رفتار سازه ای Al–Si و ریخته گرى حدیده اى: آزمایش و شبیه سازی عددی

Axial crushing and three-point bending tests have been performed in order to establish an experimental database of the behaviour of generic high-pressure die-cast (HPDC) Al–Si alloys. The experimental data are used to obtain a validated methodology for finite element modelling of thin-walled cast components subjected to quasi-static loading. The HPDC structural components are modelled in the non-linear explicit FE-code LS-DYNA using shell elements. The behaviour of the cast aluminium alloys are described using the classical J2J2 flow theory and the Cockcroft–Latham fracture criterion, which is coupled with an element erosion algorithm available in LS-DYNA. The comparison of the experimental and predicted behaviours of HPDC components gives promising results. The use of the fracture criterion of Cockcroft and Latham seems to be an effective approach to predict failure in HPDC components. A novel modelling approach is outlined accounting for different material properties through the thickness, thus incorporating the effects of a fine-grained surface layer and a central region that possesses lower ductility.

The use of high-pressure die-castings (HPDC) made of aluminium alloys is no longer limited to applications such as fittings and housings. The casting process and the aluminium alloys have reached a level of quality that makes manufacturing of structural components possible. However, this requires careful control of the casting process, melt quality and a correct selection of aluminium alloys. A commonly used aluminium alloy for HPDC automotive components is AlSi9Mg. The requirements for high ductility in chassis components are stringent, and a cost and time consuming solution heat treatment is often required for this type of alloys. The ductility of Al–Si castings is largely governed by the volume fraction of Al–Si eutectic and it is therefore advantageous to reduce the Si content. A reduction in the amount of Si is feasible without significantly affecting the castability. Fig. 1 shows the typical HPDC microstructure of an AlSi4Mg alloy. This alloy contains approximately 20% Al–Si eutectic in contrast to AlSi9Mg that contains about 80%, as estimated by the Alstruc solidification model (Dons et al., 1999). Therefore, the AlSi4Mg alloy gives much better ductility in the as-cast condition (Cosse et al., 2003). In addition, some microstructural features that are inherent to the HPDC process control the ductility in the die-castings. It is widely accepted that a fine-grained, defect-free surface layer is a requisite for obtaining good ductility in thin-walled HPDC aluminium alloys. Furthermore, the central region, or core, that contains macro-segregation of Al–Si eutectic in the form of bands that follow the surface contour of the casting (Laukli et al., 2005a) and a mixture of externally solidified crystals (ESCs) and fine grains (Laukli et al., 2005b), affects the properties. The segregation bands form from deformation of the mushy zone adjacent to the die wall during the HPDC process and dilatation that results in localisation of segregated liquid (Gourlay and Dahle, 2007).This article develops engineering design and modelling tools that allow the structural behaviour of thin-walled cast aluminium components to be predicted under static and dynamic loading conditions. Here, it has been chosen to model the observed behaviour using relatively simple models. The results from this work can then serve as a basis for development of more sophisticated models at a later stage. When using HPDC aluminium alloys in structural applications, it is a requisite to understand the influence of the microstructural features on the structural behaviour. Therefore, to obtain a realistic model of the structural behaviour of HPDC components, a correct interpretation of defects must be applied. Thin-walled HPDC U-profiles made of the AlSi4Mg and AlSi9Mg alloys are investigated in the present study. The material behaviour is examined using uniaxial tension tests, notched specimen tests providing near plane-strain condition, shear tests, and plate-bending tests. In addition, axial crushing and three-point bending tests of the components are performed. Only quasi-static loading conditions are considered in the material and component test programmes. The experimental data are applied to obtain a validated methodology for finite element modelling of thin-walled aluminium castings, using shell elements, the J2J2 flow theory and the Cockcroft–Latham ductile fracture criterion. Shear fracture (due to shear band localisation) has not been accounted for in the present work.

The mechanical behaviour of thin-walled HPDC AlSi4Mg and AlSi9Mg U-profiles has been investigated by material and component testing. The materials were characterised in T1 condition by means of uniaxial and notched tension tests (providing near-plane strain conditions), shear tests, and plate bending tests. The elastoplastic behaviour of the cast aluminium alloys was modelled using the classical J2J2 flow theory. Fracture was modelled by the ductile fracture criterion proposed by Cockcroft and Latham. The experimental data were applied to obtain a validated methodology for finite element modelling of thin-walled aluminium castings. The fracture parameters for the interior material in the castings were calibrated from uniaxial tensile tests, while for the surface material they were determined by simulation of the plate bending tests. The calibrated constitutive model and fracture criterion were used in finite element analysis of the deformation behaviour of the U-profiles in axial crushing and three-point bending, accounting for a high-integrity ductile surface layer and a central region containing lower ductility. The values of the Cockcroft–Latham fracture parameter were also determined from notched tension tests and shear tests. The main results are summarised as follows. 1. The fracture strains obtained from the uniaxial tensile tests of the AlSi4Mg and AlSi9Mg alloys were approximately 15 and 8%, respectively, when using specimens machined from the web of the U-profiles. 2. Simulation of the plate bending tests predicted Cockcroft–Latham fracture parameters for the surface layer corresponding to fracture strains of 22 and 15% in the web of the U-profiles made of AlSi4Mg and AlSi9Mg alloys, respectively. 3. It was suggested that the fracture elongations in the uniaxial tensile tests are controlled by crack initiation in the interior material. By comparing the results obtained from uniaxial tensile tests with results from plate bending tests, it was found that the surface material is more ductile than the interior material. 4. Numerical simulations of notched tension tests and shear tests showed that the classical J2J2 flow theory, which is based on the von Mises yield criterion, the associated flow rule and isotropic hardening, fits the experimental data very well. 5. The values for the fracture parameter WcWc obtained from simulation of shear tests were significantly higher than values found from uniaxial and notched tension tests. This could indicate that the behaviour of the cast material is less sensitive to material defects in the shear test (less material is strained in the shear test than in the tensile tests), or that the fracture behaviour of cast aluminium alloys is not accurately described by the Cockcroft–Latham fracture criterion for all deformation modes. However, as the volume of material tested in the shear tests is only a small fraction of the volume tested in the uniaxial and notched tension tests the probability of testing a part of the material having a defect of a given size and orientation is much smaller in the shear tests, which should lead to increased ductility. Thus, it is probable that size effects are important with respect to identification of fracture parameters for aluminium die-castings. Other approaches for modelling of fracture might describe the fracture behaviour more accurately and should be investigated. To provide more details about the fracture mechanisms, possible shear band formation could be investigated by optical micrographs of the investigated materials. 6. During axial crushing and three-point bending tests, the AlSi9Mg U-profiles exhibited significantly more brittle behaviour than the AlSi4Mg castings. 7. The numerical predictions of axial crushing and three-point bending tests gave results similar to the experimental measurements. In the simulations of three-point bending tests, the effects of inhomogeneous distributions of tensile ductility were successfully captured. It is especially important to address the inhomogeneous distributions of defects when studying load cases where the deformation is not controlled by buckling. A coupling between die casting process simulations and the pre-processor for the subsequent FE simulations has successfully been developed for Mg–Al alloys (Dørum et al., 2007), but application of this approach to Al–Si alloys is beyond the scope of the present study.