Search

Charge welds are unavoidable defects of the continuous extrusion process whose extension needs to be accurately predicted in order to avoid profile mechanical failures or to minimize unrequired scrap of material. The aim of this work was then to evaluate the accuracy of the charge welds FEM predictions and their applicability in the industrial field. Two different industrial cases involving A6082 and AA6063 aluminium alloys were analysed. The data of charge welds behaviours were experimentally collected, discussed and then innovatively compared to the predicted outcomes of FEM simulations performed using QForm Extrusion® software and of theoretical formulas reported in the literature. As main results, a very good numerical-experimental matching was found, with a peak discrepancy of 122 mm in terms of charge weld extent, while theoretical formulas returned a significant underestimation.

Finite elements analyses are offering to extrusions a big opportunity: to estimate accurately not only the charge weld extension but also the billet skin contamination. The present work examined the dynamics of the back-end and front-end defect in the direct extrusion of a solid profile through the use of FE analyses. Numerical results have been compared with industrial experiments with the aim to understand the effect of the chosen boundary conditions on the accuracy of the predictions.

The reduction of scraps related to back end defects (i.e. billet skin contamination) and front end defects (i.e. charge welds) is gaining nowadays an increasing industrial interest in order to obtain greater process efficiency. Today, extrusion industrial practice faces the issue by means of technician’s experience, empirical rules or, in most critical profiles, through time consuming and expensive experimental analyses. On the other side, FEM simulation of extrusion dies is becoming a common support tool for the design of new critical dies. Stating this scenario, the possibility to include the prediction of front end and back end defects evolution as simulation output can then be easily obtained at almost comparable computational costs. In this paper the FEM code Altair HyperXtrude® is used for the simulation of 2 industrial cases made by AA6063 and AA6082 alloys following the transient moving boundaries approach. Experimentally, the profiles were extruded, sectioned, polished and etched with caustic soda in order to reveal and measure front and back end development in front of and behind the profile stop mark. The data obtained from experimental analyses are initially discussed referring to billet skin contamination and charge weld evolution, then compared to industrial experience, to theoretical and empirical methods available in literature and to FEM results in order to evaluate pros and cons of each evaluation method.

The continuous direct extrusion is an economic and efficient process to manufacture high-quality aluminium profiles. However, because of multiple billets processing, transient defects (coring and charge welds) can be generated affecting portions of the profiles and deteriorating their mechanical properties. In case of structural applications, the reliable prediction of the onset and extend of these portions is therefore mandatory in order to be scrapped by the commercialized length. In this context, the present work has been aimed at examining the dynamic evolution of the coring and charge weld defects in the direct extrusion of an industrial solid profile. Both experimental and numerical activities have been performed and results evaluated in terms of coring and charge welds defect evolution achieving a good agreement for both phenomena. Results have been also compared with the outcomes of the analytical models available in literature. For the charge weld extension, the analytical model returned a poor predictably while a very good match with numerical data was obtained for the coring defect.

Charge welds are, together with the seam welds, defects generated during the continuous extrusion of metallic materials. However, while the seam welds can exhibit the same resistance of the basic material if proper levels of pressure, contact time and temperature are applied, charge welds are always contained parts of the profile with lower mechanical properties. Indeed, at the end of each process stroke, the back end of the old billet material that completely fill the die starts to interact with the front side of the new billet loaded into the press that is usually contaminated by oxides, dust or lubricant thus producing a transition zone that extends to a variable length. It comes then clear that the length of the profile marked by the charge welds need to be scrapped and an accurate prediction and reduction of this portion become mandatory, not only for the final user of the profile, in order to avoid in-service product failures, but also for extruders and die makers to increase process efficiency. In this context, aim of the present work was to compare the experimental -numerical investigations performed by the authors in terms of charge welds extensions with the predictions of the analytical models nowadays reported in literature. The final global aim work was to assess the applicability of these models in the everyday industrial practice.

The extrusion process is an economic maunfacturing method used to produce profiles with a constant section. However, even when simple components are considered, many process and product issues can emerge such as low mechanical properties or distorsions of the exit profile, die premature failure and/or low production rates. These problems are related to the complex nature of the process that has to guarantee, at the same time, proper conditions for a good welding (in case of hollow profiles), absence of profile distorsions, an acceptable die stress and as much as possible fast extrusion. In addition, all these conditions have to be reached in the short extrusion time of each single billet and can emerge to be conflicting one each other’s. Usually, the optimization of the extrusion process is performed by the operators on the basis of consolidated experience and empirical rules. However, in this way, it is extremely difficult to control all the process and geometrical parameters involved and then to control the final outputs. This explains the number of works in literature aimed to develop numerical and analytical models for the extrusion process optimization. In this context, the authors recently presented a novel procedure for the multi-goal optimization of extrusion dies based on meta-models that was preliminary validated on the experimental results available for a simple round tube profile. In this work, two additional test cases are presented with the aim to extended the validation cases of the proposed procedure to a more complex industrial profile.

The increasing attention to magnesium alloys in extruded profiles, especially in the transportation industry, is related to their low density associated with good mechanical properties and complete recyclability. This allows to push towards both increasing efficiency and pollution restrictions. However, these advantages are negatively balanced by the production rates drop in relation to dangerous profile temperatures increasing that force to keep low velocities. In this context, a novel porthole die has been purposely designed for magnesium alloys allowing an increasing of the process velocity up to four times with respect to past solutions. The mandrel consisted of three ports made by 120° bridges that created an equal number of seam welds. The extruded tubes, made in ZM21, were 50 mm in diameter and 2 mm in thickness and were tested under different process conditions. In the present work, the quality of the seam welds has been investigated in relation to each process condition by means of the rubber plug testing method that allowed to applied an hydrostatic tensile state.

Constitutive equations relate thermo-mechanical parameters, i.e. strain (ε), strain rate (έ) and temperature (T), with flow stress (σ). The most popular constitutive relationship, among those used in the study of hot deformation, is the phenomenological sinh Garofalo equation. In recent years several papers described the hot deformation of aluminum alloys by mean of hot torsion testing. However, sinh constitutive parameters are seldomly available in literature even for the commonest aluminum alloys used today by the extrusion industry. This paper presents the result of the torsion tests providing constitutive equations for AA6005A, AA6063 and AA7020 alloys. The relative extrudabilities of these alloys were estimated by means of FEM simulation carried out by the HyperXtrude® software.

A comprehensive investigation on the correlation between die design and process parameters with extrusion welds (seam and charge) prediction and die deformation after multiple-cycles is presented. In particular, the influence of the number of legs (2,3 and 4) as well of some commonly adopted industrial best design practices in the extrusion of a round tube profile are examined by the ALE code HyperXtrude® for the AA6005A alloy. Then, for the same three designs, the die deformation after multiple-cycles is computed by applying a developed user sub-routine that account for the pressure and temperature maps and for the dwell-time that represents the time required to extrude each single billet. A sensitivity analysis on the influence of process parameters (ram speed, billet length and alloy) on the achieved die deformation after a fixed extruded amount is also performed.