A Selective Relaxation Iterative Solution Technique for Non-Linear Structural-Analysis Problems

AbstractA series of algorithms, to solve the nonlinear algebraic equations resulting from nonlinear finite element models, are proposed. These algorithms involve a selective relaxation of the corrections defined by the modified Newton method. The resulting techniques accelerate the convergence of the modified Newton method. The improved convergence rate is obtained at the expense of a small increase in storage requirements. The results of numerical experiments, on certain nonlinear elasticity problems, are presented to verify the developed techniques.