Abstract
Analysis and design of strong and durable joints of composite structural parts is recognized as primarily important problem. Most common analytical techniques for bonded, bolted, riveted and other typical joints are based on 2-D shell/plate theories which allow one to obtain reasonable results only in the zones away from the sites of high stress/strain gradients. Accordingly, these theories may provide not only quantitatively inaccurate, but even qualitatively wrong results in the zones of sharp stress/strain variation, even in the case of relatively thin adherends. High shear and peel stress concentration (possibly, even singularity) may occur along these lines causing delaminations, subsequent loss of hermeticity and destruction of the joint. Analysis of this problem shows that complete and accurate solution can be only obtained in 3-D elasticity formulation. A general methodology and some preliminary results of solving 3-D problems of bonded plates are presented in this paper. Numerical results reveal high concentration of the in-plane normal stress and transverse normal and shear stresses along the edge lines of the joint.
