The projective-iterative method and neural network estimation for buckling of elastic plates in nonlinear theory

The paper deals with a nonlinear buckling problem for von Karman elastic plates in bending. The simply supported and partially clamped plates are considered. A variational-projective approach with an iterative scheme is suggested for the calculation of eigenfunctions and the buckling loads for the studied problem. The convergence of the projective-iterative method is investigated. The bifurcation scenario is demonstrated by examples. Numerical results show the effectiveness of the proposed method. Prediction of the eigenvalues (which are the bifurcation points of the nonlinear problem) of the linearized problem with different sizes of the plate can also be done by an approximately trained neural network, as it is briefly demonstrated in this work.