Resin Transfer Molding Process:

Fundamentals, Numerical Computation

and Experiments

Felipe Ferreira Luz, Sandro Campos Amico, Jeferson Ávila Souza,

Enivaldo Santos Barbosa and Antonio Gilson Barbosa de Lima

Abstract Resin Transfer Molding (RTM) is one of the most widely known
composite manufacturing technique of the liquid molding family, being exten-
sively studied and used to obtain advanced composite materials comprised of fibers
embedded in a thermoset polymer matrix. The fibrous reinforcement is considered
a porous medium regarding its infiltration by the polymer resin. In this sense, this
chapter aims to briefly discuss multiphase flow and heat transfer theory in RTM
process, focusing on a multifluid model and the Control Volume/Finite Element
(CV/FE) method. Finally, computational analysis was developed on the basis of
ANSYS CFX and PAM-RTM commercial software’s for the investigation of the
fluid flow in RTM composite molding. In order to show the versatility and
performance of the commercial codes, RTM experiments were carried under
distinct injection pressure and fiber volume fraction conditions using plain-weave

F. F. Luz  S. C. Amico
Federal University of Rio Grande do Sul, Porto Alegre, RS 91501-970, Brazil
e-mail: [email protected]
S. C. Amico
e-mail: [email protected]
J. Á. Souza
Federal University of Rio Grande, Rio Grande, RS 96201-900, Brazil
e-mail: [email protected]
E. S. Barbosa
Federal University of Campina Grande, Campina Grande, PB 58429-900, Brazil
e-mail: [email protected]
A. G. B. de Lima (&)
Department of Mechanical Engineering, Federal University of Campina Grande,
Campina Grande, PB 58429-900, Brazil
e-mail: [email protected]

J. M. P. Q. Delgado et al. (eds.), Numerical Analysis of Heat and Mass Transfer 121
in Porous Media, Advanced Structured Materials 27,
DOI: 10.1007/978-3-642-30532-0_5,  Springer-Verlag Berlin Heidelberg 2012
122 F. F. Luz et al.

glass fiber cloth as the porous media. The transient numerical simulations provided
information about volume fraction, pressure and velocity distribution of the phases
(resin and air) inside the porous media.

Keywords LCM  RTM  Modelling  Simulation  Experimentation 

Composites

1 Introduction

1.1 Composite Materials

Composite materials are considered to be combinations of materials differing in

composition or form on a macroscale [1]. The constituents, the dispersed phase
and the matrix (continuous phase), retain their identities in the composite and they
also exhibit an interface between one another [2]. In general, the advantages of this
class of materials include weight reduction, flexibility in design, reduction of
manufacturing cost and corrosion resistance.
According to the principle of the combined action, better property combinations
are fashioned by the judicious combination of two or more distinct materials.
However, for a hybrid material to be regarded as a composite, it must satisfy at
least four criteria, including: (i) Being made up of two or more materials, inher-
ently immiscible; (ii) The properties of the composite should be markedly different
from those of the constituents; (iii) The components must be present in reasonable
proportions, and (iv) The different materials must be separated by an interface.
Technologically, the most important composites are those wherein the dis-
persed phase is in the form of a fiber, whose strength and/or rigidity characteristics
are often desired. The fibrous reinforcement is usually comprised of fabrics
produced with continuous strands or tows of a few hundred fibers. The most
common synthetic fibers currently used are glass, aramid and carbon, which can be
used in a variety of architectures such as random mat, plain weave and stitched
fabric. Natural fibers, such as hemp, flax, bamboo, sisal, coir, jute, pineapple,
among others, can also be employed due to their renewability, biodegradability,
light weight, reasonably high specific mechanical properties and low thermal
conductivity properties [3]. For the selection of the resin (matrix), reasonable cost,
durability under anticipated exposure conditions, adhesion to the aggregate,
handling characteristics and ease of curing are very important considerations [4]
and the most commonly used thermoset resins nowadays are epoxy, polyester,
phenolic and vinyl ester.
Most composites are developed in order to exploit mechanical properties such
as stiffness, strength and toughness, which are sometimes combined with other
properties such as resistance to environmental conditions and high temperature.
Furthermore, the interface is known to have large influence on the mechanical
Resin Transfer Molding Process 123

performance of the composites, especially with regard to strength and fracture

toughness [5]. The final properties of the composite materials are dependent on
several factors including the properties of the phases, size and orientation of the
fibers and fiber content. In addition, their characteristics are highly susceptible to
the manufacturing process whereby they were produced. Therefore, investigation
of the process is of uttermost importance.

1.2 Composite Manufacturing

The Liquid Composite Molding (LCM) family, or simply liquid molding,

comprises a number of manufacturing techniques for composite materials where a
liquid thermosetting resin is injected into a closed mold containing a dry fibrous
perform (the only existing absorbed moisture into the preform is due to water
vapor in the air), impregnating the reinforcement. Some of the most popular LCM
molding techniques are detailed below.

1.2.1 Resin Transfer Molding

Among the several advanced composite materials manufacturing processes, Resin

Transfer Molding (RTM) is considered state-of-the-art technology for producing
textile reinforced composite parts [6]. It was mainly developed for general
aero-space applications in the 1980s, following some limited earlier applications
such as aircraft radomes [7]. Nowadays, RTM is used by many industrial sectors
such as automotive, aerospace, civil and sporting equipment.
RTM processing consists in positioning the preform (i.e., fibrous reinforcement
preform in the shape of the piece) within the mold. This preform is a dry porous/
fibrous medium through which fluid flow occurs. The mold is closed and resin is
injected into the mold cavity until the fibrous reinforcement is fully impregnated.
After resin curing, the mold is opened and the part is removed from the mold. This
part may still require finishing operations and/or to go through a post-curing process.
The RTM mold must have at least one inlet port for resin injection and one outlet to
enable, during resin injection, the output of air from inside the mold. However, in
larger pieces, the presence of multiple inputs and outputs is common [8].
The main advantages of RTM include: non-expensive process equipment,
excellent control on mechanical properties, closed mold process, low filling
pressures, ability to incorporate metal inserts and attachments, possibility of
producing large and complex parts with low labor costs [9], use of different weave
and mat fabrics architectures, high fiber volume content, use of various types of
resins, low injection pressure (10 bar max.), good dimensional tolerances,
relatively low equipment cost, short production cycle times, process automation
and low solvent emission. RTM molding is successfully used in medium-to-high-
volume applications, but the need for tooling of relative high cost and difficulty of
124 F. F. Luz et al.

controlling fiber orientation limits their application in non-structural situations.

RTM was identified with the potential to minimize/eliminate manufacturing
difficulties involved in processing parts with medium and high fiber content, being
a processing route with acceptable cost [10].
The fiber volume content and the fibrous preform architecture influence the
mechanical properties of the composites manufactured by liquid molding. In many
cases, the mechanical properties can be improved by increasing the fiber volume
fraction. Higher fiber volume fraction, however, requires higher injection
pressures, which may adversely affect the properties of the final molded compo-
nent [11]. The injection pressure is between 1 to 10 bar, since higher values can
cause fiber movement inside the mold (i.e., fiber wash), deflection of the mold and
the appearance of voids in the composite. For a good impregnation and to reduce
injection time, the resin must have low viscosity (typically B1 cP).
To avoid defects (i.e., voids due to partial impregnation) in the molded
components, the resin should fill all spaces not occupied by the fibers and the air
within the reinforcement must be totally eliminated [12]. Some works in the
literature discuss a few techniques for void reduction in the composites through the
study of different reinforcement architectures [12, 13] and injection parameters,
especially fluid pressure and front flow orientation [14].
RTM is still considered underutilized and reproducibility of the finished parts [13]
and the need to understand the resin flow through the preform are among the major
barriers. For example, if the input and output channels are not properly positioned in
the mould, some perform areas may not be reached by the resin being injected.
Indeed, understanding of the flow process is of uttermost importance for RTM
and all liquid molding processes. However, it is very difficult to intuitively
determine the behavior of the resin flow, so there is the need to perform a
numerical analysis of the mold filling process, contemplating the development of a
mathematical model and the simulation of this model [15]. The behavior of a fluid
in a porous media has been studied for many years and numerical analysis is
recognized as an effective method to analyze it and to propose optimum processing
conditions [16]. However, to allow a proper prediction of the mold filling, these
simulation codes need correct input values such as mold geometry, fluid viscosity,
reinforcement porosity and reinforcement permeability [17].

1.2.2 Vacuum-Assisted Resin Transfer Molding

In the Vacuum-Assisted Resin Transfer Molding (VARTM) process, dry fabrics

are impregnated with a thermoset resin using vacuum bag and a vacuum pump.
VARTM is widely used for molding of complicated composite structures because
of its high drapability and low cost, being applied to marine vessels and next
generation airplanes [18], refrigerated cargo boxes, etc. [19]. The VARTM
process offers several advantages over the competing composite manufacturing
processes such as: (i) low tooling cost; (ii) low emission of volatile organic
chemicals; (iii) process flexibility; (iv) low void-content in the molded parts, and
Resin Transfer Molding Process 125

(v) potential for the production of relatively large (surface area *150–200 m2)
and thick (0.1–0.15 m) composite parts, containing high reinforcement content
(75–80 wt%) [19].

1.2.3 Infusion

In this process, the reinforcement is laid on the rigid mold with a layer of removable
flow-enhancement medium, used to reduce filling time, along with a peel ply to
facilitate disposal. Inlets and outlets are placed and the vacuum bag is sealed to the
mold using a sealant tape. Air is extracted from the cavity to compact the preform
between the vacuum bag and the mold and to subject the resin to a pressure differ-
ential that drives resin flow through the compacted preform. Infusion is a closed mold
resin infiltration production process that presents many similarities with RTM. The
most striking difference, however, is that the fully enclosing, two-part rigid mold
used in RTM is replaced with a one-part rigid mold sealed with a vacuum bag [20].

1.2.4 RTM Light

In RTM Light, a variant of traditional RTM, a more flexible mold material (e.g., a
composite) and an upper translucent component are used, allowing observation of
the progress of the resin during impregnation. For LCM processes utilizing flexible
molds such as RTM light, local laminate thickness evolution is governed by the
compaction characteristics of the reinforcement, and improved compaction models
are vital for enhanced filling simulations [21]. RTM light uses a peripheral vacuum
to seal the mold, whereas a low positive pressure along with partial vacuum in the
mold cavity drives resin flow inside the cavity industries [22]. RTM Light has been
widely used in automotive, aerospace, civil and sports [9].

1.2.5 Seemann Composites Resin Infusion Molding

In this process, a network of grooves or channels is used to distribute the resin, and
reduce flow resistance and filling time. The resin fills those grooves/channels first by
vacuum pressure and then infuses into the fiber preform. The Seemann Composites
Resin Infusion Molding (SCRIMP) is often used to produce large composites of
complex geometry such as boat hulls and decks [23]. Considering factors like
environmental protection, relatively inexpensive tooling and the ability to fabricate
large parts with complex geometry, the SCRIMP process may be used to replace
traditional open molding processes [24].
Now that we have discussed about composite materials and manufacturing of
these materials, the next step is to address discussions about modeling and
simulation applied to LCM process. Following a mathematical modeling to predict
fluid flow and heat transfer inside porous media with particular reference to RTM
process on an overall basis is presented.
126 F. F. Luz et al.

2 Multiphase Flow and Heat Transfer Theory

in RTM Process

2.1 General Mathematical Modeling in Porous Media

2.1.1 Macroscopic Conservation Equation

Non-isothermal fluid flow in porous media can be modeled on the basis of fluid
velocity, pressure and temperature into the media. For very small velocity, many
researchers have utilized the so called Darcy’s model, which will be presented and
discussed later.
In macroscopic scale, the porous model is a generalization of both, the Navier–
Stokes equations and Darcy’s law. The model retains both advection and diffusion
terms and can therefore be used for flows in which such effects are important. In
deriving the continuum equations, it is assumed that control volumes are infinitesi-
mal and that surfaces are large in relation to the interstitial spacing of the porous
medium but small relative to the scales in study. Thus, given control cells and control
surfaces are assumed to contain both solid and fluid regions. This is a concept of
continuum mechanics. This way, the particular details of the porous structure such as
dispersion, tortuosity and interfacial transfer between phases at the pore level are
neglected.
Dispersion is the transport due to hydrodynamic mixing of the interstitial fluid
at the pore scale (spreading phenomena). At this point, it is important to distin-
guish between dispersion and diffusion. Diffusion occurs due to random molecular
motion whereas dispersion is caused by fluctuations of the bulk flow. Then,
dispersion occurs when convection occurs [25]. Discussion about this topic can be
found in related works [25–29].
Tortuosity is a very important property of a porous medium. It is defined as the
ratio between the straight line distance and the curved path length between two
points [25]. Allen III et al. [27] defined this parameter as the average ratio between
the distance traveled by the fluid in the microscopic pores of the medium and the
net macroscopic distance traveled. Other definitions of tortuosity are given by
Vallabh et al. [30] and Epstein [31]. According to the former author, higher
tortuosity value indicates longer, more complicated and sinuous path, resulting in
greater resistance to fluid flow.
In this context, by assuming that the continuum approach provides the necessary
information, the general scalar advection–diffusion equation in a porous medium is
given by [32]:
oðq/UÞ
þ r  ðqK  UUÞ  r  ðCU K  rUÞ ¼ /S ð1Þ
ot
where: t is the time, q is the density, U is the velocity vector, U is the potential of
transport, CU represents effective transport property and S the source term.
Resin Transfer Molding Process 127

In Eq. (1), the volume porosity, /; at a point is the ratio between the volume
0
V available to flow in an infinitesimal control cell surrounding the point, and the
physical volume, V; of the cell. Hence:
0
V ¼ /V ð2Þ
Nield and Bejan [33] and Bear [26] defined porosity as the fraction of the total
volume of the porous media that is occupied by pores. However, disconnected
pores can exist in the material, therefore the term effective porosity has been
defined as the ratio between the volume occupied by connected pores and overall
volume of the porous material.
0
In this formulation, it is assumed that the vector area available to flow, A ;
through an infinitesimal planar control surface of vector area A is given by:
0
A ¼ K:A ð3Þ
where: K is a symmetric second rank tensor called the area porosity tensor. The
inner product of a symmetric rank two tensor and a vector produces a vector.
In addition to the usual production and dissipation terms, the source term S may
contain transfer terms from the fluid to the solid parts of the porous medium.

2.1.2 General Initial and Boundary Conditions

For a problem that involves partial differential equations to be well-posed, the

solution to the problem must exist, must be unique and must depend on the initial
and/or boundary data [34, 35]. In general, the following initial and boundary
conditions can be specified for the unknown U.

Initial conditions

U ¼ f1 ðD; t ¼ 0Þ ð4Þ

where D is the studied domain.

Boundary conditions

• Prescribed value (Dirichlet or first kind condition)

U ¼ f2 ðSÞ ð5Þ
where S is the boundary of the domain.
• Prescribed flux (Newmann or second kind condition)
128 F. F. Luz et al.

oU
¼ f3 ðSÞ ð6Þ
on
oU
¼ f4 ðSÞ ð7Þ
os
where n and s represent the normal and tangential directional vectors to the
boundary S, respectively.
• Mixed (Robin or third kind condition)

oU
f5 ðSÞ þ f6 ðSÞU ¼ f7 ðSÞ ð8Þ
on
where f5 ðSÞ 6¼ 0 and f6 ðSÞ 6¼ 0 on the boundary S.
The choice of one or more conditions is dependent on the problem to be solved.
Solution of a partial differential equation applied to physical problems requires
physically realistic initial and boundary conditions. Fluid flow equations in porous
media comprise a complicated formulation that involves elliptic, parabolic and
hyperbolic terms. These different behaviors affect the way like the initial and
boundary conditions in the domain must be specified.
For a physical problem where fluid-fluid (liquid–gas) interface exists, the
momentum flux and hence the velocity gradient in the liquid is negligible com-
pared to that in the gas phase (because liquid viscosity is much greater than that of
the gas) [36]. In addition, in transient problems the initial condition, i.e., the values
of the variables, must be defined at the start of the process (t = 0).

2.2 Heat Transfer and Fluid Flow in RTM

The reinforcement (porous) media inside the mold is initially filled with air. Resin
is injected into the mold through one or more points (inlets) and air is forced out of
the mold through one or more outlets. In standard RTM, the use of a single resin
inlet and a single air/resin outlet is common. Resin injection through the rein-
forcement media can be mathematically described as a viscous fluid moving
through a porous medium. The problem consists of two inviscid fluids (resi-
n ? air) flowing through a porous media contained inside the mold cavity. Dif-
ferent approaches can be used to model the problem. In the present work, three
formulations will be discussed and compared:
1. Multifluid model;
2. Control Volume/Finite Element (CV/FE) based model;
3. Volume of Fluid (VOF) model.
Resin Transfer Molding Process 129

2.2.1 Multifluid Model

Fluid Flow Model

From the general equation, the following macroscopic equations can be

derived:
1. Mass conservation equation:
o
ð/qÞ þ r  ðqK  UÞ ¼ 0 ð9Þ
ot

2. Momentum conservation equation:

oðq/UÞ   
þ r  ½qðK  UÞ  U  ¼ r  le K  rU þ ðrUÞT
ot ð10Þ
/R  U  /rp þ B
In these equations, U is the true velocity, B is the volumetric body force, p is the
pressure, le is the effective viscosity—either the laminar viscosity or a turbulent
quantity, and R represents the resistance to flow in the porous medium. This is, in
general, a symmetric positive definite second rank tensor, in order to account for
possible anisotropies in the resistance.
Modeling of the fluid flow strives to obtain the velocity, pressure and volume
fraction (saturation) distribution of the fluid phases within the porous medium. If
there is a large resistance to flow, a large pressure gradient must be set up to
balance that resistance. In an extreme situation, both terms on the right-hand side
of Eq. (10) are large and of opposite sign, whereas the convective and diffusive
terms on the left-hand side are negligible. Neglecting body force effects, Eq. (10)
reduces to Eq. (11). Hence, in the limit of large resistance, an anisotropic version
of Darcy’s law is obtained, with permeability proportional to the inverse of the
resistance tensor.

U ¼ R1  rp ð11Þ

Heat Transfer Model

Heat transfer into the porous media can be modeled with an equation similar to:
o
ð/qHÞ þ r  ðqK  UHÞ ¼ r  ðke K  rTÞ þ /SH ð12Þ
ot
where: H is the specific enthalpy, T is the temperature, ke is an effective thermal
conductivity and SH contains a heat source or sink to or from the porous medium.
130 F. F. Luz et al.

A heat transfer model attempts to obtain the temperature distribution within the
porous medium. It is important to understand the temperature profile during RTM
process for different reasons: (a) temperature for resin cure is elevated, (b) the
curing process is usually very exothermic, (c) resin viscosity is dependent on the
temperature thus, influencing mold filling time, and (d) depending of the nature of
the preform (e.g., vegetal fibers are very sensitive to heat), high temperature may
affect the final composite.
Momentum sources can be used to model isotropic losses in porous regions,
directional losses in porous regions, or other processes. For a better understanding,
consider the general momentum equation (Eq. 8) for a fluid domain in the form:
~
oðqUÞ ~ UÞ
þ r  ðqU ~ ¼ rp þ qg þ r  s þ SM ð13Þ
ot
where s represents the stress tensor, and g is the gravity acceleration vector.
The momentum source, SM ; can be represented by:
~  CQR jUj
SM ¼ CLR U ~ U~ þ Sspec ð14Þ

where: CLR is a linear resistance coefficient due to viscous losses, CQR is a qua-
dratic resistance coefficient referring to inertial losses; Sspec contains other
momentum sources (which may be directional) and U ~ is superficial velocity
(Darcy velocity). The relationship between true (interstitial) velocity and super-
ficial velocity, at any point in the porous medium, is given as follows:
~ ¼ /U
U ð15Þ
True velocity corresponds to the mean fluid velocity through the pores whereas
superficial velocity is the velocity that a fluid would exhibit if no reduction in
cross-sectional area of flow occurred due to the presence of the solid material.
Usually, the superficial velocity is expressed by the ratio between volumetric flow
rate and cross-sectional area of the porous medium. The particulate material within
the porous medium reduces the area available for fluid flow, therefore true velocity
is greater than superficial velocity.
On the basis of Eq. (11), Darcy’s law is only valid when the pore Reynolds number
is of the order of 1 [37]. When the microscopic Reynolds number is greater than that
(flow occurs at high velocity), Darcy’s equation is supplemented on the right-hand
side by a quadratic non-linear term corresponding to form drag within the medium to
correct the advection term (Forchheimer term) [25, 33, 38–40]. Other versions of the
Darcy’s equation exist; for example in Liu and Masliyah [25], Nield and Bejan [33],
Magyari et al. [38] and Hsu [40].
The following generalized form of Darcy’s law has been reported in the
literature:
l ~ q ~ ~
rp ¼ U þ KEL jUj U ð16Þ
KP 2
Resin Transfer Molding Process 131

where: l is the dynamic viscosity; KP is the permeability and KEL is the empirical
loss coefficient.
Permeability is a measure of how easily a single-phase fluid moves through a
porous medium under the influence of a dynamic pressure gradient [41]. Perme-
ability, defined by Ye et al. [42] as a measure of the resistance to fluid flow into a
porous material, can be obtained through experiments or analytical predictions and
different equations to calculate this parameter have been report. According to Hsu
[40], Nield [43], Lage [44], Chen and Hsiao [45], Vafai and Amiri [46], and Nield
and Bejan [33], permeability may be obtained as follows (Kozeny equation):

/3 dp2
KP ¼ ð17Þ
að1  /Þ2
^

where: dp is particle diameter and a ^ is a constant to characterize the microscopic

geometry of the porous material (morphology of the pores). When the porous
medium is made of particles or fibers, ^ a is usually set to 180 [47].
When two or more distinct fluid phases (in contact) simultaneously flow
through a porous medium, the medium exhibits distinct permeability for the
different fluids. This dependence reduces the effective permeability of each phase
depending on the degree to which each fluid phase affects the flow of the other.
Thus, fluid permeability is different of the porous medium permeability, intro-
ducing the concept of relative permeability. This parameter, which a value
between 0–1, appears in Darcy equation multiplying the porous medium perme-
ability and therefore alters this equation. More details about this topic can be found
in Allen III et al. [27].
According to the cited literature, the empirical loss coefficient in Eq. (16) can be
obtained as follows:
2CFD
KEL ¼ pffiffiffiffiffiffi ð18Þ
KP
where: CFD is a dimensionless form-drag constant.
Comparing Eq. (14) with Eq. (16), the following coefficients can be obtained:
l
CLR ¼ ð19Þ
KP
KEL q
CQR ¼ ð20Þ
2
Equations (19) and (20) use superficial velocity. However, data is often
expressed in terms of true velocity and, in this case; the coefficients would be
better represented by:
l
CLR ¼ ð21Þ
/KP
132 F. F. Luz et al.

KEL q
CQR ¼ ð22Þ
2/2
Considering KEL ¼ 0 and comparing Eq. (16) with Eq. (11), one can write:
l
R¼ ð23Þ
KP
when superficial velocity is used in Eq. (16), and
l
R¼ ð24Þ
/KP
when true velocity is used.
In practical terms, heat and mass transfer phenomena dominate resin impreg-
nation and curing (exothermic chemical reaction) in RTM. Therefore, the process
can be treated as non-isothermal reactive flow [48]. In Eq. (12), the source term
represents the energy generation by volume unit that occurs into the porous
material, e.g., heat release due to the curing process. Sometimes, impregnation of
resin into the porous material is comparatively fast and resin curing has not yet
started to alter the temperature of the fluid and, in such case, the RTM process may
be considered isothermal.
In this chapter, the medium is considered rigid, i.e., relative positions of solid
structure are constant and dilation or contraction within the porous material is
negligible. Besides, no sink phenomenon (absorption of liquid by the pores of the
fiber) occurs within the porous medium. More information about sink models can
be found in Parnas and Phelan [49] and Pillai [50].
The transient formulation presented here can be applied to 1D, 2D and 3D situ-
ations and also to all phases present into the preform, for example, resin and air. Thus,
there is a local thermal non-equilibrium condition between the phases. This math-
ematical treatment yields more accurate results than the homogeneous treatment
under local thermal equilibrium conditions. More discussion on local thermal
equilibrium condition can be found in previous works [25, 33, 40, 46, 51–59].
In addition, since the volumetric heating rate release due to curing is a function
of the degree to which the chemical reaction occurred, a curing model is required
for a more complete formulation and to obtain more realistic results. Discussion on
the cure kinetic model can be found in Advani and Hsiao [48].

2.2.2 Control volume/Finite element Based Model

In this formulation, the problem is simplified by only considering the resin. The
influence of the air flow is assumed negligible and is not taken into account in the
model. Another simplification regards the fact that the transient problem of
determining resin position as a function of time inside the mold is obtained by
successive solutions of a simpler steady-state problem.
Resin Transfer Molding Process 133

The resin velocity is formulated using Darcy’s Law, which can be mathematically
described as shown in Eq. (11). Typical polymeric resins used in the RTM process
show a non-Newtonian behavior. However, for the purpose of this chapter
(description of the basic formulation of the problem), a Newtonian approach is
suitable, being, in fact, largely used in the numerical investigation of RTM [60–62].
Thus, assuming a Newtonian incompressible fluid we can write:
rU ¼0 ð25Þ

Combination of Eqs. (11), (23) and (25) results in

 
KP
r rP ¼ 0 ð26Þ
l
The boundary conditions to be used with Eq. (26), (see Fig. 1), are given by
1. p = P0 at the injection point;
2. op=on ¼ 0 at the mold walls (n is the direction normal to the wall), and
3. p = Pf at the resin front line, where Pf is the front line preset pressure which is
usually zero (manometric).
The numerical solution of Eq. (26) is quite simple, but it must be solved many
times during the simulation. Since Eq. (26) does not include a transient term, the
transient problem is solved by finding a steady state solution for this equation for
each time step. The solution procedure for a rectilinear flow can be summarized in
five steps schematically shown in Fig. 1. The methodology used to track the resin
front position is presented in Phelan [63], being called Flow Analysis Network
(FAN).
The first step in solution consists on the discretization of the computational
domain. Figure 1a shows a simple computational domain discretized with irregular
quadrangular elements. In this method, besides the elements, it is also necessary to
create control volumes (detailed in Fig. 1a) around the nodes. The control volumes
are used to evaluate resin flow advance.
The second step consists in using Eq. (26) to calculate the pressure gradient
between the inlet section and the flow front volumes. As shown in Fig. 1b, for a
partially filled domain, a prescribed pressure P0 is assigned to the inlet section (left
wall) and p = Pf is set to all elements not yet filled with resin. Neumann boundary
condition (op=on ¼ 0) is assigned to all others frontiers of the computational domain
(top, bottom and right wall). The gray gradient in Fig. 1b is a schematic represen-
tation of the pressure field obtained with the solution of Eq. (26). Figure 1b also
shows the flow front volumes and the adjacent (to the front) volumes.
Solution of Eq. (26) is usually obtained with a finite element method [10, 64].
However, other methods like finite volume and finite differences are also used.
Detailing of such solutions will not be included here, but can be found in refer-
ences [10, 64–67].
In the third step, the pressure gradient determined in step two is used to
calculate the velocity field. In the simplified solution presented in Fig. 1c, the
134 F. F. Luz et al.

Fig. 1 CV/FE solution. a Step 1 Discretization b Step 2 Pressure field calculation c Step 3 Velocity
field calculation d Step 4 Filling time calculation and e Step 5 Determination of the new flow front
Resin Transfer Molding Process 135

velocity field is parallel to the x axis, thus only the u component of the velocity
vector is presented, however depending on the problem and geometry, both u and v
components of U are calculated with Eq. (11). It is also possible for the resin to
enter the control volume by two or more faces and leave the control volume
through one or more faces when unstructured grids are used. The same analysis is
valid for 3D solutions, though all three velocities components (u, v and w) must be
considered in the formulation. In Fig. 1c, only the velocity vectors that are
entering the volumes adjacent to the flow front are shown. These velocities are
needed in the calculations, but it is possible to use Eq. (11) to calculate the flow
velocity for all volumes.
The filling time for all volumes adjacent to the flow front is calculated in the
fourth step (Fig. 1d). With the velocity field U and the volume V of each control
volume, the filling time of volume i can be calculated by

Vi  Vif ðtÞ
Dti ¼ ð27Þ
V_ ðtÞ

where: Vi is the overall volume of the control volume i, Vif ðtÞ ¼ f V i ðtÞ is the filled
volume at time t (current time step), and V_ t ðtÞ ¼ Unormal A is the volumetric flow
rate into volume i. Additionally, Unormal is the velocity component in the direction
normal to the face with area A.
The advance in the flow front position can be calculated by determining the
control volume that will be the first to be completely filled with resin. This can be
done by using Eq. (27) to identify the control volume with minimum Dt: Calcu-
lation of Dtmin guarantees that at least one control volume is completely filled and
that none of them will overflow (to receive more resin than its volume) in every
solution step.
All other volumes adjacent to the flow front will be partially filled. The filling
factor of these volumes for the next time step is calculated by

Vif ðtÞ  Dtmin V_ i ðtÞ

fi ðt þ DtÞ ¼ ð28Þ
Vi
Finally, in the fifth step, the new flow front position is determined. All adjacent
volumes (Fig. 1e) with f = 1 will be considered as flow front volumes in the next
time step. Figure 1e shows a particular case where only one new volume is added
to the flow front and the flow front line (dashed line) moves to include this volume.
The CV/FE formulation is quite easy to understand and of simple computa-
tional implementation. It has a major advantage in the fact that only one differ-
ential equation (Eq. 26) must be solved. Eq. (26) is linear and, even though it may
present problems related to the discretization of complex geometries, its solution is
fast and stable even for coarse grids. However, accuracy in flow front position and
filling time predictions will increase for a larger number of elements in the
discretization.
136 F. F. Luz et al.

The fast solution of Eq. (26) will not always represent an overall fast solution of
the problem. Depending on the discretization, it is possible that only one new
volume is completely filled in every time step. When this happens, the number of
time steps needed to solve the problem will be equal to the number of volumes in
the grid (Nv), implying that Eq. (26) will be solved Nv times. In an attempt to
increase the number of volumes completely filled in every time step, it is
recommended to use grids with elements of the same size and, if possible, aligned
with the resin flow. The use of rectangular (2D) or hexahedral (3D) elements is
also recommended.

2.2.3 Volume of Fluid Model

In general, the VOF method [68] can be used to model multiphase flows with two
or more inviscid fluids. In this formulation, all phases are well defined and the
volume occupied by one phase cannot be occupied by the others. The volume
fraction concept is used to represent the existence of different phases inside each
control volume. All fractions are assumed continuous in space and time and their
sum, inside each control volume, is equal to one. If the volume fraction of a
particular phase inside a control volume is denoted as fi, then the following three
conditions are possible:
1. fi = 0: the cell is empty with fluid of phase i;
2. fi = 1: the cell is full with fluid of phase i;
3. 0 \ fi \ 1: the cell contains the interface between phase i and one or more
other phases.
For the particular case of modeling resin and air flowing through a porous
media (e.g., in RTM), only two phases are considered in the formulation. In the
VOF method, only a single set of momentum and continuity equations is applied to
both fluids, and the volume fraction of each fluid in every computational cell
(control volume) is tracked throughout the domain. The model is composed of
continuity, volume fraction and momentum equations as follows:
oq
þ r  ðqUÞ ¼ 0 ð29Þ
ot
oðqf Þ
þ r  ðqfU Þ ¼ 0 ð30Þ
ot
oðqU Þ
þ r  ðqU Þ ¼ rp þ r  s þ qg þ B ð31Þ
ot
where: f is the volume fraction of resin and s is the stress tensor.
Since a single set of equations is used for both phases, average properties for q
and l need to be defined. These properties can be approximated as [69]
Resin Transfer Molding Process 137

q ¼ fqresin þ ð1  f Þqair ð32Þ

l ¼ flresin þ ð1  f Þlair ð33Þ

The porous media (reinforcement) resistance to the flow is modeled by adding a
source term to the standard momentum equation such as
l
B=  U ð34Þ
KP
Combining Eqs. (31) and (34), and considering that KP is very small
(*1 9 10-10), it may be inferred that B is very large, allowing simplification of
Eq. (31) as shown in Eq. (35). This equation may be written as Eq. (11) which is
actually the Darcy’s Law.
l
rp =  U ð35Þ
KP
In contrast to the CV/FE formulation where only one partial differential equation
must be solved, in the VOF model, four partial differential equations (five in a 3D
formulation) must be simultaneously solved. Due to that, solution normally suffers
from convergence issues, what brings grid refinement and time discretization
constraints.
An important characteristic of the VOF method is that it can be used to solve
problems whose computational domains are divided into two or more regions,
some of them filled with the reinforcement and the others are empty. This is
possible by setting the permeability close to infinity (high value) in Eq. (35) for the
regions without reinforcement. This will result in a resistance term B in Eq. (31)
equal to zero. This kind of solution is of particular interest for the solution of
RTM-Light problems, due to the presence of empty channels, and cannot be
performed with the CV/FE method.

2.3 Preform Geometry and Grid Generation

When numerical methods are applied to solve a particular physical problem, it is

necessary to define the grid (mesh), i.e., the numerical representation of the
physical domain in the computational domain where the conservation equation can
be applied. When the domain presents a complex geometry, grid generation can be
a very time-consuming task. However, a well-constructed grid improves the
quality of the results.
According to Shafer [70] and Ferziger and Perić [71], the grid can be classified
as: boundary-fitted, Cartesian and overlapping grid. Another well known classi-
fication is based on the logical arrangement of the control-volume as structured
and unstructured grids (Fig. 2) [34, 70, 71]. Structured grids are characterized by a
regular arrangement of grid control-volumes. For unstructured grids, generally
138 F. F. Luz et al.

Fig. 2 Types of grids:

a structured, b unstructured
and c hybrid

used for very complex geometries, there is no regularity in the arrangement of the
control-volumes (cells). Methods of grid generation can be found in related books
[71, 72]. The grid can present uniform or non-uniform size distribution. For higher
accuracy of a particular physical problem in regions where a large gradient is
expected, the non-uniform grid size distribution is more effective [36].
For liquid composite molding, flow is transient, and mold and preform may
present a very irregular shape. In this case, an adequate grid is essential to the
accuracy of the solution, mainly when we apply a three-dimensional treatment. In
addition, depending on the specific liquid molding technique and the particular
mold design and construction being investigated, the mold may display defor-
mation (change in relative position between the upper and lower parts of the mold)
during infiltration. This brings significant challenges related to grid generation
along the transient process and, because of that, researches usually consider the
mold and grid to be rigid during RTM process. In spite of the importance, just a
few works on this the research area are available [73–75].

3 RTM Applications: Numerical Computation

and Experiments

3.1 RTM Experimentations

Luz [76] and Luz et al. [77] conducted several experiments of resin injection into a
parallelepipedic porous preform (300 9 300 9 2 mm). The Newtonian fluid used
in the experiments was a commercial vegetable (soybean) oil with density of
914 kg/m3. The fluid viscosity at 23 C (same temperature of the experiments),
measured in a Brookfield viscometer HBDV-II ? C/P with the S40 spindle, was
37.1 cP. As the fibrous reinforcement, plain-weave (0/90) E-glass fiber cloth from
Owens Corning (300 g/m2) was used.
Resin Transfer Molding Process 139

Fig. 3 Radial RTM system available at UFRGS (Brazil): a pressure pot, b strengthened-glass
top mold, c steel bottom mold, d pressure controller, e pressure transducers, f data acquisition
system and g camera

Fig. 4 Geometry of the injection mold with the dimensions of the studied area of the mold
(300 9 300 9 2 mm

The radial experiments were conducted in a RTM system (shown in Fig. 3)

with one injection point and four outlet points (vents) (Fig. 4). For radial infil-
trations, data comprising the time required (tff) for the fluid which passes through
an inlet to fill a region of radius rff in the mold may be used to calculate perme-
ability of the medium according to Eq. (36). This equation may be applied only
until the flow reaches the wall of the mold (twall).
140 F. F. Luz et al.

Table 1 Permeability and mold filling time for several experiments

Case Vf (%) Vf (%) Pinj (bar) Pinj (Pa) KP (9 10-11 m2) twall (s)
(pre-set) (real) (pre-set) (real)
P01V20 20 24.0 0.10 7420 136.8 130
P01V30 30 30.3 0.10 7780 30.9 420
P01V40 40 43.4 0.10 7460 4.4 2280
P03V40 40 41.7 0.30 28570 3.0 860
P05V40 40 41.6 0.50 48260 2.5 700
P07V40 40 42.2 0.70 67240 2.2 640
P09V40 40 41.7 0.90 88450 2.0 560
P10V40 40 41.9 1.00 99800 3.7 380

 
/l 2 rff 1
K¼ r ln  r2ff  r2inj ð36Þ
2Pinj tff ff rinj 2

where: rinj is the radius of the injection port, / is the porosity of the porous fibrous
media (/ = 1-Vf, Vf is the fiber volume fraction) and Pinj is the pressure (gauge)
in the resin injection gate. During infiltration, position (radius) of the flow front
was recorded with a camera at different times.
The injection pressure of the resin in the mold was monitored with pressure
transducers during each experiment. Several experiments were performed by
varying the volumetric fiber content and injection pressure. The conditions of each
experimental run, and the observed mold filling time and the calculated perme-
ability data (Eq. 36) are presented in Table 1.

3.2 RTM Simulations

In the LCM process, simulations are being used mainly to verify the trial-and-error
approach, which is still prevalent in the manufacturing industry. However, as the
composite structures being manufactured by LCM become larger and more
complex, use of process simulation may aid in improving the process design and
increase the yield by counterbalancing any unforeseen disturbances that may arise
during the impregnation phase [48].
For the application of the models and equations presented, it is necessary the
use of a software to solve them. In modeling by RTM, simulation software are
used to (i) predict the resin injection profile and then be able to evaluate the filling
time, that should remain below the resin curing time, (ii) predict impregnation
deficient (dry) points, and (iii) determine the more suitable injection and outlet
points of air/resin.
Flow simulation software may be used, increasing the confidence in the mold
design and fabrication process parts. There are different software dedicated
exclusively to study RTM process such as PAM-RTM from ESI Group, the
RTM-WORX from Polywork, the LIMS from the University of Delaware that are
Resin Transfer Molding Process 141

Fig. 5 a The grid used in the ANSYS CFX simulations, b Details of the outlet port

very commonly used by industry for having a simple usage and focused only on
this process. However, dedicated software may end up limiting some more pro-
found studies on the process behavior in adverse situations. In fact, it is possible to
use commercial software, not only dedicated to the study of composite processing,
but computational tools that allow customization of the problem being studied.
However, adapting such software to solve specific problems of RTM can be quite
laborious. The non-dedicated commercial software commonly used to study CFD
are Ansys CFX and FLUENT, both from ANSYS and Abaqus CFD from Simula
Abaqus.
Herein, RTM modeling was applied to the manufacturing of a composite with
paralellepipedic shape by considering the isothermal system, i.e., no heat transfer.
The geometry and dimensions of the mold are shown in Fig. 4. For the 3D
numerical analysis, it was used two commercial softwares: ANSYS CFX (non
RTM-dedicated) and PAM-RTM (RTM-dedicated).

3.2.1 ANSYS CFXÒ Model

In the ANSYS CFX model, the system is composed of air, resin and solid
material (glass fiber). Simulations were performed in a structured 3D mesh with
17,532 elements and 23,876 nodes. It was built using ANSYS ICEM CFD software
version 12.0.1 and time step 0.05 s. Figure 5 illustrates the mesh used in the
142 F. F. Luz et al.

Table 2 Constants used in Eq. (37)

Case te (s) a1 (mbar s-0.25) a2 (mbar) a3 (s-1) a4 (-) a5 (s-1)
-4
P01V20 40 13.61 6.01 9 10 7.09 1.29 0.62
P01V30 85 20.21 1.62 9 10-4 6.76 1.55 0.56
P01V40 35 25.21 2.74 9 10-5 8691.63 1232.42 637.45
P03V40 132 29.19 1.30 9 10-1 7.02 9.93 0.87
P05V40 168 42.05 3.68 9 10-2 1.42 1.39 0.14
P07V40 240 30.41 5.29 9 10-1 -72.28 -178.62 -9.67
P09V40 294 45.20 0.18 -3099.14 -6960.16 -352.70
P10V40 325 45.87 3.99 4.54 32.35 0.75

Fig. 6 The grid used in the PAM-RTM simulations

simulations. From the collected pressure data, a non-linear regression was made
using the Mathematica software, yielding a pressure equation as a function of
the process time (Eq. 37):
(
P0 þ a1 t0:25 þ a2 exp a4aþa
3t
; for 0  t  te
Pinj ðtÞ ¼ 5t ð37Þ
Pe ; for t [ te

where: P0 is the atmospheric pressure, Pe is the final pressure (stable pressure)

achieved in the experiment and te is the instant of time where P is now considered
Pe. The parameters of Eq. (37) are presented in Table 2.

3.2.2 PAM-RTM Model

In the PAM-RTM model, the system is composed of resin and solid material (glass
fiber). It was used a 2D mesh with 14,819 elements and 16,775 nodes. Figure 6
illustrates the mesh used in the simulations. PAM-RTM, unlike the ANSYS CFX,
is a commercial software dedicated to the study of RTM process, so it was not
Resin Transfer Molding Process 143

Table 3 Boundary conditions used

Model Boundary fa fr P (mbar) Ua Ur
(m/s) (m/s)
Multifluid Inlet 0 1 P(t), Eq. (37) – –
(ANSYS CFX) Outlet – – 1013.25 – –
Walls – – – 0 0
CV/FE Inlet – 1 P(t), Collected data table – –
(PAM-RTM) (manometric)
Flow front – – 0 (manometric) – –
Walls – – oP – –
on ¼ 0

fa and fr are the volumetric fractions of air and resin, respectively; Ua and Ur velocity vector of air
and resin, respectively; P is the static pressure at the inlet and outlet boundaries of the preform;
n is the normal direction to the boundaries

necessary make adjustments and create regression equations for the pressure data.
The collected pressure data was directly implemented in the software.

3.2.3 Initial and Boundary Conditions

In all simulations, the preform was initially considered as having pressure (Pi),
temperature (Ti), and air saturation (Sai) homogeneously distributed throughout the
mold with the following values: Pi = 1013.23 mbar, Ti = 300 K and Sai = 1.0
(consequently, resin saturation is Sri = 0.0). The adopted boundary conditions are
presented in Table 3.

3.3 Results Analysis

The variation of the volume fraction of fluid in the experiments and simulations in
ANSYS CFX and PAM-RTM was analyzed. Figure 7 shows a comparison of the
evolution of the volume fraction of fluid in the case P09V40 obtained by exper-
iment and numerical simulation generated by the ANSYS CFX and the PAM-
RTM at three different times 60, 150 and 360 s.
It is observed that the flow front moves in radial direction in the experiments
and in the simulations the same pattern is observed, but the impregnated region is
further defined and there is a regular circle. This is because in the simulations is
considered a homogeneous distribution of fibers in the cloth, and constant prop-
erties (permeability, porosity and viscosity), which are not always the cases,
experimentally. Where there is a local variation of the concentration of fibers
generates a forward heterogeneous flow. This problem can be corrected by
including in the simulation permeability as a function of saturation resin, for
example, or yet porosity as a function of position into the preform.
144 F. F. Luz et al.

®
Experimental ANSYS CFX PAM -RTM
(a)

(b)

(c)

Fig. 7 Evolution of the fluid flow front (resin) in the case P09V40 in a 60 s, b 150 s and c 360 s

It is shown in Fig. 8 a comparison between the rff experimental, the rff predicted
by ANSYS CFX and PAM-RTM at the moment twall for the experimental cases
P07V40, P09V40 and P10V40. Table 4 presents a comparison between the twall
experimental and twall predicted in ANSYS CFX and PAM-RTM for all the
simulated cases, showing their percentage deviations compared to twall
experimental.
Comparing the twall, it is observed that there is no errors exceeding 10 %,
indicating a strong relationship between the numerical solution of the equations of
porous medium and the experimental procedure. Considering the cases P01V20,
P01V30 and P01V40 observes a relation between the error and the change in Vf,
because as Vf is increased while keeping Pinj constant, the error is increased. This
behavior is not observed in other cases, when increasing the Pinj and Vf remains
constant.
As seen in Table 1, the KP highest values are found in these same three cases,
thus the deviations of comparing the numerical simulations may be related to an
experimental errors, as in the KP determining, which is a crucial property to the
success of the numerical simulation. Therefore, the deviations indicated in Table 4
can be linked to the experimental procedure and, for example, the following
Resin Transfer Molding Process 145

®
Experimental ANSYS CFX PAM -RTM
(a)

(b)

(c)

Fig. 8 Comparison between rff experimental and rff simulated by ANSYS CFX and PAM-RTM
at the moment twall experimental for the cases a P07V40, b P09V40 and c P10V40

Table 4 Comparison between the twall experimental and twall simulated in ANSYS CFX and
PAM-RTM for all the simulated cases
Case Experimental ANSYS CFX PAM-RTM
twall (s) twall (s) Deviation (%) twall (s) Deviation (%)
P01V20 130 138 6.15 131 0.77
P01V30 420 450 7.14 430 2.38
P01V40 2280 2506 9.91 2389 4.78
P03V40 1000 1098 9.80 1031 3.10
P05V40 700 768 9.71 722 3.14
P07V40 640 700 9.38 648 1.25
P09V40 560 588 5.00 568 1.43
P10V40 380 405 6.58 378 -0.53

sources of potential experimental errors, may be mentioned (i) the lack of

homogeneity in the cloth properties, (ii) the operator sensitivity to start the time
measurements and (iii) the readings of the radius of the flow front. The deviation
146 F. F. Luz et al.

of the reading of PAM-RTM twall for the case P10V40 accused a value below zero,
indicating that this time was less than the experimental time, as this was the only
case where this occurred, it is evident that these errors are related to experimental
errors.
By observing the deviations shown in Table 4, it is noted that they are higher in
ANSYS CFX simulations in all cases, ranging from 5.0 to 9.91 %, while in the
PAM-RTM these deviations were in the range -0.53 and 4.78 %. In both, the
higher deviations was found in the case of higher twall, the P01V40 case, as
mentioned before, is the case with most experimental error. Moreover, since the
slightest deviation found for ANSYS CFX was in the case P09V40 and for PAM-
RTM in the P10V40 case, there is no apparent relationship of deviation and the
time of molding.
Figure 9 shows the pressure distribution along the mold in the instant twall for
the case P09V40 simulated in ANSYS CFX (Fig. 9a) and simulated with the
PAM-RTM (Fig. 9b). As expected, the maximum pressure is located in the fluid
injection port, and decreases as it moves away from this point, establishing into a
regular field of pressure. In the simulation by using ANSYS CFX, the total
pressure corresponding to the sum of Pinj plus the P0, so the high pressure value
indicated in Fig. 9a; in PAM-RTM, the pressure gauge is displayed. The pressure
field seen in a transverse plane is shown in Fig. 10 generated by the ANSYS
CFX. It is noteworthy that this level of detail can not be observed in the 2D
model of the PAM-RTM. The pressure in most part of the mold is 1.01 bar,
42.19 % less than the Pinj (1.89 bar) showing quantitatively the pressure drop over
the mold.
Comparing the images generated by ANSYS CFX and by PAM-RTM in all
images presented above, is observed a better quality in the ANSYS CFX images,
indicating a concern in smooth the variation of values in the range of colors on this
software. Therefore, analyse the results is easier because it’s possible to observe
more details in these regions, especially in the analysis of fluid volumetric fraction.

4 Concluding Remarks

In this chapter advanced topics related to composite manufacturing by RTM

process, including multiphase mathematical modeling (fluid flow and heat transfer
in porous media) and computer simulation is presented and discussed. Appropriate
numerical analysis on the basis of the ANSYS CFX and PAM-RTM commercial
softwares, and experiments of RTM process with radial infiltration of resin into a
mould is performed. Different examples dealing with 2D and 3D treatment illus-
trated the effectiveness, versatility and performance of the commercial codes to the
solution of RTM practical problems. Since the formation and migration of gas
bubbles near the flow front occurs in practical situation of RTM process, the
mathematical modeling presented herein predict the distribution of saturation,
pressure and velocity of the fluid phases (resin and air) inside the mold in any
Resin Transfer Molding Process 147

Fig. 9 Total pressure field from the case P09V40 in twall simulated in a ANSYS CFX and
b PAM-RTM

Fig. 10 Total pressure field simulated in ANSYS CFX in a transversal plane shown the
injection port for the case P09V40

moments. According to observations, the ANSYS CFX software requires greater

training in their applications more than the PAM RTM software.
The fluid flow governing equations applied to porous media proved again to be
suitable to the study of the RTM process. The developed numerical analysis
showed to be very reliable, with errors below 10 % in the measurement of the flow
front radius for the various studied cases. The numerical simulation enabled
greater understanding of the behavior of the fluid flow during the RTM process,
allowing the observation of events that are not possible only by performing lab-
oratory experiments.
148 F. F. Luz et al.

Acknowledgments The authors thank to the Brazilian agencies CNPq (Conselho Nacional de
Desenvolvimento Científico e Tecnológico) and CAPES (Coordenação de Aperfeiçoamento de
Pessoal de Nível Superior) for supporting this work, and are also grateful to the authors of the
references in this chapter that helped in the improvement of quality. Sincere thanks to Prof. João
M.P.Q. Delgado (Editor) by the opportunity given to present our research in this book.

References

1. Military Handbook—MIL-HDBK-17-1F: Composite Materials Handbook, vol. 1—Polymer

Matrix Composites Guidelines for Characterization of Structural Materials. U.S. Department
of Defense (2002)
2. Gay, D., Suong, V.H., Tsai, S.W.: Composite Materials: Design and Applications. CRC
Press, Boca Raton (2003)
3. Liu, K., Takagi, H., Osugi, R., Yang, Z.: Effect of physicochemical structure of natural fiber
on transverse thermal conductivity of unidirectional abaca/bamboo fiber composites.
Compos. Part A (2012, in press)
4. Asaad, J.N., Tawfik, S.Y.: Polymeric composites based on polystyrene and cement dust
wastes. Mater. Des. 32, 5113–5119 (2011)
5. Gamstedt, E.K., Sandell, R., Berthold, F., Pettersson, T., Nordgren, N.: Characterization of
interfacial stress transfer ability of particulate cellulose composite materials. Mech. Mater.
43, 693–704 (2011)
6. Morren, G., Bottiglieri, M., Bossuyt, S., Sol, H., Lecompte, D., Verleye, B., Lomov, S.V.: A
reference specimen for permeability measurements of fibrous reinforcements for RTM.
Compos. Part A 40, 244–250 (2009)
7. Potter, K.D.: The early history of the resin transfer moulding process for aerospace
applications. Compos. Part A 30, 619–621 (1999)
8. Lawrence, J., Simacek, P., Advani, S.: New advances in modeling and simulation of liquid
composite molding processes. In: Proceedings of the International SAMPE Symposium and
Exhibition, pp. 1–12, Long Beach (2006)
9. Haider, M., Hubert, P., Lessard, L.: An experimental investigation of class a surface finish of
composites made by the resin transfer molding process. Compos. Sci. Techol. 67, 3176–3186
(2007)
10. Rudd, C.D., Lac, K.N., Mangin, C.G.E.: Liquid moulding, structural reaction injection
moulding and related processing techniques. Woodhead Publishing Limited, UK (1997)
11. He, F., Wang, Y., Huang, Y., Wan, Y.: Preparation and mechanical properties of 3-D braided
glass fiber reinforced light-cured resin composites. Mater. Lett. 60, 3339–3341 (2006)
12. Leclerc, J.S., Ruiz, E.: Porosity reduction using optimized flow velocity in resin transfer
molding. Compos. Part A 39, 1859–1868 (2008)
13. Li, J., Zhang, C., Liang, R., Wang, B.: Statistical characterization and robust design of RTM
processes. Compos. Part A 36, 564–580 (2005)
14. Gourichon, B., Deléglise, M., Binetruy, C., Krawczak, P.: Dynamic void content prediction
during radial injection in liquid composite molding. Compos. Part A 39, 46–55 (2008)
15. Gokce, A., Chohra, M., Advani, S.G., Walsh, S.M.: Permeability estimation algorithm to
simultaneously characterize the distribution media and the fabric preform in vacuum assisted
resin transfer molding process. Compos. Sci. Techol. 65, 2129–2139 (2005)
16. Han, K., Jiang, S., Zhang, C., Wang, B.: Flow modeling and simulation of SCRIMP for
composites manufacturing. Compos. Part A 31, 79–86 (2000)
17. Hoes, K., Dinescu, D., Sol, H., Vanheule, M., Parnas, R.S., Luo, Y., Verpoest, I.: Compos.
Part A 33, 959–969 (2002)
18. Matsuzaki, R., Kobayashi, S., Todoroki, A., Mizutani, A.: Full-field monitoring of resin flow
using an area-sensor array in a VaRTM process. Compos. Part A 42, 550–559 (2011)
Resin Transfer Molding Process 149

19. Grujicic, M., Chittajallu, K.M., Walsh, S.: Non-isothermal preform infiltration during the
vacuum-assisted resin transfer molding (VARTM) process. Appl. Surf. Sci. 245, 51–64
(2005)
20. Correia, N.C., Robitaille, F., Long, A.C., Rudd, C.D., Simacek, P., Advani, S.G.: Analysis
of the vacuum infusion moulding process: I. Analytical formulation. Compos. Part A 36,
1645–1656 (2005)
21. Buntain, M.J., Bickerton, S.: Modeling forces generated within rigid liquid composite
molding tools. Part A: experimental study. Compos. Part A 38, 1729–1741 (2007)
22. Garay, A.C., Heck, V., Zatera, A., Amico, S.C.: Influence of the calcium carbonate (CaCO3)
in the infiltration process and properties of composites molded by RTM and RTM light.
In: 198 Brazilian Congress of Materials Sciences and Engineering, Campos do Jordão, Brazil
(2010) (In Portuguese)
23. Han, K., Jiang, S., Zhang, C., Wang, B.: Flow modeling and simulation of SCRIMP for
composites manufacturing. Compos. Part A 31, 79–86 (2000)
24. Hudson, A.: Vacuum-assisted RTM increasingly popular. Compos. Techol. March/April,
24–30 (1996)
25. Liu, S., Masliyah, H.: Dispersion in porous media. In: Vafai, K. (ed.) Handbook of Porous
Media, 2nd edn. CRC Press, Boca Raton (2005)
26. Bear, J.: Dynamics of Fluid in Porous Media. Dover Publications Inc, New York (1972)
27. Allen III, M.B., Behie, G.A., Trangenstein, J.A.: Multiphase Flow in Porous Media. Springer-
Verlag, Berlin (1988)
28. Delgado, J.M.P.D.: Longitudinal and transverse dispersion in porous media. Trans. IChemE
Part A, Chem. Eng. Res. Des. 85(A9), 1245–1252 (2007)
29. Hunt, A.G., Skinner, T.E., Ewing, R.P., Ghanbarian-Alavijeh, B.: Dispersion of solute in
porous media. Eur. Phys. J. B 80(4), 411–432 (2011)
30. Vallabh, R., Banks-Lee, P., Seyam, A.: New approach for determining tortuosity in fibrous
porous media. J. Eng. Fiber Fabr. 5(3), 7–15 (2010)
31. Epstein, N.: Tortuosity and the tortuosity factor in flow and diffusion through porous media.
Chem. Eng. Sci. 44(3), 777–779 (1989)
32. ANSYS Inc.: CFX-Theory manual (2010)
33. Nield, D.A., Bejan, A.: Convection in Porous Media. Springer-Verlag, New York (2006)
34. Tannehill, J.C., Anderson, D.A., Pletcher, R.H.: Computational Fluid Mechanics and Heat
Transfer. Taylor & Francis, Philadelphia (1997)
35. Fletcher, C.A.J.: Computational Techniques for Fluid Dynamics: Fundamental and General
Techniques, vol. 1. Springer-Verlag, Berlin (1991)
36. Majumdar, P.: Computational Methods for Heat and Mass Transfer. Taylor & Francis, New
York (2005)
37. Mojtabi, A., Charrier-Mojtabi, M.: Double-diffusive convection in porous media. In: Vafai,
K. (ed.) Handbook of Porous Media, 2nd edn. CRC Press, Boca Raton (2005)
38. Magyari, E., Rees, D.A.S., Keller, B.: Effect of viscous dissipation on the flow in fluid
saturated porous media. In: Vafai, K. (ed.) Handbook of Porous Media, 2nd edn. CRC Press,
Boca Raton (2005)
39. Hugo, J., Topin, F.: Metal foams design for heat exchangers: structure and effectives
transport properties. In: Delgado, J.M.P.Q. (ed.) Heat and Mass Transfer in Porous Media.
Springer-Velag, Berlin (2012)
40. Hsu, C.: Dynamic modeling of convective heat transfer in porous media. In: Vafai, K. (ed.)
Handbook of Porous Media, 2nd edn. CRC Press, Boca Raton (2005)
41. McKibbin, R.: Mathematical modeling for heat and mass transport in geothermal systems. In:
Ingham, D.B., Pop, I. (eds.) Transport Phenomena in Porous Media. Elsevier, Oxford (1998)
42. Ye, X., Qiu, J., Zhang, C., Liang, R., Wang, B.: A finite element-based heuristic estimation of
local perform permeability for resin transfer molding. Transp. Porous Media 76, 247–263
(2009)
43. Nield, D.A.: Modeling fluid flow in saturated porous media and at interfaces. In: Ingham,
D.B., Pop, I. (eds.) Transport phenomena in Porous Media II. Elsevier, Oxford (2002)
150 F. F. Luz et al.

44. Lage, J.L.: The fundamental theory of flow through permeable media from Darcy to
turbulence. In: Ingham, D.B., Pop, I. (eds.) Transport Phenomena in Porous Media. Elsevier,
Oxford (1998)
45. Chen, C.K., Hsiao, S.W.: Transport phenomena in enclosed porous cavities. In: Ingham, D.B.,
Pop, I. (eds.) Transport Phenomena in Porous Media. Elsevier, Oxford (1998)
46. Vafai, K., Amiri, A.: Non-Darcian effects in confined forced convection flows. In: Ingham, D.B.,
Pop, I. (eds.) Transport Phenomena in Porous Media. Elsevier, Oxford (1998)
47. Merrikh, A.A., Lage, J.L.: From continuum to porous-continuum: the visual resolution
impact on modeling natural convection in heterogeneous media. In: Ingham, D.B., Pop, I.
(eds.) Transport Phenomena in Porous Media III. Elsevier, Oxford (2005)
48. Advani, S.G., Hsiao, K.: Transport phenomena in liquid composites molding processes and
their roles in process control and optimization. In: Vafai, K. (ed.) Handbook of Porous Media,
2nd edn. CRC Press, Boca Raton (2005)
49. Parnas, R.S., Phelan Jr, F.R.: The effect of heterogeneous porous media on mold filling in
resin transfer molding. SAMPE Q. 22(2), 53–60 (1991)
50. Pillai, K.M.: Modeling the unsaturated flow liquid composite molding processes: a review
and some thoughts. J. Compos. Mater. 38(23), 2097–2118 (2004)
51. Cheng, P., Hsu, C.: Heat conduction. In: Ingham, D.B., Pop, I. (eds.) Transport Phenomena in
Porous Media. Elsevier, Oxford (1998)
52. Wang, C.Y.: Modeling multiphase flow and transport in porous media. In: Ingham, D.B., Pop, I.
(eds.) Transport Phenomena in Porous Media. Elsevier, Oxford (1998)
53. Bradean, R., Heggs, P.J., Inghan, D.B., Pop, I.: Convective heat flow from suddenly heated
surfaces embedded in porous media. I. In: Ingham, D.B., Pop, I. (eds.) Transport Phenomena
in Porous Media. Elsevier, Oxford (1998)
54. Kaviany, M.: Principles of Convective Heat Transfer, 2nd edn. Springer, New York (2001)
55. Rees, D.A.S., Pop, I.: Local thermal non-equilibrium in porous medium convective. In:
Ingham, D.B., Pop, I. (eds.) Transport Phenomena in Porous Media III. Elsevier, Oxford
(2005)
56. de Lemos, M.J.S.: The double-decomposition concept for turbulent transport in porous
media. In: Ingham, D.B., Pop, I. (eds.) Transport Phenomena in Porous Media III. Elsevier,
Oxford (2005)
57. de Lemos, M.J.S.: Mathematical modeling and applications of turbulent heat and transfer in
porous media. In: Vafai, K. (ed.) Handbook of Porous Media, 2nd edn. CRC Press, Boca
Raton (2005)
58. Nield, D.A., Kuznetsov, A.V.: Heat transfer in bidisperse porous media. In: Ingham, D.B.,
Pop, I. (eds.) Transport Phenomena in Porous Media III. Elsevier, Oxford (2005)
59. de Lemos, M.J.S.: Turbulence in Porous Media: Modeling and Applications. Elsevier,
Amsterdan (2006)
60. Hu, J.L., Liu, Y., Shao, X.M.: Study on void formation in multi-layer woven fabrics.
Compos. Part A Appl. Sci. Manuf. 35, 595–603 (2004)
61. Shojaei, A.: A numerical study of filling process through multilayer preforms in resin
injection/compression molding. Compos. Sci. Techol. 66, 1546–1557 (2006)
62. Yang, J., Jia, Y.X., Sun, S., Ma, D.J., Shi, T.F., An, L.J.: Enhancements of the simulation method
on the edge effect in resin transfer molding processes. Mater. Sci. Eng. A 478, 384–389 (2008)
63. Phelan, F.R.: Simulation of the injection process in resin transfer molding. Polym. Compos.
18, 460–476 (1997)
64. Reddy, J.: Introduction to the Finite Element Method. McGraw-Hill Science/Engineering/
Math, New York (2005)
65. Patankar, S.V.: Numerical Heat Transfer and Fluid Flow. Hemisphere Publishing
Corporation, New York (1980)
66. Maliska, C.R.: Computational Heat Transfer and Fluid Mechanics. LTC, Rio de Janeiro
(2004)
67. Versteeg, H., Malalasekra, M.: An Introduction to Computational Fluid Dynamics: The Finite
Volume Method. Prentice Hall, London (2007)
Resin Transfer Molding Process 151

68. Hirt, C.W., Nichols, B.D.: Volume of fluid (VOF) method for the dynamics of free
boundaries. J. Comput. Phys. 39, 201–225 (1981)
69. Srinivasan, V., Salazar, A.J., Saito, K.: Modeling the disintegration of modulated liquid jets
using volume-of-fluid (VOF) methodology. Appl. Math. Model. 35, 3710–3730 (2001)
70. Shäfer, M.: Computational Engineering. Springer-Verlag, Berlin (2006)
71. Ferziger, J.H., Perić, M.: Computational Methods for Fluid Dynamics. Springer-Verlag,
Berlin (2002)
72. Fletcher, C.A.J.: Computational Techniques for Fluid Dynamics: Specific Techniques for
Different Flow Categories, vol. 2. Springer-Verlag, Berlin (1991)
73. Gutowski, T.G., Morigaki, T., Cai, Z.: The consolidation of laminate composites. J. Compos.
Mater. 21(2), 172–188 (1987)
74. Kokx, V.S. Jr.: Modeling the filling phase of liquid composite moulding processes, including
dynamic mould deformations. Ph.D. thesis, University of Auckland, New Zealand (2002)
75. Kokx, V.S. Jr., Bickerton, S.: Modeling liquid composite molding processes involving
dynamic cavity thickness changes. In: 48th International SAMPE Symposium and
Exhibition, Long Beach, 1690–1701 (2003)
76. Luz, F.F.: Comparative analysis of fluid flow in RTM experiments utilizing commercial
softwares. Master dissertation, Materials Engineering Department, Federal University of Rio
Grande do Sul, Brazil (2011)
77. Luz, F.F., Amico, S.C., Cunha, A.L., Barbosa, E.S., de Lima, A.G.B.: Applying computational
analysis in studies of resin transfer moulding. Defect Diffus. Forum 326–328, 158–163 (2012)