A new mathematical foundation for contact interactions in continuum physics (Q880934)

This paper is a further contribution to the debated problem of a rigorous definition of contact interactions between parts of a continuous body. If the interactions are exerted across smooth surface elements, and forces and heat fluxes, are continuous there, then Cauchy's celebrated argument proves that these interactions depend linearly on the unit vector normal to the piece of surface on which they act. But this classical theory of stresses is insufficient for handling singularities due to corners, concentrated forces, and shocks. The history of the attempts of extending the theory starts with a paper by \textit{W. Noll} [The foundations of mechanics. C.I.M.E. 1. Ciclo Bressanone 1965, Nonlinear Continuum Theor., 159--200 (1966; Zbl 0202.56101)] who, in order to have a divergence theorem available, assumed that the parts of a body (subbodies) may be defined as the closures of open sets with piecewise smooth boundaries. But this family of sets is not closed under intersections. Therefore the theory was further generalized. \textit{C. Banfi} and \textit{M. Fabrizio} [Atti Accad. Naz. Lincei, VIII. Ser., Rend., Cl. Sci. Fis. Mat. Nat. 66, 136--142 (1979; Zbl 0446.73006)] and \textit{W. P. Ziemer} [Arch. Ration. Mech. Anal. 84, 189--201 (1983; Zbl 0531.73005)] introduced the notion of sets of finite perimeter as a better characterization of subbodies, a definition improved by \textit{W. Noll} and \textit{E. G. Virga} [Arch. Ration. Mech. Anal. 102, No. 1, 1--21 (1988; Zbl 0668.73005)] by the additional requirement that subbodies are bounded regulary open sets with finite perimeter. The historical account is enriched by Noll's ``Buridean'' problem on how to cut a piece of cheese. Then other, more recent, contributions are quoted, but are not commented upon critically. The author accepts the axiom that physically contact interaction must obey a principle of local action. Then he proves that, even in the case of high singularities of these interactions, they can be represented by a tensor field. The paper does not add substantial progress with respect to the results obtained in the last forty years, but some examples are very illuminating in order to understand the theory.