Simulation Modelling Practice and Theory 14 (2006) 82–94

www.elsevier.com/locate/simpat

Layered fluid model and flow simulation for

microchannels using electrical networks
Manisah Aumeerally *, Renate Sitte
School of Information and Communication Technology, Gold Coast Campus, Griffith University,
PMB50, Gold Coast Mail Centre, Qld 9726, Australia

Received 4 November 2004; received in revised form 27 June 2005; accepted 30 June 2005
Available online 16 September 2005

Abstract

In this paper, we present the modelling of the flowrate of a circular and a rectangular
microchannel using an electrical network. The aim of this study is to produce a fast first
approximation of the flowrates of microchannels for the design of microfluidic devices. It con-
tributes to the physical component of our virtual reality-prototyping computer-aided design
tool for microelectromechanical systems, with emphasis on fast calculations for virtual reality
representations. In our model, the flow is segmented into layers and the pertinent models
derived. We have achieved this by solving the Navier–Stokes equation, obtaining an analytical
model for the circular and a numerical model for the rectangular channels. The resistances of
the layers are obtained from the velocity profile of the flow. The electrical network model is
implemented in Matlab Simulink. The results are compared with finite element model software
(ANSYS) and experimental data.

2005 Elsevier B.V. All rights reserved.

Keywords: Microfluidics; Microchannel flow; Lumped parameter model; Dynamic system; Equivalent
network model

*
Corresponding author. Tel.: +61 755528503.
E-mail address: m.aumeerally@griffith.edu.au (M. Aumeerally).

1569-190X/$ - see front matter
2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.simpat.2005.06.002
 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94 83

1. Introduction

Microelectromechanical systems (MEMS) devices such as microchannels, micro-

nozzles, micropumps, micromixers, microfilters and microvalves involve the flow of
fluid. These devices are often the components of an integrated microfluidic chip or
‘‘lab on a chip’’, forming part of a miniaturized chemical analysis system, a micro
total analysis system or a micro dosing system as in [1–3]. These devices have dimen-
sions in the range of 10 lm and are capable of delivering fluid in the nanolitres scale.
Their application area includes the computer industry whereby smaller devices are
essential to improve the resolution of inkjet printers. In the medical industry, they
are used in laser ablation and drug delivery. Other application areas include pharma-
ceuticals, environmental maintenance, health care and analytical system for DNA
sequencing. These microdevices have the advantage of producing high throughput,
minute amount material requirement, low manufacturing, operating and mainte-
nance cost, low power consumption, reduced waste, disposability and reduced risk
of contamination, increased precision and accuracy.
MEMS design is costly because it is a lengthy and tedious process, especially the
fine-tuning of the geometric sizes, which are critical for the correct functioning of the
device and its manufacturability. For a device such as the micropump, this involves
lengthy hours of finite element analysis calculations just to get the approximate geo-
metric dimensions of the structures. In our work, we try to shorten this time drasti-
cally, by producing fast first order approximations that lead to the approximate
design of the device. From then on, fine-tuning is not only faster but also cheaper,
because fewer prototypes need to be built. For a more efficient prototyping phase,
a cyclic sequence of steps of first order approximation modelling, followed by ani-
mated time scaled virtual reality simulations, followed by readjustments of the geo-
metry, helps to drastically reduce the cost and time for MEMS design.
With the rapid growth in the demand for such devices, there is a need to develop
CAD tools that provide for quick and yet accurate design tool [4]. Finite element
modeling software such as ANSYS FlotranTM provides detailed analysis of flow. This
often requires hours of work involving the creation of the physical model, meshing,
convergence monitoring and several iterations before satisfactory solution is ob-
tained. The objective of this paper is to develop a quick first approximation model
for the flow of fluid in a microchannel which will be used in an interactive virtual
reality prototyping computer-aided design tool [5]. The underlying fluid model is
based on the Navier–Stokes equations.
Tchikanda et al. [6] derived analytical approximations for the mean velocity of li-
quid flow in an open rectangular microchannel for a wide range of aspect ratios and
obtained relative errors less than 15%. Damean and Regtien [7] developed approxi-
mate analytical formula for the velocity field inside hexagonal duct using a point-
matching method which gives maximum errors of 5% and using a generalized
integral transform method which gives maximum errors of 12%. Bendib et al. [8]
made approximation for rectangular microchannels with aspect ratio close to 1 by
replacing the diameter of a circular channel by an equivalent hydraulic diameter.
This technique however is subject to errors of order 10–20% [9].
84 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94

Equivalent electrical circuit has been used in the modeling and simulation of
microchannels in micropumps as in Morris and Forster [10] and Voigt et al. [11].
In [9], the concept of electrical equivalent circuit is applied to the whole microchannel.
In [10] the channel is divided into small slices of equal length and each is described by
an RLC circuit.
In our paper, the electrical circuit concept is applied to layers of fluid within the
channel, so that the different flowrates can be visualized and the transient starting
flow in the microchannel can be demonstrated [12,13]. The error for the flowrate
in circular microchannel is less than 10%.

2. Layered flow model

Fluid flow in MEMS is of low Reynolds number (Re) and at these miniscule
dimensions, viscous force dominates in relation to inertia while centrifugal forces
are negligible. In the following analysis, we assume that the flow is laminar and
not turbulent, i.e., the flow is smooth with no eddies and the fluid particles move
in definite paths. In a long microchannel the flow is uniaxial and parallel to the wall.
Important flow characteristics are viscosity, pressure loss and velocity distribution.
In order to drive the fluid around the chip, a micropump is usually employed to pro-
vide the pressure driven flow. For our fast models that are suitable for visualizations
we have derived an analytical model for the circular channel and a numerical model
for the rectangular channel.

2.1. Circular microchannel

For the computer-aided design tool to display an animation of the flow for
each layer of fluid in the microchannel, the different flow rate of each layer needs
to be calculated. Laminar flow in a horizontal circular microchannel can be derived
from the Navier–Stokes equation in cylindrical coordinates. This flow is given by
[14]

2 
d ux 1 dux 1 dp
þ ¼ ð1Þ
dr2 r dr l dx

The solution to the above differential equation gives the velocity distribution


a2
r 2 dp
ux ¼
ð2Þ
4l dx

where a is the radius of the channel, l is the viscosity of the fluid and p is the hyd-
rostatic pressure. The volumetric flow rate Q is found by integrating the velocity
distribution across the cross-sectional area of the microchannel
Z a
Q¼ ux ð2prÞ dr ð3Þ
0
 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94 85

The microchannel of length l and radius a is divided into annular sections of width
dr, as shown in Fig. 1. The flow rate Q of each section is given by

 Z r2
 r
p dp 2 2 p dp r2 a2 r4  2
Q¼
rða
r Þ dr ¼

 ð4Þ
2l dx r1 2l dx  2 4 r1

where r1 and r2 are the inner and outer radii of the sections. For the lumped-element
model of the mechanical effect of Poiseuille flow, using e ! V convention, the resis-
tance is given by [15]
across DP
R¼ ¼ ð5Þ
through Q
The resistance for each section can be determined by applying the above equation
and is given by
8l‘
Ro ¼ ð6Þ
p½2a2 ðr22
r21 Þ
ðr42
r41 Þ

The flow rate for laminar flow in a circular pipe is given by the Hagen–Poiseuille
equation [16]

pa4
Q¼ DP ð7Þ
8lL

2.2. Rectangular microchannel

To determine the equivalent resistance for laminar viscous flow in a rectangular

microchannel, a numerical approach is required. The Navier–Stokes in Cartesian
coordinates reduces to [17]

o2 u o2 u 1 oP
þ ¼ ð8Þ
oy 2 oz2 l ox

Fig. 1. Sections in circular microchannel.

86 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94

By substituting the following variable into Eq. (8)

" #
 l
u ¼  dP  u; ð9Þ

dx
we obtain the PoissonÕs equation:
o 2 u o2 u
þ 2 ¼
1 ð10Þ
oy 2 oz
for 0 6 y 6 A and 0 6 z 6 B, where A is the width of the cross-section and B is the
height of the cross-section of the channel.
To obtain the values of u*, Eq. (10) is discretized using finite differences and the
Dirichlet boundary condition. It is assumed that the no-slip boundary condition ap-
plies at the walls, where the velocity of the fluid is equal to zero. The program is writ-
ten in MATLAB 6.5 and iterated 250 times with step-size 0.7 and error tolerance
0.001. The velocity profile for a rectangular microchannel of aspect ratio
a = 0.285, where a is the ratio of the height over the width of the cross-section of
the microchannel is shown in Fig. 2.
The average velocity  u is Q/area, where Q is the volumetric flowrate. Using Eq.
(9), we obtain
Z Z
dP 1
Q¼
u dy dz ð11Þ
dx l
Resistance is given by Eq. (5) and by rearranging Eq. (11), we obtain
ll
R¼RR  ð12Þ
u dy dz
where l is the viscosity of the fluid and l is the length of the microchannel.

Fig. 2. Velocity profile for a microchannel of aspect ratio 0.285.

 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94 87

To determine the resistances of each layer, the cross-sectional area is divided into
three equal parts. The denominator in Eq. (12) is calculated using the trapezoidal
rule. The limits of integration depend on the cross-sectional area of each layer.
The code is implemented in MATLABTM 6.5.

3. Equivalent electrical model

To represent the flow of the different layers in the channel, we use parallel circuit
components. The idea is to incorporate the whole as a Simulink model into our
MEMS computer-aided design tool, where Simulink provides a convenient design
interface.
In our model, the current represents the volume flow, the voltage is the pressure
difference, the resistance is the viscous force, and the inductance depends on the mass
of the fluid.
The resistance is given by Eq. (12) and the inductance (L) is given by [8]
ql
L¼ ð13Þ
S
where q is the density of the fluid, l is the length of the microchannel and S is the
cross-sectional area.
A branch consisting of a resistor and inductor in series represents each layer of the
fluid. For the three layers of fluid, the model is given by Fig. 3.

4. Simulation of starting flow

Our aim is to show dynamic visualization in time-scaled virtual reality. The steady
velocity profiles derived in Section 2 are considered as the limiting solutions for large
times after a steady pressure gradient is applied. We want to show how these steady-
state velocity profiles evolve in time. The fluid is initially motionless at the instant the
pressure gradient is applied, i.e., at time t = 0. When the pressure, which is constant
and uniform, is applied the fluid begins to flow. This flow gradually approaches the

Fig. 3. Resistance and inductor in series.

88 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94

steady Poiseuille flow as described in Section 2. In this analysis we assume that the
microchannel is very long and entrance and exit effects are not considered. Thus at
times t P 0, the velocity is dependent only on the radius or height of the microchan-
nel and time (Fig. 4).
Since the pressure-gradient is applied suddenly at time t = 0, achieving a constant
and uniform value instantaneously, a step-function input is used to represent this
pressure-gradient. For each branch of the RL circuit in Fig. 3 the current is given
by the equation [18]
Z
1 t
iðtÞ ¼ ðV
iRÞ dt ð14Þ
L 0
where V is the voltage increase, i is the current, L the inductance and R the resis-
tance. The integral Eq. (14) is implemented in Simulink (MatlabTM) as shown in
Fig. 5.
Fig. 6 shows the simulation result for a rectangular microchannel with dimensions
48080 lm · 57 lm · 200 lm and pressure difference of 6250 Pa [19]. In this figure,
the topmost curve is the response for the microchannel as a whole, the second curve
from the top is the response for the innermost fluid layer, the third curve shows the
response of the middle fluid layer and the lowest curve shows the response of the out-
ermost fluid layer. The innermost fluid layer has the fastest flowrate while the outer-
most layer, which is closest to the walls of the channel, has the slowest flowrate.
From the results of the calculations the total flowrate is 5.13e8 (lm)3/s, the inner-
most layer is 2.72e8 (lm)3/s, the middle layer is 1.79e8 (lm)3/s and the outermost
layer is 0.64e8 (lm)3/s.
The transient part of the response is a very small portion of the total time. For the
total flow, the transient time is 4.0 ms, the inner layer is 3.3 ms, the middle layer is
2.4 ms and the outer layer is 0.9 ms. The outermost layer, nearest to the walls,
achieves its steady state much faster than the innermost layer. The innermost layer
reaches its steady state at a longer time than the outermost layer. Initially the viscous
shear stress affects mainly the particles of the fluid near the wall, retarding their flow,
while those in the central core region are able to accelerate uniformly. As time goes

Fig. 4. Velocity profiles at different values of t for start-up flow in pressure-driven flow.
 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94 89

Fig. 5. Simulink model for the total response of the flow to the increase in pressure.

Fig. 6. Total response to increase in pressure.

on, more particles are affected by viscous forces, causing the velocity profile to
evolve into the steady-state shape. The response shown in Fig. 6 reflects the transient
90 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94

behavior of the flow when a sudden, uniform and constant pressure-gradient is ap-
plied to an initially motionless fluid in a microchannel, away from the entrance or
exit.

5. Validations and discussion

In order to validate the model, the flowrates for the layers of fluid are compared
with results obtained by using numerical technique, finite element method (ANSYS)
or finite difference in Matlab.

5.1. Circular microchannel

A finite element model of the microchannel of length 4.8 cm and diameter 152 lm
is created using ANSYS. Meshing is done using Fluid element 141 and the axial
velocities for the laminar flow of water in the channel are obtained. This is shown
in Fig. 7. Flowrate is dependent on the cross-sectional area and the average velocity
of the fluid, see Eq. (3). In this analysis the areas of the layers are made equal so the

Fig. 7. Contour plot for the axial velocity in a circular microchannel using ANSYS.
 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94 91

flowrates are dependent on the velocity only. The flowrate for each layer obtained
from our electrical model is compared with those obtained using ANSYS and the
Hagen–Poiseuille equation as given in Eq. (7). This is shown in Fig. 8. It can be seen
that the flowrate for the innermost layer is the fastest and the outer layer is the slow-
est. The flowrates obtained using our electrical model is lower than those obtained
using the other methods. The percentage difference between our model and ANSYS
and the Hagen–Poiseuille equation are shown in Table 1. When compared with cal-
culations using Hagen–Poiseuille equation, the electrical model gives an error rang-
ing from 3.2% to 5.22%. When compared with finite element model ANSYS, an
error range of 3.05–6.58% is obtained.

5.2. Rectangular microchannel

We use the experimental data from Park et al. [19] to validate the total volumetric
flowrate obtained from our electrical model. The rectangular microchannel has a

Fig. 8. Comparison of sectional flow rates using the Simulink-electrical model and ANSYS.

Table 1
Comparison between Simulink-electrical model and ANSYS for the layered flow rates for circular
microchannel
Circular microchannel flowrates in (lm)3/s
Electrical Finite element Hagen– % Difference % Difference (with
model model (ANSYS) Poiseuille (with ANSYS) Hagen–Poiseuille)
Outermost layer 1.45E+10 1.53E+10 1.53E+10 4 5.22
Middle layer 4.35E+10 4.49E+10 4.49E+10 3.05 3.2
Innermost layer 8.30E+10 8.88E+10 8.73E+10 6.58 4.97
92 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94

length of 48,050 lm, height 57 lm and width 200 lm. A pressure difference of
6250 Pa is applied across the channel. The working fluid is deionized water at
25.5 C at which the density is 996 kg/m3 and viscosity is 0.0009 Pa s. The mass flow-
rate is 0.5 mg/s.
The experimental volumetric flowrate is compared with the total flowrate ob-
tained from our model and an error of 2.19% is obtained (see Table 2). The total
flowrate is also calculated by using the hydraulic diameter concept, where hydraulic
diameter is defined as the ratio of 4 times the cross-sectional area to the wetted
perimeter, and using Eq. (7). This gives an error of 19.32%.
To validate the flowrates of the layers of fluid, a finite difference technique using
Matlab is used. The results are shown in Fig. 9. As before, the flowrates for the
innermost layer is the fastest while the outermost layer has the slowest flowrates
(Table 3).

Table 2
Comparison of total flowrates between electrical model, Hagen–Poiseuille approximation and exper-
imental data for rectangular microchannel
Rectangular microchannel flowrates in (lm)3/s
Electrical Hagen–Poiseuille, Experimental % Difference % Difference
model DH (electrical) (Hagen–Poiseuille
and experiment)
Total flowrate 5.13E+08 4.05E+08 5.02E+08 2.19 19.32

Fig. 9. Comparison of layered flow rates using the electrical model and finite difference calculations for
rectangular microchannel.
 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94 93

Table 3
Comparison between electrical model and numerical calculation using Matlab
Rectangular microchannel flowrates in (lm)3/s
Numerical model Electrical model % Difference
Outermost layer 6.34E+07 6.39E+07 0.79
Middle layer 1.78E+08 1.80E+08 0.73
Innermost layer 2.72E+08 2.72E+08 0.18

6. Conclusion

In this paper we presented an electrical model for the fast calculation of the flow
rates of fully developed laminar fluid flow in a microchannel. We presented the cases
of circular and rectangular channel sections. For the circular channel an analytical
model was derived and for the rectangular channel a numerical model was derived.
The model was implemented in Simulink and the results compared with analytical
calculations as well as with finite element model (ANSYS) or finite difference method
using Matlab.
For the circular microchannel, the values for the resistances are obtained analyt-
ically. Compared with ANSYS, the electric network model for the circular micro-
channel gives percentage errors up to 6.6% and compared with Hagen–Poiseuille
equation, the error is below 5.22%. For the rectangular microchannel, the resistance
is obtained numerically. The electrical model provides a fast first approximation of
the flow rate. There is no physical modeling or meshing to be done. This method
is restricted to microchannels of regular shapes in which the velocity profile can be
obtained analytically from the Navier–Stokes equation.
The electrical network concept of parallel resistances provides a suitable model
for the determination of flowrates for the layers of fluid in a laminar flow.

References

[1] K. Ikuta, T. Hasegawa, T. Adachi, S. Maruo, Fluid drive chips containing multiple pumps and
switching valves for biochemical IC family. in: Proceedings of IEEE Conference on MEMS (Japan),
IEEE, 2000, pp. 739–744.
[2] M. Koch, C.G.J. Schabmueller, A.G.R. Evans, A. Brunnschweiler, Micromachined chemical reaction
system, Sensors and Actuators A: Physical 74 (1–3) (1999) 207–210.
[3] N.T. Nguyen, S. Schubert, S. Richter, W. Dötzel, Hybrid-assembled micro dosing system using
silicon-based micropump/valve and mass flow sensor, Sensors and Actuators A: Physical 69 (1998)
85–91.
[4] R. Sitte, Introductory physics based visual simulation, in: V. Skala, (Ed.), Proceedings of the C# and
NET Technologies 2003, Plzen (Czech Republic), ISBN 80-903100-3-6, 2003, pp. 63–69.
[5] M. Aumeerally, R. Sitte, Design and modelling of microactuators in MEMS virtual reality
prototyping CAD tool, in: N. Bergmann (Ed.), Electronics and Structures for MEMS II, Proceedings
of SPIE, vol. 4591, SPIE, The International Society for Optical Engineering, USA, 2001, pp. 163–171.
[6] S.W. Tchikanda, R.H. Nislon, S.K. Griffiths, Modeling of pressure and shear-driven flows in open
rectangular microchannels, International Journal of Heat and Mass Transfer 47 (2004) 527–538.
94 M. Aumeerally, R. Sitte / Simulation Modelling Practice and Theory 14 (2006) 82–94

[7] N. Damean, P.P.L. Regtien, Velocity field of the fully developed laminar flow in a hexagonal duct,
Sensors and Actuators A 92 (2001) 144–151.
[8] S. Bendib, O. Français, P. Tabeling, W. Hervé, Analytical study and characterization of microchannel
and passive microdiode, in: 12th Micromechanics Europe Workshop MME2001, Cork, Ireland,
September 16–18, 2001.
[9] K.V. Sharp, R.J Adrian, J.G. Santiago, J.I. Molho, Liquid flows in microchannels, in: M. Gad-
el-Hak (Ed.), The MEMS Handbook, CRC Press, London, 2002.
[10] C. Morris, F.K. Forster, Microelectromechanical systems (MEMS), ASME 2 (2000) 473–479.
[11] P. Voigt, G. Schrag, G. Wachutka, Microfluidic system modelling using VHDL-AMS and circuit
simulation, Microelectronics Journal 29 (1998) 791–797.
[12] M. Aumeerally, R. Sitte, Modelling of layered fluid flow in a circular microchannel, in: Proceedings of
the European Simulation and Modelling Conference ESMC, 2003 (Naples, Italy, October 15–17).
[13] M. Aumeerally, R. Sitte, Simulation and modelling of layered fluid flow in a rectangular
microchannel, in: Proceedings of the Middle East Simulation and Modelling Conference MESM
2003 (Sharjah, UAE, January 5–7, 2004).
[14] J.A. Fay, Introduction to Fluid Mechanics, MIT Press, USA, 1994.
[15] S.D. Senturia, Microsystem Design, Kluwer Academic Publishers, USA, 2001.
[16] R. Zengerle, M. Richter, Simulation of microfluid systems, Journal of Micromechanical Microen-
gineering 4 (1994) 192–204.
[17] R.L. Panton, Incompressible Flow, John Wiley & Sons, USA, 1984.
[18] C.-M. Ong, Dynamic Simulation of Electric Machinery Using Matlab/Simulink, Prentice-Hall, New
Jersey, 1998.
[19] H. Park, J.J. Pak, S.Y. Son, G. Lim, I. Song, Fabrication of a microchannel integrated with inner
sensors and the analysis of its laminar flow characteristics, Sensors and Actuators A 103 (2003)
317–329.