It is important that all designers, faced with profiles showing innovative shapes, have the tools needed to design the optimal matrix at the first attempt. The strong interaction among process variables makes it at any rate difficult to formulate correct assumptions when designing; in order to obtain a good result a close communication between designers and technologists is essential
by Maurizio Grillo
One of the strengths of designing using extruded profiles is the freedom in shaping the section: this however clashes against a series of technological requirements which should be known in order to obtain an optimal design.
The variables available for designers, which must be associated with technological limits, are the following:
• Type of alloy
• Circumscribed circle and weight per meter
• Wall thickness
• Length of the profile
• Number and type of hollows
Matrices deserve a particular attention, since they are an essential element of success or a source of high risks, with consequent extra costs and delivery delays. For the designer it is very important to know the impact of design on the matrix: premature breakages must be completely avoided, allowing it to work ton after ton, until the limits of natural wear are reached, when thicknesses grow on average by about 5 to 10% with respect to the nominal values and it is replaced.
The type of alloy
It is well known that the choice of the alloy considerably affects extrusion speed and therefore costs. Table 1, with a value of 100 for relative extrusion capability attributed to alloy 6060, sums up the issue.
Circumscribed circle and weight per metre
The value of the circumscribed circle is a fundamental measurement in extrusion: from the designer’s standpoint it is a clear indication as to the type of press required and therefore the extruder to contact. From the diameter of the circumscribed circle, with an increase of about 10%, the diameter of the required billet may be determined; in order to foresee the pressure required there are three methods, one is rather approximate but simple, the second is more complex and provides visibility as to the effect of some process variables, the third, which should be used in complex and important cases, envisages the use of numeric simulation. The complexity and the cost of the third method make it practically inapplicable in the phases of the preliminary design of the extrusion. The first method simply implies multiplying the area to be extruded by the maximum extrusion pressure (600-800 MPa for 6xxx alloys, higher values for alloys which are harder to extrude).
The second method, useful for the designer who only needs to make a rough forecast, includes the use of empirical formulae which however have the advantage of being simple and immediate. This simplified and approximate approach also provides the advantage of quantifying the weight of some of the fundamental variables linked to the process. By using, for instance, a rather old formula suggested by Professor Carlo Panseri , it is possible to obtain immediately a better understanding of the elements which altogether determine the value of the maximum extrusion pressure and even their weight in the final value. The effect of the material is summarized by the term s, which the author suggests to determine experimentally working backwards starting from the pressure indicated by the extrusion press with the actual process parameters (temperature, extrusion speed, type of alloy). By using numerical simulation the same result may be obtained.
The expression which provides the value of the maximum extrusion pressure according to  is the following:
p= extrusion pressure (MPa)
s= yield strength corresponding to the material compression, temperature and deformation speed present at the extrusion conditions (MPa)
ß = shape coefficient, generally equal to 1.5
A= area of the container (mm2)
a = area of the profile to be extruded (mm2)
µ = friction coefficient with the walls of the container (0,07÷0,12)
L = length of the billet (mm)
D = diameter of the container (mm)
The formula includes within the term s the material’s extrusion properties, the temperature of the process and the extrusion speed; considering this value as a constant pertaining to the material is a considerable simplification, therefore even results must be evaluated carefully. The bibliographical reference  provides, for three frequently used alloys and for a certain set of parameters, the following values for extrusion pressure, p:
Alloy 6063: p = 440 MPa
Alloy 6005A: p = 520 MPa
Alloy 7020: p = 630 MPa
By estimating the area of the profile (4150 mm2) used in  and with the value provided of the billet’s diameter (11”) the extrusion ratio was calculated as E=(2792/4)/4150≈ 15; using in formula  the values shown in  it is possible to calculate backwards the yield strength of the material in those extrusion conditions (for the three alloys shown, the following values are obtained: 37.1 – 43.9 – 53.2 MPa). The last variable which must be defined is the length of the billet: even in this case formula  may be used considering as fixed the extrusion pressure p, the billet’s diameter and the type of alloy. In this way we may obtain the length which may in theory be extruded: percentage allowances for scrap varying from 30 to 50% depending on the different cases must then be applied. Making preliminary forecasts on the maximum length which may be extruded in the design phase prevents surprises, especially in applications which require considerable lengths (railway and naval industries and, for some applications, civil engineering). By means of the estimates previously shown it is therefore possible to establish the maximum value for the extrusion’s weight per metre; in the design phase such considerations, knowing the properties of the presses, allow to establish, by means of successive attempts, the optimal size of the profiles.
Limits in profile length
This is a value which, when designing, is defined in terms of finished product: it is normally up to the production managers to provide the values of metal allowances required by intermediate machining. Values which are not negligible are requested, for instance, by stretch bending in order to allow a grip on the ends of the profile which needs to be bent; in the case of large extrusions used in the railway industry (lengths of up to 26 metres) used as straight rods, metal allowances are definitely lower (about 10 cm), just to make up for cutting and positioning errors on the mounting stops. The technological limit is provided by the length of the stretching counters: at the end of the extrusion process, the profile, having been cooled, is grabbed at both ends and subjected to a traction force which is slightly higher than the material’s yield strength. This operation, which of course involves discarding the ends, allows to obtain profiles which are sufficiently straight as well as having mechanical properties which are more uniform across the profile with respect to the longitudinal properties. Regarding the length limit it should be noted that in many cases this is quite irrelevant since, for transportation convenience reasons, profiles are supplied using bar lengths of 6 or 12 metres which fully respond to the requirements of road transport. In some cases, however, such as the railway sector, since very long extrusions are needed, transportation occurs by sea or rail, if need be using even exceptional road transport. Profile length, besides requiring stretching counters of adequate length, also causes limitations in terms of the profile’s maximum weight per metre. Press containers normally envisage limited lengths for the billets (1-1.5 metres approximately). These, which are much longer to begin with, are cut to the desired length and inserted into the press. To foresee the maximum length which will be extruded, the volume of the billet must be calculated and divided by the area of the profile, making allowances for scrap at the beginning and end. For instance, with reference to a direct extrusion 5000-ton press, the weight of the largest billet (420 mm in diameter by 1180 mm in length) is equal to 440 kg; if we wish to extrude a profile weighing 30 kg per metre, the length we can theoretically obtain is equal to about 14 metres; considering scrap, for an open shape the maximum useful length will be about 11 metres, for a hollow shape it may reach 9 metres. In general the percentage of scrap with respect to the initial volume of the billet is 30 to 50% depending on the case being examined. This simple example shows how the design must evaluate from the very first phases the technological domain to avoid unpleasant surprises in the end.
Profiles with open, semi-hollow and hollow shapes
As we know, extruded profiles are classified as open, semi-hollow and hollow profiles based upon their shape. Hollow profiles are manufactured using bridge matrices: open profiles may be made using simpler matrices, practically a slab with an opening whose external contour corresponds to the profile which must be produced. The cost increases with size and varies more or less from 1,500 to 12,000 euro (presses from 1,600 to 6,000 tons).
For hollow extrusions, the shape no longer has a single connection: in order to obtain the entire shape of the profile it is necessary to add to the external line as many closed internal lines as the hollows required. In this case the matrix becomes considerably more complex, since in order to obtain the internal hollows, devices are required having the purpose of separating the flow of metal upstream of the hollow, reconnecting the flow just afterwards, thereby obtaining the desired shape. In this case the cost varies from 2,200 to 27,000 euro depending on size.
It is important for semi-hollow extrusions to have specific proportions: the part of the matrix which allows the hollow to form is none other than a bracket subject to pressure by the metal. The bending generated on the minimum resistant section increases along with the hollow’s section and decreases along with its length. Figure 2 shows the highest allowed values.
Hollow extrusions are characterized by the presence of so-called welding seams, present along the whole length of the profile. The supports of the needles cause the separation, first, and then the reunion of the flow of metal downstream. The convergence of the flows causes the occurrence of hot-heading which reinstates the metal continuity of the profile. This occurs if the process variables (temperature, extrusion speed…) have been properly gauged. However, the presence of lubricants, gas trapped in the container or other instances may contribute towards creating a poor welding quality, difficult to evaluate in objective terms. It is a defect which may show up during monitoring downstream of the extrusion, but also, blatantly, after the delivery of the profiles. The most critical extrusions in this respect are those mainly subjected during use to torsional stress. Figure 3 shows two welding seams highlighted by a micro-graphic attack. The huge variety of shapes does not allow adopting standard control methods; the only method which may definitely be applied to all types of extrusions is macro/micro-graphic observation, to verify with high magnification the metal continuity, at least in the section being examined. It would be desirable to be able to carry out a more objective evaluation, such as an evaluation of plastic strength of the seam by means of stretching, bending or deep drawing tests; unfortunately the variety of shapes practically prevents the execution of such tests in many cases.
The total freedom of form of the section
One of the most popular themes used to describe the properties of the extrusion process is the totally free form of the sections. In actual fact this freedom is conditioned, given the technological limitations. A first and evident limit is geometrical, given by the billet diameter: in the case of open sections, the ring margin with respect to the container is about 10% of the radius; it is slightly larger for hollow sections. Only presses with a widening device, at least in one dimension, may overcome this limit: this is a sort of funnel which, starting from the container’s circular section, evolves by widening in one direction.
Such a device allows to obtain shapes which are highly developed in one dimension and poorly developed in the orthogonal dimension: such is the case, for instance, with truck sides and tailgates (Figures 4 and 5). A second technological constraint is the so-called extrusion ratio, E, between the section of the billet and the section to be extruded. It is evidently not possible to have E<1, nor is it convenient to bring its value down to a few units. On the other hand, very high levels of E imply considerable extrusion forces if not the impossibility of extruding. In the case of hard alloys (2xxx, 5xxx, 7xxx and 6xxx alloys with a high silicon and magnesium content), E varies approximately between 10 and 35. In the case of soft alloys (1xxx, 3xxx and 6xxx with low magnesium and silicon content) between 10 and 100.
Another technological limitation which should be considered is the maximum tensile force which may be applied by the stretcher. Following extrusion the profile, having been cooled, is griped at the ends: the load applied is calculated so as to produce a residual lengthening of 1 to 2%. Especially in the case of hard alloys, even if they are in the physical state of raw extrusion material, the yield strength takes on values which are not negligible (135 MPa for a 2014 alloy and 165 MPa for a 7075 alloy) which may create a limit for the maximum size of the profile.
A value which is very interesting for designers is the minimum thickness which may be obtained using extrusion: this is a very precise boundary for structural optimization which modern calculation methods easily allow. By observing profiles present on the market, it is immediately possible to note the link between the typical size of the section (circumscribed circle) and the minimum thickness which may be extruded: at the lower end we may find small special extruded profiles for heat exchangers (with rectangular section, nine chambers, external dimensions 30 by 1.8 mm, minimum wall thickness 0.3 mm) while the higher end is occupied by profiles extruded by large 20- 24,000 ton presses, with circumscribed circles of 900 mm. Profiles easier to find (circumscribed circle 400-600 mm) have minimum thicknesses of at least 7-8 mm in the case of tubes without internal septa; for more complex sections with many internal septa (such as, large extrusions for railway use) thickness may be as low as 2.7-3 mm or even less. In all the cases mentioned alloys used are 6xxx with good extrusion properties (6063-6060-6005A). Moving on to so-called hard alloys (2xxx e 7xxx), or even 6xxx alloys with higher Mg and Si contents (for instance, 6082), even with the same geometry the minimum thickness required is greater, by about 15 to 30%. Designers will need to accurately consider the effect of the alloy and the consequent increase in the thickness required for technological reasons, verifying its effect on dimensions. An exam of data available in literature shows a certain dispersion of the declared values of the minimum thicknesses which may be obtained; as may easily be imagined, these are higher than the minimum values which are actually obtained. The reason lies with the technological complexity and the risk of declaring beforehand a certain thickness which may be obtained by considering only the alloy and circumscribed circle: it is much more appropriate to evaluate case by case with reference to previous experiences and to the shape requested. In any case the following points should be highlighted:
A linear relationship between the value of the circumscribed circle and the minimum thickness for a given alloy
An alloy effect (from the easier to extrude to the hardest ones, to be evaluated on a case-by-case basis)
Going from open extrusions to hollow profiles implies an increase in minimum thickness with a percentage value which grows with the increase of the value of the circumscribed circle. On small extrusions (50-70 mm) the increase is about 15%; on medium-sized extrusions (about 200 mm) it may reach 100%.
A hot plastic deformation is all that is required to obtain extruded profiles, even with very complex shapes; the simplicity of the process does not of course correspond to an extremely precise result in terms of size allowances, such as those commonly detained by machining involving chip removal. Designers should be aware of this inevitable fact, since they are entrusted with finding the right solutions so as to consider the technological limitations. The basic document where all information may be obtained is norm EN 755-9 which contains a rather exhaustive list of geometric allowances which normally occur; it should be consulted for further investigation. Especially in architecture, the EN 12020 norm is used which applies to alloys which may be extruded easily (6060 and 6063), with a diameter of the circumscribed circle smaller than 350 mm; size allowances are more restrictive and more accurate surfaces are expected to facilitate the successive anodic oxidation.
To have a first rough idea of allowances, the following rules of thumb may be useful:
• Linear dimension: ± 1% of the nominal value
• Wall thickness: ± 5÷10% of the nominal value;
greater values on inferior thicknesses;
• Tolerance on corners: ± 2°
The issue which often arises, starting with values taken from norms, is the excessive amplitude of values of the various allowances; in such cases there is nothing else to do except discuss the matter with the extruder. The norm has to have general contents and therefore includes a few margins, therefore, faced with significant design requirements and possibly considering the larger amount of scraps as regards costs, a fair mediation might be found. Another way out is to use wits: in the railway industry, for instance, there is the problem of obtaining the width of a carbody with just a few mm of allowance. But how can this be achieved considering that the under frame, for instance, is formed by at least 5 extrusions, 400 mm long? With 1% of incremental allowance of the five extrusions the allowance on the total would be exceeded; the solution lies with lap instead of butt joints, with reduction of the tensile and fatigue strenght.
Optimized design As mentioned before, absolute freedom of shape practically does not exist, however actually extruded shapes show how much this process allows to do. The best results may be obtained following some simple basic rules: Symmetrical shapes, better still with double symmetry, allow a balanced flow of metal without having to adopt empirical expedients on the matrices (to slow down the flow of the metal if its exit speed is not uniform). For the same reason it would be better for the thicknesses not to differ too widely: high thicknesses allow an easier flow of the metal, which avoids narrow passages (high friction). If thicknesses are equal, or at any rate uniform in some lengths, also helps welding: usually welding parameters are set up beforehand and remain constant (apart from starting and shutdown of the arc) during welding. Extrusion does not allow the creation of sharp corners or pointed shapes; it is sufficient to foresee a tip radius of 0.5-1 mm to improve the situation considerably while prolonging the working life of the matrix. If design requirements demand changes in thickness, envisage appropriate transitions. The internal cores of a hollow profile must have thicknesses comparable to those of the external surfaces to encourage filling; it is important to remember that the mechanical properties of the internal cores are always lower than those of external surfaces: these receive more efficient hardening. Deep hollows must have proportions such as to respect the criteria shown in figure 1. If deeper hollows are required the expedient of a flap engraved at the ends which may be removed mechanically after extrusion may be used.
If slots for screws are necessary in a direction parallel to extrusion, closed slots should be avoided preferring open slots with an angle of 60°. There are three types of hollows: for feedthrough screws (with a smooth inside face), for self-tapping screws (with a grooved inside face) and for screws requiring tapping (of adequate thickness).