M. C. Oliveira

Improving Computational Performance through HPC Techniques: case study using DD3IMP in‐house code

The computational efficiency of the FEA is strongly dependent on the algorithmic and numerical efficiency of the FE solver. This is particularly important in case of implicit FE codes, such as DD3IMP, the in‐house static implicit FE solver under analysis in this work. This study describes the procedure adopted to identify the main computational bottlenecks of the FE solver in order to introduce the OpenMP directives and, consequently, to achieve a major speedup of the whole algorithm. The different parallelized branches of the code are tested using the well‐known square cup deep drawing example, considering different FE discretizations. The analysis of the preliminary results, concerning the CPU wall time, allows to demonstrate that the adoption of HPC techniques, such as the abovementioned OpenMP directives, enables to: (i) achieve a speedup factor close to the number of cores (in a single computer); (ii) solve a problem in a shorter time; (iii) solve a bigger problem in the same amount of time and, thus...

Improving Nagata patch interpolation applied for tool surface description in sheet metal forming simulation

The contact surface description is a very important field in the numerical simulation of problems involving frictional contact, which are among the most difficult ones in continuum mechanics, as is the case of sheet metal forming simulation. In this paper, a methodology to control the Nagata patch interpolation of piecewise linear meshes is proposed, in order to improve its applicability for tool surface description used in the numerical simulation of sheet metal forming processes. The interpolation can be applied either to triangular and quadrilateral Nagata patches, as well as structured and unstructured patches. The normal vectors needed for the Nagata interpolation are obtained through two distinct strategies. The first uses the information available in the CAD surface model, while the second resorts only to the piecewise linear mesh model information. In order to evaluate the interpolation accuracy, the Nagata patch is applied to describe a sheet metal forming complex shape part tool geometry. The results obtained show that, regardless of the strategy used to evaluate the surface normal vectors, the use of the proposed Nagata patch interpolation enables a large improvement in the geometric accuracy when compared with the models composed by piecewise linear elements. The use of CAD surface geometry to evaluate the surface normal vectors leads to the best Nagata patch interpolation in terms of shape and normal vector field accuracy.

Study on the Influence of the Refinement of a 3‐D Finite Element Mesh in Springback Evaluation of Plane‐Strain Channel Sections

Springback phenomenon associated with the elastic properties of sheet metals makes the design of forming dies a complex task. Thus, to develop consistent algorithms for springback compensation an accurate prediction of the amount of springback is mandatory. The numerical simulation using the finite element method is consensually the only feasible method to predict springback. However, springback prediction is a very complicated task and highly sensitive to various numerical parameters of finite elements (FE), such as: type, order, integration scheme, shape and size, as well the time integration formulae and the unloading strategy. All these numerical parameters make numerical simulation of springback more sensitive to numerical tolerances than the forming operation. In case of an unconstrained cylindrical bending, the in‐plane to thickness FE size ratio is more relevant than the number of FE layers through‐thickness, for the numerical prediction of final stress and strain states, variables of paramount importance for an accurate springback prediction. The aim of the present work is to evaluate the influence of the refinement of a 3‐D FE mesh, namely the in‐plane mesh refinement and the number of through‐thickness FE layers, in springback prediction. The selected example corresponds to the first stage of the “Numisheet’05 Benchmark♯3”, which consists basically in the sheet forming of a channel section in an industrial‐scale channel draw die. The physical drawbeads are accurately taken into account in the numerical model in order to accurately reproduce its influence during the forming process simulation. FEM simulations were carried out with the in‐house code DD3IMP. Solid finite elements were used. They are recommended for accuracy in FE springback simulation when the ratio between the tool radius and blank thickness is lower than 5–6. In the selected example the drawbead radius is 4.0 mm. The influence of the FE mesh refinement in springback prediction is discussed, for this example where the drawbead restraining force results in a non‐symmetrical through‐thickness stress gradient.

Inverse Strategies for Identifying the Parameters of Constitutive Laws of Metal Sheets

This article is a review regarding recently developed inverse strategies coupled with finite element simulations for the identification of the parameters of constitutive laws that describe the plastic behaviour of metal sheets. It highlights that the identification procedure is dictated by the loading conditions, the geometry of the sample, the type of experimental results selected for the analysis, the cost function, and optimization algorithm used. Also, the type of constitutive law (isotropic and/or kinematic hardening laws and/or anisotropic yield criterion), whose parameters are intended to be identified, affects the whole identification procedure.

Comparison of two LCA Methodologies in the Machine-Tools Environmental Performance Improvement Process

This paper presents the results of a comparison between a simplified LCA analysis and a detailed LCA carried out under a Ecodesign project of a machine-tool, a commercial CNC press-brake. The aim was to assess to what degree, the most detailed information resulting from a detailed LCA, allows decisions qualitatively superior in terms of environmental efficiency of the equipment to redesign. This work was also intended to initiate a process of establishing guidelines to support these studies, particularly in the case of its application to complex mechanical and electromechanical systems such as machine-tools.

Springback Evaluation with Several Phenomenological Yield Criteria

In the last decades, several orthotropic phenomenological yield criteria were proposed to accurately describe the anisotropic behaviour of the rolled metallic sheets that are widely used in the production of sheet metal formed parts. In this work, the authors evoke the implementation of several yield criteria in the implicit finite element code DD3IMP, namely the Hill 1948 [1], the Cazacu & Barlat 2001 [2] and the isotropic Drucker’s yield criterion mixed with a linear transformation [2]. The Numisheet’2002 benchmark “Unconstrained Cylindrical Bending” was selected to evaluate the influence of the yield criteria on the springback evaluation. A 6111-T4 aluminium alloy was used. Its mechanical characterization was performed by Alcoa [3]. The results show that only the Cazacu and Barlat 2001 yield criterion fits very accurately both the uniaxial tensile yield stresses and the r-values. All the other yield criteria fits rather well the experimental r-values. Concerning the influence of the yield criteria on springback, only a minor influence was found. However, the results obtained from the Cazacu and Barlat 2001 yield criterion are clearly the closest to the experimental ones [4]. Introduction At the present time, the reliability of finite element method simulations is still a dilemma, which easily explains the reason of the continuous efforts and developments that have been done by all the scientific community in the last decades. In sheet metal forming, numerous works have been published concerning the improvement of both the behaviour laws and the initial yield locus description. In fact, for computer simulation of sheet metal forming processes, a quantitative description of plastic anisotropy by the yield locus of the material is required. The accuracy of the numerical results is obviously correlated with the accuracy of this description. Several phenomenological yield criteria are implemented in the implicit finite element code DD3IMP (contraction of ‘Deep Drawing 3-D IMPlicit code’) [5]. DD3IMP is a 3-D elastoplastic finite element code with an updated Lagrangian formulation, following a full implicit time integration scheme, large elastoplastic strains and rotations, with several isotropic and anisotropic constitutive models (8 yield criteria and 7 isotropic/kinematic hardening laws). The Coulomb’s law models the frictional contact problem, which is treated with an augmented Lagrangian approach. This code has a 3-D finite element library with different types of elements and integration schemes. The main goals of this work are to compare the fitting accuracy of the proposed yield criteria to the experimental results, and to present a comparative study of the influence of those yield criteria in the evaluation of the springback effects. Numisheet’2002 benchmark “Unconstrained Cylindrical Bending” [4] was select to evaluate the influence of the yield criteria on springback. A 6111-T4 aluminium alloy was used. The mechanical characterization was performed by Alcoa [3]. A new methodology (the minimization of a functional) to the identification of the anisotropy parameters of the yield criteria is also followed and presented. Materials Science Forum Online: 2004-05-15 ISSN: 1662-9752, Vols. 455-456, pp 732-736 doi:10.4028/www.scientific.net/MSF.455-456.732

Towards standard benchmarks and reference data for validation and improvement of numerical simulation in sheet metal forming

Abstract The last decade has witnessed many advances and a lot of improvement in FE codes for simulation of sheet metal forming processes. Such advances could be followed mainly by benchmarks proposed in Numisheet conferences. It was possible to notice that the scatter of results among numerical codes has decreased so significantly that recently scattering of experimental results among different corporations was evident. However in order to pursue further developments and validate numerical results it is fundamental to have reliable reference experimental data. This is one of the objectives of a current IMS project called 3DS-Digital Die Design System. In this paper such objectives are presented as well as some of the proposed benchmarks. It is intended to show part of the developed work concerning tool design and manufacturing methodology. Also an experimental case study about the use of piercing holes in parts and the use of counter-punch is presented. Finally some simulation results are also shown concerning one of the proposed benchmarks.

Young's modulus of thin films using depth-sensing indentation

A new methodology for the determination of Youngs modulus of thin films, using a single hardness test measurement of film/substrate composites, has been developed. It is based on a recently proposed weight function, the reciprocal of the Gao function. Firstly, results of three-dimensional numerical simulation of the Vickers hardness tests on several fictitious composites are considered. Then, the methodology is checked experimentally by using depth-sensing indentation results on real composite materials.

Application of the Incremental Volumetric Remapping Method in the Simulation of Multi‐Step Deep Drawing Processes

Since sheet metal forming has a high percentage contribution in the overall design costs of a new car, this engineering area assisted in the last decades to considerable development efforts. The present challenge is to simulate all the production stages, from the initial blank sheet to the final part ready to assembly. On this particular issue of multi‐step deep‐drawing simulation, this work presents a new remapping method called Incremental Volumetric Remapping (IVR) developed to minimize the error that occurs, when performing the variable transfer operation between two different meshes. The IVR method is based in a volumetric approach where the calculus of the remapped state variables is obtained by means of a weighted average of the intersection volume between the meshes. The method performance is tested and compared with a standard extrapolation‐interpolation, by applying a numerical example of the Numisheet’005 Conference, “The Channel Draw/Cylindrical Cup Benchmark”.

Mechanical Behaviour and Springback Study of an Aluminium Alloy in Warm Forming Conditions

This study deals with the mechanical behaviour and material modelling of an AA5754-O alloy at elevated temperature. Experimental shear tests were performed from room temperature up to 200°C, and the material behaviour has been identified with both shear and tensile tests, as a function of temperature. To analyse the influence of temperature during forming over springback, a split-ring test is used. Experimental results are obtained and compared to numerical simulations performed with the finite element in-house code DD3IMP. The numerical process of ring splitting is performed with the in-house code DD3TRIM. The main observed data are force-displacement curves of the punch during forming, cup thickness at the end of forming, and ring gap after splitting. It is shown that all these parameters are strongly dependent on the forming temperature. A correlation is obtained between experimental data and numerical simulation for the evolution of punch force and opening after springback as a function of temperature. The distribution of the tangential stress in the cup wall is the main factor influencing the springback mechanism in warm forming condition.