DEVELOPMENT OF A PLUG ASSISTED THERMOFROMING

SIMULATION.

R.McCool1*, P.J.Martin2 and Eileen Harkin Jones 3

1
[email protected]; 2 [email protected]; 3 [email protected]
Queen’s University of Belfast, School of Mechanical Engineering, Ashby Building, Stranmillis Road, Belfast, Northern
Ireland, BT9 5AH

Abstract - Thermoforming is a very important process in the manufacture of thin gauge food packaging products.
Industry methods currently rely on trial end error, however there is a growing need to develop robust finite element
simulations of the process. This paper describes the development of a simulation of the plug assisted thermoforming of
high impact polystyrene (HIPS) products, using the commercially available finite element software ABAQUS. The
simulation takes into account the key elements associated with the process which include the deformation response of
the material, and the friction and heat transfer between contacting surfaces. The evolution of the simulation from an
isothermal condition with friction between contacting surfaces, to a non-isothermal simulation with friction and heat
transfer effects is presented. The output of the simulation is detailed in the paper and its performance is correlated with
experimental observations.

Introduction
Thermoforming is widely employed in the manufacture of thin gauge packaging for the food industry. Essentially in
the thermoforming process a pre-extruded thermoplastic sheet is clamped, heated to a desired softening temperature and
shaped into a cool mould by a combination of mechanical movement and air pressure. The desired output from
thermoforming is to have a controlled product wall thickness distribution, as in many products excessive material
thickness in certain regions is common, and in addition companies rely on trial and error methods in achieveing an
exceptable product. However the possibilities to simulate the process with the aid of computer software have been
investigated in recent years, and the most common approach is to use the finite element method, which is also the
chosen method in this work. A major stumbling block in the development of such simulations is the need for
researchers to understand the key elements of the thermoforming process which are the high strain/strain rate material
deformation and the tool sheet contact1. And a successful process simulation must include these aspects through an
accurate material deformation model, a temperature dependant friction model and a model which is capable of
predicting the heat transfer effects in the process.

Plug assisted thermoforming and key process elements

In the plug assisted thermoforming process there are two distinct stages in which the movement of a mechanical plug is
followed by the application of positive air pressure. A process diagram is shown in Figure 1. Initially the heated sheet
is clamped followed by the advancement of the plug into the mould thus pre-stretching the sheet to form a preform. Air
is applied to the top surface of the sheet at the point when the plug is fully advanced, the sheet is then forced to take up
the shape of the mould and the forming is complete.

Figure 1 – Plug Assisted Thermoforming Process.

The interaction between the plug, sheet and mould material strongly affects the final part wall thickness distribution1-4.
These are known as the three key elements, the first key element is the mechanical stress/strain response of the material
and its ability to stretch through different modes of biaxial stretching5. During the mechanical pre-stretching of the
sheet by the movement of the plug, the slip along the surface of the plug is related to the coefficients of friction between
The Polymer Processing Society 23rd Annual Meeting
the contacting surfaces. However sheet slip is only achievable through extension of the sheet, therefore this effect is
highly dependant on the ability of the sheet to stretch. This ability for the sheet to slip is the second key element.
Finally the ability for heat to flow from the sheet affects its local temperature and consequently its ability to stretch by
either free or sliding extension, therefore the third key element is the heat transfer effects. To develop a simulation of
the plug assisted thermofroming process all the key elements associated with the process must be included, also
accurate replication of the geometry and process conditions.

Experimental
Experimental thermoforming tests were carried out on a purpose built laboratory thermoforming machine using tooling
and forming parameters that were designed to mimic industrial conditions. The product shape was a typical conical pot,
as shown in Figure 2. The tooling consisted of an aluminium alloy mould and a closely matched flat bottomed plug
made from syntactic foam6. All experimental work was carried out on 1.31 mm thick extruded HIPS (Nova 643N7)
sheet. In this study the commercial Finite Element package ABAQUS was used to create the thermoforming model.
ABAQUS has been widely employed in other polymer process simulations and it has been successful in modeling
stretch blow moulding8,9. Initially an isothermal simulation of the process was developed and the capabilities of the
model were increased by systematically adding the effects of heat transfer through conduction and convection.

Modeling of the Geometry

The single shot experimental thermoformer at Queen’s University of Belfast3 was used to produce test pots to allow
comparison with simulations. The tooling used has the major dimensions as outlined in Figure 2, and produces a pot of
major dimensions 75mm base diameter, 100mm lip diameter and 90mm in height from an initial sheet thickness of
1.31mm. Taking advantage of the tooling symmetry resulted in the ability to model the tooling axi-symmetrically
within ABAQUS. Subsequent wall thickness measurements were recorded from a point on the centre of the pot base
(0mm) to the outer point on the pot lip (133mm).

Major Plug Dimensions

(units mm):
Height 89, base Φ47, corner
radius 7, side angle 6º

Major Mould Dimensions with

Ejector (units mm):
Height 96, base Φ75, side angle 6º

Ejector

Figure 2 – Thermoforming Tooling.

Modeling of the Material
Accurate material models are a fundamental part of any FE simulation and construction of such models requires data
obtained from material tests carried out under conditions comparable to actual processing conditions10. A purpose built
biaxial test rig has been developed by Queen’s University Belfast (QUB)10 and work has focused on accurate
measurement of the deformation behaviour of polymer materials4,10,11. Research carried out by Tshai11, into the biaxial
stretching of HIPS has led to the development of a HIPS deformation model. It is based on a Van der Waals model
combined with a Prony series relaxation function, and uses the Williams, Landel and Ferry, WLF ref, equation to
introduce temperature dependency. Killian and Vilgis12 developed the Van der Waals strain energy function. For
further details of the material model readers are referred to reference1. The models performance in simulating the
biaxial deformation response of the HIPS material used in this study was in very good agreement over a range of strain
rates.

Modeling of Contact
Friction
Research at QUB has employed a sled test mechanism to identify friction coefficients 2,3,13-15. In these experiments
samples of plug material were pulled across the surface of sheet materials at varying contact conditions. Frictional
shear stresses may develop at a contact interface if the two contacting bodies have rough surfaces. If the frictional shear

The Polymer Processing Society 23rd Annual Meeting

stress reaches a critical value, the bodies will slide. For the purpose of this work a temperature dependant friction
model measured at QUB3 is used for the combination of plug and sheet material used in this study. Within ABAQUS
isotropic Coulomb friction using the penalty method was selected in this work. In the initial isothermal simulation
single values of friction were tested, as the simulation icorporated heat transfer the temperature dependant friction
model was implemeneted and its performance was evaluated.

Heat Transfer
To accurately model the thermal effects in the plug assisted thermoforming process, all components in the system, i.e.
the plug, the sheet and the mould are assumed to be in a non-isothermal condition. Heat transfer by conduction depends
on the conductivity of the material within an interface as well as the temperatures of the respective surfaces, and is
modelled in ABAQUS by Gap Conductance. Investigations into the effects of heat tranfer were conducted using
thermal imaging equipment and the experimental thermoformer which was setup to allow plug only tests and
subsequent measurement of sheet temperature during plugging, similar tests were carried out at QUB3. Further details
of the approach used in modeling of heat transfer and friction, are detailed in paper P02-006: MODELING OF
CONTACT IN THERMOFORMING.

Results and Discussion

Figure 3 shows the experimental wall thickness distribution as measured from a cup produced at the base settings (Sheet
Temperature 120ºC, Plug Temperature 100ºC). Also shown in figure 3 is the output from an isothermal simulation
showing how the wall thickness distribution varies by changing the coefficient of friction between the plug and the
sheet and for simplicity the contact between the sheet and the mould was assumed “rough”. When a high value of
coefficient of friction was chosen between the plug and the sheet more material resulted in the base of the cup (0 to
35mm) where as the material in the lip region (120 to 133mm) was reduced. A major area of discontinuity is in the side
wall of the cup from 50 to 90mm where the simulation under predicts the thickness, this may be due to the contact
between the sheet and the plug, or material behaviour, and further investigations are being carried out to establish the
cause of this.

1400
Base Setting
1200 Frictionless
0.5
1000 1.0
Thickness (µ m)

800

600

400

200

0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
Distance from Base Centre (mm)
Figure 3 – Effect of Plug Sheet Friction on Wall Thickness under Isothermal Conditions.

Figure 4 shows the results of increasing the sheet temperature up to 140ºC on the wall thickness distribution. In this
case the fully coupled temperature dependant friction model is incorporated and heat flow is permitted between
contacting surfaces by using a fixed arbitrary gap conductance value of 500. The results show that as the sheet
temperature is increased the amount of material in the base region increases, this maybe due to the fact that at higher
temperatures the friction coefficients also increase.

The Polymer Processing Society 23rd Annual Meeting

 1400

1200 Sheet 120ºC

Sheet 130ºC
1000
Thickness (µ m)

Sheet 140ºC
800

600

400

200

0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
Distance from Base Centre (mm)
Figure 4 – Effect of Plug and Sheet Temperature on Wall Thickness.

Conclusion
A finite element simulation of the plug assisted thermoforming process was constructed using ABAQUS, the model was
initially constructed isothermally and was subsequently improved to incorporate the effects of heat transfer. It was
shown that even under isothermal conditions the model was capable of predicting relatively accurately the profile of the
final wall thickness distribution as achieved experimentally by systematically changing the level of friction between the
sheet and the plug. By adding a temperature dependant friction model and the effects of heat transfer the capabilities of
the simulation were improved. It is important to stress however that even though the simulation incorporates all the key
elements associated with the process, it does not automatically result in accurate simulation. A balance has to be struck
between all key elements so that they perform in an inter-related manner, for example, if the sheet temperature cools
during forming this has a bearing on the contact friction and the ability of heat to flow from the sheet, and it is these
complex interactions that must be accurately modeled.
Acknowledgements

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

References
1. McCool, R., Martin,P.J., Harkin-Jones,E. Plastics, Rubber and Composites 2006, 35, pp1-8.
2. Collins, P.; Harkin-Jones, E. M. A.; Martin, P. J. ANTEC, 2002, pp 531-547.
3. Collins, P.; Harkin-Jones, E. M. A.; Martin, P. J. PPS XVII,2002, 17,pp 361-369.
4. Martin, P. J.; Tan, C. W.; Tshai, K. Y.; McCool, R.; Menary, G. H.; Armstrong, C. G.; Harkin-Jones, E. M. A.
Plastics, Rubber and Composites 2005, 34,pp 276-282.
5. Tshai, K. Y.; Harkin-Jones, E. M. A.; Martin, P. J.; Menary, G. H. ANTEC 2004 , pp 919-923.
6. NewMet. www.newmet.co.uk/.
7. Chemicals, N. www.novachem.com 2006.
8. Menary, G. H.; Armstrong, C. G.; Crawford, R. J.; McEvoy, J. P. Plastics, Rubbers and Composites 2000,
29,pp 360-370.
9. Yang, Z. J.; Harkin-Jones, E. M. A.; Menary, G. H.; Armstrong, C. G. Journal of Materials Processing
Technology 2004, pp153-154.
10. Tshai, K. Y.; Harkin-Jones, E. M. A.; Martin, P. J., ANTEC 2002, pp 463-476.
11. Tshai, K. Y.PhD Thesis, Queen's University of Belfast, 2004.
12. Kilian, H. G.; Vilgis, T. Colloid & Polymer Science 1984, 262, pp15-21.
13. Laroche, D.; Collins, P.; Martin, P. J. ANTEC, 2001, pp 94-98.
14. Collins, P.; Lappin, J. F.; Harkin-Jones, E. M. A.; Martin, P. J. Plastics, Rubber and Composites 2000, 29,
pp349-359.
15. Collins, P.; Martin, P. J.; Harkin-Jones, E. M. A. ANTEC, 2001, pp 630-634.

The Polymer Processing Society 23rd Annual Meeting