The paradigmatic example of a twisted square plate is here considered. An equivalent partition of the plate in a grid of beams a la Grashof is found such that, as the number of beams tends to infinity, the grid exhibits the same deflection of the plate. This scheme is used to interpret, through the distinction between Euler‐Bernoulli and Timoshenko beam theories, the different types of natural boundary conditions that can arise in the Kirchhoff‐Love and Mindlin‐Reissner theories of plates. A physically based interpretation for the occurrence of lumped forces at the plate corners through the formation of a boundary layer is provided. It is well known that the solution of the biharmonic equation governing the bending of plates in Kirchhoff‐ Love theory is compatible with only two distinct conditions at each boundary point, whereas in general three boundary data can be independently assigned on an unconstrained border. This contradiction for the order of the equation is a two-hundred-year-old problem. The paradox arose when the three-boundarydata statement by Poisson [1829] was criticized by Kirchhoff [1850], who obtained only two natural conditions at the border within a variational framework, using a static equivalence sometimes referred to as the “Kirchhoff transformation” [Vasil’ev 2012]. This result arose from the first variation of the energy functional, but it was not corroborated by any physically based interpretation. A long discussion ensued among the most eminent scientists of the period with the purpose of reconciling the Poisson and Kirchhoff theories. The dispute culminated with the elementary interpretation by Thomson and Tait [1883], who showed how to reduce the torque per unit length on the contour to a shear transverse force. Friedrichs and Dressler [1961] and Gol’Denveiser and Kolos [1965] have independently shown that the plate theory is the leading term of the expansion solution (in a small thickness parameter) for the linear elastostatics of thin, flat, isotropic bodies. As expected, this leading term alone is unable to satisfy arbitrarily prescribed edge conditions. There has been a renewed interest during the last years in the fundamental problem of understanding the relationship between the three-dimensional elasticity theory and theories for lower-dimensional objects (plates, shells, rods). Due to the availability of sophisticated methods of variational convergence [Ciarlet 1997], important achievements have been obtained by showing that various theories of plates arise as a rigorous variational limit (or 0-limit) of the equations of three-dimensional elasticity as the
Read full abstract