A computational strategy for making complicated structural problems simple

Abstract An effective computational strategy is presented for the analysis of large and complex structures. The strategy is based on generating the response of the complex structure using large perturbations from that of a lower-order (simpler) model associated with a simpler structure (or a simpler mathematical/discrete model of the original structure). The three key elements of the strategy are: (a) mixed (or primitive variable) formulation with the fundamental unknowns consisting of generalized displacements and stress parameters; (b) operator splitting, or a reduction method to relate the arrays and degrees of freedom of the original complex structure to those of the simpler system; and (c) efficient iterative process for the generation of the response of the complex structure starting from that of the simpler system. The effectiveness of the proposed strategy is demonstrated by means of two numerical examples.