Gurson model parameters for ductile fracture simulation in ASTM A992 steels

Abstract

This study is concerned with the modelling the ductile fracture in ASTM A992 steels using the Gurson‐Tvergaard‐Needleman (GTN) model for high stress triaxiality regime. The GTN model for ASTM A992 structural steels is calibrated from the experiments performed on axisymmetrically notched tensile specimens. The experiments are designed to obtain a range of stress triaxiality and different fracture initiation locations. The non‐uniqueness in the constitutive parameters of the GTN model is illustrated in this study. The choice of a unique set of GTN constitutive parameters is made by choosing the nucleation strain (ɛN) as a material constant. The process of estimating this material specific nucleation strain is provided. All the other GTN model parameters corresponding to the material specific nucleation strain (ɛN) are evaluated to best fit the experimental results. The calibrated GTN model is shown to predict the load displacement behaviour, ductility and fracture initiation locations in the notched specimens. The calibrated GTN parameters are used to successfully predict the ductility of structural components: (a) bars with a hole; (b) plate with reduced section and (c) plate with holes; that are typically found in structural engineering applications.

Ductile tearing simulation of Battelle pipe test using simplified stress‐modified fracture strain concept

“…As calculated values of Δε e p and σ m /σ e depend on the element size in FE analysis, it is obvious that the element size in FE damage analyses can affect simulation results. As explained in the Introduction, such element‐size‐dependence on damage simulation results has been also pointed out by other researchers . Typically an element size ranging from 50 µm to 200 µm is used in ductile tearing simulation to reflect a micro‐structural fracture zone size (such as void spacing).…”

“…The number of parameters can vary depending on the background of a model and the more sophisticated model tends to include more parameters. Another aspect is that simulation results tend to depend on the element size used in simulations . These two aspects can be important in practical application of numerical ductile tearing simulation for cracked piping components.…”

“…Another aspect is that simulation results tend to depend on the element size used in simulations. 16,19,20 These two aspects can be important in practical application of numerical ductile tearing simulation for cracked piping components. Damage model parameters must be determined from material test data, prior to numerical simulation.…”

“…Damage model parameters must be determined from material test data, prior to numerical simulation. As available material data can be limited particularly for operating plant components and typically include only tensile and fracture toughness (J-R) data (sometimes only initiation toughness data), an issue related to reliable determination of damage model parameters should be addressed.. [19][20][21] Furthermore, determination procedures should be clear and transparent so that the parameters can be reproduced independently.…”

Recent Trends in the Development of Gurson’s Model

“…Multi-objective evolutionary algorithms [90] can be used instead to overcome such difficulties. To handle problems of multi-objective optimization, the Pareto optimal solutions can be used, while it is very important to be aware of the non-uniqueness in the constitutive parameters of Gurson-type models [91,92]. A structural design is Pareto optimal, if there is no any other design that satisfies all of the objectives better.…”