Implementation of a Fatigue Analysis Program Based on the Dual Formulation of the Boundary Element Method

Resumo

The study and analysis of fatigue are extremely important, given the countless financial and social losses caused by this damage process. For this reason, a complete computational program is implemented for two-dimensional fatigue analysis in cracked components, coded in Fortran 95 language. The Dual Boundary Element Method (DBEM) is used to evaluate the elastic fields of displacement and traction, through the placement of Boundary Integral Equations (BIE) of different natures on each of the crack faces. Stresses and strains at boundary points are evaluated through the hypersingular BIE. Once the variables of interest are obtained in the boundary, the fatigue subcritical propagation of cracks is evaluated based on the concepts of Linear Elastic Fracture Mechanics (LEFM). J-Integral is used to measure the stress intensity factors at the crack tip, due to its effectiveness, precision and excellent coupling with the dual formulation of the Boundary Element Method (BEM). The direction of crack propagation is defined by the Maximum Circumferential Stress criterion. Finally, the lifespan of the cracked component is determined using an adapted Paris Equation, in order to correctly evaluate mixed mode fracture and non-zero mean stresses. Furthermore, using the Rainflow Cycle Counting algorithm, the program is capable of evaluating components under the action of cyclic loads of variable amplitude. In order to evaluate the implemented computational program, two examples are presented. The first considers a Compact Mixed Mode (CMM) specimen under mixed mode fracture. The second example evaluates the driven wheel of an iron ore pelletizing furnace under cyclic loading of variable amplitude. According to the results presented, it is proved the effectiveness and reliability of the developed computational program, which allows to accurately analyze real situations of components under mode I, II or mixed mode fatigue.