J. Biomechanics Vol, 26, No. 9, pp. 1129 1141, 1993. 0021-9290/93 $6.00+.

00
Printed in Great Britain 9 1993 Pergamon Press Ltd

A NON-NEWTONIAN FLUID MODEL FOR BLOOD FLOW

THROUGH ARTERIES UNDER STENOTIC CONDITIONS
J. C. MISRA, M. K. PATRA and S. C. MISRA*
Department of Mathematics, Indian Institute of Technology, Kharagpur-721302, India

Abstract--This paper presents an analytical study on the behaviour of blood flow through an arterial
segment having a mild stenosis. The artery has been treated as a thin-walled initially stressed orthotropic
non-linear viscoelasticcylindricaltube filledwith a non-Newtonian fluid representing blood. The analysis is
restricted to propagation of small-amplitude harmonic waves, generated due to blood flow whose wave
length is large compared to the radius of the arterial segment. For the equations of motion of the arterial
wall consideration is made of a pair of appropriate equations derived by using suitable constitutive
relations and the principle of superimposition of a small additional deformation on a state of known finite
deformation. It has been shown through numerical computations of the resulting analytical expressions
that the resistance to flow and the wall shear increase as the size of the stenosis increases. A quantitative
analysisis also made for the frequencyvariation of the flow rate at differentlocations of the artery, as well as
of the phase velocities and transmission per wavelength.

INTRODUCTION maximum where the stenosis is of maximum height.

At this point the hydrostatic pressure is minimum.
Although the aetiology of the initiation of a stenosis is The pressure tirop is partially recovered in the di-
not completely understood, it is apparent that the verging area of the stenosis. However, the pressure
presence of a stenosis can lead to serious circulatory drop depends not only on the geometry of the region
disorders. In order to have a fuller understanding of but also on the flow pattern. Several other investiga-
the development of stenosis, an accurate knowledge of tors (Caro et al., 1971; Deshpande et al., 1976; Kang et
the mechanical properties of the vascular wall to- al., 1976; Lee et al., 1970; Morgan et al., 1974; Rod-
gether with that of the flow characteristics of blood kiewicz, 1974; Young et al., 1973) studied the steady
are indispensable. The relevant informations are and unsteady flow through stenosed segments of
deemed to be of great help in the treatment of vascular blood vessels and tried to explain the various
diseases and also to bioengineers who are engaged in anomalies associated with blood flow by proposing
the design and construction of improved artificial different models.
organs. Various experimental observations indicate that
To understand the effects of stenosis in the lumen of the flow characteristics of blood are considerably in-
an artery, a number of studies (cf. McDonald, 1979; fluenced by the vessel wall stress and strain distribu-
Young, 1968, 1979) on the flow through stenosed tions. Morgan and Young (1974) obtained theoretical
arteries have been performed. Most of these studies estimates for velocity profiles, pressure drop and wall
are based on the assumption that blood behaves like shear stress for blood flow through arteries; they also
a Newtonian fluid. However, experimental observa- studied the separation phenomena for specific geo-
tions reveal that blood does not behave as a Newto- metries by considering a Newtonian model of blood.
nian fluid under certain conditions. It is generally A consistent mathematical analysis for the unsteady
accepted that blood, being a suspension of red cells in flow of blood through arteries in the presence of
plasma, behaves like a non-Newtonian fluid at a low stenosis was put forward by Misra and Chakravarty
shear rate (Hart and Barnett, 1973; Whitmore, 1968), (1986) in which the blood was treated as a Newtonian
the shear rate being usually low in the downstream viscous incompressible fluid. The study considered the
side of the stenosis (Nakamura and Swada, 1988). pre-stressed condition as well as the orthotropic
Thus, it is considered that the analysis of the flow material behaviour of the wall tissues. They derived
pattern near the stenosis should include the non- a frequency equation by employing the equation of
Newtonian property of blood. There are still more motion of the fluid and those for the wall, together
(Finav et al., 1975; Horimoto et al., 1979) experimental with the equation of continuity and finally derived
findings which indicate the importance of the non- analytical estimates for the velocity of wave propaga-
Newtonian effects of blood. Misra and Chakravarty tion, the rate of flow and the resistance to flow at
(1986) noted that the mean flow velocity attains its different points of the arterial wall.
An attempt is made here to address the question
Received in final form 11 September 1992.
* Present address: Department of Electronics and Com- whether the viscoelasticity of blood and the wall tis-
munications Engineering, Andhra University College of sues (as observed experimentally by previous investi-
Engineering, Waltair-530003, India. gators) bear the potential to influence significantly the
1129
1130 J.C. MISRA et al.

flow picture in the vicinity of a stenosis developed in Owing to these observations, blood will be treated
an artery. here as a viscoelastic fluid and the following equations
derived on the basis of long-wave approximations (cf.
Misra et al., 1989) will be used to describe the motion
FORMULATION OF THE PROBLEM of blood:

Let us consider an axially symmetric laminar and ['63~u 1 63u u']

8p = -63S
p~7 t3p r t~r2 + r ~ r , ~ S ] , (3)
unsteady flow of blood through a circularly cylin-
drical arterial segment with a mild stenosis developed 630 63p /'632V 1 8v'~
in an axially symmetrical manner (Fig. 1). The P 63r f* (4)
geometry of the stenosis is described (cf. Young, 1968)
mathematically as together with the equation of continuity
63u u 63v
= 1-~aa l l +cos ~o z - d - , ~+7+Tzz=0, (5)
d<z<Lo§ where f * is the complex viscosity of the fluid. Equa-
tions (3) and (4) bear the potential to account for the
= 1, otherwise, (1)
influence of the viscoelastic relaxation phenomena of
where Lo and e are, respectively, the length and the blood on the wave propagation in a blood vessel of
maximum height of the stenosis, d indicates its loca- the circulatory system. Hence, the model used here to
tion and R(z) represents the radius of the arterial represent the non-Newtonian nature of blood is real-
segment under consideration at an axial distance istic. These equations are particularly applicable for
z from one of its ends. Here e is taken to be quite small the case when the shear rate is small or when the
compared to the radius a of the normal artery (outside relaxation and retardation times are almost equal
the stenotic region). I,such situations were discussed by Cokelet et al.
We shall investigate here the propagation of forced (1963) Chien et al. (1966)].
waves which are harmonic in z and t. Hence, we
consider the velocity components and pressure in the
WALL MOTION AND BOUNDARY CONDITIONS
form
u = t2(r) exp Jim(t- z/c)], On the basis of experimental findings, the arteries
will be considered to be in a prestressed state and the
v = ~(r) exp I-ico(t- z/c)], wall tissues to be orthotropic non-linear and vis-
p =/3(r) exp lifo(t- z/c)], (2) coelastic (cf. Patel and Vaishnav, 1972; Young et al.,
1977). Taking into account the inertial forces, the
where u and v denote the components of the velocity surface forces and the forces of constraint which are
of blood in the increasing directions of r and z, respec- the reaction of the surrounding tissues, Misra et al.,
tively, p is the pressure, co is the circular frequency and (1981) derived the equations of motion of the arterial
c is the (complex) wave propagation velocity. wall by employing the principle of superimposition of
Experimental observations reported in the literat- a small additional deformation (which may be at-
ure (cf. Lessner et al., 1971; Thurston, 1973, 1981) tributed to be due to the relatively small fluctuations
clearly indicate that whole blood possesses significant of the arterial pressure from its mean value) upon
viscoelastic properties in the frequency range of a known state of finite deformation of the arterial
physiological importance. The Viscoelastic properties wall. These equations are
of blood are supposed to be due to the viscoelastic
properties of the individual red cells and the internal 2ahoq% c32~ ~ho
structures formed by cellular interactions. It is the ,t2 ~z2 R~ (/300-24~o)
opinion of many investigators (cf. Chein et al., 1975; ho2~ ~ hoPo 632~
Copley et aL, 1975; Stoltz et al., 1978) that blood Ro21 (fl~176 v. +X 2j22 t3t2' (6)
viscoelasticity and the non-Newtonian behaviour of
whole blood are quite prominent in large blood ho 63~{~== 4,o'X h0
q/o)
vessels. Ro 63z
c32rl , y = h o P o ~2y] (7)
tr d Lo
L- ~i= ~1 i"<>
In these equations ( and t/are the physical compon-
ents of the superimposed displacement in the radial
and axial directions, Ro and ho denote the mean
L
surface radius and the wall thickness of the segment of
the blood vessel, 21 and •2 are the axial and circum-
Fig. 1. Geometry of the arterial stenosis. ferential stretch ratios and P0 is the volume density of
 Non-Newtonian blood flow through arteries 1131

the wall tissues. The specific expressions for (11) into equations 0)-(7) and (9) and then equate the
~o, q~o,flo0, rio,, fl,, and fl=0 are given in Appendix A. coefficients of the corresponding powers of 6. In first-
X and Y are the radial and longitudinal components order approximation, the above equations assume the
of the external forces acting on the wall tissues and following forms:
have the following expressions:
dut Opl _ ~_[02u t 10ul ul'~
p 0--~-=-0-7+2s-t--~-~ +r ~ ~-), (12)
Ovl Opl ~, [02Vl 1 ~vl"~
The boundary condition used for the velocity field , o-7=-77z +s ), (13)
v(r, z, t) is the traditional 'no slip" condition from
viscous fluid theory. These conditions require that the Ou~ u~ Ovl
~ r +--~-+--~z = O, (14)
velocity of blood at the inner surface of the arterial
segment may be taken to be equal to the velocity of 2~hoWo 02~1 ho
the inner surface of the arterial wall. These may be ,h Oz2 ~ (/~0o-2~o)~,
stated mathematically as
ho~.2 . . . . 0r/l _ ~, ORl //0V 1 OUl'~I
u(r, z, t)l,:R~:)=~

and hoPo d2~t

-~.~2~ a T ' (15)
~q
v(r, z, t)l,=m:~= ~ - (9) ho/B:= ~o'X04, ho . . . . . . 02~i
Since all these equations have been derived on the
basis of experimental observations (mentioned + 2 f * OR, Or1 ., RI c~vt
above), the model used for the vessel wall is realistic. 0z 0z , = . - J -a- -g-r ,=.
_hoPo O2rll (16)
METHOD OF SOLUTION 2122 63t2 '
OG
In order to solve equations (3)-(7) subject to the ul(r, z, t)l,=, = ~ - , (17)
conditions (8) and (9), all the dependent variables
involved in these equations are expanded in power
series in terms of a small parameter 6(=~/a) around tgq~
vl(r, z, t)l,=,= 0--t-" (18)
a known solution of the problem. The parameter 6 is
so chosen that 6 = 0 will reduce all the dependent Solutions of equations (12)-(14) are first deter-
variables of the problem to their known values. We mined. Assuming that a harmonic wave propagates
expand the dependent variables as follows: due to the flow of blood through the arterial segment,
U = U 1 6 "q- U 2 6 2 "t- . . . . we can write
1) = 1316 -'~ 1)262 -'~ . . . , [ul, v l, P l] = [rio(r), ~o(r),/~o(r)] exp Eico(t-z/c)]. (19)
p=po+pt6+p262 + . . . . Substitution of equation (19) into equations
~=G~+G62+ .... (12)-(14) yields three equations involving the three
unknowns rio, vo and Po- Eliminating the variables
r/=~h6 + r/26z + . . . . suitably one after another, the general solutions of
equations (12)-(14) corresponding to the forced har-
R(z) = a + R~ (z) 6 + R2(z) 6 2 + . . . . (10) monic oscillations have been found in the following
where Po and a are constants which define the initial forms:
state of the system. In order to linearize the equations,
we calculate the value of a function f(r, z, t) for [A,o jx(ctoi3/2r)+;j,(flor)]
ul = L~toi3/2
r = R(z)-ho/2. Using Taylor's theorem together with
the power series expansion around a and employing x exp [ito(t- z/c)],
the last of equations (10), we have
f(r, z, t)l,=R~z)_~=fo(a, z, t)
I- .... Ofo(a,z, x exp [ico(t- z/c) ],
+6Lf~(a'z'tJ+l%tz'-ff-r t)]
px=BJo(,or)exp[iw(t-z/c)], (20)
+ o(62). (11)
We now substitute equation (10) and use the result where A and B are arbitrary constants and ao, flo are
1132 J.C. MISRAet al.

dimensionless parameters defined by where

a2 =pa)aZ/f *, fio =iae)/c (21)
Z = Loo z - d - . (29)
and Jo, J~ denote, respectively, the zeroth and first-
order Bessel functions of the first kind. Similarly, Using expressions (20) and (22) in equations (25) and
corresponding to the forced harmonic oscillations (27) we get
given by equation (19), the solutions of the linear
equations governing the motion of the arterial wall h //,~ltlJo~2 fioo-2d~o poO~2~
may be taken as
1 = C exp [ k o ( t - z/c)], ~1 = D exp [ k o ( t - z/c)], (22) icoho22
C (fioz--~oo)D=O, (30)
C and D being arbitrary constants. Using equations
cRo21
(20)-(22) in the boundary conditions (17) and (18), we f *R]0~0i3/2 j "3/2 f*R]fio
obtain a2 l(~ol ) A + ~ J t ( f i o ) B
Afio "3/2 B
o~oi3/2J1(~r ) +)-c J1 (fio)- koC = 0, (23) ic~
cRo
3/2
AJo(ctoi ) + -B- J l ( f i o ) - k o D = O . (24) o)2ho//po fizO+ ~O'\
pc 7 )o=0. (31)
Writing R 1= R~ + R~, where R c is independent of the
axial coordinate z and R d is a function of z, in equa- Similarly, using relations (20) and (22) in equations
tions (15) and (16), we get the following four equa- (26) and (28) and keeping in mind the choice (29), we
tions: find the following pair of equations:

21hoWo (~2~1 ho ho,~2 2f*ne sin Z . . . . . . //Otoi3/2 icofio "~ .

22 OZ2 ~o (]3oo- 2q~o)(a - - Ro21
-- Lo J,,c~o,~ a +~)A

&h ho Po ~2 ~1 4ie~enf* sin Z Jl(fio)B

• 02 2122 Ot2 , (25) q Lopc 2
.'tlhoWo t92~ ho ho).2 //r~O2,21tlJ0 floo-- 2(fio O)2pO')c
~o (fioo-24'o) ~ ~- - -
22 OZ2 Ro21
~ _ .~ __,OR~//&~ eu~'~ koho22
cRo,~l (fl~176 (32)

hop o c32~1 and

- (26)
2122 Ot2 ' F2ico~ sin Z . cos Z(oto i3/2) q
8f* L- cg~ J~ (~176 q -2a~ "Jl(c~~
ho f/i= C])o']Or ho .. . .~zt/,
,/-2i~_~ sin Z cos ZfloJ1 (fio)] B
_ f , R ~c 8v~ hoPo Ozt/1 (27)
• [- LopC2 Jo(fio)- j
V r=a
- a Or 2122 Ot~ ' icoho//fizz C~o'~
t,g - g ) c+ ho(~2
fi,o+Vo po'
" ~
D
2~/ =0. (33)
_ ...OR~ Or,
_f,R~Ov, =hoPo O2q,
+2J~'-~-z ~-~ r=a a dr r=. 2~22 63t2 " Equations (23), (24) and (30) (33) are all homogene-
(28) ous in the four unknowns A, B, C and D which are, of
course, different in stenosed and non-stenosed re-
Equations (26) and (28) govern the motion in the
gions. In order that there may exist a non-trivial
stenosed area, while equations (25) and (27) depict the
solution of the said equations in both the regions
motion in the arterial segment outside the stenosis.
(inside as well as outside the stenosed portion of the
Since ~R~/~z = ~R~/~z (R~ being independent of z),
arterial segment under consideration), the fourth-or-
unlike equations (26)-(28), equation (25) will not con-
der determinants of their coefficients must vanish.
tain the fluid stress term. In view of equation (1)
This yields the frequency equation which involves the
describing the geometry of the stenosed area, we
unknown complex wave propagation velocity c. We
choose
can simplify the above equations for the case when
8 Ioa/c[~,l, i.e. Ifi0l~l whereby Jo(fio)~l and
R{ = - ~ cos Z,
J l (fio)~ fio/2..With these approximations the derived
 Non-Newtonian blood flow through arteries 1133

frequency equation for the non-stenotic region is

given by equation (35),

floJ l (o~oi3/2) flO

-io~
CtOi3/2 2pc
1
Jo(~o i3/2) -- 0 -io~
pc
. /~1q%~o2 iwho22/a.
0 0 h o t ~ + =0. (34)

floo-2CPo pO(D2~
xT2-;)
f,o~~;3/2
9-1 ~o, j1(~oi3/2)
f*R~fl 2 iwho (~bo ]7::~
<: 2.ca 2

Using the second relation of equation (21) the expan-

sion of equation (34) yields a biquadratic equation in This expression yields also a biquadratic equation in
c given by c given by
Al c* + A2c2 + A3=O, (35) Blcr + B2c3 + Bac2 + B4c + Bs =O, (37)
where the coefficients Ai (i= 1, 2, 3) have expressions where the expressions for all the coefficients
as shown in Appendix B. The coefficients of this Bi (i = 1-5) are presented in Appendix B. Here too, the
biquadratic equation are all complex and hence the coefficients being complex, the roots are also complex.
roots must be complex. Out of the four complex roots, Writing the roots of equations (35) and (37)
two roots differ from the other two only in sign. Two as c = q +ic2, we have exp[ico(t-z/c)] =
of these complex roots in which the real parts are exp[iw(t- Z/Vp)] exp(kz/2), where vv =(c~ + c~)/cl is
positive correspond to the outgoing waves, while the the phase velocity, k = - 2n(c2/c0 is the attenuation
other two complex roots represent incoming waves. In constant and ,~ is the wavelength. Let us denote by
the sequel, we shall restrict our consideration to the Vp, and vp2 the phase velocities of the first and second
outgoing waves. Of the two types of outgoing waves, kinds of outgoing waves, respectively.The shear stress
the waves having smaller velocities of propagation may be calculated as
will be referred to as the waves of the first kind, while
the other type of waves will be called the waves of the 3= _ f , ~ v
second kind. ~r r~R

Similarly the frequency equation for the stenotic

region is

floJ1 (cr i3/2) flo

-io9 0
0~0i3/2 2pc
1
Jo(~o i3/2) 0 --iw
pc
'[2f*rtg sin ZJ 1(0~0i3/2) 2f*mflg sin Z h [-~ 1klJ~ iwho22
L- -~o Lopac
Xo~3/2 i~oflo\-] /~oo-24'o cO2po-I =0. (36)
Rg A
2f* ineco sin Z jo(aoi3/:) f*eflo (2n sin Z +
cLo pac \ Lo
% cos Z'~
)+
f*/3~X0i3/2 COS Z J1 (~xoi3/2)q
2a 2
J
1134 J.C. MISRAet al.

e x p ( - k ) is called the transmission per wave length The resistance to flow, 2, is defined as
and is usually denoted by Z.
In the absence of stenosis, equations (23), (24), (30)
2=(pi-po)/Q, (43)
and (31) can be rewritten as where pi and po are, respectively, the pressures for the
incoming and outgoing waves. Using the expressions
dlA'-d2C' +d3D'=O, -d,A' +d2D'=KI, for vl from equation (41) in equation (42) we get the
dsC'-d6D'=O, dTh'-dsC' +d9D'=O, (38) following expressions for Q:
where A ' = A/B, C'= C/B, D' =D/B and the coeffic- Qpre(Q in the prestenotic region)
ients di (i= 1-9) and K1 are presented in Appendix C. = 2nBa 2 exp [icn(t - z/c)],
By employing the principle of least squares, we find
A=(x+iy)B/(J2+O2), where the expressions for (x + iy) J1 (~o i3/2)
x, y, 61 and 32 are given in Appendix C. Using the r 1, o~<z~<d,
above relations in equations (20), the expressions for
the velocity components and the pressure are found to Qs (Q in the stenotic region)= 2nBa 2
be x exp [io9(t - z/c) ]

] •
[ , . , R

iA-T+A~
/ .3/2R'~

2pc\a/I J '
]

x exp [ico(t-z/c)],
d<~z<~d+Lo,
F x+iy . / .~/2~
Qps, (Q in the poststenotic region)
= 2gBa 2 exp [io~(t- z/c)],
and • exp [ko(t-z/c)]
I(x + iy) Ji (0~0 i3/2)
-~-~2~ +2~ 1, d+Lo <~z<<-L"
Consequently, the expressions for the resistance can
Now equations (23), (24), (32) and (33) can be rewritten be expressed from equation (43) in stenotic as well as
as non-stenotic regions. Since Iflo R(z)/alr 1,
di A'-d2C' + d3D'=O, ).pre= (1 -- e-i~ 2Q1 e- i~
d,A' + d2D' = Ki,
-

(1 --e-itaL~ -it~d/c
fi A' +f2 C' +f3O" = O,
f,A' +fsC' +f6D'=O, (40) ,2.ps` =(1 --e-io~a/e) e-i~176 Q1 e-ie~z/e),
where the derived expressions of the coefficients where the expressions of Q1 and Q2 are presented in
f~ (i= 1-6) are shown in Appendix C. By using the Appendix C. The axial component of the wall shear
method mentioned earlier, here we find
A=(x'+iy')B/(A2+A2), where x',y', A~ and A2 are stress is given by Zw= --f* ~ and the wall shear
as shown in Appendix C. Thus, equations (20) will
stress ratio by
have the following analogous forms:
[" 2 , " , //.3/2R'~ ]
[fl~176 1 Zs I lz~ ~ i~/~a) ao92 ['R~
u'=BL ~oi3/2(A2 +A~) ~-lJt(fl~
ri~o2(x'+iy) Jl(~oi 3/2) aoj 2"-]
l
x exp [ k o ( t - z/c)],

(The suffixes S and N indicate, respectively, the corres-

Vl=B[(x'+iY')J~176176176 +A22) pc ponding values in the stenotic and normal artery.)

x exp [io~(t-z/c)], (41)

RESULTS

With a view to examining the validity of the math-

The flow flux Q is defined as ematical model presented in this paper, we have
R (zl undertaken a specific numerical example and have
Q= I
do
2~rv dr. (42) performed relevant computational work. In agree-
ment with the observation of Young et al. (1977) that
 Non-Newtonian blood flow through arteries 1135

the four function theory for the relaxation functions distance is concerned, in the prestenotic and pos-
describing the viscoelastic properties of the wall tis- tstenotic regions blood flow rate remains almost con-
sues is good enough for most practical applications, stant, but in the stenotic region it decreases monotoni-
the numerical results presented here have been cal- cally in the upstream side and after crossing the peak
culated using the following four relaxation functions position of the stenosis it increases monotonically in
appearing in the expressions for ~bo,tPo, floo, flo~, flzo the downstream side. As expected, the blood flow rate
and fizz (shown in Appendix A), in units of 102 N m - 2: decreases with an increase in the size of the stenosis.
The nature of variation of the resistance to flow at
K~(t) = 282 + 22 e x p ( - 0.47t~ different locations of the arterial segment (Fig. 4)
K3(t) = 179,
K32(t)=270+ 16 e x p ( - 1.67 t~
I00
K3(t)=267+28 exp(-0.51 t~
Four sets of values have been obtained by taking
two sets of stretch ratio parameters (21=1.28,
I
o
80

22 = 1.55) and (21 = 1.4, 22 = 1.61) as well as two sets of x

values of the non-dimensional height of the stenosis, -tt~ 60
viz. (e/a)=0.t and 0.2. For the values of the other
E
physical and material constants, we have taken E /..0
ho=0.16 mm, Ro =0.45 mm, p = 1056 kgm -3,
po= 1100 kg m -3 , #=0.004 N s m -z, co=7.854 rad s -1,
L=30mm, a=2mm, L0=15mm, d=7.5mm,
ao = 0.02 and bo = 0.005. b
In order to examine the extent to which the stretch I l
ratios of the wall tissues affect the unknown quantit- 3 5 7 9 ]0
ies, two different sets of values of the stretch ratios
have been used in the computational work. Our Fig. 2. Variation of the blood flow rate in the stenotic region
choice of the stretch ratio values of the wall tissues is with frequency.
based upon the experimental observation that the
average range for the circumferential stretch ratio is (o): h~=1.28, XZ=I.5S, E / o = 0 . 1
~2
from 1.17 to 1.32, while that for the longitudinal (b): xt=1.28, x z = l . S S , E l 0 : 0 . 2
stretch ratio is from 1.28 to 1.61 (cf. Young et al., It):
(d):
;~1:1./-
~.1=1."
, X2=1,61~
~ 7",Z:1.61,
E/o=O.1
E/o =0,2
1977). Numerical results obtained through the use of
the above-mentioned data are exhibited in Figs 2-10
and Tables l-4. In all the figures (a) refers to the case
when 21=1.28, 22= 1.55, e/a=O.1, (b) to 21=l.28, %
)~2= 1.55, e/a=0.2, (c) to ; q = 1.4, ~.2= 1.6l, e/a=0.1 ~E 6
and (d) to 21 = 1.4, 22 = 1.61, e/a=0.2.
The four expressions for Kz(t) and K3(t) are in- ' ' ' ~ L ~ ' ' ' 2' ~ ' '~' , ,
/. 8 2 16 2 28 30
volved in the expressions for ~bo,Wo, floo, fle~, flzo and z i m m ) --'-P
fl=z (whose expressions are given in Appendix A)
which appear in the wall equations of motion (6) and Fig. 3. Variation of the rate of blood flow at different
(7). So they are all involved implicitly in each of the locations of the artery.
cases (a)-(d).
Numerical values of the rate of blood flow are 5c
calculated for different values of the frequency para-
meter c~0 and for different values of z, i.e. at different
locations of the artery (cf. Figs 2 and 3). These results
indicate that there is a sharp reduction in the blood ~3o
flow rate as the value of ct0 rises from 1 to 3. Sub- un
= ~RE- ~TENOTIC pOST-STE NOTIC
sequently, it maintains almost a steady value in the z20
case 2~ = 1.28, 22 = 1.55 for both 19% (e/a=0.l) and
36% ff,/a=0.2) stenosis. But in the case 21=1.4, 10
21 = 1.61, the blood flow rate suddenly decreases with
an increase in ao from 1 to 3.8 in the ease of 36% 0 I I l I I I I
stenosis and from 1 to 4.7 in 19% stenosis, then it o ,'6 ' ; 1o
2 | m m ) -----I~
increases slowly with the increase in C~o;the increase is
least in the neighbourhood of the peak of the stenosis. Fig. 4. Variation of the magnitude of resistance to flow at
So far as the variation of blood flow rate with the axial different locations of the artery.
1136 J.C. MISRAet al.

reveals that the flow resistance curve is symmetric in ratios 21 and 22 and is a weak function of the size of
the stenosed area about an axis along which the the stenosis. For 21 = 1.28, 22 = 1.55 the phase velocity
constriction of the arterial lumen is maximum 9 The increases with Cto from 1 to 2 and attains its maximum
resistance becomes higher and higher till the constric- at C~o= 2 and then diminishes gradually until cto = 4.75;
tion of the vessel wall attains its maximum and then, after this, it gradually increases for 4.75 ~<Cto~<8 and
with the diminution of the constriction in the lumen of then again it decreases. However, for 21=1.4,
the arterial segment, the flow resistance gradually 22=1.61 the phase velocity steadily decreases for
diminishes. Finally, as anticipated, t h e resistance re- 1 ~<cto ~<2.75 and then gradually increases. At the peak
duces quite appreciably at an axial distance position of 19 and 36% stenosis for the set 21 = 1.28,
z=22.5 mm, where there is an end to the stenotic 2z = 1.55, the phase velocity for the waves of the sec-
region9 The symmetry in the flow rate curve and the ond kind initially increases as Cto increases its value
flow resistance curve in the stenosed area may be from 1 to 2 then decreases steadily upto =o = 6; sub-
attributed to be due to the assumed symmetry in the sequently, it increases for cto>6. But for the set
geometrical configuration of the stenosis. 21 = 1.4, 2z = 1.61 it gradually increases with the in-
The frequency variation curves (Figs 5 and 6) of the crease of Cto in 1 ~<Cto~<10. In the stenotic region for
phases velocities of the wave propagation of blood for the waves of the first kind, the phase velocity changes
both the first and second kind of waves illustrate that in a very irregular m a n n e r (cf. Fig. 5). It is noted that
in the non-stenotic region the phase velocity for the while the phase velocity for the waves of the first kind
waves of the second kind decreases as the frequency has a maximum at Cto=2 when 2 , = 1 . 4 , 22=1.61, it
parameter ~o increases from 1 to 5 and remains nearly has a m i n i m u m for the same value of ct0 when
steady for 5 ~<~o ~<I0 in all the four selected cases (Fig. 2, = 1.28, 22 = 1.55. It may also be pointed out that the
6). The observation for the waves of the first kind (Fig. phase velocity of the waves of the second kind is larger
5) in the non-stenotic zone is quite different. Here the than that for the first kind of waves9
phase velocity is strongly influenced by the stretch The observations for the transmission coefficient
per wavelength (Figs 7-10) are found to be similar to
those for the phase velocity reported earlier. However,
"~ S T E N O T I C REGION
the stretch ratios of the wail tissues are found to affect
..... NON-S TENO[IC REGION

t 6

E 4 re ,-~,,\\d STENOTICREGION
V,/~ 'C,A\ d
0.8

0.6
I I I I I I I I 1
3 5 7 9 'I0 C
I I | I I I I I 1
o(o-.--,,- 0 3 5 7 9 10
(X 0----~
Fig. 5. Variation of phase velocity for the waves of the first
kind with frequency. Fig. 7. Variation of the transmission per wavelength for the
waves of the first kind with frequency.

30
..... S
NTO
EN
N-5
TT
O EIC
NORT
EIC
GR
IOE
NGION 0.8~ STENOT

I i,'",~
20
~ 0.(~ IC REGqON

E
# ',.%', ~
10

',dr -.d d

i I i "?'-:r-c-"V " ~ i "i 0 i i

1 3 5 7 9 10
~ ~ 0
(Xo----~

Fig. 6. Variation of phase velocity for the waves of the Fig. 8. Variation of the transmission per wavelength for the
second kind with frequency, waves of the first kind with frequency.
 Non-Newtonian blood flow through arteries 1137

Table 1. Values of the wall shear stress ratio ~s/rN

Axial
NON-STENOTIC REGION distance
z (ram) (a) (b) (c) (d)
~N 7.50
8,25
1.00000
1.00246
1.00000
1.00489
1,00000
1,00255
1.00000
1,00489
9,00 1.00965 1.01910 1,00965 1.01910
9,75 1.02064 1.04122 1.02066 1.04122
10.50 1.03456 1.06910 1.03465 1.06910
11,25 1.05002 1.100(20 1.05020 1.10000
12,00 1.06545 1.13090 1.06585 1.13090
I I I ,I I I I l I 12,75 1.07989 1.15878 1,07989 1.15878
3 5 7 g 10 13.50 1.09045 1.18090 1.09046 1.18090
14,25 1.09755 1.19511 1.09755 1.19511
15.00 1.10000 1.20000 1.10000 1.20000
Fig. 9. Variation of the transmission per wavelength for the 15.75 1.09755 1.19511 1.09755 1.19511
waves of the second kind with frequency. 16.50 1.09045 1.18090 1.09045 1.18090
17.25 1.07939 1.15878 1.07939 1.15878
18.00 1.06545 1.13090 1.06545 1.13090
1.8 18.75 1.05000 1.10000 1.05000 1.10000
t6 19.50 1.03455 1.06910 1.03455 1.06910
20.25 1.02061 1.04122 1.02061 1.04122
21.00 1.00955 1.01910 1.00955 1.01910
1.2 21.75 1.00245 1.00489 1.00245 1.00489
22.50 1.00000 1.00(O 1.00000 1.00000
t
0.8

influence on various wave propagation charac-

0.4 teristics, especially on the phase velocity and the
transmission per wavelength.
A mathematical model has been developed here in
3 5 7 9 10 order to have a theoretical estimate of the influence of
the viscoelasticity of blood as well as the wall tissues
on the various flow characteristics of blood in
Fig. I0. Variation of the transmission per wavelengthfor the
waves of the second kind with frequency. a stenosed artery. The major assumptions under
which the model has been formulated are that the
artery is circularly cylindrical in shape, the shape of
the wave transmission for the first kind of waves to the stenosis is symmetric about the axis of the artery
a greater extent. and that the wavelength of the propagating waves is
Variation of wall shear stress ratio ZS/ZNat different large compared to the arterial radius.
locations in the stenotic region for different stretch A comparison of our results with the existing liter-
ratios 2~ and 22 and stenosis size is displayed in Table ature reveals (Table 2) that the flow resistance is
1. The wall shear stress in the stenotic region is larger increased owing to the viscoelasticity of blood and the
than that in the non-stenotic region and in the wall tissues. For eto > 1 the aforesaid factors are re-
stenotic region, the wall shear stress gradually in- sponsible for increasing the flow flux (Table 3). The
creases and attains its maximum value at the peak of non-linear viscoelasticity of the wall tissues reduces
the stenosis. Also the values of the stretch ratio para- the phase velocity of the outgoing waves of the first
meters do not have an appreciable influence on the kind (Table 4). Also blood viscoelasticity bears the
wall shear stress ratio. potential to enhance the aforesaid phase velocity of
the wave propagation in arterial blood flow whenever
DISCUSSION the wave propagation parameter ~to exceeds unity.
Our observation (Figs 5 and 6) that the phase velocity
The development of haemodynamic models is quite of the waves of the second kind is greater than that of
important from the viewpoint of physiology. How- the waves of the first kind is in conformity to the
ever, while developing a mathematical model for findings of Cox (1970). Moreover, our observation
studying the mechanics of blood flow through any that the resistance to flow and the wall shear stress
segment of the arterial tree, one has to employ a set of increase as the size of the stenosis increases (cf. Fig.
simplifying assumptions in order to represent reality 4 and Table 1) agrees with the results reported by
in terms of equations which can be conveniently Young (1968). Our results for the variation of blood
manipulated. Past experimental studies of haemo- flow rate (Fig. 3) agree with that of Young's (1979) in
dynamics dearly indicate that the viscoelastic behavi- that the flow rate decreases as the severity of the
our of the wall tissues as well as blood has a marked stenosis increases.

BM 26:9-H
1138 J. C. MISRA et al.
Table 2. Values of 2(N sm -5) x 10s at different locations of the artery

Axial distance z(mm) 0 3 6 9 12 15 18 21 24 27 30

Results of the present

study (non-Newtonian) 10.0 12.5 15.0 22.1 32.5 36.3 29.0 17.4 13.0 11.5 7.5
Results of Misra and
Chakravarty (1986)
(Newtonian) 1.6 4.3 2.9 4.5 20.7 21.8 19.3 5.5 0.3 2.1 1.0

Table 3. Values of Qs (mm3 s- l) x 103 for different %

Frequency parameter % 1 2 3 4 5 6 7 8 9 10

Results of the present

study (non-Newtonian) 47.0 13.0 8.0 7.0 7.0 7.0 6.7 6.4 6.3 6.2
Results of Chakravarty
and Ghosh Chowdhury (1988)
(Newtonian) 8.7 16.5 21.2 22.0 22.2 22.4 22.7 23.0 23.2 23.4

Table 4. Comparison of the phase velocity of outgoing waves of the first kind in the present analysis (non-linear viscoelastic
model) with the linear elastic model of the arterial wall tissues and with the results of Misra et al. (1986), who considered the
wall to be linear elastic and blood to be Newtonian

Frequency parameter cto 1 2 3 4 5 6 7 8 9 I0

Results of the present study

(non4inear viscoelastic model) 3.24 3.10 2.88 1.75 2.12 2.30 1.45 1.43 1.40 1.42
Results of the present study
(linear elastic model) 3.42 3.43 3.55 3172 4.01 4.33 4.68 5.11 5.47 5.84
Results of Misra and
Chakravarty (1986) 3.74 3.06 2.21 1.36 2.04 2.89 3.57 3.62 3.63 3.64

APPENDIX A where
K2(t)=282+22 exp(-0.47t~ K23(t)= 179,
The expressions for q~o, teo, tee, ro~, fl~eand r~z appearing
in equations (6) and (7) are K2(t)= 270 + 16 e x p ( - 1.67 t~
~o = (1 + 2ao) [aoK~(oo) + boK2(oo)], K 3(t) = 267 + 28 e x p ( - 0.5 lt~
Wo=(1 +2bo) [aoK32(~)+boKaa(~)], and ao, bo,represent the constant finite strains.

APPENDIX B
rio, = ( 1 + 2bo) (~bb,+ q~b~),
The expressions for A1, A2, and A 3 appearing in equation
r,0 = (i + 2bo) ( ~ + 't'b,),
(36) are
r~, =(1 + 2ao) (V,~ + ~P~,),
where
Aa=coho[(floo~2(ao) P0092~2122
J

~b,,~= (I + 2ao) K~ (0), ~aoghoPo(Jl(otoi3/2) iJo(2i3/z!)

4,..=(l+2ao)fo~e-i~"ds, X L 2122 ~x ~0 i3/2
"+f* Rl(ot~176176
c 3/2 3/2)l
q~ =(1 + 2ao) K~(0), q~,~=0, 2a _[
9 a, =(1 + 2bo) K](0), 09hof*22 R~ (fl0z-- ~bo)(~oi3/2)Jx (~oi3/2)
q'~ =(1 +2bo) --['~dK](~) e_i, ' ds,
a2Ro21
20 ds A2=aco2ho(.flee~ 22dPo Pot~ +Wo)Jo(aoia/2)
qJb, = (1 + 2bo) K3a(0), \ Ro ,h,h ) k 222

"e,,,=(l+2bo)ff~e-'~"ds, J'(%i3/--~2)
§ '/z) ~, (h~17622
 Non-Newtonian blood flow through arteries 1139

c0321ho~Po FaoghoPoJ1(%is/2) a(2/c~o.i2f* sin Z)2 Jl(~o i3/2)

+
Lo~(~oi 3:~)
iao~hopoJo(%i3/2) kf* R](%ia/2)Jl(%i3/2)] 4"io%h~ ~2f* cos Z (flo~- 4,o)Jo(% i3/2)
22, & " -~ J 4Ro,l.1
im2ho22(/7o,-4,o)Fiwf*R~.,~ ;sn,~ hoa
-t Ro2s L 2 ~o~o, ~-r Ro \ Zo I
x(8= 4,o'~/~,(~oe/b 2f*nae/ieho sin Z . . . . . . [-- - .3/2- J1(%i312!]

aco~2xhoIPo[ihoJo(%i 3/2)(flzo+ It~o) nm2oYihof * sin Z {/7~ 4,o'\

LoRo(O~o13/2) tA22 ~12)J 1(~01312)
nmm'i21hoq%f* sin Z I jo(%i3/2) 2j_1_(~0i3/2)1
+ L622 (~tois/2) _]
The calculated coefficients of the biquadratic equation (37) inao~(eco:f*) 2 sin Z cos Zll(%i sn)
are +
2Lo(%isn)

(%i3/2) J~~o -]
aco4t21h~l'IJ0 . . . . FJ1 (~0i3'2) 9.
B1 && \ Ro ~ / B~ ~ (/&o+'Po, L
x[Jo(~is/2) Jt(%i3/z)]~
.j iaecos 21ho~Pof * cos ZJl (~Xois/2)
42z(%i ~/2)
iegehof* cos Z(~(oiSl2)j t (~oi3/2)
4 2a 2

APPENDIX C
L RoX, 2\ Ro2 2~2,]]
( 9<o.,i*sinZV
~ ) \---~-) J,(~oe/2), iaco I-J1 (0~oi3/2)
dl = - - ~ L. ~ J~ d2=io9, da=-ato2/2,
2tohonef* sin Z 71 (% i 3n) 21 hotttoo)2 ho hoPo ~2
B2 d+=Jo(~oistz), d~ c222 ~-~oo(~oo--24,o) 2112 '
~: Lo
xrfl'o+~Fo,~ is/z,• a~3po ]
t~0 ST 21 (~0i3/2)j ~ ( 80= - 4,o),
L a
2amco2hof* sin Z Jl(%i 3/2)(800--24,0 oa2po~ r , Re V a2(o2 "1
dT=J-I~21L(~oi3/2)Jt(~oia/2)+-~c2 Jo(os
2aenellhoPof * sin Z Jo(%i312)
Lo2122
koSne2f .2 sin Z cos Z(%ian)Jl (%i a/2) oj2hoPo m2ho imSf* R]
4 d9 ~122 ~ (flzO+qlO) 2C2 '
2aLo
net~176 4,~162 i3/2xJ '0( i 3/2"~ 1
K 1 ~-~.--~
LoRo ~"~---"~"
22 2iJ t o # it o # pc
"4ac~ sinZ \(8~ 224,~ ~,~o,2//J~176 el=d~ +d~+d~, e2=-(dld2+d7ds),
es=dlda-d2d4 +dTd9,
aoJXh~Jo(%i3/2)rm2PoWo ,^ +lio, e,=d~ +d~ +d~,
es = -(d2da + dsd6 + dsdg),
x(8.-24,0 r -] aco2h~22. . . . e6=d
.o e~ and e2 are the real and imaginary parts of
(8,, 4,0] [Jo(~oP/~) Jx(~o?/2)]
x k~-'~2) L 7 (%is:2---~i
a02 h2J, {0{0i3]2)FS.0"t"kI/0 (800--24,0 0)2jO0~
"~ (%i 3/2) t L d2K1 % %

[i"'}
and ~l and 6~ are the real and imaginary parts of
cozpo~Po.] iacoaehof* cos ZJI(%P n)
222 J 4(~tois/2)
e 2 e,, es
3 e5 e6
koae2t hoWof* cos Z(%i a/~)Jl(%i a/2) x=el61 +~z62,
4a22 y=s --f162,
1I40 J.C. MISRAet al.

2f*ne sin Z r (aoi3i2 aol2 "~ Copley, A. L., King, R. G., Chien, S., Usami, S., Skalak, R.
f'= ~o L\ a C2aoi3/2J and Huang, C. R. (1975) Microscopic observations of
visco-elasticity of human blood in steady and oscillatory
shear. Biorheology 12, 257-263.
X d 1(~t0i3/2)+ ~ Jo (~toi3/2)~, Cox, R. H. (1970) Wave propagation through a Newtonian
c d
fluid contained within a thick-walled visco-elastic tube:
O2'~1h~176 R~ ('~ h~P~ the influence of wall compressibility. J. Biomechanics 3,
A= c ~ ~- (/7oo-2,;bo) ,h~t2 ' 317-335.
(i oI ?a.,<:i'sin z Deshpande, M. D., Giddens, D, P. and Mabon, R. F. (1976)
Steady laminar flow through modelled vascular stenosis.
A=-t-T,SL ~ ~ J J. Biomechanics 9, 165-174.
Finav, S., Berman, H. J., Fuhro, R. L., Digiovanni, P. R.,
f, 8f* cos z [(~oi~/~)So(~oi~/~)
o~2Jo(~oi~/~)] Fine, S. and Fridman, J. D. (1975) Measurement of velo-
2 L ~ ~ ~ J' city profiles of red blood cells in the microcirculation by
io~ho//7,, r laser Doppler anemometry (LDA). Biorheolooy 12,
f5 = - S~---~ , 207-210.
cRo t~, 2 21) Han, C, D. and Barnett, B. (1973) Measurement of the
~02ho ( flzo+ W o rheological properties of biological fluids. In Rheolooy of
Biological System (Edited by Henry, L G. and Mitchel, L.),
pp. 195-217. Charles C. Thomas Publishers, IL.
(2i.m z cosZ Horimoto, M., Koyama, T., Mishina, H. and Asakura, T.
c \ Lo 4c ]' (1979) Pulsatile blood flow in arteriole of frogweb. Bior-
heology 16, 163-170.
g,--dl +d~+f? +#, Kang, C. K. and Fringen, A. C. (1976) The effect of micro-
o 2 --- - d~ d2 +ftf2 + f , fs, structure on the theological properties of blood. Bull.
Math. Biol. 38, 135-159.
~'~=dl d3 -d2 d4 +fjf~ +f4f6, Lee, J. C. and Fung, Y. C. (1970) Flow in locally contracted
94 -- d~ +f22 +f2, tubes at low Reynolds number. J. appl. Mech. 37, 9-16.
Lessner, A., Zahavi, J., Silberberg, A., Frei, E. H. and
9s = -dzd3 +f2f3 +fsf6, Dreyfus, F. (1971) The visco-elastic properties of whole
o~=d ~,+d~ +f~ +#. blood. In Theoretical and Clinical Hemorheology (Edited
by Hartert, H. and Copley, A. L.), pp. 194-205. Springer,
e~ and e~ are the real and imaginary parts of New York.
--Kid4 02 g3 ] McDonald, D. A. (1979) On steady flow through modelled
vascular stenosis. J. Biomechanics 12, 13-20.
o g* 9s Misra, J. C. and Chakravarty, S. (1986) Flow in arteries in
the presence of stenosis. J. Biomechanics 19, 907-918.
Kld2 9s g6
Misra, J. C. and Roychoudhury, K. (1981) Non-linear stress
and Ax and A2 are the real and imaginary parts of field in blood vessels under the action of connective tis-
sues. Blood Vessels 19, 19-29.
Misra, J. C. and Sahu, B. K. (1989) Propagation of pressure
g2 g,~ gs 9 waves through large blood vessels: a mathematical model
of blood visco-elasticity. Math. Comput. Model. 12,
g3 g5 g6 333-349.
• =~A1 + s~A2, Morgan, B. F. and Young, D. F. (1974) An integral method
for the analysis of flow in arterial stenosis. Bull. Math. Biol.
y' =e~A1 --e~ A2,
36, 39-53.
^ x+iy Jl(~xoi3/2) 1 Nakamura, M. and Sawada, T. (1988) Numerical study on
121= ~2--~-2 (ixoi3/2) 2t0C the flow of a non-Newtonian fluid through an axisymmet-
ric stenosis. J. biomech. Engng 110, 137-143.
x'+iy' {R\Jl(%i3/2R/a) 1 fR'~ 2 Patel, D. J. and Vaishnav, R. N. (1972) The rheology of large
blood vessels. In Cardiovascular Fluid Dynamics (Edited
by Bergel, D. H.), pp. 1-64. Academic Press, New York.
Rodkiewicz, C. M. (1974) Atherosclerotic formations in the
light of fluid mechanics. Fluid Engng Conf., Montreal,
Jointly sponsored by the American and Canadian Soc.
REFERENCES Mech. Engrs.
Stoltz, J. F,, Galliard, S. and Guillot, M. (1978) An approach
Caro, C. G., Fitz-Gerald, J. M. and Schroter, R. C. (1971) of the visco-elastic properties of blood in transient flow. In
Atheroma and arterial wall shear observation, correlation Biorheology: AICHE Symposium Series 182 (Edited by
and proposal of a shear dependent mass transfer mechan- Huang, C. R. and Copley, A. L), pp. 4-9. AIChE, New
ism for artherogenesis. Proc. R. Soc. Lond. B 177, 109-159. York.
Chakravarty, S. and Ghosh Chowdhury, A. (1988) Response Thurston, G. B. (1973) Frequency and shear rate dependence
of blood flow through an artery under stenotic conditions. of visco-elasticity of human blood. Biorheolooy 10,
Rheol. Acta 27, 418-427. 375-381.
Chien, S., King, R. G., Skalak, R., Usami, S. and Copley, A. Thurston, G. B. (1981) Significance and methods of measure-
L. (1975) Visco-elastic properties of human blood and red ment of visco-elastic behaviour of blood. In The Rheology
cell suspension. Biorheology 12, 341-346. of Blood, Blood Vessels and Associated Tissues (Edited by
Chien, S., Usami, S., Taylor, M., Lundberg, J. L. and Greger- Cross, D. R. and Hwang, N. H. C.), pp. 236-256. Alphen
sen, M. I. (1966) J. appl. Physiol. 21, 81-87. aan den Rijn Sijthoff, Noordhoff.
Cokelet, G. R., Merrill, E. W., Gilliland, E. R., Shin, H., Whitmore, R. L. (1968) Rheolooy of the Circulation. Per-
Britten, A. and Wells, R. E. (1963) Trans. Soc. Rheol. 7, gamon, New York.
307-317. Young, D. F. (1968) Effects of a time dependent stenosis on
 Non-Newtonian blood ttow through arteries 1141

flow through a tube. J. Enon9 Ind. Trans. ASME 90, Young, D. F. and Tsai, F. Y. (1973) Flow characteristics in
248-254. models of arterial stenosis-l: steady flow. d. Biomechanics
Young, D. F. (1979) Fluid mechanics of arterial stenosis. 6, 395-410.
J. biomech. Enon9 Trans. ASME 101, 157-175. Young, J. T., Vaishnov, R. N. and Patel, D. J. (1977) Non-
Young, D. F., Cholvin, N. R., Kirkeeide, R. L. and Roth, A. linear anisotropic visco-elastie properties of canine arter-
C. (1977) Hemodynamics of arterial stenoses at elevated ial segments. J. Biomechanics 10, 549-559.
flow rates. Circ. Res. 41, 99-107.