Development of the DYNA3D simulation code with automated fracture procedure for brick elements

Abstract

SUMMARYNumerical simulation of cracked structures is an important aspect in structural safety assessment. In recent years, there has been an increasing rate of development of numerical codes for modelling fracture procedure. The subject of this investigation is implementing automated fracture models in the DYNA3D non-linear explicit ÿnite element code to simulate pseudo 3D crack growth procedure. The implemented models have the capabilities of simulating automatic crack propagation without user intervention. The implementation is carried on solid elements. The methodology of implementing fracture models is described. An element deletion-and-replacement remeshing procedure is proposed for updating the explicit geometric description of evolving cracks. Fracture parameters such as stress intensity factors, energy release rates and crack tip opening angle are evaluated. The maximum circumferential stress criterion is used to predict the direction of crack advancement. Seven crack problems are presented to verify the e ectiveness of the methodology. Mesh sensitivity and loading rate e ects are studied in the validation of the presented procedure.

Automated dynamic fracture procedure for explicit time integration brick finite elements

“…The current model uses the same methodology to model actual 3-D crack shapes in finite element models as the earlier LEFM model, [11]. The crack shapes are modeled with edges and surfaces of brick elements.…”

“…Though the domain integral expression used in this model is capable of predicting energy release rates for cracks propagating at speeds up to one-half the shear wave speed in the material, currently, it is not suited for simulations of propagating cracks. The automatic remeshing strategy implemented by Tabiei and Wu [11] to model the crack growth explicitly uses the strategy of delete-andfill process: first, a group of elements in a region around crack front is deleted, then the crack is extended and finally, the local domain is refilled with new elements. Since this process involves the instantaneous release of some nodal constraints, spurious high-frequency oscillations are observed in the finite element solutions making the model unsuitable for simulations of crack propagation problems.…”

“…Other input data are as follows: Simulation time in DYNA3D input file = 0.085 ms The variation of dynamic energy release rate with time for this problem, obtained using the three finite element models with 8000, 2000, and 500 elements are shown in figure (11) along with the exact solution. Two major observations can be made from the curves.…”

“…Furthermore, the stress based failure methodology is not able to describe failure accurately. Therefore, it is important to implement fracture theories for solving crack propagation problems [11]. The present work follows the implementation of LEFM models by Tabiei and Wu [11].…”

“…Therefore, it is important to implement fracture theories for solving crack propagation problems [11]. The present work follows the implementation of LEFM models by Tabiei and Wu [11]. Their models have the capabilities to simulate automatic crack propagation without user intervention.…”