Computational Aspects of Time Integration Procedures in Structural Dynamics—Part 1: Implementation

Abstract

A unified approach for the implementation of direct time integration procedures in structural dynamics is presented. Two key performance assessment factors are considered, viz., computational effort and error propagation. It is shown that these factors are strongly affected by details in the reduction of the second-order equations of motion to a system of first-order equations, and by the computational path followed at each time step. Part 1 is primarily devoted to the study of the organization of the computational process. Specific implementation forms derivable from the unified approach are studied in detail, and rated accordingly. An analysis of the computational error propagation characteristics of these implementations is presented in Part 2.