On the correctness of the basic problem of dynamics in systems with friction (Q581119)

The paradoxical situation whereby the motion of a mechanical system with friction is not unique is studied on the basis of Chetayev's stability postulate, according to which the motion must have a ``particular kind of stability''. As the disturbing factor we consider impacts which arise with the relative motion of rough surfaces and which lead to displacements, normal to the plane of frictional contact. We demand of the true solution that it depends continuously on a small parameter, defining the size of these disturbances. Painlevé's principle is usually employed to select the true solution from all possible solutions: ``two rigid bodies, which under given conditions would not produce any pressure on one another, if they were ideally smooth, whould likewise not act on one another if they were rought''. This a priori principle has so far not obtained experimental confirmation, and its validity is uncertain. Moreover, the solution obtained using Painlevé's principle does not have the property of continuous dependence on the initial conditions. Our present approach leads to one way of resolving the situation whereby the basic problem of dynamics ceases to be correct in cases where the motion does not exist or is not unique in the given initial conditions: in either case we use the hypothesis that the impact is tangential. The resulting solution then differs from that obtained by Painlevé's principle.