RESEARCH ARTICLE | JULY 13 2005

A mathematical model for blood flow in magnetic field

E. E. Tzirtzilakis

Physics of Fluids 17, 077103 (2005)

https://doi.org/10.1063/1.1978807

CrossMark

 
View Export
Online Citation

Articles You May Be Interested In

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

Biomagnetic fluid flow in an aneurysm using ferrohydrodynamics principles
Physics of Fluids (June 2015)

Velocity distribution of biomagnetic blood flow through a cosine-shaped stenosed artery

AIP Conference Proceedings (August 2019)

Biomagnetic flow in a curved square duct under the influence of an applied magnetic field
Physics of Fluids (July 2004)
 PHYSICS OF FLUIDS 17, 077103 共2005兲

A mathematical model for blood flow in magnetic field

E. E. Tzirtzilakisa兲
Department of Mathematics, Section of Applied Analysis, University of Patras, 26500 Patras, Greece
共Received 4 November 2004; accepted 31 May 2005; published online 13 July 2005兲
In the present study a mathematical model of biomagnetic fluid dynamics 共BFD兲, suitable for the
description of the Newtonian blood flow under the action of an applied magnetic field, is proposed.
This model is consistent with the principles of ferrohydrodynamics and magnetohydrodynamics and
takes into account both magnetization and electrical conductivity of blood. As a representative
application the laminar, incompressible, three-dimensional, fully developed viscous flow of a
Newtonian biomagnetic fluid 共blood兲 in a straight rectangular duct is numerically studied under the
action of a uniform or a spatially varying magnetic field. The numerical results are obtained using
a finite differences numerical technique based on a pressure-linked pseudotransient method on a
collocated grid. The flow is appreciably influenced by the application of the magnetic field and in
particularly by the strength and the magnetic field gradient. A comparison of the derived results is
also made with those obtained using the existing BFD model indicating the necessity of taking into

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

account the electrical conductivity of blood. © 2005 American Institute of Physics.
关DOI: 10.1063/1.1978807兴

I. INTRODUCTION Tesla兲 were also used on a living rat and the consequence
was the reduction of the blood flow and the temperature of
Biomagnetic fluid dynamics 共BFD兲 is a relatively new the rat.18 Also experiments have shown that for a magnetic
area in fluid mechanics investigating the fluid dynamics of field of the same strength 共8 Tesla兲, the flow rate of human
biological fluids in the presence of magnetic field. During the blood in a tube was reduced by 30%.19
last decades extensive research work has been done on the In order to investigate the flow of a biomagnetic fluid,
fluid dynamics of biological fluids in the presence of a mag- the BFD model has been developed by Haik et al. The math-
netic field. Numerous applications have been proposed in ematical formulation of BFD is analogous to the one of fer-
bioengineering and medical sciences. Among them is the de- rohydrodynamics 共FHD兲, which deals with no induced elec-
velopment of magnetic devices for cell separation, targeted tric current, and considers that the flow is affected by the
transport of drugs using magnetic particles as drug carriers, magnetization of the fluid in the magnetic field. Thus, the
magnetic wound or cancer tumor treatment causing magnetic resulting equations of BFD take into account the magnetiza-
hyperthermia, reduction of bleeding during surgeries or tion of the fluid, as opposed to the formulation of the well-
provocation of occlusion of the feeding vessels of cancer known magnetohydrodynamics 共MHD兲, which deals with
tumors and development of magnetic tracers.1–8 conducting fluids, and the corresponding mathematical
A biomagnetic fluid is a fluid that exists in a living crea- model ignores the effect of polarization and magnetization.
ture and its flow is influenced by the presence of a magnetic The arising force due to magnetization depends on the exis-
field. The most characteristic biomagnetic fluid is blood, tence of a spatially varying magnetic field and in a uniform
which behaves as a magnetic fluid, due to the complex inter- magnetic field this force vanishes.1,15,19–21
action of the intercellular protein, cell membrane and the Thus, according to the existing BFD model of Haik
hemoglobin, a form of iron oxides, which is present at a et al., biofluids are considered electrically poor conductors
uniquely high concentration in the mature red blood cells, and the flow laminar, Newtonian and affected only by the
while its magnetic property is affected by factors such as the magnetization of the fluid in a spatially varying magnetic
state of oxygenation.9 It is found that the erythrocytes orient field. However, blood, in particular, exhibits considerably
with their disk plane parallel to the magnetic field9–13 and high static electrical conductivity which is hematocrit and
also that blood possesses the property of diamagnetic mate- temperature dependent. Over the above, the electrical con-
rial when oxygenated and paramagnetic when ductivity of blood varies as the flow rate varies.22–24
deoxygenated.14 Measurements have also been performed for In the present study the existing mathematical model of
the estimation of the magnetic susceptibility of blood which BFD,1,15,19–21 is extended. According to the presented model,
was found to be 3.5⫻ 10−6 and −6.6⫻ 10−7 for the venous both magnetization of BFD, which is consistent with the
and arterial blood, respectively.15,16 Experiments have been principles of FHD,25–31 and the Lorenz force, due to the in-
made using a relatively weak magnetic field 共1.8 Tesla兲 and duced electric current of MHD,33–35 are taken into account.
low temperatures 共75–295 K兲.17 Strong magnetic fields 共8 Various equations describing the magnetization according to
FHD are also given.
a兲
Author to whom correspondence should be addressed. 18 Ersis Str., 11146 As a representative application the laminar, incompress-
Galatsi, Athens, Greece. Electronic mail: [email protected] ible, three-dimensional, fully developed viscous flow of a

1070-6631/2005/17共7兲/077103/15/$22.50 17, 077103-1 © 2005 American Institute of Physics

077103-2 E. E. Tzirtzilakis Phys. Fluids 17, 077103 共2005兲

Newtonian biomagnetic fluid 共blood兲 in a straight rectangular Continuity equation:

duct is numerically studied. Two cases of the applied mag-
netic field are taken into account, uniform and spatially vary- ⵱ · V = 0; 共1兲
ing. For the spatially varying magnetic field calculations with Momentum equations:
different magnetic field gradients are also performed. The
system of the partial differential equations, resulting after the DV
␳ = − ⵱ p + ␳F + ␮⵱2V + J ⫻ B + ␮0M ⵱ H; 共2兲
introduction of appropriate nondimensional variables, is Dt
solved applying an efficient finite differences numerical tech-
Magnetic field equations:
nique based on a pressure-linked pseudotransient method on
a collocated grid. ⵱ ⫻ H = J = ␴共V ⫻ B兲,
The same physical problem using the above-mentioned 共3兲
numerical method and for a specific magnetic field gradient ⵱ · B = ⵱ · 共H + M兲 = 0;
has been studied by adopting the existing model of BFD of
Haik et al. in Ref. 36. The biomagnetic fluid flow in a curved Energy equation.
duct adopting the same model of Haik et al. and by the use In the nonisothermal case, it is also necessary to consider
of the SIMPLE method has also been studied in Ref. 37. in the mathematical model, the energy equation containing
The obtained results, for different values of the param- the temperature T of the fluid. This equation can be written

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

eters entering into the problem under consideration, show as
that the flow is considerably influenced by the presence of
DT ⳵ M DH J · J
the magnetic field. In the presence of spatially varying, as ␳C p + ␮ 0T − = kⵜ2T + ␮⌽. 共4兲
well as uniform magnetic field, the axial velocity component Dt ⳵ T Dt ␴
considerably reduces. Especially in the case of spatially In the above equations V = 共u , v , w兲 is the velocity field,
varying magnetic field secondary flow arises. It is also ob- D / Dt = ⳵ / ⳵t + V⵱ is the Stokes tensor, ⵜ = 共⳵ / ⳵x , ⳵ / ⳵y , ⳵ / ⳵z兲
tained that the dominant factor for the formation of the flow with x , y , z the axis of a three-dimensional system, ␳ is the
field is the form of the magnetic field gradient. A comparison fluid density, p is the pressure, F is the body force per unit
of the obtained results is also made with the existing model volume, ␮ is the dynamical viscosity, ␮0 is the magnetic
of BFD and it is apparent that the inclusion of the Lorentz permeability of vacuum, M is the magnetization, H is the
force of MHD is a necessity for constant and relatively magnetic field intensity, B is the magnetic induction, ␴ is the
smooth magnetic field gradients. These results, of the effect electrical conductivity of the fluid, J is the density of the
of an applied magnetic field on the flow of a biomagnetic electric current, T is the temperature, k is the coefficient of
fluid, encourage further studies for possible useful medical thermal conductivity of the fluid, C p is the specific heat at
and engineering applications. constant pressure and ⌽ is the dissipation function which for
the three-dimensional case has the form

II. MATHEMATICAL MODEL ⌽=2 冋冉 冊 冉 冊 冉 冊 册 冉

⳵u
⳵x
2
+
⳵v
⳵y
2
+
⳵w
⳵z
2
+
⳵v ⳵u
+
⳵x ⳵y
冊 2

The laminar incompressible flow of a homogeneous

Newtonian and electrically conducting biofluid is considered.
As the blood flows under the influence of a magnetic field,
+ 冉 ⳵w ⳵v
+
⳵y ⳵z
冊 冉
2
+
⳵u ⳵w
+
⳵z ⳵x
冊 冉 2
−
2 ⳵u ⳵v ⳵w
+ +
3 ⳵x ⳵y ⳵z
冊 2
.

two major forces will act upon it. The first one is the mag- The magnetic force due to magnetization per unit volume is
netization force due to the tendency of the erythrocytes to generally ␮0共M · ⵜ 兲H.25–31 Under the assumption of equilib-
orient with the magnetic field and the second one is the Lor- rium magnetization 共M and H parallel兲 and keeping in mind
entz force which arises due to the electric current generating that the used magnetic field is solenoid 共ⵜB = 0兲 , J = ⵱ ⫻ H
from the moving ions in the plasma. is the induced current and B = ␮0H, it is valid, as mentioned
According to the existing mathematical formulation of in Ref. 30, that
BFD,1,15,19–21 which is consistent with the principles of
FHD,25–31 the biofluid is subject to equilibrium magnetiza- M
tion and the apparent viscosity due to the application of the ␮0共M · ⵱ 兲H = ␮0 共H · ⵱ 兲H
H

冉 冊
magnetic field is considered negligible. The validation of the
assumption of equilibrium magnetization has been proven in M 1
= ␮0 ⵱ 共H · H兲 − H ⫻ 共⵱ ⫻ H兲
Ref. 15. Consequently, the existing model of BFD as well as H 2
the model presented in this paper, are both valid for the flow M1 M
like the one in large blood vessels, where the blood can be = ␮0 ⵱ H2 − ␮0 共H ⫻ J兲
considered as a homogeneous and Newtonian fluid.32 The H2 H
contribution of the Lorentz force can be incorporated in the M
mathematical model adopting the principles of MHD.33–35 = ␮0 M ⵱ H − 共B ⫻ J兲
H
According to the above mentioned considerations the
governing equations of flow for an incompressible, homoge- M
= ␮0 M ⵱ H + 共J ⫻ B兲.
neous, Newtonian biofluid are H
077103-3 A mathematical model for blood flow Phys. Fluids 17, 077103 共2005兲

Thus, the magnetization force along with the Lorenz force field strength intensity 共see the following section兲, and the
共superposition兲 is governing equations 共1兲–共4兲 are fully coupled. This fact in
combination with the computational difficulty investigating
M
␮0 M ⵱ H + 共J ⫻ B兲 + J ⫻ B = ␮0M ⵱ H such small temperature variations in the flow field may lead
H to a significantly complicated problem. Thus, depending on

冉 冊
+ 1+
M
H
共J ⫻ B兲.
the difficulty of a specific physical problem one may initially
omit the energy equation, as is also made in the application
presented in Sec. III. Although, this is a simplification, it
It also holds, as mentioned, for example, in Ref. 30, that H provides a good initial approximation in order to investigate
Ⰷ M thus M / H Ⰶ 1 and the force due to the imposed mag- the major effect of the magnetic field in the flow pattern.
netic field is finally ␮0M ⵱ H + 共J ⫻ B兲.
Thus, the term ␮0M ⵱ H of Eq. 共2兲, represents the com- A. Magnetization equations
ponent of the magnetic force, per unit volume and depends The behavior of a biomagnetic fluid when it is exposed
on the existence of the magnetic gradient, whereas the term to magnetic field 共magnetized兲 is described by the magneti-
J ⫻ B appearing in 共2兲, represents the Lorentz force per unit zation property M. Magnetization is the measure of how
volume and arises due to the electrical conductivity. These much the magnetic field affects the magnetic fluid.
two terms are generally of the same order of magnitude and

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

In an equilibrium situation the magnetization property is
in MHD the first one is discarded, whereas the second one is generally determined by the fluid temperature, density and
discarded in FHD.25–35 magnetic field intensity. Various equations, describing the
The term ␮0T共⳵ M / ⳵T兲共DH / Dt兲 of Eq. 共4兲, represents dependence of M on these quantities, are given in Refs. 25,
the thermal power per unit volume due to the magneto- 27, 28, 30, 36, 37, and 43–48. The simplest is the linear
caloric effect. This term arises due to the FHD,25–31 whereas equation for isothermal cases:19,27,30,36,37
the term J · J / ␴ represents the Joule heating and arises due to
the MHD.33–35 M = ␹H, 共5兲
The initial model of BFD of Haik et al.19–21 was devel- where ␹ is a constant called magnetic susceptibility. The fol-
oped only for isothermal cases. In their manuscripts they lowing relation is the linear equation of state, given in Ref.
referred that they expected temperature changes of about 2 or 46:
3 °C which they considered negligible. However, this varia-
tion of temperature is in reality of great importance for a M = K共Tc − T兲, 共6兲
biological system. The physiological temperature of blood, where K is a constant called pyromagnetic coefficient and Tc
for example, is about 37 °C. When the temperature of blood is the Curie temperature. Above the Curie temperature the
rises above 41 °C, irreversible damage occurs in the proteins biofluid is not subject to magnetization.
of plasma and that is the reason that one cannot survive after Another equation for magnetization, below the Curie
such high fever.38 temperature Tc is given in Refs. 44 and 47:

冉 冊
Moreover, hyperthermia or hypothermia is extensively
␤
used for various purposes like cancer tumor treatment, or Tc − T
M = M1 , 共7兲
open heart surgeries.39,40 Especially for the tumor treatment, T1
the role of the temperature is considerably significant. For
increments of 1 °C the time of treatment reduces to the half where ␤ is the critical exponent for the spontaneous or satu-
for a specific biological result like the reduction to one-third ration magnetization. For iron, ␤ = 0.368, M 1 = 54 Oe and
of cancer cells of a tumor.39–41 T1 = 1.45 K. An equation involving the magnetic intensity H
Furthermore, hyperthermia caused by the application of and the temperature T is given in Refs. 45 and 48:
magnetic field on injected magnetic fluid has been used for M = K⬘H共Tc − T兲. 共8兲
the treatment of eye injuries. By using this method the treat-
ment of a group of patients was possible without using anti- where K⬘ is a constant.
inflammatory medication.28 Use of hyperthermia, resulting Finally, Higashi et al.9 found that the magnetization pro-
by the application of a magnetic field, has also been reported cess of red blood cells behaves like the following function,
for the treatment of acid burn necrotic skin wounds of ani- known as the Langevin function, which also describes the
mals. When a magnetic field was applied the temperature variation of magnetization for a magnetic fluid:20,25,27,30
raised more than 3 °C and the wounds of 12 and 20 cm2
closed after 21–26 days, whereas similar wounds, that were
not treated magnetically, showed ulcers and scabs, even after
冋 冉 冊
M = mN coth
␮0mH
␬T
−
␬T
␮0mH
, 册 共9兲

50 days.42 where m is the particle magnetization, N is the number of

Thus, the temperature changes that might occur are im- particles per unit volume and ␬ the Boltzmann’s constant.
portant especially for a biofluid like blood and the inclusion The most accurate of the above expressions is 共9兲. How-
of the energy equation is necessary at least at an initial math- ever, Eqs. 共7兲 and 共8兲, which have been calculated experi-
ematical model of BFD. On the other hand, the adoption of a mentally in Refs. 47 and 48, respectively, are very good ap-
nonisothermal case leads to the use of the magnetization proximations. Moreover, Eq. 共5兲 constitutes a good
equation that depends on both temperature and magnetic approximation as experiments show,15 because especially for
077103-4 E. E. Tzirtzilakis Phys. Fluids 17, 077103 共2005兲

FIG. 1. The flow configuration of case I. The contours

of the spatially varying magnetic field of strength H are
shown in the ABCD plane of the cross section.

blood, the variations of the temperature in the flow field are section ABCD, where the flow is studied, the way the mag-
not so significant to influence its magnetization. The choice netic field is applied as well as the vectors of the magnetic

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

of one of the above-mentioned expressions can be made de- field induction B, are shown in Fig. 2共b兲. For both cases the
pending on the physical problem under consideration. temperature changes, for simplification, are considered insig-
nificant and the magnetic field is assumed to be sufficiently
strong to saturate the biofluid 共equilibrium magnetization兲,
III. APPLICATION the walls of the duct are assumed electrically conducting.
As a simple but representative application the fully de-
A. Case I
veloped, laminar, incompressible, three-dimensional, viscous
flow, of an electrically conducting biomagnetic fluid 共blood兲, For the case I, according to the above-mentioned math-
in an impermeable rectangular duct of square cross section of ematical model, and adopting the simplifications of the hy-
side h̄ is considered. Two cases 共I and II兲 concerning the drodynamic fully developed flow as in Refs. 36, 37, and 50,
application of the magnetic field are taken into account. In the dimensional 共bar above the quantities兲 governing equa-
case I a spatially varying magnetic field is generated by an tions of the fluid flow, are
electric current going through a thin wire placed parallel to ⳵ ū ⳵ v̄
the axis of symmetry of the bottom plane of the duct, at + = 0, continuity, 共10兲
⳵ x̄ ⳵ ȳ
distance c below it, as shown in Fig. 1共a兲. The flow is studied
at a representative cross section ABCD with the origin of the
coordinate system placed at the point A. Consequently, the
points A, B, C, and D are A共0,0兲, B共h , 0兲, C共h , h兲 and D共0 , h兲
⳵ ū
⳵ t̄
+ ū
⳵ ū
⳵ x̄
⳵ ū
+ v̄ = −
⳵ ȳ
1 ⳵ P̄
¯␳ ⳵ x̄
+ ¯␯ 再
⳵2ū ⳵2ū
+
⳵ x̄2 ⳵ ȳ 2
冎
respectively. The wire is placed at the point 共h̄ / 2 , −c兲 and
␮
¯ 0M̄ ⳵ H̄
thus, the axis of symmetry is x̄ = h̄ / 2. The contours of the + , x̄-momentum, 共11兲
¯␳ ⳵ x̄
intensity of the magnetic field H are shown in Fig. 1共b兲.
In case II a constant magnetic field is applied from be-
low the duct 共x-z plane兲 all over the area where the duct is
situated and perpendicularly to the flow 共parallel to z-y
⳵ v̄
⳵ t̄
+ ū
⳵ v̄
⳵ x̄
⳵ v̄
+ v̄ = −
⳵ ȳ
1 ⳵ P̄
¯␳ ⳵ ȳ
+ ¯␯ 再
⳵2v̄ ⳵2v̄
+
⳵ x̄2 ⳵ ȳ 2
冎
plane兲, as shown in Fig. 2共a兲. The flow is studied in the same
with case I representative cross section ABCD with the ori- ␮
¯ 0M̄ ⳵ H̄
+ , ȳ-momentum, 共12兲
gin of the coordinate system placed at the point A. The cross ¯␳ ⳵ ȳ

FIG. 2. The flow configuration of case II. The contours

of the constant magnetic field of strength H are shown
in the ABCD plane of the cross section.
077103-5 A mathematical model for blood flow Phys. Fluids 17, 077103 共2005兲

⳵ w̄
⳵ t̄
+ ū
⳵ w̄
⳵ x̄
+ v̄
⳵ w̄
⳵ ȳ
=−
1 ⳵ P̄
¯␳ ⳵ z̄
+ ¯␯ 再
⳵2w̄ ⳵2w̄
+
⳵ x̄2 ⳵ ȳ 2
冎 ⳵ ū ⳵ v̄
+
⳵ x̄ ⳵ ȳ
= 0, continuity, 共17兲

−
¯␴0B̄2
¯␳
w̄, z̄-momentum. 共13兲 ⳵ ū
⳵ t̄
+ ū
⳵ ū
⳵ x̄
⳵ ū
+ v̄ = −
⳵ ȳ
1 ⳵ P̄
¯␳ ⳵ x̄
+ ¯␯
⳵2ū ⳵2ū
+
⳵ x̄2 ⳵ ȳ 2
再, 冎
The boundary conditions are x̄-momentum, 共18兲
ȳ = 0

x̄ = 0
or ȳ = h̄

or x̄ = h̄
and

and
0 艋 x̄ 艋 h̄:

0 艋 ȳ 艋 h̄:
ū = v̄ = w̄ = 0,

ū = v̄ = w̄ = 0, 共14兲
⳵ v̄
⳵ t̄
+ ū
⳵ v̄
⳵ x̄
⳵ v̄
+ v̄ = −
⳵ ȳ
1 ⳵ P̄
¯␳ ⳵ ȳ
+ ¯␯
⳵2v̄ ⳵2v̄
+
⳵ x̄2 ⳵ ȳ 2
再, 冎
ȳ-momentum, 共19兲
where B̄ = ␮ ¯ 0H̄ is the magnetic field induction and ¯␯ = ␮
¯ / ¯␳ is
the kinematic viscosity. The variation of the magnetic field is
at the x̄-ȳ plane so the terms, arising due to FHD, 共␮ ¯ 0M̄ / ¯␳兲
⫻共⳵H̄ / ⳵x̄兲 and 共␮0M̄ / ␳兲共⳵H̄ / ⳵ȳ兲 represent the magnetization
¯ ¯
⳵ w̄
⳵ t̄
+ ū
⳵ w̄
⳵ x̄
+ v̄
⳵ w̄
⳵ ȳ
=−
1 ⳵ P̄
¯␳ ⳵ z̄
+ ¯␯
⳵2w̄ ⳵2w̄
+ 再
⳵ x̄2 ⳵ ȳ 2
冎

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

force per unit mass in the x̄ and ȳ directions, ¯␴B̄2
respectively.25–31 In both cases the variation of the magnetic − w̄, z̄-momentum, 共20兲
¯␳
field toward the z̄ direction is zero. Thus, the magnetization
force is zero at the z̄ momentum equation whereas the term The boundary conditions remain the same, namely 共14兲, as
共¯␴B̄2 / ¯␳兲w̄ arises in the direction of the flow and perpendicu- well as the magnetic field intensity and the variation of mag-
larly to the direction of the application of the magnetic field, netization which are given by the relations 共15兲 and 共16兲,
due to MHD.33–35 The components of the magnetic field in- respectively.
As far as it could be investigated in the references pre-
tensity H̄x and H̄y along the x̄ and ȳ coordinates 关H
sented in this paper, the blood in some studies has been
= 共H̄x , H̄y兲兴 are given, respectively, by treated as an electrically conducting fluid.49 Thus these stud-
␥ x̄ − a ies depended on the principles of MHD and the analogous
H̄x = system of governing equations was that of case II. From this
2␲ 共x̄ − a兲2 + 共ȳ − b兲2
point of view, the polarization 共magnetization兲 of blood in
spatially varying magnetic fields had been regarded. The
␥ ȳ − b force due to magnetization is generally significant depending
H̄y = − ,
2␲ 共x̄ − a兲2 + 共ȳ − b兲2 also on the magnetic field gradient, as it will be noted below.
According also to the existing model of BFD by Haik et
where 共a , b兲 is the point where the magnetic field is applied al. the blood is treated as an electrically nonconducting
and ␥ is the magnetic field strength at this point 共x̄ = a , ȳ fluid.15,20 Thus, according to this formulation the analogous
= b兲. For the problem under consideration a = h̄ / 2 and system of governing equations is that of case I without the
b = −c 共see Fig. 1兲. corresponding terms to MHD discarding in that way the elec-
The magnitude H̄, of the magnetic field intensity, is trical conductivity of blood. Using the existing BFD model
given by of Haik et al., a uniform magnetic field will have no effect on
the flow because the arising force due to magnetization de-
␥ 1
H̄共x̄,ȳ兲 = 关H̄2x + H̄2y 兴1/2 = . 共15兲 pends on the existence of the magnetic field gradient. Using
冑
2␲ 共x̄ − a兲 + 共ȳ − b兲2
2
the model presented in this paper it is possible to study the
effect of uniform as well as the effect of spatially varying
For the variation of magnetization, Eq. 共5兲 for isothermal
magnetic field, taking into account all the magnetic proper-
cases is used:
ties of blood.
M̄ = ¯␹H̄, 共16兲
IV. TRANSFORMATION OF EQUATIONS
where ¯␹ is the magnetic susceptibility of the biofluid.
The following nondimensional variables are introduced:
B. Case II
x̄ ȳ z̄ ūh̄ v̄h̄ w̄h̄
For the case II where the magnetic field is uniform, the ␰= , ␩= , z= , u= , v= , w= ,
h̄ h̄ h̄ ¯␯ ¯␯ ¯␯
force due to magnetization does not exist and as a conse-
quence the blood is affected only due to its electrical con-
ductivity. The problem in this case is a trivial problem of p̄ H̄ t̄␮
¯
p= , H= , t= , 共21兲
MHD and the Lorentz force will arise only at the axial di- ␳␯ /h̄
¯¯ 2 2
H̄0 ¯␳h̄2
rection in the perpendicular plane of the application of the
magnetic field. Thus, the system of the governing equations where H0 = H̄共ā , 0兲. Also the pressure is split as in Refs. 36,
for this case is 37, and 50:
077103-6 E. E. Tzirtzilakis Phys. Fluids 17, 077103 共2005兲

FIG. 3. Grid stretching, the real plane

is the ␰-␩ one, whereas the computa-
tional is the x-y plane.

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

P共␰, ␩,z兲 = P1共z兲 + p共␰, ␩兲 共22兲 h̄2␮
¯ 0K̄H̄20 h̄2B̄0M̄ 0 h̄2¯␴␮¯ 20H̄20 h̄2¯␴B̄20
MnF = = , MnM = = ,
for ¯␯2¯␳ ¯␯2¯␳ ¯␯¯␳ ¯␯¯␳
共30兲
⳵ P ⳵ P1
= = Pz = const,
⳵z ⳵z where H̄0 = H̄共ā , 0兲, B̄0 = B̄共ā , 0兲 = ␮¯ 0H̄共ā , 0兲 and M̄ 0
= M̄共ā , 0兲 = ¯␹H̄共ā , 0兲. Especially the MnM number is the
⳵P ⳵P ⳵P ⳵p square of the widely known MHD Hartmann number.33–35
= = P ␰, and = = P␩ . 共23兲
⳵␰ ⳵␰ ⳵␩ ⳵␩ Increment of the above mentioned numbers, for a specific
fluid 共¯␯ , ¯␳ , M̄ 0 = const兲 and for a specific flow problem 共h̄
A. Case I = const兲 means increment of the magnetic field strength B̄0. It
is noted also that the two magnetic numbers are also depen-
By substitution of the relations 共15兲, 共16兲, and 共21兲–共23兲
dent on the height of the duct. Consequently, for a specific
to the governing equations 共10兲–共13兲, the following nonlin-
magnetic field the magnetic numbers can significantly
ear system of partial differential equations is obtained:
change by changing the geometric characteristics and this
⳵u ⳵v dependence could be of significant importance from a bio-
+ = 0, 共24兲 logical view point.
⳵␰ ⳵ ␩
The physical problem described by Eqs. 共24兲–共27兲 under
the boundary conditions 共28兲 and 共29兲 with MnM = 0 is the
⳵u ⳵u ⳵u ⳵ 2u ⳵ 2u ⳵H
+u +v = − P␰ + 2 + 2 + MnFH , 共25兲 one obtained by the application of the existing BFD model of
⳵t ⳵␰ ⳵␩ ⳵␰ ⳵␩ ⳵␰ Haik et al.20 This problem of flow in a straight rectangular
duct using the BFD model of Haik et al. has also been stud-
⳵v ⳵v ⳵v ⳵ 2v ⳵ 2v ⳵H ied in Refs. 36 and 37.
+u +v = − Py + 2 + 2 + MnFH , 共26兲
⳵t ⳵␰ ⳵␩ ⳵␰ ⳵␩ ⳵␩
B. Case II
⳵w ⳵w ⳵w ⳵w ⳵w 2 2
For the case II where a constant magnetic field is taken
+u +v = − Pz + 2 + 2 − Mn M H2w 共27兲
⳵t ⳵␰ ⳵␩ ⳵␰ ⳵␩ into account, the magnetic force due to FHD vanishes. Thus,
the system of equations governing the flow for the case II is
under the boundary conditions the same with that of case I setting MnF = 0. For the case II,
in the same way as before, the governing system takes the
␩ = 0 or ␩ = 1 and 0 艋 ␰ 艋 1: u = v = w = 0,
form
共28兲
⳵u ⳵v
+ = 0, 共31兲
␰=0 or ␰ = 1 and 0 艋 ␩ 艋 1: u = v = w = 0. 共29兲 ⳵␰ ⳵ ␩
The two parameters appearing in the problem under consid-
eration are the magnetic numbers MnF and MnM due to FHD ⳵u ⳵u ⳵u ⳵ 2u ⳵ 2u
+u +v = − P␰ + 2 + 2 , 共32兲
and MHD, respectively, ⳵t ⳵␰ ⳵␩ ⳵␰ ⳵␩
077103-7 A mathematical model for blood flow Phys. Fluids 17, 077103 共2005兲

⳵v ⳵v ⳵v ⳵ 2v ⳵ 2v TABLE I. Magnetic field strength and corresponding magnetic numbers.

+u +v = − Py + 2 + 2 , 共33兲
⳵t ⳵␰ ⳵␩ ⳵␰ ⳵␩ B 共Tesla兲 MnF 共Deoxygenated兲 MnF 共Oxygenated兲 MnM

2 6.90⫻ 105 −1.3⫻ 105 0.61

⳵w ⳵w ⳵w ⳵ 2w ⳵ 2w
+u +v = − Pz + 2 + 2 − MnM H2w. 共34兲 4 2.76⫻ 10 6
−5.21⫻ 105 2.46
⳵t ⳵␰ ⳵␩ ⳵␰ ⳵␩ 6 6.21⫻ 106 −1.17⫻ 106 5.53
The boundary conditions remain the same, namely 共28兲 and 8 1.10⫻ 107 −2.08⫻ 106 9.83
共29兲. 10 1.73⫻ 107 −3.25⫻ 106 15.36

C. Grid configuration
For case I, grid stretching is used.51 The stretching is
pictured in Fig. 3, the computational is the x-y plane and the
real one is the ␰-␩ plane. The relation between the coordi-
+ 冉
⳵ w ⳵ w ⳵ x ⳵ 2x
u − 2 +
⳵ t ⳵ x ⳵␰ ⳵␰
冊 冉
⳵ w ⳵ y ⳵2y
v −
⳵ y ⳵ ␩ ⳵ ␩2
冊
nates is given below:
= − Pz + 冉 冊 冉 冊
⳵ 2w ⳵ x 2
+
⳵ 2w ⳵ y 2
− MnM H2w, 共42兲
␰ ⬅ ␰共x兲 = a 1 +冉sinh关␶1共x − ␭兲兴
, 冊 ⳵ x2 ⳵␰ ⳵ y2 ⳵ ␩

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

sinh关␶1␭兴 where
共35兲 ⳵x sinh关␭␶兴 ⳵y 1
␩ ⬅ ␩共y兲 = ␶2e␬y − ␶2 , =
⳵␰ a␶冑1 + 共␰/a − 1兲2sinh关␭␶兴2
, =
⳵ ␩ ␬ 共 ␩ + ␶ 2兲
,

x ⬅ x共␰兲 = ␶1 +
1
␶1
冋冉 冊
sinh−1
␰
a
− 1 sinh共␭␶兲 , 册 ⳵ 2x 共␰/a − 1兲sinh关␭␶兴3
共43兲

共36兲
冋 册
= − ,
⳵␰2 a2␶共1 + 共␰/a − 1兲2sinh关␭␶兴2兲3/2
1 ␩
y ⬅ y共␩兲 = ln +1 , 共44兲
␬ ␶2
⳵2y 1
2 =− .
where ⳵␩ ␬ 共 ␩ + ␶ 2兲 2

␭=
1
2␶1
ln 冏
1 + 共e␶1 − 1兲a
1 + 共e−␶1 − 1兲a
, 冏 ␶2 = ␬
1
e −1
and ␶ 1, ␬
The boundary conditions are
y=0 or y = 1 and 0 艋 x 艋 1: u = v = w = 0, 共45兲
are constants that control the stretching at the ␰ and ␩ direc-
x=0 or x = 1 and 0 艋 y 艋 1: u = v = w = 0. 共46兲
tions, respectively. The transformation of the derivatives is

冉 冊
For case II of constant magnetic field a uniform grid is used.
⳵ ⳵ ⳵x ⳵2 ⳵2 ⳵ x 2
⳵ ⳵ 2x Thus the system of the governing equations is the one of
= , = + , 共37兲
⳵␰ ⳵ x ⳵␰ ⳵␰2 ⳵ x2 ⳵␰ ⳵ x ⳵␰2 共31兲–共34兲 replacing ␰ with x and ␩ with y, under the bound-
ary conditions 共45兲 and 共46兲. This system of equations is
⳵
=
⳵ ⳵x
⳵␩ ⳵x ⳵␩
,
⳵2
⳵␩ 2 =
⳵2 ⳵ y
⳵ y2 ⳵ ␩
冉 冊 2
+
⳵ ⳵2y
⳵ y ⳵ ␩2
. 共38兲
equivalent, for uniform grid, with that of case I by setting
MnF = 0.

Thus, the system of equations 共24兲–共27兲 by using the rela-

tions 共35兲–共38兲, is transformed to the following: V. NUMERICAL SOLUTION AND RESULTS

⳵u ⳵x ⳵v ⳵y For the study of the steady state of the flow, for both
+ = 0, 共39兲 cases under consideration, a pseudotransient method is used
⳵ x ⳵␰ ⳵ y ⳵ ␩
where the time t plays the role of an iteration parameter until

冉 冊 冉 冊
the steady state is reached. Thus, the system of equations
⳵ u ⳵ u ⳵ x ⳵ 2x ⳵ u ⳵ y ⳵2y 共39兲–共42兲, subject to the boundary conditions 共45兲 and 共46兲,
+ u − 2 + v −
⳵ t ⳵ x ⳵␰ ⳵␰ ⳵ y ⳵ ␩ ⳵ ␩2 is solved applying a numerical technique based on the

=−
⳵x
⳵␰
⳵ 2u ⳵ x
Px + 2
⳵ x ⳵␰
冉 冊 冉 冊
2
+
⳵ 2u ⳵ y
⳵ y2 ⳵ ␩
2
+ MnF H
⳵x ⳵H
⳵␰ ⳵ x
, 共40兲
pseudotransient pressure linked equation method 共PLEM兲.36
The PLEM scheme is used on a collocated orthogonal
grid. The advantage is that the complexity of the discretized

冉 冊 冉 冊
differential equations using a staggered grid is avoided. In a
⳵ v ⳵ v ⳵ x ⳵ 2x ⳵ v ⳵ y ⳵2y collocation grid all the variables are determined at the same
+ u − 2 + v −
⳵ t ⳵ x ⳵␰ ⳵␰ ⳵ y ⳵ ␩ ⳵ ␩2 grid nodes whereas, in a staggered grid each variable is de-

=−
⳵y
⳵␩
⳵ 2v ⳵ x
Py + 2
⳵ x ⳵␰
冉 冊 冉 冊
2
+
⳵ 2v ⳵ y
⳵ y2 ⳵ ␩
2
+ MnF H
⳵y ⳵H
⳵␩ ⳵y
,
termined at different grid points.50
The magnetic susceptibility of blood
= 1050 kgr/ m , v̄ = 3.1⫻ 10 m s 共Ref. 32兲兴 has already
3 −6 2 −1
关¯␳

共41兲 been measured to be −6.6⫻ 10−7 for the oxygenated and

077103-8 E. E. Tzirtzilakis Phys. Fluids 17, 077103 共2005兲

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

FIG. 4. Axial velocity profiles and corresponding contours for different magnetic field strengths for the oxygenated blood.

3.5⫻ 10−6 for the deoxygenated blood, respectively.14–16 0.7 sm−1.24 The electrical conductivity of flowing blood is
Thus the constant ¯␹ of the equation of magnetization 共16兲 always greater than that of the stationary blood. The incre-
takes the above-mentioned values depending on which con- ment for medium shear rates is about 10% and increases with
dition of blood is considered. Blood, in particular, also ex- the increment of the hematocrit.22 In the current study the
hibits considerably high static electrical conductivity, which electrical conductivity of blood is assumed, for simplicity,
depends on the hematocrit and the temperature. The electri- temperature independent and equal to 0.8 sm−1.
cal conductivity ¯␴ of stationary blood was measured to be From the definition of the MnF it is apparent that
077103-9 A mathematical model for blood flow Phys. Fluids 17, 077103 共2005兲

h̄2␮
¯ 0¯␹H̄20 h̄2B̄20¯␹ W共x , y兲 ⬅ w, the contours of this velocity and the stream
MnF = = 2 共47兲 function ⌿共x , y兲 in the transverse plane, are pictured in three
¯␯2¯␳ ¯␯ ¯␳␮
¯ 0,
columns, for each magnetic field strength case. The values
where ␮¯ 0 is the magnetic permeability of vacuum equal to printed in the x-y axes are the number of grid points used for
4␲ ⫻ 10−7 Hm and B̄0 is the magnetic field induction at the the calculation which in the present case is 75⫻ 75. It is
obtained that the axial velocity W共x , y兲 is reduced as the
point 共0.5,0兲. For h̄ = 2.5⫻ 10−2 m, which corresponds to
magnetic field strength increases. From the second column it
cross-section of a large vessel, it is apparent that MnF ⬇ −3
is also clear that the maximum of the axial velocity is some-
⫻ 106 for the oxygenated and MnF ⬇ 1.5⫻ 107 for the deoxy-
how “attracted” toward the area where the magnetic source is
genated blood, respectively.
From the definition also of MnM it is apparent that located. Finally, from the third column it is obtained that a
secondary flow is generated in the transverse plane as the
h̄2¯␴␮¯ 20H̄20 h̄2¯␴B̄20 magnetic field increases. The secondary flow occurs in the
MnM = = . 共48兲 form of two vortices rotating from the outer walls to the
¯␯¯␳ ¯␯¯␳
center of the duct as shown in the aforementioned Fig. 4. The
As it was already mentioned, the blood exhibits magnetiza- flow at the transverse plane is obviously symmetric with re-
tion and also holds the property of an electrically conducting spect to the point where the magnetic field is applied.
fluid. Most of the biofluids due to the existing ions in the In Fig. 5 the profiles and corresponding contours for

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

body may be influenced by the magnetic field only due to different magnetic field strengths for a spatially varying mag-
their electrical conductivity. In order to study the effect of netic field and for the deoxygenated blood are pictured. The
the magnetic field due only to the electrical conductivity, applied magnetic field intensity H共x , y兲 is the same with that
MnF can be set to zero and MnM may vary. If the opposite is pictured in Fig. 4 for B = 0. For this case it is also obtained
done then it is possible to study the effect of the magnetic that the axial velocity W共x , y兲 is reduced as the magnetic
field due to the magnetization of the fluid. In the present field strength increases. However, in this case the magnetic
study when the blood is exposed to the uniform magnetic field seems to “turn away” the blood. The flow at the trans-
field the MnF is set to zero. Thus, studying the blood in verse plane is again symmetric with respect to the point
uniform magnetic field is qualitatively equivalent with study- where the magnetic field is applied and it appears in the form
ing one biofluid which exhibits only electrical conductivity. of two vortices. In this case these two vortices are rotating in
It is also reminded that the existing model of BFD of Haik an opposite way with that of the previous case and they are
et al. is obtained by the present one by setting MnM = 0. rotating from the center plane toward the two outer walls.
The magnetization of blood can be increased by adding From Fig. 5 it is also obtained for the case of deoxygenated
artificially created nanoparticles.21 As a consequence, the blood that the magnetic field affects stronger the axial veloc-
MnF for a specific magnetic field strength can be increased ity W共x , y兲 than that of the oxygenated blood.
several orders of magnitude. Equivalently, a specific MnF, The aforementioned results for the deoxygenated blood
when the magnetic particles are added, may correspond to in spatially varying magnetic field are qualitatively similar to
much lower 共even to some thousand gauss兲 magnetic field those obtained for a magnetic fluid in varying magnetic field
intensity. This practically means that blood except from para-
in Refs. 25, 36, and 37. The behavior also of oxygenated
magnetic or diamagnetic material, can be considered as fer-
blood, which at the transverse plane seems to be “attracted”
romagnetic fluid when the magnetic particles are added.
from the magnetic source, whereas the deoxygenated seems
In the present study, normal blood 共without artificial
to be “turned away,” is justified due to its diamagnetic and
nanoparticles兲 is considered and the corresponding Mn num-
paramagnetic nature, respectively.
bers, for a specific magnetic field of strength intensity B, are
calculated from the relations 共47兲 and 共48兲. For both cases I B. Case II and results in conjunction with case I
and II the B0 in the relations 共47兲 and 共48兲 is considered the
same with the imposed magnetic field strength B̄ = ␮ ¯ 0H̄. The The axial velocity profiles for uniform magnetic field are
difference between the two cases is in the calculation of the shown in Fig. 6. This case qualitatively corresponds to
MnF = 0 as mentioned above. Under the action of the uniform
dimensionless magnetic field strength intensity H̄ in the flow
magnetic field no secondary flow is formed and the magnetic
field. For the case I H is calculated by the relations 共15兲 and
field in that case affects only the axial velocity. As the mag-
共21兲, whereas for the case II H is constant all over the flow
netic field strength increases the axial velocity is reduced.
domain. Representative values of the magnetic numbers
The mean velocity w̄av passing through the cross section
used, for h̄ = 2.5⫻ 10−2 m, are shown in Table I. ABCD 共see Fig. 1兲 is given by
A. Case I

冕冕
The profiles and corresponding contours for different ¯h ¯h
1
magnetic field strengths, for varying magnetic field strengths w̄av = w̄ dx̄ dȳ, 共49兲
and for the oxygenated blood are shown in Fig. 4. Especially h̄2 0 0
for B = 0 共hydrodynamic case兲, the magnetic field strength
H共x , y兲 for the varying case is also pictured. For the other
cases of the magnetic field strength, the axial velocity profile which, using 共21兲, can be written as
077103-10 E. E. Tzirtzilakis Phys. Fluids 17, 077103 共2005兲

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

FIG. 5. Axial velocity profiles and corresponding contours for different magnetic field strengths for the deoxygenated blood.

w̄av =
v̄
h̄
冕冕
0
1 1

0
w dx dy. 共50兲 Re = 冕冕
0
1 1

0
w dx dy 共51兲

and is calculated after the solution of the problem under con-

Thus, the Reynolds number defined by Re= w̄avh̄ / v̄, can be sideration. From 共51兲 it can be observed that the Re number
written as for the specific problem represents actually the flow rate
077103-11 A mathematical model for blood flow Phys. Fluids 17, 077103 共2005兲

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

FIG. 6. Axial velocity profiles for different magnetic field strengths for the oxygenated blood and uniform magnetic field.

passing through the cross section ABCD 共see Figs. 1 and 2兲. sequently of flow rate, due to the presence of the magnetic
The number Re* = 100关共Re− Re0兲 / Re0兴 is defined in order to field.
investigate the influence of the magnetic field in the flow rate The variation of Re* with the magnetic field intensity B,
for different values of the magnetic numbers MnM and MnF, is pictured in Fig. 7. It can be observed that the least influ-
where Re0 is the Re number for MnM = MnF = 0. Conse- ence occurs for the varying magnetic field and oxygenated
quently, Re* represents the percentage change of Re and con- blood. The deoxygenated blood in varying magnetic field is

FIG. 7. Percentage variation of Re with the magnetic field strength for FIG. 8. Percentage variation of f Re with the magnetic field strength for
constant and varying magnetic field for oxygenated and deoxygenated constant and varying magnetic field for oxygenated and deoxygenated
blood, respectively. blood, respectively.
077103-12 E. E. Tzirtzilakis Phys. Fluids 17, 077103 共2005兲

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

FIG. 9. Axial velocity and contours of the stream function at the transverse plane for deoxygenated blood for MnM = 0 and for the present model for different
forms of magnetic field intensity.

affected more than blood in constant magnetic field for mag- Pz

netic field strength up to 7.5 Tesla. For greater values of the f Re = − . 共53兲
Re
magnetic field, the blood in constant magnetic field is af-
fected more. For magnetic field of 10 T the flow rate is
In order to investigate the influence of the magnetic field in
reduced by 42%, 32% and 10% for constant magnetic field,
the blood flow we define, for different values of the magnetic
varying magnetic field and deoxygenated blood and oxygen-
number MnF or MnM , the number f Re* = 100关共f Re
ated blood in varying magnetic field, respectively.
− f Re0兲 / f Re0兴, where f Re0 is the f Re number for MnM
An important flow characteristic is the dimensionless
= MnF = 0. Consequently, f Re* represents the percentage
skin friction coefficient f of the flow, given by the expression
change of f Re due to the presence of the magnetic field. The
P̄zh̄ variation of f Re* with the magnetic field intensity is shown
f=− . 共52兲 in Fig. 8. The smallest variation is observed for the varying
¯␳w̄2av
magnetic field and oxygenated blood where the increment of
With the use of 共21兲 and 共49兲–共52兲 the following nondimen- f Re is 10% for magnetic field 10 T. The highest increment is
sional product is obtained: observed for varying magnetic field and deoxygenated blood
077103-13 A mathematical model for blood flow Phys. Fluids 17, 077103 共2005兲

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

FIG. 10. Percentage variation of f Re with the magnetic field strength for different kids of varying magnetic field for deoxygenated blood for MnM = 0 and for
the present model.

untill 7.5 T where the increment is by 40% and becomes the Figure 9 shows the axial velocity W共x , y兲 共columns 2 and 3兲
same with blood in constant magnetic field. For higher mag- and contours of the stream function at the transverse plane
netic field the f Re increases with the highest rate and at 10 T 共4th column兲 for deoxygenated blood for different forms of
magnetic field intensity reaches increment of almost 75% magnetic field intensity 共1st column兲 and for B = 10 T at the
whereas the increment for deoxygenated blood is 45%. magnetic source. It is obtained that, qualitatively speaking,
It is noted that the percentage variation of f Re, f Re*, as the results are similar and discarding the electrical conduc-
well as the percentage variation of Re, Re*, with the mag- tivity 共MnM = 0, 2nd column兲 leads to overestimation of the
netic field intensity are independent of the Pz and thus of the axial velocity mostly when the magnetic field gradient is
Re number. The results also presented in Figs. 4–6 are quali- smaller 共1st and 2nd row兲. For sharp magnetic fields tending
tatively the same regardless of the Pz number. The presented to zero very fast such as 3rd and 4th row, there is almost no
results are obtained for Pz = 15 000, which results to Re difference. The pattern of the flow field in the transverse
= 526 for B = 0. Calculations where also performed for Pz plane 共4th column兲 remains the same for each case of mag-
= 5000 and 25 000. netic field gradient regardless of the electrical conductivity.
Another result that can be obtained from Fig. 9 is the effect
C. Comparison with the BFD model of Haik et al. and of the magnetic field gradient on the flow. The sharper mag-
effect of the magnetic field gradient
netic field leads to stronger secondary flow and greater re-
In order to investigate the contribution of the electrical duction of the axial velocity and consequently of the flow
conductivity in the derivation of the mathematical model, rate.
calculations where performed setting MnM = 0 for varying The reduction of the flow rate with the magnetic field
magnetic field 关Eqs. 共39兲–共42兲兴. For MnM = 0 the mathemati- strength and for different kinds of magnetic field gradients,
cal model is that of BFD proposed by Haik et al. in Ref. 20. for deoxygenated blood, is pictured in Fig. 10. For each type
077103-14 E. E. Tzirtzilakis Phys. Fluids 17, 077103 共2005兲

of magnetic field gradient the calculations are performed magnetic drug targeting,” J. Biomech. 35, 813 共2002兲.
4
E. K. Ruuge and A. N. Rusetski, “Magnetic fluid as drug carriers: Targeted
twice. One using the present model and one the existing
transport of drugs by a magnetic field,” J. Magn. Magn. Mater. 122, 335
model of Haik et al. 共MnM = 0兲. 共1993兲.
Consideration of electrically nonconducting fluid 共MnM 5
V. Badescou, O. Rotariu, V. Murariu, and N. Rezlescu, “Transverse high
= 0兲 for the smoothest magnetic field gradient leads to reduc- gradient magnetic filter cell with bounded flow field,” IEEE Trans. Magn.
33, 4439 共1997兲.
tion of flow rate almost by 10%, whereas consideration of 6
W. Andra and H. Nowak, Magnetism in Medicine 共Wiley VCH, Berlin,
the electrical conductivity leads to flow reduction almost by 1998兲.
7
23%. The difference of the estimations between the two J. Liu, G. A. Flores, and R. Sheng, “In-vitro investigation of blood embo-
models reduces as the magnetic field gradient is sharper and lization in cancer treatment using magnetorheological fluids,” J. Magn.
Magn. Mater. 225, 209 共2001兲.
it is getting greater for greater values of the magnetic field 8
J. Plavins, and M. Lauva, “Study of colloidal magnetite binding erythro-
intensity. For the sharpest magnetic field there is almost no cytes: Prospects for cell separation,” J. Magn. Magn. Mater. 122, 349
difference between the two models. Analogous variation is 共1993兲.
9
observed for the oxygenated blood. T. Higashi, A. Yamagishi, T. Takeuchi, N. Kawaguchi, S. Sagawa, S.
Onishi, and M. Date, “Orientation of erythrocytes in a strong static mag-
Summarizing, the electrical conductivity of blood leads netic field,” J. Blood 82, 1328 共1993兲.
to the increase of significant Lorentz force which should be 10
C. Gasparovic, and N. A. Matweiyoff, “The magnetic properties and water
included, together with the magnetization force, in the gov- dynamics of the red blood cell,” Magn. Reson. Med. 26, 274 共1992兲.
11
erning equations of the corresponding mathematical model A. N. Shaylgin, S. B. Norina, and E. I. Kondorsky, “Behaviour of eryth-

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

rocytes in high gradient magnetic field,” J. Magn. Magn. Mater. 31, 555
for blood flow in spatially varying magnetic field. For the 共1983兲.
case of sharp magnetic field gradients the existing model of 12
T. Takeuchi, T. Mizuno, A. Yamagishi, T. Higashi, and M. Date, “Orien-
BFD of Haik et al. could be used as a good approximation tation of red blood cells in high magnetic field,” J. Magn. Magn. Mater.
whereas for smoother magnetic field gradients the present 140-144, 1462 共1995兲.
13
T. Higashi, N. Ashida, and T. Takeuchi, “Orientation of blood cells in
model is proposed. Finally, for uniform magnetic field, static magnetic field,” Physica B 237–238, 616 共1997兲.
where the model of Haik et al. cannot be applied, the blood 14
L. Pauling, and C. D. Coryell, “The magnetic properties and structure of
can be treated as a magnetic fluid in MHD, which is a case hemoglobin, oxyhemoglobin and carbonmonoxy hemoglobin,” in Proc.
that is also covered by the present model. Natl. Acad. Sci. U.S.A. 22, 210 共1936兲.
15
Y. Haik, V. Pai, and C. J. Chen, “Biomagnetic fluid dynamics,” in Fluid
Dynamics at Interfaces, edited by W. Shyy and R. Narayanan 共Cambridge
VI. CONCLUSIONS University Press, Cambridge, 1999兲, pp. 439–452.
16
M. Motta, Y. Haik, A. Gandhari, and C. J. Chen, “High magnetic field
In this work a mathematical model of BFD is proposed. effects on human deoxygenated hemoglobin light absorption,” Bioelectro-
As an application the 3D blood flow in a rectangular duct chem. Bioenerg. 47, 297 共1998兲.
17
M. Bartoszek and Z. Drzazga, “A study of magnetic anisotropy of blood
under the action of a magnetic field is numerically studied.
cells,” J. Magn. Magn. Mater. 196-197, 573 共1999兲.
Generally, the magnetic field reduces the flow rate. The 18
S. Ichioka, M. Minegishi, M. Iwasaka, M. Shibata, T. Nakatsuka, K. Harii,
greater effect of the magnetic field occurs for deoxygenated A. Kamiya, and S. Ueno, “High-intensity static magnetic fields modulate
blood in spatially varying magnetic field and the flow rate skin microcirculation and temperature in vivo,” Bioelectromagnetics
共N.Y.兲 21, 183 共2000兲.
can be reduced, for strong magnetic fields, even by 40%. 19
Y. Haik, V. Pai, and C. J. Chen, “Apparent viscosity of human blood in a
Secondary flow occurs in the transverse plane only for spa- high static magnetic field,” J. Magn. Magn. Mater. 225, 180 共2001兲.
20
tially varying magnetic field in the form of two rotating vor- Y. Haik, J. C. Chen, and V. M. Pai, “Development of bio-magnetic fluid
tices. The form of the magnetic field gradient plays an im- dynamics,” in Proceedings of the IX International Symposium on Trans-
port Properties in Thermal Fluids Engineering, Singapore, Pacific Center
portant role and substantially determines the flow field. The
of Thermal Fluid Engineering, Hawaii, June 25-28, edited by S. H. Wi-
consideration of the electrical conductivity is also necessary noto, Y. T. Chew, and N. E. Wijeysundera, pp. 121–126 共1996兲.
21
especially for the cases of relatively smooth magnetic field Y. Haik, C. J. Chen, and J. Chatterjee, “Numerical simulation of biomag-
gradients and uniform magnetic field. The Lorentz force can netic fluid in a channel with thrombus,” J. Visualization 5, 187 共2002兲.
22
R. A. Frewer, “The electrical conductivity of flowing blood,” Biomed.
be omitted, as it happens according to the model of BFD of
Eng. 9, 552 共1974兲.
Haik et al., only for the regions of application of very sharp 23
F. Jaspard and M. Nadi, “Dielectric properties of blood: An investigation
magnetic field gradients. For the uniform magnetic field only of temperature dependence,” Physiol. Meas 23, 547 共2002兲
24
the axial velocity reduces and no secondary flow occurs. The S. Gabriel, R. W. Lau, and C. Gabriel, “The dielectric properties of bio-
logical tissues: III. Parametric models for the dielectric spectrum of tis-
decrement of flow rate is less than that of the spatially vary-
sues,” Phys. Med. Biol. 41 2271 共1996兲.
ing magnetic field and reaches the 10% for strong magnetic 25
V. G. Bashtovoy, B. M. Berkovsky, and A. N. Vislovich, Introduction to
field. Thermomechanics of Magnetic Fluids 共Hemisphere, Springer-Verlag, Ber-
lin, 1988兲.
26
B. Berkovski and V. Bashtovoy, Magnetic Fluids and Applications Hand-
ACKNOWLEDGMENT
book 共Begell House Inc., New York, 1996兲.
27
E. Blums, A. Cebers, and M. M. Maiorov, Magnetic Fluids 共Walter de
This work was financially supported by the Pythagoras II
Gruyter, Berlin, 1997兲.
research program co-funded of the EU program Epeaek II. 28
V. E. Fertman, Magnetic Fluids Guidebook: Properties and Applications
共Hemisphere, New York, 1990兲.
1 29
Y. Haik, V. Pai, and C. J. Chen, “Development of magnetic device for cell J. L. Neuringer and R. E. Rosensweig, “Ferrohydrodynamics,” Phys.
separation,” J. Magn. Magn. Mater. 194, 261 共1999兲. Fluids 7, 1927 共1964兲.
2 30
J. M. R. Carlton, C. A. Yowell, K. A. Sturrock, and J. B. Dame, “Biom- R. E. Rosensweig, Ferrohydrodynamics 共Cambridge University Press,
agnetic separation of contaminating host leukocytes from plasmodium- Cambridge, 1985兲.
infected erythrocytes,” Exp. Parasitol. 97, 111 共2001兲. 31
R. E. Rosensweig, “Magnetic fluids,” Annu. Rev. Fluid Mech. 19, 437
3
P. A. Voltairas, D. I. Fotiadis, and L. K. Michalis, “Hydrodynamics of 共1987兲.
077103-15 A mathematical model for blood flow Phys. Fluids 17, 077103 共2005兲

32
T. J. Pedley, The Fluid Mechanics of Large Blood Vessels 共Cambridge ited by W. Andra and H. Nowak 共Wiley, Berlin, 1998兲, pp. 489–494.
43
University Press, Cambridge, 1980兲. E. E. Tzirtzilakis and N. G. Kafoussias, “Mathematical models for biom-
33
K. R. Cramer and S. I. Pai, Magnetofluid Dynamics for Engineers and agnetic fluid flow and applications,” in Proceedings of 6th National Con-
Applied Physicists 共Scripta, Washington, D.C., 1973兲. gress on Mechanics, Thessaloniki, Greece, 19–21 July 共2001兲, pp. 227–
34
W. F. Hughes and F. J. Young, The Electromagnetodynamics of Fluids, 1st 232.
ed. 共Wiley, New York, 1966兲. 44
E. E. Tzirtzilakis and N. Kafoussias, “Biomagnetic fluid flow over a
35
G. W. Sutton and A. Sherman, Engineering Magnetohydrodynamics stretching sheet with non linear temperature dependent magnetization,”
共McGraw-Hill, New York, 1965兲. ZAMP 54, 551 共2003兲.
36 45
E. E. Tzirtzilakis, V. D. Sakalis, N. G. Kafoussias, and P. M. Hatzikon- E. E. Tzirtzilakis and G. B. Tanoudis, “Numerical study of biomagnetic
stantinou, “Biomagnetic fluid flow in a 3D rectangular duct,” Int. J. Nu- fluid flow over a stretching sheet with heat transfer,” Int. J. Numer. Meth-
mer. Methods Fluids 44, 1279 共2004兲. ods Heat Fluid Flow 13, 830 共2003兲.
37 46
P. K. Papadopoulos and E. E. Tzirtzilakis, “Biomagnetic flow in a curved H. I. Andersson and O. A. Valnes, “Flow of a heated ferrofluid over a
square duct under the influence of an applied magnetic field,” Phys. Fluids stretching sheet in the presence of a magnetic dipole,” Acta Mech. 128, 39
16, 2952 共2004兲. 共1998兲.
38 47
G. R. Cokelet, “The rheology and tube flow of blood,” in Handbook of A. S. Arrott, B. Heinrich, and T. L. Templeton, “Phenomenology of fer-
Bioengineering, edited by R. Skalak and S. Chien 共McGraw-Hill, New romagnetism: I. Effects of magnetostatics on susceptibility,” IEEE Trans.
York, 1987兲, pp. 14.1–14.17. Magn. 25, 4364 共1989兲.
39 48
A. S. Ahuja and W. R. Hendee, “Thermal design of a heat exchanger for H. Matsuki, K. Yamasawa, and K. Murakami, “Experimental consider-
heating or cooling blood,” Phys. Med. Biol. 23, 937 共1978兲. ations on a new automatic cooling device using temperature sensitive
40
W. L. Lin, J. Y. Yen, Y. Y. Chen, K. W. Jin, and M. J. Shieh, “Relationship magnetic fluid,” IEEE Trans. Magn. 13, 1143 共1977兲
49
between acoustic aperture size and tumor conditions for external ultra- P. K. Suri and P. R. Suri, “Effect of static magnetic field on blood flow in

Downloaded from http://pubs.aip.org/aip/pof/article-pdf/doi/10.1063/1.1978807/16748639/077103_1_online.pdf

sound hyperthermia,” Med. Phys. 26, 818 共1999兲. a branch,” Indian J. Pure Appl. Math. 12, 907 共1981兲.
41 50
X. Q. Lu, E. C. Burdette, and G. K. Svensson, “A dual-frequency ultra- C. A. J. Fletcher, Computational Techniques for Fluid Dynamics 1, 2nd
sonic system for breast cancer treatment,” Acta Phys. Sin. 共Overseas Ed.兲 ed. 共Springer-Verlag, Berlin, 1991兲.
共Suppl. S兲 8, S345 共1999兲. 51
J. D. Anderson, Jr., Computational Fluid Dynamics 共McGraw-Hill, New
42
O. Stemme, “Magnetic wound treatment,” in Magnetism in Medicine, ed- York, 1995兲.