K-V is a linear viscoelastic model, so will likely only work for very small strains for something quite brittle like ice.  You don't mention how hard you are hitting the plate, but I would suggest that you look at metal plasticity theory as an alternative. The simplest of these theories just have a yield stress and are not rate dependent, but there are more complex models that include those kinds of effects.

No Adam, I think for the impact study the plasticity isn't a relevant process: it's far too slow (with dislocation movement etc). Impact is generally a very different regime from a classical stress-relaxation (which is what Voigt model is designed for): in impact it is all about high-frequency Fourier modes arising from the steep front (near-infinite strain rate).

Nina: generic viscoelasticity is governed by the retarded constitutive relation: stress(t) = [integral]_0^t  E(t-t') strain-rate(t') dt'. The simplest impact model is strain rate = e*delta(t'). There are various issues the integral with a delta-function at the edge of interval, but very crudely, this gives: stress(t) = E(t)*e. The relaxation modulus E(t) is usually something like E= Eo + Er*exp[-t/T], The single relaxation time T in such a trivial construct is always the ration of viscosity to Er.  However. the literal Voigt model does not have stress relaxation in it -- you need a Zener model to account for it, see this link, for instance:

http://arxiv.org/pdf/1110.3400.pdf

Thank you very much for your answer. However, the Kelvin-Viogt model is widely used for particle collisions because of its simplicity and applicability. The system I want to study are ice ball (with a diameter of a few centimeters) impacts of velocities of several cm/s. As long as the impact velocity is below a certain value the model should be valid. I would be interested in a limit of the strain rate or the impact velocity below which the Kelvin-Voigt model is still adequate.

I think that Eugene has a good point about the speed of the impact with respect to the speed of the collision.  

Just to clarify, I wasn't suggesting that the plasticity approach would give an appropriate description of the micro-mechanics, only that it provided a relatively simple mathematical framework for dealing with recoverable and non-recoverable deformation on relatively long timescales i.e. between approach and 'bounce' off the plate.

To try and address the question of strain rate and velocity limits of the K-V equation, you need to remain within  the linear viscoelastic regime - essentially this requires that the strain rate range is such that the material has a constant retardation time, and the strain is within the Hookean region where the elastic modulus is constant.  These limits are material dependent, and I have no experience for ice, so difficult to judge.  Having said that I would be surprised if the linear strain region is more than a few percent though for a solid like material.

Have you considered the effects of fracture, or does the ice ball remain homogeneous? 

No, fracture is not considered.

maxvel-viogt only support limited materila, best model for all materials is general maxvel model such as many maxvel-voigt with maxvel itself all together in one model, or an another way many kelvin-voigt parralel to maxvel model :) wish u best ...