Search

This paper presents a methodology based on the NICE integration scheme [1, 2] for simple and rapid numerical implementation of a class of plasticity constitutive models. In this regard, an algorithm is purposely developed for the implementation of newly developed advanced constitutive models into explicit finite element framework. The methodology follows the organization of the problem state variables into an extended form, which allows the constitutive models' equations to be organized in such a way, that the algorithm can be optionally extended with minimal effort to integrate also evolution equations related to a description of other specific phenomena, such as damage, distortional hardening, phase transitions, degradation etc. To confirm simplicity of the program implementation, computational robustness, effectiveness and improved accuracy of the implemented integration algorithm, a deep drawing simulation of the cylindrical cup is considered as the case study, performed in ABAQUS/Explicit. As a fairly complex considered model, the YLD2004-18p model [3, 4] is first implemented via external subroutine VUMAT. Further, to give additional proof of the simplicity of the proposed methodology, a combination of the YLD2004-18p model and Gurson-Tvergaard-Needleman model (GTN) is considered. As demonstrated, the implementation is really obtained in a very simple way.

The paper deals with constitutive modelling of highly anisotropic sheet metals. It presents FEM based earing predictions in cup drawing simulation of highly anisotropic aluminium alloys where more than four ears occur. For that purpose the BBC2008 yield criterion, which is a plane‐stress yield criterion formulated in the form of a finite series, is used. Thus defined criterion can be expanded to retain more or less terms, depending on the amount of given experimental data. In order to use the model in sheet metal forming simulations we have implemented it in a general purpose finite element code ABAQUS/Explicit via VUMAT subroutine, considering alternatively eight or sixteen parameters (8p and 16p version). For the integration of the constitutive model the explicit NICE (Next Increment Corrects Error) integration scheme has been used. Due to the scheme effectiveness the CPU time consumption for a simulation is comparable to the time consumption of built‐in constitutive models. Two aluminium alloys, namely AA5042‐H2 and AA2090‐T3, have been used for a validation of the model. For both alloys the parameters of the BBC2008 model have been identified with a developed numerical procedure, based on a minimization of the developed cost function. For both materials, the predictions of the BBC2008 model prove to be in very good agreement with the experimental results. The flexibility and the accuracy of the model together with the identification and integration procedure guarantee the applicability of the BBC2008 yield criterion in industrial applications.

The paper focuses on a reliable identification of the hardening curve of sheet metal after the onset of necking by means of a purposely developed shear test. The main comparative advantage of the test is its simple implementation needing no further development of the experimental equipment used in standard tensile testing. In addition, in order to achieve more flexible hardening curve description from those given by the classical analytical hardening laws, the paper presents the hardening curve approximation based on the cubic spline technique. The identified hardening curves are validated and compared to the analytically extrapolated ones obtained by a continuation of classical hardening laws for large plastic strain. It is shown, that due to large difference in response the analytical extrapolation is seriously disputable.

Strain recovery after removal of forming loads, commonly defined as springback, is of great concern in sheet metal forming, in particular with regard to proper prediction of the final shape of the part. To control the problem a lot of work has been done, either by minimizing the springback on the material side or by increasing the estimation precision in corresponding process simulations. Unfortunately, by currently available software springback still cannot be adequately predicted, because most analyses of springback are using linear, isotropic and constant Young’s modulus and Poisson’s ratio. But, as it was measured and reported, none of it is true. The aim of this work is to propose an upgraded mechanical model which takes evolution of damage and related orthotropic stiffness degradation into account. Damage is considered by inclusion of ellipsoidal cavities, and their influence on the stiffness degradation is taken in accordance with the Mori‐Tanaka theory, adopting the GTN model for plastic flow. Wit...

Strain recovery after removal of forming loads, commonly defined as springback, is of great concern in sheet metal forming, in particular with regard to proper prediction of the final shape of the part. To control the problem a lot of work has been done, either by minimizing the springback on the material side or by increasing the estimation precision in corresponding process simulations. Unfortunately, by currently available software springback still cannot be adequately predicted, because most analyses of springback are using linear, isotropic and constant Young’s modulus and Poisson’s ratio. But, as it was measured and reported, none of it is true. The aim of this work is to propose an upgraded mechanical model which takes evolution of damage and related orthotropic stiffness degradation into account. Damage is considered by inclusion of ellipsoidal cavities, and their influence on the stiffness degradation is taken in accordance with the Mori‐Tanaka theory, adopting the GTN model for plastic flow. In order to improve the numerical springback prediction, two major things are important: first, the correct evaluation of the stress‐strain state at the end of the forming process, and second, correctness of the elastic properties used in the elastic relaxation analysis. Since in modelling of the forming process we adopt a damage constitutive model with orthotropic stiffness degradation considered, a corresponding damage parameters identification upon specific experimental tests data must be performed first, independently of the metal forming modelling. An improved identification of material parameters, which simultaneously considers tensile test results with different type of specimens and using neural network, is proposed. With regard to the case in which damage in material is neglected it is shown in the article how the springback of a formed part differs, when we take orthotropic damage evolution into consideration.