113 Narowski1:Layout 1 2015-09-08 17:36 Strona 1

Study on Filling Patterns of Engineering Polymers

in Geometrically Balanced Injection Molds
Przemysaw Narowski, Krzysztof Wilczyski
Warsaw University of Technology,
Faculty of Production Engineering,
Polymer Processing Department,
Narbutta 85, 02-524 Warsaw, Poland,
k.wilczynski@wip.pw.edu.pl

Filling patterns and imbalances which occur during injection molding of engineering plastics, e.g. Polyoxymethylene (POM), have been studied.

Nanotechnologies and Materials

Four different geometries of runner systems have been applied for experimentation. Autodesk Moldflow software has been used for finite element
method (FEM) simulations of imbalance phenomenon. Evolution strategies have been suggested for optimization of runner systems in multi-cavity
injection molds.
Key words: Injection Molding, Filling Imbalance, FEM Simulations

Introduction fully studied for engineering material Polyoxymethylene

POM with totally different thermo-rheological character-
Multi-cavity injection molding has been regarded as a low
istics and high requirements for processing.
cost mass-production method for producing plastic parts.
The greatest effort in explaining the phenomenon of
However, the flow imbalance during filling a multi-cavity
filling imbalance in multi-cavity injection molds has been
mold using symmetrical runner system is a serious problem
taken by Beaumont [1]. He revealed that the flow imbal-
[1]. It causes non-fills, dimensional variations, warp, flash,
ance is closely related to the three-dimensional thermo-me-
and many more product inconsistences. The filling imbal-
chanical history of the material melt flow in the runners,
ance in geometrically balanced molds is always difficult to
and he proposed a novel apparatus called MeltFlipper to
handle in injection molding. Previous studies have shown
overturn the melt stream to avoid the imbalance problem.
that the most common filling pattern in geometrically bal-
Various configurations of the MeltFlipper system are shown
anced feeding systems is faster filling of inner cavities, and
in Fig.2.
slower filling of outer cavities which is presented in Fig.1.
The polymer melt flow in the injection molds is a com-
There is a little literature concerning the filling imbalance
plex fluid mechanics problem. This is a three-dimensional,
phenomenon. Recently, an extensive study has been per-
unsteady, compressible, and non-newtonian as well as non-
formed for several commodity plastics using various runner
systems [1]. In this paper, the phenomenon has been care-

Figure 2: Runners configurations: a) standard - G_S,

b) single MeltFlipper - G_1, c) double Melt Flipper
Figure 1: Example of mold filling imbalance (Polystyrene) G_2, d) peripheral MeltFlipper - G_3

15
113 Narowski1:Layout 1 2015-09-08 17:36 Strona 2

Przemysaw Narowski, Krzysztof Wilczyski

isothermal flow, and sometimes even viscoelastic flow. the glass transition point Tg = 323K, and data-fitted coef-
Molten polymers are the pseudoplastic fluids, and their vis- ficients
cosity is strongly dependent on the shear rate and temper- D1=5.321023 Pas,
ature. The viscosity decreases with an increase of the shear A1=59.833,
rate and temperature. A2=51.6 K.
Filling imbalance results from a non-linear velocity pro- The effect of using of an overturn apparatus (MeltFlip-
file in the runners which generates a non-linear shear rate per) on the mold filling pattern has been presented in Fig.3.
profile, and finally a non-linear temperature profile, which For a standard runner system (Fig.2a) the material of high
strongly influence the polymer melt viscosity. And, shear viscosity flows into outer cavities, and the material of low
rate and viscosity determine the heat amount produced viscosity flows into inner cavities. These lead to faster filling
during the flow. of inner cavities and slower filling of outer cavities (Fig.3a).
It can be written [3] that For a MeltFlipper system (Fig.2b) the filling pattern
changes, and faster filling of outer cavities is observed
(1) which results from lower viscosity of the material as well as
slower filling of inner cavities is seen which results from
where is the heat amount produced in an elemental volume higher viscosity of the material (Fig.3b).
of the flowing material, W/m3, is the apparent shear vis-
cosity of the material, Pas, is the temperature, oC, and is Experimental
the shear rate, 1/s.
Such a complex phenomenon can be studied using so- An extensive experimentation has been made to study the
phisticated numerical methods, and/or advanced specifi- filling imbalance problem and investigate an effect of using
cally oriented software, like Moldflow, Cadmould and of the MeltFlipper apparatus on the phenomenon. Eight-
Moldex3D [3-6]. This software is specially dedicated to cavity mold with changeable runner inserts has been de-
simulate the injection molding process. signed to this aim.
Four different runner layouts have been tested:
the geometrically balanced standard runner system G_S
FEM Simulations
(Fig.2a), and three MeltFlipper systems:
In order to study the filling imbalance in multi-cavity in- the classical single system G_1 (Fig.2b),
jection molds FEM simulations have been made using an the double system G_2 (Fig.2c), and
advanced injection molding system Autodesk Moldflow the peripheral system G_3 (Fig.2d).
[5]. Polybutylene Terephthalate (PBT) Valox 337 has been Each system has been tested for Polyoxymethylene
used in the study. The Cross-WLF rheological model of the (POM) Hostaform C 9021. The material properties have
material has been applied for simulations. been presented in Tabl.1.
The Cross-WLF viscosity model describes the temper- Three injection rates have been applied, 20%, 50%,
ature, shear rate, and pressure dependency of the viscosity and 80% of the maximum available injection rate equal to
of flowing material. The model is given by the following Vmax = 100 mm/s. The processing parameters have been
equation: shown in Tabl.2.

(2)
Table 1. Material properties
where:
is the melt viscosity, Pas,
0 is the zero shear viscosity or the Newtonian limit the
viscosity approaches at a very low shear rate,
is the shear rate, 1/s,
* is the critical stress level, Pa, at the transition to shear
thinning behaviour, determined by curve fitting, and
n is the power law index in the high shear rate polymer flow
region, determined by curve fitting. Table.2. Processing parameters
The parameters of the model we used in the study were the
following:
the power-law flow index n=0,2139,
the zero viscosity ,

the critical stress * = 353100 Pa,

16
113 Narowski1:Layout 1 2015-09-08 17:36 Strona 3

Study on Filling Patterns of Engineering Polymers in Geometrically Balanced Injection Molds

Figure 6. POM molding with double MeltFlipper (Geo-

metry_2)

Nanotechnologies and Materials

Figure 3. PBT simulations viscosity plots: a) Standard
configuration, b) MeltFlipper configuration

Figure 7. POM molding with peripheral MeltFlipper (Geo-

metry_3)

Some results of experimentations have been presented

in Figs.4-7. It is clearly seen that standard runner geometry
causes faster filling of inner cavities (Fig. 4), and single
as well as double MeltFlipper geometries noticeably bal-
ance the filling pattern (Figs. 5-6). Using of a peripheral
Figure 4. POM molding with standard runners (Geome- MeltFlipper geometry leads to the filling pattern similar to
try_S) that one specific for standard geometry (Fig. 7).
Mold filling imbalance has been evaluated using a mass
filling imbalance factor defined by the following relation

(3)

where:
Im is the mass filling imbalance factor,
m1 is an average mass of the material from inner cavities,
and
m2 is an average mass of the material from outer cavities.
Mass filling imbalance factors Im have been plotted
against injection rate for all the geometries used in the
study (Fig. 8). These show the differences in mold filling
patterns. Geometry_S and Geometry_3 cause positive im-
Figure 5: POM molding with single Melt Flipper (Geo-
metry_1)
balance which means that inner cavities are filling faster,

17
113 Narowski1:Layout 1 2015-09-08 17:36 Strona 4

Przemysaw Narowski, Krzysztof Wilczyski

Conclusions
Filling imbalance in multi-cavity geometrically balanced
injection molds has been studied using several feeding sys-
tems. It can be stated that any clear regularities in filling
have not been noticed when changing the system, and ob-
viously there is no universal design for the runner system.
The individual simulation has to be applied for each case.
Meshing, cooling conditions, material characteristics are of
high importance for these simulations. Evolution strategies
(Genetic Algorithms) are suggested for optimization of
runner systems in multi-cavity injection molds to solve the
problem.
Genetic Algorithms are search and optimization meth-
Figure 8: Flow imbalance for POM ods that imitate the natural evolution process. They start
from a set of randomly generated points and apply genetic
operators like crossover and mutation to confine the region
where the optimum is located. It prevents the algorithm
from being trapped in a local minimum. And, this does not
require any derivatives and does not impose any restrictions
on the convexity of the search space. It is a very powerfool
tool for optimization in polymer processing [7].

References
[1] Beaumont J.P.: Runner and Gating Design Handbook,
Hanser, Munich 2004
[2] Wilczyski K., Wilczyski K.J., Narowski P.: Badania symu-
lacyjno-dowiadczalne nierwnomiernego wypeniania form
wtryskowych zrwnowaonych geometrycznie, Polimery 2015
(accepted for publication).
[3] Wilczyski K.: Reologia w przetwrstwie tworzyw sztucz-
nych, WNT, Warszawa 2001.
[4] Wilczyski K.: Cadmould-3D System for Computer Model-
ing of Injection Molding of Polymeric Materials (in Polish),
Polimery 1999, p.407.
[5] Cook P., Yu H., Kietzmann C., Costa F. (2005, May 1). Pre-
diction of Flow Imbalance in Geometrically Balanced Feed Sys-
tems [online]. Available: www.moldflow.com
[6] Chien C. C., Chiang C. C., Yang W. H., Tsai V., Hsu D. C.
Figure 9. POM simulations filling time plots: a) standard (2005, May 1). True 3D CAE visualization of filling imbal-
configuration, b) MeltFlipper configuration ance in geometry-balanced runners [online]. Available:
www.moldex3D.com
[7] Wilczyski K., Nastaj A.: SSEM-AG Computer Model for
Optimization of Polymer Extrusion, Proceedings of IMECE,
and it is opposite to Geometry_1 and Geometry_2 which Chicago 2006, p.13074.
induce negative mass filling imbalance which results in
faster filling of outer cavities. Authors:
Numerical simulations for mold filling have been de- MSc Eng. Przemysaw NAROWSKI PhD student at the
picted in Fig.9. Faster filling of inner cavities is clearly seen Polymer Processing Department at the Warsaw University of
for standard geometry (Fig.9a) which is consistent with ex- Technology, graduated in Mechanical Engineering at the War-
perimental data (Fig.4). MeltFlipper geometry obviously saw University of Technology at the Faculty of Production En-
balances the filling pattern and faster filling of outer cavities gineering.
is observed (Fig.9b) which is comparable with experimental Prof. Krzysztof WILCZYSKI professor at the Polymer Pro-
data (Fig.5). cessing Department at the Warsaw University of Technology, a
supervisor of the doctoral candidate.

18