Womersley Arterial Flow - Simple

4/16/2014 MBW:Womersley Arterial Flow - MathBio
http://mathbio.colorado.edu/mediawiki/index.php/MBW:Womersley_Arterial_Flow 1/10
April 16, 2014, Wednesday
Log in / create account
Go Search
MBW:Womersley Arterial Flow
From MathBio
Author: Tracey Morland
This article summarizes the results and approach introduced by Womersley in his famous paper on pulsatile flow in
arteries
[1]
. His approach uses concepts from fluid mechanics, including Poiseuille Flow
(http://en.wikipedia.org/wiki/HagenPoiseuille_equation) , to model the pressure gradient and flow velocity in an arterial
pulse. The model considers the flow of blood in a rigid tube. Womersley's number is also defined, and the usefulness of
Womersley flow is discussed.
Contents
1 Overview
1.1 Biological Context
1.2 History
1.3 Mathematical Background
1.4 Notation
2 Womersley Flow Defined
2.1 Using Fourier Series to Represent Pressure Gradient
2.2 Solving for Flow Velocity (u)
2.2.1 Solving the Bessel Differential Equation
2.3 Womersley Number
2.4 Flow Rate (Q)
3 Analysis and Conclusions
4 Recent Extension
5 References
Overview
This is a model of blood flow as a Newtonian fluid through a rigid tube propagated only by a pressure gradient. The
Newtonian fluid is described by the Navier-Stokes second-order PDE's, and simplified by assuming a Poiseuille flow.
A Fourier series accounts for the periodicity of the pressure gradient, and the coefficients are determined by the Fast
Fourier Transform (FFT). Simplifications and substitutions allow for the system of PDE's to be approximated by
Ordinary Differential Equations with solutions in the form of Bessel functions.
Biological Context
The human heart beats over 2.5 billion times in an average lifetime and about 100,000 times per day, and in one day
your blood travels 12,000 miles,
[2]
or roughly the distance of traveling from Denver to Tokyo and back again. The
heart is composed of four chambers. The sinoatrial nodes (SA) nodes in the right atrium (RA) initiate the electric pulse
and cause the right ventricle (RV) to fill with blood. The action potential is propagated through the atria via the atrial
cells
[3]
. The RV then contracts, sending blood to the pulmonary artery where it is then sent to the lungs. The blood,
now fully oxygenated, now returns to the heart and fills the left atrium (LA). The LA contracts sending blood into the
left ventricle (LV). Blood is then pumped from the LV into the aortic artery which sends the blood to the rest of the
body. For an animation of the contraction of the heart valves, see NOVA: Map of the Human Heart
(http://www.pbs.org/wgbh/nova/body/map-human-heart.html) .
Figure 1: The Human Heart, Nucleus Medical Art, Inc./Getty Images
[2]
How fast does the blood flow through the arteries with each beat? Is there any phase-lag between the shock wave (the
pulse you feel in your wrist or neck) and the flow of blood? These are questions addressed in Womersley's paper
reviewed in this article. The more we can understand the nature of how the arterial pulse works and perform accurate
calculations of arterial blood flow, the more we can detect heart disease and defects, such as in patients with diabetes
and atherosclerosis
[3]
.
History
Most of the information on the nature of the arterial pulse has evolved from the study of fluid mechanics. Various
aspects of the pulse can be included in a model, including the elasticity of the artery, fluid viscosity, pressure gradient,
and the presence of arterial bifurcations (branches) to name a few. In 1808, Thomas Young, a British scientist and
physician, connected the elastic nature of the arteries to pulse wave velocity
[4]
. Then, in 1878, Moens and Korteweg
independently derived a mathematical model relating arterial elasticity, or stiffness, to pulse wave velocity. Today it is
called the Moens-Korteweg equation (http://en.wikipedia.org/wiki/MoensKorteweg_equation) and it is dervied from
Newton's second law of motion .
The simplest model for pulsatile flow is the [Windkessel model (http://en.wikipedia.org/wiki/Windkessel_effect,) ],
developed by Otto Frank in 1899
[5]
. This approach is not accurate enough to be used for quantitative analysis, but it
provides a simple foundation on which to build more complicated models. In the model, blood storage is simplified to a
single chamber, called a Windkessel, and the pressure of this chamber varies periodically over time
[6]
. Inflow into the
chamber is from the heart, and outflow is to the outer arteries, veins and capillaries and is represented as a simple
resistance vessel
[3]
. There are two parts to this model: compliance (representing the elastic nature of arteries) and
resistance. The resistance in the system causes blood to enter the arteries at a higher rate than it flows out. Thus, there is
storage of blood in the arteries
[7]
.
Figure 2: Windkessel Model
[8]
Figure 2: Windkessel Model
[8]
In 1970, a three-part Windkessel model was created by Westerhof
[9]
. Westerhof's model incorporates impedence
which, combined with Frank's model, includes aspects of wave propagation. Westerhof's work was inspired by the
previous work on pulsatile flow established by Womersley in his 1955 paper, as well as McDonald in his classic book
[10]

[11]
.
Mathematical Background
As mentioned above, Womersely's model makes use of Poiseuille flow, and is a simplification of the Navier-Stokes
equations. Therefore, it is important to have some background on these approaches. The Navier-Stokes equations can
be used to completely model the motion of incompressible, Newtonian fluids
(http://en.wikipedia.org/wiki/Newtonian_fluid) . However, these equations are very difficult to analyze since they are
non-linear, second order partial differential equations, and only in a few special cases can their exact solutions be found
[12]
. The equations, simplified using the continuity equation, for the directions are listed below
[12]
. In the
right-hand side of Equation (1), and similarly for Equations (2) and (3), the term represents the pressure force, the
term represents the weight of the fluid, and the second-order partials in parentheses represents the viscous forces.
Navier-Stokes Equations:

These equations are frequently written in cylindrical form. For example, Equation (1) can be written as:
(4)
A simplification of the Navier-Stokes equations can be made assuming a Poiseuille flow in which the velocity of the fluid
is described by the following equation.
(5)
A Poiseuille flow assumes that the flow is steady, uniform (over a cross-section), laminar
(http://en.wikipedia.org/wiki/Laminar_flow) , and axially symmetric within a cylindrical tube. Under these assumptions,
the term (since there is no change in velocity over time). Furthermore, the terms and equal zero.
Figure 3: Poiseuille Flow in a Rigid Tube
[13]
(For another project using Poiseuille flow, see MBW:Optimum Design of Blood Vessel Bifurcation)
Notation
Before examining Womersley's model, it is important to define some notation that will be used.
liquid density
viscosity
kinematic viscosity
pressures at ends of pipe
radius of pipe
distance from center of pipe
length of pipe
longitudinal velocity of liquid
fundamental frequency (typically the heart rate in
radians)
period of the pressure gradient wave
Womersley Flow Defined
We now take an in-depth look at Womersly's paper and model for pulsatile flow. To provide the basis for
Womersley's model, we begin with a more complete derivation of Poiseuille's formula for steady flow and also include
the pressure gradient. A constant pressure gradient throughout a pipe of length is defined as:
(6)
After accounting for the simplifications of the Navier-Stokes equations based on Poiseuille flow, the equation of motion
is then:
(7)
This has Equation (5) as its solution.
Now, Womersley expresses the pressure gradient as a periodic function of time with frequency to
represent the arterial pulse. The representation of this function is done using Fourier series.
Using Fourier Series to Represent Pressure Gradient
Since the change in pressure gradient is periodic, it can be expressed using the following function:
(8)
where .
Using Euler's formula, this can be expressed as a Fourier series with Fourier coefficients
.
(9)
Navigation
Main Page
Math Bio Wiki
Tools
Group Schedule
Funding
CU Courses
APPM 4390/5390
CSCI 7000
Researchers
David Bortz
Aaron Clauset
Vanja Dukic
Sam Flaxman
Manuel Lladser
Bret Melbourne
Juan Restrepo
Sarthok Sircar
Students
Anna Broido
Victoria Gershuny
Sam Hsu
Dustin Keck
Eric Kightley
Stephen Kissler
Vicky Li
Inom Mirzaev
John Nardini
Jay Stotsky
Miscellaneous
Alumni
The coefficients are calculated as follows:
(10)
(11)
These coefficients can be easily calculated using Matlab or Mathematica. A Fast Fourier Transform (FFT) algorithm
can also be used instead to calculate these coefficients. See
[14]
page 196 for more information and sample Matlab
code.
Solving for Flow Velocity (u)
Since we are assuming a Poiseuille flow that changes over time, we have that and . After some
algebra we get that for a single harmonic, n:
(12)
Let's consider the following simple solution to Equation (12) since the velocity also changes periodically over time:
(13)
Next we substitute this equation for into Equation (12) and divide both sides by to get the following ordinary
differential equation that is not dependent on time:
(14)
This equation can be rewritten to be a Bessel zero-order differential equation using the fact that :
(15)
Solving the Bessel Differential Equation
The general form for a zero-order Bessel differential equation (http://en.wikipedia.org/wiki/Bessel_function) is:
(16)
In our case note that . So the homogeneous version of Equation (15) fits the above form and its solution
is of the general form:
(17)
Alumni
How to Edit
(17)
The term The term must be discarded since has the requirement that it must be finite at the center (origin) of
the pipe. Therefore, . Since Equation (15) is non-homogeneous, we use the technique described in
[14]
and try
the simple solution . Then we have that the terms involving the derivative of are zero, and thus we have
that . Therefore, using the fact that , we have that:
(18)
Then the solution looks like:
(19)
We can now solve for by using the no-slip condition
(http://scienceworld.wolfram.com/physics/NoSlipCondition.html) that at the boundary . Therefore we
have:
(20)
(21)
Then finally we get that:
(22)
The last step is to add the steady flow velocity term and thus we have:
(23)
Womersley Number
The quantity in Equation (19) is called the Womersley Number, . It is a dimensionless parameter that
represents the ratio of transient forces, originating from the pulse wave, to the viscous force, or shear force. To get a
feel for the magnitude of , an example problem is presented.
Example 1: A 20 kg dog has an average heart rate of 90 beats per minute (bpm) and a 70 kg human has an average
of 70 bpm
[15]
.The density of blood is 1060 kg/ . Using the following calculation (
[15]
Equation 3.7, p. 27) we
calculate the diameter of the aorta to be 2.03 cm for humans, and 1.33 cm for dogs.

where W is the weight of the animal. From
[16]
we get that the blood viscosity for a human is 0.006 Ns/ and for a
dog is 0.0056 Ns/ .
Now we calculate Womersley's number: . Note that we need to convert bpm to rad/s so we
multiply the bpm by .
For a human, we have:
For a dog, we have:
This tells us that the oscillatory inertial forces become more important than the viscous force as the size of the animal
increases, or particulary as the size of the blood vessels increases. (Note that this example is adapted from a similar
example found in
[14]
, p. 30).
Flow Rate (Q)
In the next part of the paper, Womersley derives the flow rate (Q) of the fluid passing through a cross-sectional area of
the pipe, or artery in this case. This is accomplished by the integrating the velocity over a differential area:
(21)
For steady flow, recall that . After integrating, we get that:
(22)
which is Poisueille's formula. Now, we can substitute in the formula we found for , and then use the fact that
, then after some calculus and algebraic simplification we get that:
(23)
Recall that the need to be summed for each , and added to the average flow rate given by the
constant term in the Fourier series
[14]
to get:
(24)
(24)
Womersley now derives a way to calculate Q without Bessel functions, perhaps because back in 1955 these functions
were numerically more challenging to calculate. He uses modulus and phase functions [1] (http://dlmf.nist.gov/10.18) to
accomplish this. The details of these functions are out of the scope of this review, but the following relations are used:
(25)
(26)
He also uses the fact that the real part of is where
since
[14]
. Then, after some simplifications Womersley gets the
following, simplified formula for Q (not including the steady flow term ):
(27)
where and the values for , , , , and are all given in a table in Womersley's
paper. Let's look at an example problem presented in the paper to get more of an idea of the values for Q.
Example 2: The radius of an artery is 0.15 cm, the viscosity is , and the pulse frequency is 180 bpm.
The value for , , , , and are given in the following two tables:
The expression for Q is then then sum of the six terms:
This result for Q is plotted against the pressure gradient below:
Analysis and Conclusions
Womersley's arterial flow model gives formulas to calculate the flow rate of a viscous fluid through a rigid tube under a
periodic pressure gradient, described by a Fourier series. The solution is then extended to calculate the flow rate over a
cross-sectional area of the tube.
The graph in Example 2 depicts the pressure gradient (P) juxtaposed with the flow rate (Q), and clearly shows the
phase-lag between the two curves. This implies that the pulse wave is first sent through the body (seen as an increase in
the pressure gradient) and then the blood flow follows. Typically the phase-lag is about 90 degrees, except at the
boundary layer where it is about 45 degrees
[17]
. Also noteworthy is that this phase-lag is only present in large arteries;
in smaller arteries less than 4mm in diameter, the flow behavior closely follows Poiseuille's formula
[17]
. Interestingly, the
pulse wave travels about 5 times the maximum blood velocity
[17]
.
Also evident from the graph is that the direction of flow velocity is actually reversed as seen when the velocity becomes
negative. However, this is the topic of another paper written by Womersley et. al in 1955
[18]
.
Womersley's number has provided both fluid mechanics and biological sciences with a means to measure the inertial
forces versus the viscous forces. It is as significant in analyzing unsteady flow as the Reynolds number is in measuring
steady flow
[19]
.
Womersley's model has been foundational to many models of arterial blood flow, but itself is limited to modeling laminar
flow through a rigid, cylindrical tube. In reality, there are elastic effects in the blood vessel and the pressure gradient
may depend on other factors besides time. In addition, Womersley's model considers only longitudinal velocity,
whereas there may also be a radial part.
Recent Extension
In 2011, the Womersley article was cited in a study on Effects of vessel wall elasticity and non-Newtonian rheology on
blood flow regime and hemodynamic parameters distribution
(http://www.sciencedirect.com/science/article/pii/S1350453311002633) by Foad Kabinejadian and Dhanjoo N.
Ghista. This paper is a follow up of an earlier study (http://link.springer.com/article/10.1007%2Fs10439-010-0068-5)
, wherein they computationally simulated blood flow using Womersley's model. They were focused on vessel
intersections and junctions. The current paper revisits the computational simulation with compliant walls and non-
Newtonian fluid. They compared the results of their two studies and found that the compliant and non-Newtonian
model was consistent with the results of earlier studies and observational data. They conclude that the inclusion of wall
compliance and non-Newtonian rheology in flow simulation of blood vessels can be essential in quantitative and
comparative analyses
[20]
. For more information on human blood flow rheology, read Viscoelastic Versus Newtonian
Behavior (http://prl.aps.org/abstract/PRL/v110/i7/e078305) from the Physical Review Letters.
References
1. Womersley, J.R. Method for the calculation of velocity, rate of flow and viscous drag in arteries when the
pressure gradient is known, J. Physiol. 127 (1955) 553-563.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1365740/.
2.
2.0

2.1
NOVA. Amazing Heart Facts. http://www.pbs.org/wgbh/nova/heart/heartfacts.html.
3.
3.0

3.1

3.2
Keener, J. and J. Sneyd. Mathematical Physiology II: Systems Physiology. Second Edition. Springer,
3.
3.0

3.1

3.2
Keener, J. and J. Sneyd. Mathematical Physiology II: Systems Physiology. Second Edition. Springer,
New York, 2009.
4. Young, Thomas. The Croonian Lecture: On the Functions of the Heart and Arteries. Phil. Trans. R. Soc.
Lond. 1809 99, 1-31. http://rstl.royalsocietypublishing.org/content/99/1.2.full.pdf+html.]
5. Frank, Otto. Die Grundform Des Arteriellen Pulses Erste Abhandlung: Mathematische Analyse. Zeitschrift f
ur Biologie , 37 (1899). 483526.
6. Ethier, C. Ross , Craig A. Simmons. Introductory Biomechanics: From Cells to Organisms. Cambridge
University Press. 2007.
7. Wikipedia. Windkessel Effect. http://en.wikipedia.org/wiki/Windkessel_effect.
8. Saouti, N., et al. The arterial load in pulmonary hypertension. Eur Respir Rev 2010; 19: 117, 197203.
http://err.ersjournals.com/content/19/117/197.full.pdf+html.
9. Westerhof, Nico., Noodergraaf, A. Errors in measurement of hydraulic input impedence. J Biomech. 1970,
3: 351-356.
10. McDonald, DA. Regional Pulse Wave Velocity in the Arterial Tree. Journal of Applied Physiology. 1968,
24:73-78.
11. Gross, D.R., Animal Models in Cardiovascular Research. 3rd Edition. Springer Science + Business Media,
LLC. New York: 2009.
12.
12.0

12.1
Young, Donald F., Munson, Bruce R., and Theodore H. Okiishi. A Brief Introduction to Fluid
Mechanics. Second Edition. John Wiley & Sons, Inc. New York, 2001.
13. La Barck Jr., Anthony Joseph. A Computational Study of Blood Flow and Oxygen Transport during
Reperfusion and Post-conditioning. Thesis. Rowan University, 2010.
14.
14.0

14.1

14.2

14.3

14.4
Waite, Lee, and Jerry Fine. Applied Biofluid Mechanics. McGraw-Hill Professional.
Chicago. 2007.
15.
15.0

15.1
Li, John K-J. Comparative Cardiovascular Dynamics of Mammals. CRC Press. Boca Raton, 1996.
16. Windberger, U. et. al. Whole blood viscosity, plasma viscosity and erythrocyte aggregation in nine mammalian
species: reference values and comparison of data. The Physiological Society. Experimental Physiology (2003)
88.3, 431440.
17.
17.0

17.1

17.2
Lighthill, James (Sir). Mathematical Biolfluiddynamics. Society for Industrial Mathematics,
Philidelphia: 1975.
18. Womersley, J.R., Hale, J.F., and McDonald, D.A. Velocity profiles of oscillating arterial flow, with some
calculations of viscous drag and the reynolds number. J. Physiol, 128:629-640. 1955.
19. Chandran, K. B., et al. Biofluid Mechanics: the Human Circulation. CRC: Taylor and Francis, Boca Raton:
2007.
20. Kabinejadian, Foad and Ghista, Dhanjoo. Compliant model of a coupled sequential coronary arterial bypass
graft: Effects of vessel wall elasticity and non-Newtonian rheology on blood flow regime and hemodynamic
parameters distribution. Elsevier Ltd, 2011.
http://www.sciencedirect.com/science/article/pii/S1350453311002633.
Retrieved from "http://mathbio.colorado.edu/mediawiki/index.php/MBW:Womersley_Arterial_Flow"
MBW
Discussion
View source
History
What links here |
Related changes |
Upload file |
Special pages
| Permanent link
This page was last modified 05:52, 1 March 2013. This page has been accessed 9,251 times.
Designed by Paul Gu
Contact Applied Mathematics
Campus Directory Information:
(303) 492-1411

Womersley Arterial Flow - Simple

Uploaded by

Copyright:

Available Formats

Womersley Arterial Flow - Simple

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Womersley Arterial Flow - Simple

Uploaded by

Copyright:

Available Formats

4/16/2014 MBW:Womersley Arterial Flow - MathBio

You might also like