Modelling of the Through-air Bonding Process

M. Hossain1, M. Acar, Ph.D. 2, W. Malalasekera2

1
School of Engineering, The Robert Gordon University, Aberdeen, UNITED KINDOM
2
Mechanical and Manufacturing Engineering, Loughborough University, Leicestershire, UNITED KINGDOM

Correspondence to:
Memis Acar, Ph.D. email: [email protected]

ABSTRACT distribution of velocity, pressure and temperature

A computational fluid dynamics (CFD) modelling of inside the bonding chamber and optimized the
the through-air bonding process of nonwoven fabric design. Tafreshi and Pourdeyhimi [3] and Tafreshi et
production is reported in this article. In the through- al. [4] investigated the flow dynamics inside hydro-
air process, hot air is passed through the fibrous web entangling nozzles using CFD modelling. More
to heat and melt polymer fibers. Molten polymer recently, Hossain et al. [5, 6] reported a CFD model
subsequently flows to the point of contact between of the airflow and heat transfer through fibrous webs
any two fibers to produce a bond. Two different as well as a simulation of the fiber bonding process.
modelling strategies are adapted to produce a These examples alone show that the CFD method can
comprehensive understanding of the through-air play a vital role in analyzing and improving the
bonding process. In macroscale modelling, a CFD nonwoven manufacturing processes.
model is developed treating the whole web as a
porous media in order to investigate the effect of Manufacturing of nonwoven fabric begins by laying
process parameters. Results reveal that the time fibers to form a web. The through-air bonding
required to heat and melt the fibers decreases with the process, involves passing hot air through an
increasing porosity of the web and the velocity of hot unbonded web to heat and melt the polymer fibers.
air. The CFD modelling technique is then used to To create a bond of sufficient strength, the web is
analyze the bonding process at a more fundamental kept at an elevated temperature for certain amount of
level by considering the bonding of individual fibers time, known as the dwell time. Then, the molten
at microscale. The effects of the fiber diameter, polymer, aided by the reduced viscosity at high
bonding temperature and contact angle between two temperatures, flows to the contact point of fibers by
fibers on the bonding time are investigated. Results the surface tension force. Most through-air bonded
show that the time required to bond fibers is weakly webs consist of bicomponent sheath-core type
related to bonding temperature and fiber diameter. polymer fibers, while the core material with a higher
Fiber orientation angle, on the other hand, has melting point provides the structural integrity to the
significant effect on the progression of bond web whereas the sheath polymer with a lower
formation. melting point enables the bonding of the fibers
together at the point of contact.
INTRODUCTION
The Computational Fluid Dynamics (CFD) technique Literature survey reveals that the published work on
has become a major tool for the design and the through-air bonding process is mostly concerned
development of many engineering systems involving with the experimental study investigating the effects
fluid flow and heat transfer. It has been extensively of process parameters on the properties of the
used in the field of aerospace, automotive, nonwoven. Randall [7] reported an experimental
biomedical, chemical and many other industrial study of the effects of process variables such as air
applications. Its use in textile research and temperature, air velocity, dwell time, restraining and
development is however not very common. There are carrying mesh size on the mechanical properties of
only very few examples of such computational the through-air bonded nonwovens. Randall [8] also
modelling in textile research and development. Gong studied the effects of binder type, binder fiber
et al. [1] developed an experimental process for content, web basis-weight on the product properties.
producing through-air bonded three dimensional Kim et al. [9] investigated experimentally the
nonwoven web by applying the CFD technique. formation of bond and the bonding time for different
Gong et al. [2] in a subsequent article investigated the fiber diameters between two fibers laying orthogonal

Journal of Engineered Fibers and Fabrics 1 http://www.jeffjournal.org

Volume 4, Issue 2 - 2009
to each other. Kim et al. [10] in a follow up article
investigated the bond formation and development k
eff a
[( l s
)
= φ k + ( 1 − φ ) ε γ k + ( 1 − γ )k + ( 1 − ε )k
c
] (5)
between two fibers by using a simple computational
model and predicted the characteristics shape. In addition, the relationship describing melt fraction
However, it failed to predict the bonding time γ as a function of temperature is given by,
relationship with the fiber diameters as observed in
the experiments.
⎧
⎪1 if T > T + ΔT (6)
⎪ m
In the present study, the fundamental mechanism of ⎪
γ = ⎨0 if T < Tm − ΔT
bond formation during the through-air bonding ⎪
⎪ T − Tm + ΔT
process has been investigated in greater detail using ⎪
2ΔT
if T − ΔT < T < T + ΔT
m m
⎩
the computational fluid dynamics technique. First a
macroscale model is developed in which the web as a The FLUENT CFD software was used for the
whole is modelled as a porous media. Then, a solution of the governing continuity and momentum
microscale model is developed to study the bond and energy equations.
formation between two contacting fibers.
Microscale modeling
MATHEMATICAL MODEL In microscale modeling, two-phase Volume of Fluid
(VOF) method is used for the modelling of the air
Macroscale Model and molten polymer flows. In the VOF method, a
A macroscopic formulation based on volume single set of momentum equations is shared by
averaging is used in the present study. The averaged phases (here, air and polymer) and volume fraction of
mass and momentum equations for a single domain each phase is tracked through the computational
based model are given by Beckerman [11]: domain. Molten polymer flows to the contact point
because of the surface tension force. Therefore, the
Continuity: momentum equation contains a source term
representing the surface tension force. The
∂( φρ a ) ρ
+ ∇.( ρ uφ ) = 0 (1) governing equations for the VOF model are given
∂t a
below:

Momentum: Continuity:
ρ
ρa
∂( φu ) ρ ρ ρ μρ
+ ρ ( φu .∇ )u = −φ∇P + μ∇ 2 ( φu ) − u (2) ∂ρ
∂t a k + ∇.( ρu ) = 0
∂t (7)
The last term in the momentum equation is due to
Dercy’s term representing the resistance caused by Momentum:
the fibrous web. The value of permeability k is
∂( ρ u )
calculated from a model given by Mao and Russell + ∇ .( ρ u u ) = −∇ P + ∇ .μ ( ∇ u + ∇ u T ) + S + F (8)
∂t
[12, 13].
Initially, the fibers are in solid state and obviously the
The averaged energy equation for air and the porous
velocity is zero. When the fibers are melted by heat
matrix can be written as
transfer from the flowing hot air, the molten polymer
∂T ∂γ (3) starts to flow. The term S in equation (8) works as a
ρc + ρ c ∇ .( u T ) = ∇ .( k ∇T ) − ( 1 − φ )ερ Δh
∂t a a eff l ∂t momentum sink to ensure that the fiber velocities in
the initial cold state and then in the molten hot state
The mean thermal capacitance of the mixture, ρc , is are correctly represented. The surface tension force in
defined as, equation (8) is represented by F . The force at the
surface is expressed as a volume force and is added
[( )
ρc = φρa c a + ( 1 − φ ) ε γρ l c l + ( 1 − γ )ρs c s + ( 1 − ε )ρ c c c ] (4) to the momentum equation as a source term.

Volume fraction equation:

The effective thermal conductivity is given as a
volumetric average,

Journal of Engineered Fibers and Fabrics 2 http://www.jeffjournal.org

Volume 4, Issue 2 - 2009
To track the interface between phases a volume obtained with a fixed time step of one microsecond
fraction continuity equation for the secondary phase [5].
(fiber) is solved along with the above equations:
∂α q ϖ The web is assumed to be made of sheath-core type
+ u.∇α q = 0
∂t (9) bi-component fibers of polyethylene (PE) sheath and
polypropylene (PP) core with 68 mm web thickness.
where subscript q represents each phase component This is a high loft web which represents and extreme
Air volume fraction is obtained from the relation thickness. The calculations are carried out for the air
2 inlet velocity of 1.5, 2.0, 2.25, 2.5 and 3.0 m/s and
∑ αq = 1 . the web porosity of 0.5, 0.7 and 0.9.
q =1
Figure 2(a) gives the simulation results of the
The properties appearing in the transport equations temperature distribution inside the web and Figure
are determined by the presence of the component 2(b) shows the corresponding distribution of the melt
phases in each control volume. For example, the fraction of fibers. The predicted results are for the
density is considered to be: porosity 0.9 and air velocity 1.5 m/s. The hot air
transfers heat to the web as it flows through it. The
2
fiber temperature rapidly increases to the melting
ρ = ∑α q ρ q point of sheath polyethylene fiber and the fiber starts
q =1
(10) to melt. The simulation result shows that it takes
approximately 6.6 seconds to heat and melt fibers
RESULTS AND DISCUSSION throughout the web thickness and to reach a steady
state condition.
Macroscale analysis of through-air bonding
A schematic drawing of a through-air bonding system Hossain et al. [5] showed that the time required to
is shown in Figure 1. A nonwoven web is wrapped melt the fibers through the thickness of the web
around a perforated rotating drum, through which the shows a diminishing linear relationship with the web
hot air passes radially. To simplify the development porosity. This relationship is quite realistic as the
of the model, the CFD calculation was carried out on increased porosity of the web implies that there is
a 60° sector of the drum. less volume of fibers within the same volume of web
to be heated and melted. The volume of the hot air
flowing through the web increases with the
increasing air velocity which in turn leads to higher
convective heat transfer rates to the web
consequently resulting in a reduced heating and
melting time.

Although, the macroscopic model is very useful in

providing information related to the heating and
melting time, this model can not provide information
regarding the actual bond formation process. A
microscale model, which considers the bonding
process of individual fibers, is reported in the next
section. A detailed account of the micro level
modelling can be found in Hossain et al. [6].
FIGURE 1. A Schematic Drawing of the Through-Air Bonding
Process

The computational domain is divided into

computational mesh with an average grid size of 1
mm in both the radial and circumferential directions.
A uniform inlet velocity and temperature is specified
at the inlet boundary. Pressure boundary condition is
used at the outlet boundary. The inlet temperature is
kept fixed at 140° C. The transient solution is

Journal of Engineered Fibers and Fabrics 3 http://www.jeffjournal.org

Volume 4, Issue 2 - 2009
FIGURE 2. (a) Contour plot of temperature of web at different time steps (b) corresponding contour plot of the melt fraction of fibers for a
porosity of 0.9 and air velocity of 1.5 m/s.

Microscale analysis of through-air bonding circumferentially to the contact point. As the bond
Stochastic nature of fiber distribution in a web makes formation progresses, the molten polyethylene begins
it very complicated and resource intensive to generate to flow along the fiber axis towards the intersection,
a realistic computational model for a web specimen. which then continues to flow to the contact point.
Instead, the formation of a single bond between two This causes a reduction in the fiber diameter in the
fibers is investigated in the present study. In the first vicinity of the intersection.
instance, bonding of two fibers at a contact angle of
90°, as shown in Figure 3, is studied at a bonding
temperature of 140°C. Then the effects of the air
temperature and the fiber diameter are investigated.
Finally, the effect of contact angle between two fibers
is investigated.

The development of bond formation between two

polyethylene sheath/polypropylene core bi-
component fibers is shown in Figure 4 as a function
of time. The molten polyethylene flows to the
intersection of two fibers to form a bond. The flow of
molten polymer is mainly driven by the surface
tension force. In the initial stages of the bond
formation, molten polyethylene flows

Journal of Engineered Fibers and Fabrics 4 http://www.jeffjournal.org

Volume 4, Issue 2 - 2009
FIGURE 3. Computational domain showing two fibers at 90°
contact angle

FIGURE 4. Computed shape of bond at different times for the bonding temperature of 140°C.

In order to quantify the progress of bonding, a Effect of the fiber contact angle
dimensionless characteristic bond size is proposed as The effect of the angle between two fibers on the
the ratio of the area of the section of the bond on a progression of bonding is also investigated. Figure 5
plane parallel to the axes of the fibers through their shows the computational model involving two fibers
contact point to the cross-sectional area of the fibers. with orientation angle of 10°, 20° and 30°. Figures 6
This characteristic bond size provides a quantitative and 7 show the progress of bonding for these contact
comparison of the bonding. Bonding proceeds at a angles at 1 and 5 seconds respectively. The
faster rate at the initial stages. This is because the characteristic bond size defined earlier has been used
molten polymer flows rapidly to fill up the narrow to compare the bond size for different fiber contact
gap at the contact point of two fibers driven by the angles. Figure 8 shows the development of the bond
high surface tension force. As the bond grows filling size with time for different contact angles. The effect
the gap at the contact point between the two fibers, of the contact angle on the progress of bond
the surface curvature increases, consequently development is found to be significant.
reducing the surface tension, which in turn slows
down the rate of the bond development.

Comparison of the bond size development with time

shows that the air temperature in the range of [130-
150°C] and fiber diameter in the range of [15.1-
30.3μ] made very little effect on the bonding process.
Bonding appears to proceed at a slightly faster rate at
higher temperatures. Surface tension force and
viscosity are the two most important parameters that
control the flow of the molten polymer. When the
bonding temperature increases, both viscosity and
surface tension force decrease. Overall the reduction
of viscosity at higher temperature over compensates
the reduction in surface tension force. This results in
a small increase of the bonding rate at higher FIGURE 5. Computational domain showing fibers at contact
temperatures. The effect of the fiber diameter on the angles of (a) 10° (b) 20° and (c) 30°
bond size development shows a slight decrease with
the increasing fiber diameter. This is in line with the
experimental finding of Kim et al. [9].

Journal of Engineered Fibers and Fabrics 5 http://www.jeffjournal.org

Volume 4, Issue 2 - 2009
The overlap of the fibers at low fiber contact angles orientation angle increases this rate becomes slower.
is much greater than the overlap at high contact At high fiber orientation angles, the initial fiber to
angles. The bonding at smaller contact angles fiber overlap is small and therefore the bond needs to
proceeds at a much faster rate because the low grow with polymer flow not only in the
contact angle causes a greater amount of polymer to circumferential direction but also in the longitudinal
flow in the circumferential direction to the contact direction which slows down the bond growth rate.
point due to the greater fiber overlap. This causes a These trends are observed in Figure 8.
rapid development of the bond size. As the

a) (b) (c)

FIGURE 6. Progress of bonding after 1 second. The top row shows three-dimensional view and the bottom row shows cross-sectional view. (a)
10°, (b) 20° and (c) 30° contact angle.

(a) (b) (c)

FIGURE 7. Progress of bonding after 5 seconds. The top row shows three-dimensional view and the bottom row shows cross-sectional view.
(a) 10°, (b) 20° and (c) 30° contact angle.

Journal of Engineered Fibers and Fabrics 6 http://www.jeffjournal.org

Volume 4, Issue 2 - 2009
 area between the fibers at low fiber orientation
angles.

ACKNOWLEDGEMENT
The financial support for this research from the
Nonwovens Cooperative Research Center (NCRC),
North Carolina State University, Raleigh, is
gratefully acknowledged.

REFERENCES

[1] Gong, R. H., Fang, C. and Porat, I. Single

process production of 3D nonwoven shell
structures – part 1: web forming system design
using CFD modelling. Int. Nonwovens J.,
2000, 9(4), 20-24.
[2] Gong, R. H., Dong, Z. and Porat, I. Single
FIGURE 8. The effect of contact angle at different contact angles process production of 3D nonwovens shell
on the progression of bonding.
structures – part 2: CFD modelling of thermal
bonding process. Int. Nonwovens. J., 2001,
10(1), 24-28.
CONCLUSION
A computational fluid dynamics study of through-air [3] Tafreshi, V. H. and Pourdeyhimi, B.
bonding process of nonwoven fabrics has been Simulating the flow dynamics in
presented in this article. The modelling strategy hydroentangling nozzles: effect of cone angle
involves macroscale and microscale modelling to and nozzle aspect ratio. Text. Res. J., 2003,
analyze the bonding process. In the macroscale 78(3), 700-704.
modelling, the web has been treated as a porous
[4] Tafreshi, V. H., Pourdeyhimi, B., Holmes, R.,
media and the heating and melting time of fibers
and Shiffler, D. Simulating and characterizing
within a nonwoven web have been computed. The
water flows inside hydroentangling orifices.
relevant operating parameters such as the effect of air
Text. Res. J., 2003, 73(3), 256-262.
velocity and web porosity on the melting time of the
fibers have been studied. The computational model [5] Hossain, M., Acar, M., and Malalasekera, W.,
results indicate that the melting time decreases A mathematical model for airflow and heat
linearly with the increase of web porosity and transfer through fibrous webs, Proc. IMechE,
decreases nonlinearly with the increase of air Part E: J. Process Mechanical Engineering,
velocity. Although the macroscale model has 2005, 219, N. 4, 357-366.
provided useful information, it could not provide
information regarding the bond formation process. [6] Hossain, M., Acar, M. and Malalasekera, W.,
Then, a microscale model has been developed to Computational Analysis of Fiber Bonding in
study the formation of bond between two fibers. In the Through-Air Process, Proceedings of the
the first instant, the effect of bonding time on the IMechE, Part E: Journal of Process
progression of bond between two fibers at 90° Mechanical Engineering, Volume 221, No. E2,
contact angle has been studied. The characteristic 2007, pp. 69-75
bond size has been defined and used in analyzing the [7] Randall, K. R. The influence of process
process. The computed results show that the variables on the properties of thermofusion
formation of bond starts rapidly, but slows down bonded fabrics. In Proceedings of the
gradually. The effects of temperature and fiber Nonwoven Symposium, TAPPI, 1984.
diameter on the growth of the bond have been shown
to be small. However, the effect of orientation angle [8] Randall, K. R. Through air bonding of
between the two fibers on the growth of bond has nonwoven fabrics. In Proceedings of the 13th
been observed to be significant. The rate of bonding Annual Technical Symposium –Association of
decreased sharply with the increase of the fiber the Nonwovens Fabric Industry, Boston, MA,
contact angle. This is attributed to the greater contact 1985, pp. 203-219.

Journal of Engineered Fibers and Fabrics 7 http://www.jeffjournal.org

Volume 4, Issue 2 - 2009
[9] Kim, H. S., Ito, H., Kikutani, T. and Okui, N., Δh heat of melting [J/kg]
The thermal bonding behavior of ε volume fraction of sheath fiber in the
polyethylene/poly ethylene terephthalate bicomponent fiber
bicomponent fibers. J. Text. Inst., 1997, 88 (3),
37-50. μ viscosity [kg/ms]
[10] Kim, H. S., Ito, H., Kikutani, T. and Okui, N. ρ density [kg/m3]
Computational analysis on the thermal bonding φ volume fraction of air (porosity)
behavior of bicomponent fibers. J. Text. Inst.,
1999, 90 (1), 508-525.
Subscripts
[11] Beckermann, C. and Viskanta, R. Natural a air
convection solid/liquid phase change in porous c core fiber
media. International Journal of Heat and Mass f fiber
Transfer, 1988, 31(1), 35-46.
l sheath fiber in liquid state
[12] Mao, N. and Russell, S.J. Directional s sheath fiber in solid state
permeability in homogonous nonwoven
structures Part 1: The relationship between eff effective
directional permeability and fiber orientation. m melting
Journal of the Textile Institute, part 1, 91(2),
235-243. AUTHORS’ ADDRESSES
[13] Mao, N. and Russell, S.J. Directional M. Hossain
permeability in homogonous nonwoven School of Engineering
structures Part 2: permeability in idealized The Robert Gordon University
structures. Journal of the Textile Institute, Part Schoolhill, Aberdeen, AB10 1FR
1, 2000, 91(2), 244-258. UNITED KINGDOM

NOTATIONS M. Acar, Ph.D.; W. Malalasekera

c specific heat [J/kgK] Mechanical and Manufacturing Engineering
Loughborough University
df diameter of fiber [m]
Leicestershire, LE11 3TU
Ea activation energy [J/mole] UNITED KINGDOM
F surface tension force [N/m3]
h, H enthalpy [J/kg]
k permeability (m2)
k eff effective thermal conductivity [W/m2K]
P pressure [N/m2]
R universal gas constant [J/kgK]
r radius (m/s)
S source or sink term [N/m3]
Sh source or sink term in the energy equation
[W/m3]
t time [s]
T temperature [K]
u velocity [m/s]
x, y co-ordinate location

Greek
α volume fraction
β fraction of melt of sheath fiber
γ fraction of melt of sheath fiber

Journal of Engineered Fibers and Fabrics 8 http://www.jeffjournal.org

Volume 4, Issue 2 - 2009