Printer Friendly

The role of plug design in determining wall thickness distribution in thermoforming.


Thermoforming is one of the oldest industrial polymer processing techniques. At its simplest it uses a combination of heat and pressure to create final products from preformed polymer shapes. Most commonly the feedstock is in the form of an extruded sheet, which is heated until it becomes soft and pliable, and then it is deformed by air pressure into the shape of a mold. Typically thermoforming creates products of varying wall thickness distribution because of the different levels of deformation that the parts of the sheet must undergo to reach the furthest extremities of the mold. As products become deeper this increasingly leads to excessive thinning in the furthest corners and it then becomes necessary to alter the deformation process so that more material is drawn downwards. Techniques that may be employed include addition of an intermediate draping stage to the process, nonuniform sheet heating, and the use of a mechanical plug to prestretch the sheet prior to the application of pressure. This latter process is known as plug-assisted thermoforming and it is the most common industrial thermoforming process [1]. Whilst it is desirable in most thermoforming processes to achieve a near even wall thickness distribution, this is almost impossible to achieve in practice, and the final result is usually a compromise.


The plug-assisted thermoforming process is illustrated in Fig. 1, which shows the typical arrangement for the manufacture of thin-gauge packaging. Initially the plastic sheet is heated then transported to the mold where it is clamped. A preheated mechanical plug then moves down into the sheet (the Plugging Phase) and prestretches it up to 90% of its final deformation into the mold. Finally in the Pressure Phase air pressure (typically 5-9 bar) is applied from above, stripping the plastic from the plug, and into the wall of the mold where it cools rapidly to create the final part.

For packaging applications the plugs are most commonly made from polymer materials and are unheated. Instead, they pick up heat gradually from contact with the hot sheet during an initial start-up period. From previous work it has been estimated that the equilibrium temperature of a working plug is around 100[degrees]C for a sheet temperature of 150[degrees]C [2]. Common materials for the plugs are syntactic foam, acetal (or POM), and lacquered felt. In each case they are selected primarily for their low heat transfer and low surface friction properties, and different materials are preferred for different sheet materials.

A large number of variables influence the way that the sheet material is distributed locally during the process [2, 3]. Of greatest importance is the forming temperature, which must be optimized for the particular polymer [4]. Other major operating variables are the speed and displacement of the plug, the air pressure and the timings of the different process stages [5, 6]. The research literature concerning the practical operation of plug-assisted thermoforming process is very limited, but clearly demonstrates the importance of the initial mechanical stretching of the sheet [7]. The factors associated with the design and operation of the plug have been found to have the greatest influence on the quality of final products [3, 8, 9]. It has also been shown that the thermal and frictional properties of the plug materials [3], and the temperature and speed of the plug are very important [10].


For industry one of the most challenging decisions is the design of a plug to suit a particular mold shape. Unlike other variables, this is difficult, expensive and time consuming to alter when in production, so there are substantial economic benefits to be gained by getting it right first time. Normally plugs are designed to be smaller rounded versions of the mold shape and where possible previous designs are adapted. However, with many new products, or where there are significant changes in materials or operating conditions, entirely new plug designs must be developed for optimum performance. In these circumstances, and despite the economic drawbacks, it is still very common for companies to employ trial error methods whilst in production to finalize a new plug design. Initially past experience is used to create a first design that is progressively modified by experimentation until an acceptable product is obtained. At the same time suitable operating variables may be developed for the process, and to some extent these can compensate for inadequacies in the chosen design.

Analysis of the Role of the Plug

In industry the effects of plugs on thermoforming processes are well understood in practical terms but most of this knowledge has not been translated into scientific understanding. Thermoforming is still widely regarded as a black art and the research literature on many aspects of the process is very limited. This partially reflects the difficulty in carrying out research investigations using industrial thermoforming equipment. However, it is also the case that much of the accrued processing expertise that companies possess is not held in a quantifiable form and is regarded as commercially sensitive. The greatest weaknesses are in the fundamental understanding of the physical mechanisms that underlie the process. Some of the most important of these are illustrated in Fig. 2, which shows the position of the plug part way through the forming process. At this point the center of the clamped sheet is firmly pressed into the surface of the plug, whereas the material to either side is subjected to rapid extension with the downward motion of the plug. The division between these two regions is commonly referred to as the "plug mark" as it generally leaves a visible discontinuity in the wall thickness distribution. For effective thermoforming the material cannot simply adhere to the plug, as this would otherwise lead to an excessively thick base in the product. Instead low friction plug materials are deliberately selected to permit the sheet in contact to gradually slide back up and return material to the sidewall. However, the mechanism of sliding is not free as deformation can only occur through extension of the sheet across the contacting surface. Therefore the effect of friction is to reduce the magnitude of extension for the sheet in contact.

The modes and magnitudes of deformation are complex across a typical product and have been well documented in previous research [11]. Local strains and strain rates vary considerably across the deformed sheet and typically the strain is largely unidirectional in the sidewall, but equal biaxial in the base. The situation is further complicated because the initial sheet temperature is reduced by local conductive and convective cooling effects. These are illustrated in Fig. 2. Losses due to conduction are expected to dominate where the sheet is clamped to the cool mold and where the sheet contacts the warm, but lower the temperature plug. In the other areas, which make no contact with the plug, the rapid increase in the surface area of the sheet coupled with the rapid decrease in thickness is expected to lead to significant losses due to convection. Previous work using thermal imaging techniques has reported substantial falls in sheet temperature during the plugging stage of the thermoforming process [2].

There are therefore three main effects that combine to determine the local deformation of the sheet during thermoforming. These are as follows:


1. The stress/strain response of the sheet polymer when subjected to appropriate levels of biaxial tensile strain, strain rate, and temperature.

2. The frictional properties of the plug and sheet materials when in contact. For the areas in contact this will reduce the local deformation, but away from the plug its effect will be to introduce additional material as slip takes place.

3. The thermal properties of the materials that lead to heat transfer and reductions in the local temperature. These in turn will cause the sheet to stiffen because of the changes in its stress/strain response [11] and will lower its coefficients of friction [2].

These effects are interdependent, conflicting, and extremely difficult to isolate in practice.

Process Simulation

Efforts to develop thermoforming simulations have been greatly hampered by the lack of detailed scientific understanding of the process and for this reason modeling has made much less impact in thermoforming than with rival processes like injection molding. This is despite considerable reported research [12-15] and the development of a small number of commercial thermoforming simulation packages [16, 17]. From the literature it is clear that very different approaches have been taken in the representation of process parameters in simulations. In many cases the sheet materials are modeled using data obtained in conventional low strain rate uniaxial tensile tests and a very wide range of equations have been used to represent the stress/strain response of the various polymers [12-14, 17-19]. Contact friction behavior is most commonly represented as a single value in a simple Coulomb friction relationship and in some models heat transfer effects are entirely ignored [17, 20-22]. These weaknesses have been increasingly recognized by researchers who have attempted to measure behavior under more realistic conditions. This has included the development of an impressive range of innovative biaxial testing machines, which are specifically for characterization of materials used in thermoforming and blow molding [11, 23, 24]. Attempts to measure coefficients of friction for soft polymers have been reported using conventional test apparatus such as a torsional rheometer [21] and a moving sled device [2]. These methods have given comparative measurements at low temperatures, but it becomes increasingly unreliable at higher temperatures. It is accepted that none of these methods truly replicate the extensional slip mechanism of the actual process. Measurement of heat transfer in rapidly moving thin polymer sheets is also a difficult task, although use of thermal video imaging has yielded interesting results [14].


The overall objective of this work was to investigate in practice the influence that a number of key plug design features have on the final wall thickness distribution of products manufactured by plug-assisted thermoforming. Such effects are already broadly understood in industry but they have not been formally quantified. In addition, it was planned to use the results of the study to look for evidence as to how heat transfer and frictional effects influence the local distribution of the material in both the final product and at the intermediate plugging stage. It was hoped that the effects observed would enhance the understanding of these in-process phenomena and thereby assist in the further development of process simulations. However, it was not planned to develop simulations as part of this work and no simulation software was available with the capacity to fully simulate the products and materials used in this research.



The main apparatus used in the investigation was a purpose-built small laboratory thermoforming machine [25]. This was constructed specifically at Queen's University to allow the thermoforming of materials on a single cavity mold, which was designed to match the typical tooling found in large industrial thermoforming machines. However, unlike its industrial counterparts the machine is fully instrumented with temperature, pressure, and force transducers, and there is fine control of all of the major processing variables. The operation of the thermoforming test apparatus is illustrated in Fig. 3. Sheet samples were clamped in a movable frame and then heated in the convection oven (right). The samples were rotated slowly during heating to ensure an even temperature distribution, and the sheet temperature was monitored remotely using a noncontact IR sensor, located above a window at the top of the oven. At the same time the plug was heated in a separate oven to the required test temperature and inserted into the machine immediately before the forming operation. When the appropriate test temperature was reached the samples were moved out of the oven and then clamped between the upper and lower halves of the mold cavity. In this arrangement the upper cavity, containing the plug, is fixed, and the lower cavity, containing the mold, may be driven upwards to clamp the sheet and seal the cavity. Once the sheet was clamped the plug was driven downwards to a fixed displacement and then positive air pressure was applied in the upper cavity to blow the sheet outwards into the extremities of the mold. Finally, after a short pause to ensure final cooling, the completed molding was ejected and its wall thickness distribution was measured.

Experimental Parameters

The major dimensions of the part mold insert used in this investigation are shown in Fig. 4. It is basically a standard conical pot, typical of the design used in the packaging of dairy produce, such as yoghurts. The mold was manufactured in industry and was machined from an aluminum alloy.

The effect of plug geometry on part wall thickness distribution was investigated by gradually changing three main geometrical features. These are as follow:

1. Sidewall taper: The angle that the sidewall makes with the vertical.

2. Base radius: The radius on the corner between the base and the sidewall.

3. Plug diameter: The overall diameter of the plug, measured along its top surface.

Initially three large, flat-bottom plugs, with dimensions as illustrated in Fig. 5 were produced. Each of these was then progressively altered by machining to create the changes in geometry described earlier. The detailed geometry of all of the plugs created is given in Table 1. Two further shapes were created for each of the variables investigated. The initial sidewall taper was increased from 4 to 8 and then finally 12 degrees. The base radius was increased from 6, to 20, and then 34 mm. Finally; the outside diameter of the plug was reduced from 75, to 67.5, and then 60 mm. The effect of these changes on the overall plug shape is illustrated in Figs. 6-8. In all cases the variables led to changes in the diameter of the flat base of the plugs (quantified in Table 1). The effects of these changes on the process are illustrated in Fig. 9, which shows the position of plugs 1, 2(b), 3(b), and 4(b) within the mould at the end of the plugging step. For each of the changes in plug shape the major effects on the sheet are to firstly change the proportion of material contacting the plug and secondly to leave the material further away from the corners of the mold.


Two sets of tests were carried out using the different plug designs. Firstly all of the plugs were used to create formed parts in the standard mold (Fig. 4) and the average wall thickness of the resulting parts was recorded. In all cases at least 5 parts were manufactured at each setting and the wall thickness was measured at 10 mm intervals from base to lip along 4 lines separated by 90[degrees] around the periphery of the part. The measurements were then averaged to create the final results.

In addition to the full part thermoforming tests a smaller number of tests were carried out specifically to assess the effect of the plug stage of the process alone on the material distribution. In these tests the cavity test mold was removed and replaced by an open ring, which was supported on three legs. This enabled the heated sheet to be stretched into free space by the action of the plug, and then by cooling in a jet of air, the material distribution following plugging alone could be frozen and recorded using the same procedures described for the parts. For these test only the final plug shapes in each category [2(b), 3(b) & 4(b)] were employed.


The polymer material used throughout the tests was an isotactic homopolymer thermoforming grade of polypropylene (Novolen 1184L), which was supplied as a 1.45-mm thick extruded sheet. The plug material was a common grade of syntactic foam (Emerson & Cuming SYN-TAC 350). The sheet forming temperature was 160[degrees]C, the plug speed was set to an average speed of 177 mm/s, plug displacement was 80 or 75 mm below the initial sheet surface, and the forming air pressure was 8 bar. For all tests the plug temperature was maintained at 100[degrees]C and plug surfaces were left in the unpolished as-machined condition.


The results of the investigation are presented as wall thickness distributions for pots produced by the various plug shapes. These are shown in Figs. 10-12 and on each graph the three lines plotted show the wall thickness distributions produced by the initial plug shape, compared with two subsequent changes in the geometrical feature. For each data point error bars indicate the standard deviation in the data recorded. In all cases the wall thickness is plotted from the center of the base of the final pot (left) to its top lip (right). Results from the subsequent plugging tests are shown in Figs. 13-15. These were carried out only on the final plug shapes in each case [denoted as 2(b), 3(b), and 4(b)], and again the distribution resulting from these plug shapes are plotted alongside those resulting from the initial plug shape. A similar convention is used for the x-axis, with distance measured from the base center of the part-formed "dome" shapes.



It is clear from the overall results that the shape of the plug has a very profound effect on the wall thickness distribution in the final product. Most significantly the various shape factors tend to shift the balance of material from the base to the lip and vice versa. By comparing the changes in plug shape, shown in Figs. 6-8, with the associated changes in wall thickness distribution in the pot, shown in Figs. 10-12, the major effects may be observed. In each case the solid line on each graph is the original plug shape (plug 1). As this plug has a very large flat area at its base, it draws much of the sheet material downwards, creating an overly thick base (around 0.6 mm), and its thinnest parts are in the corners of the pot and in the upper sidewall (around 0.3 mm). Overall this is a very poor distribution of the material and a processor would immediately be looking for ways to redistribute some of the excessive thickness in the base towards the thinner areas of the product. The wall thickness distribution also shows a very distinct plug mark discontinuity around 50 mm from the center of the base. To the left of this it is clear that the sheet has been in contact with the base of the plug and has retained much of its initial thickness during the plug step. However, some of this has then been redistributed into the corner of the pot by the subsequent blowing step. The overall effect is to leave some residual thickness in the lower sidewall, which may be clearly seen in a final product. To the right of the 50 mm point the sheet never contacts the plug surface and during the plug step this region has to endure the majority of the deformation, leaving it overly thin. However, as the plug is very large, this material is quite close to the sidewall of the mold (as shown in Fig. 9) and therefore has little opportunity to deform further during the subsequent blowing step.



The effect of increasing the angle of taper in the sidewall of the plug may be seen in Fig. 10. An increase from 4[degrees] to 8[degrees] almost eliminates the plug mark and flattens the distribution across the middle of the sidewall. However, this change has almost no effect on the distribution in the base, suggesting that the flat area at the base of the plug is still large enough to dominate this behavior. In addition there is a small increase in the wall thickness in the upper part of the sidewall. A further increase in taper angle to 12[degrees] leads to more dramatic change in the distribution. This time the base is considerably thinner than before and the upper sidewall is considerably thicker. In addition the corners of the product are extremely thin (around 0.2 mm) and there is a steady increase in thickness across the sidewall. This result indicates that the flat area on the base of the plug no longer draws sufficient material towards the bottom to sustain the thickness in the corners, which is now further away (Fig. 9). As a result of this the upper sidewall is now too thick.


The effect of increasing the base radius of the plug may be seen in Fig. 11. An increase from 6 to 20 mm slightly reduces the material thickness in the base and pushes more material towards the center of the sidewall. The plug mark remains very distinct and there is a slight reduction in the corner thickness, but there is no effect on the upper sidewall. A similar pattern is observed when the radius is increased again to 34 mm, although this geometry produces the thinnest corners. There is a further reduction in the base thickness and a shift of material towards the center of the sidewall. These results suggest that greater rounding of the plug permits more material to move towards the center of the sidewall, and thereby reduce the base thickness. However, the greater the radius then the further the material has to deform to reach the mold corners from the plug's surface (Fig. 9), and this tends to make the corners overly thin.


The effect of decreasing the top diameter of the plug is shown in Fig. 12. A decrease from 75 to 67.5 mm immediately reduces the thickness throughout the lower portion of the product and this is only balanced by a sharp increase in the thickness in the upper sidewall and lip. The basic pattern of the distribution remains largely the same, although the plug mark becomes much less distinct. It is clear that like the changes in sidewall taper (Fig. 10), the large flat area in the base of the plug is ensuring that much of the initial sheet material is drawn downwards. A further reduction in diameter to 60 mm further emphasizes these changes, with a slight reduction in the base thickness, further flattening of the sidewall, eliminating the plug mark, and greater thickening towards the lip.


The dominant effect of all of the changes in plug geometry is to shift material from the base of the product towards the sidewalls, and particularly the top lip. A small change in each of the three factors studied appeared to generally improve the uniformity of the distribution. However, in all cases excessive change tended to cause the corners of the product to become excessively thin and thereby negate the benefits achieved in other areas. In addition, increasing reduction in the overall size of the plug (through changes in either sidewall taper or top diameter) tended to lead to excessive thickness in the upper sidewall and lip.



Broadly similar patterns of wall thickness distribution are observed at the intermediate plugging stage, which is shown in Figs. 13-15. For plugs 2(b) (Sidewall taper) and 4(b) (Top diameter), shown in Figs. 13 and 15, respectively, the wall thickness distribution after both stages shows thicker areas in the base and upper sidewall and the thinnest areas are in the lower sidewall. All regions are thinner in the final product demonstrating that they must undergo further deformation after plugging. This varies in magnitude across each distribution, with the areas contacting the plug showing the greatest changes. The remaining areas that do not contact the plug change the least, suggesting that these highly stretched areas are less able to deform further. This is interesting as contact with the plug was expected to significantly cool the sheet and thus reduce its capacity for further deformation. Instead it appears that heat transfer and sheet stiffening are more related to the change in sheet thickness during the plug step rather than contact with the plug. It is also worth noting that after plugging the material in contact with the base of the plug is around 0.8 mm thick for 2(b) and 0.9 mm thick for 4(b). If these are compared with the initial sheet thickness of 1.45 mm, then it is clear that there has been substantial deformation in each sheet during the plug step. As the materials in these flat base regions have been in contact with the plug for almost all of its movement, this can only be attributed to a combination of slip and extension in the sheet across the contacting surface. This finding clearly demonstrates the important role that surface friction plays in determining the final wall thickness distribution.



For plug 3(b) (Base radius), shown in Fig. 14, it is clear that the highly rounded shape of the plug has had a very profound effect on the distribution of material at the intermediate stage, and the wall thickness distribution is radically different from that observed with the flat-bottomed plugs. In this case the thickness gradually reduces from base to lip, and in the upper sidewall the thickness is almost identical to that in the final product, suggesting that there is little additional deformation during the blowing step. In the base area the opposite is true, as the relatively large intermediate thickness of over 1 mm is reduced to less than 0.5 mm during the blowing stage. This again confirms that the material in contact with the plug is able to undergo substantial further deformation, whereas the much thinner regions in the sidewall are not. These results further support the hypothesis that formability in the second stage is directly related to thickness and the most likely explanation for this is that thicker regions are better able to retain their temperature, and thus their subsequent formability. It is also thought that the more rounded geometry of the plug makes it much easier for the sheet to strip from its surface during blowing, and this helps to concentrate the subsequent deformation in the thicker base material.

The results of these experiments have clearly demonstrated that the geometry of the plug plays a major role in determining the final wall thickness distribution in thermoformed products. The tests have also proven that slip during plug contact, and thus the frictional properties of the plug materials, are crucial in defining the output. Heat transfer in comparison appears to have minimal effects during plug contact, but greater effects in the highly stretched regions away from the plug. This suggests that simulations must consider convection as well as conduction for the accurate modeling of heat transfer effects in the process. Further work will be necessary to confirm these effects for other plug variables such as the plug material, surface finish, and temperature.

Comparison With Simulations

The results of running a thermoforming simulation with the approximate plug and mold dimensions used in this investigation are illustrated in Fig. 16. The graph shows the predicted wall thickness distributions for plug design variations 1, 2(b), 3(b) and 4(b). In this case the simulation used was a prototype finite element based model under development at Queen's University [26]. However, a working deformation model was not available for the polypropylene material used in this investigation and instead the results for the thermoforming of high impact polystyrene are presented. The product geometry was modeled in two dimensions using 70 membrane elements to represent the sheet. Isothermal conditions were assumed and a coefficient of friction of 0.5 was set for plug sheet contact, as this had been shown previously to give the best fit to experimental results [26]. No slip was permitted between the sheet and the mold. Other parameters used were an average plug speed of 190 mm/s and a forming air pressure of 7 bar. The results in Fig. 16 show that the largest plug (plug 1) gives the greatest thickness in the base and the lowest thickness towards the lip. Changes in any of the plug geometry parameters [plugs 2-4(b)] have broadly similar effects, shifting material from the base towards the lip. Plug 2(b) creates the thinnest base, which agrees with the experimental findings, but overall the simulation is unable to closely match the experimental wall thicknesses. In particular the minimum predicted wall thicknesses are much less than the experiments in the corner and lower sidewall, and much greater towards the lip. There is also little evidence of the plug mark feature in any of the predicted distributions. These results prove that it is relatively easy for a modeler to broadly simulate the effects of plug design changes in thermoforming with simple representations of materials and plug contact conditions. However, such simulations lack the accuracy required for industrial exploitation, which can only come from effective modeling of the effects observed in this experimental investigation of the process.



The design of the plug plays a critical role in determining the wall thickness distribution in plug-assisted thermoforming. The larger the plug relative to the mold the more likely it is for material to be distributed towards the base of the product. This effect is promoted by having a flat area on the base of the plug, but it is reduced by either increasing the base radius or the sidewall taper of the plug. For the three geometrical features investigated, small changes from the initial plug shape tended to promote more balanced wall thickness distributions, whereas large changes tended to make the base and corners of the product overly thin. From the plugging tests it is clear that contact with the sheet does not significantly reduce the ability of the material to deform further, which indicates that conductive heat losses into the plug are minimal. Thinner regions are more resistant to further deformation and this suggests that cooling by convection is a significant factor in the highly deformed areas that do not contact the plug. The initial thickness of the sheet is substantially reduced during plug contact proving that the sheet is able to deform across the contact surface by sliding. The frictional properties of the materials are therefore very important for the operation of the process. Modeling can be used to provide a broad representation of the product wall thickness distribution but its accuracy is strongly dependent on its ability to simulate the contact effects observed in this investigation.


The authors thank Wilsanco Plastics for their support of this work.


1. J. Florian, Plast. Des. Forum, 19, 55 (1994).

2. P. Collins, E. Harkin-Jones, and P.J. Martin, Int. Polym. Proc., 17, 361 (2002).

3. P.J. Martin, P. Collins, G. Harron, and E.M.A. Harkin-Jones, in the Proceedings of Polymer Processing Engineering 01, University of Bradford, UK, 108 (2001).

4. N.J. Macauley, E.M.A. Harkin-Jones, and W.R. Murphy, Polym. Eng. Sci., 38, 516 (1998).

5. A. Aroujalian, M.O. Ngadi, and J.-P. Emond, Polym. Eng. Sci., 37, 178 (1997).

6. Z. Ayhan and Q.H. Zhang, Polym. Eng. Sci., 40, 1 (2000).

7. J.L. Throne, Polym.-Plast. Tech. Eng., 30, 685 (1991).

8. G.W. Harron, E.M.A. Harkin-Jones, and P.J. Martin, SPE ANTEC Tech. Papers, 1, 5 (2001).

9. S.J. Liu, Int. Polym. Proc., 14, 98 (1999).

10. A. Aroujalian, M.O. Ngadi, and J.-P. Emond, Adv. Polym. Tech., 16, 129 (1997).

11. P.J. Martin, C.W. Tan, K.Y. Tshai, R. McCool, G. Menary, C.G. Armstrong, and E.M.A. Harkin-Jones, Plast. Rub. Comp., 34, 276 (2005).

12. M.K. Warby, J.R. Whiteman, W.G. Jiang, P. Warwick, and T. Wright, Maths. Coms. Sim., 61, 209 (2003).

13. W.N. Song, K. Kouba, F.A. Mirza, and J. Vlachopoulos, SPE ANTEC Tech. Papers, 37, 1025 (1991).

14. P. Collins, J.F. Lappin, E.M.A. Harkin-Jones, and P.J. Martin, Plast. Rub. Comp., 29, 349 (2000).

15. J.W. Lee, G.J. Nam, and K.H. Ahn, Polym. Eng. Sci., 40, 2232 (2000).

16. J.K. Lee, C.E. Scott, and T.L. Virkler, Polym. Eng. Sci., 41, 240 (2001).

17. D. Laroche and F. Erchiqui, J. Reinf. Plast. Comp., 19, 230 (2000).

18. S.R. Hummel and H.F. Nied, Exp. Mech., 44, 381 (2004).

19. F. Erchiqui, A. Gakwaya, and M. Rachik, Polym. Eng. Sci., 45, 125 (2004).

20. A. Tulsian, J. Mead, S. Orroth, and N. Tessier, SPE ANTEC Tech. Papers, 1, 904 (2004).

21. B. Hegemann, P. Eyerer, N. Tessier, T. Bush, and K. Kouba, SPE ANTEC Tech. Papers, 1, 791 (2003).

22. C.H. Wang and H.F. Nied, ASME Mat. Div. Pub., 79, 67 (1997).

23. L. Capt, S. Rettenberger, H. Muenstedt, and M.R. Kamal, Polym. Eng. Sci., 43, 1428 (2003).

24. A.M. Adams, C.P. Buckley, and D.P. Jones, Polymer, 41, 771 (2000).

25. G.W. Harron, E.M.A. Harkin-Jones, and P.J. Martin, Proc. IMechE, Part E, 217, 181 (2003).

26. R. McCool, P.J. Martin, and E. Harkin-Jones, Plast. Rubber Compos., 35, 340 (2006).

P.J. Martin, P. Duncan

School of Mechanical and Aerospace Engineering, Queen's University Belfast, UK

Correspondence to: P.J. Martin; e-mail:
TABLE 1. Plug geometry variables.

 Plug geometry
 Sidewall Base Top Base
Variable Plug taper radius diameter diameter
investigated code (degrees) (mm) (mm) (mm)

Initial shape 1 4 6 75 52.6
Sidewall taper 2(a) 8 6 75 42.1
 2(b) 12 6 75 31.3
Base radius 3(a) 4 20 75 26.5
 3(b) 4 34 75 0.0
Plug diameter 4(a) 4 6 67.5 45.1
 4(b) 4 6 60 37.6

Fixed geometry: plug height = 89 mm, sidewall flat = 9 mm.
COPYRIGHT 2007 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Martin, P.J.; Duncan, P.
Publication:Polymer Engineering and Science
Geographic Code:1USA
Date:Jun 1, 2007
Previous Article:Enhancement of wood/polyethylene composites via compatibilization and incorporation of organoclay particles.
Next Article:Optimization and scale-up of starch cationization in a twin screw extruder.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters