introduced explicitly into the expression of k_n_Hooke. Since k_n_Hooke is an input parameter for all contacts in the simulation you have to specify an estimate of the average v_0 in the system, i.e. a characteristic impact velocity.

Any tutorial script not using the Hooke contact model but specifying characteristicVelocity will just ignore that setting. It's not used for anything else.

NSalloum | Thu, 10/18/2018 - 12:38

hello Daniel,

This was really very helpful clarifying.

But is their a way I can predicts its value for my system, as I am dealing with particles at the micro-scale? or am I simply supposed to tune it through calibration?

best wishes,