You are on page 1of 13

Flow of Liquids Through "Collapsible" Tubes

By J. P. HOLT, M.D.; PH.D.


With the technical assistance of Jean Allensicorth, David Collins and John Diana
Liquid flowing through a thin-walled "collapsible" tube similar to a vein, inclined to any
angle between vertical and slightly above horizontal, differs from the flow in rigid tubes
in that: a "collapsible" tube accepts any flow presented to it without affecting the rate
of flow through the system; the lateral pressure at all cross sections throughout the tube
is zero and does not change as the flow changes; the cross section of the tube and the
mean velocity of flow increase as the flow increases; for any particular flow, the cross
section increases and the mean velocity of flow diminishes as the potential energy difference
between the upper and lower ends of the tube is decreased; for any given flow and
potential energy difference, an increase in the length of the tube causes an increase in
the cross section and a decrease in the mean velocity; also, an increase in the perimeter
of the tube causes an increase in cross section and a decrease in mean velocity, and an
increase in the viscosity of the liquid causes an increase in cross section and a reduction
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

in mean velocity. In such tubes there is no lateral pressure gradient between any two cross
sections as there is in rigid tubes; instead there is a, potential energy gradient which is
proportional to the energy loss along the tube. Under certain conditions the flow of
liquid through a collapsible tube is described with a reasonable degree of accuracy by a
modification of Poiseuille's law for tubes of elliptical cross section.

I N THE flow of liquids through "collapsi-


ble" tubes, such as veins, the cross section
of the tube is free to change as the lateral pres-
tion. When a collapsible tube becomes dis-
tended to the point, that its cross section is
circular, and its wall is stretched, it no longer
sure within the tube changes. This differs behaves as a collapsible tube but functions as
from the flow through rigid tubes where the a rigid tube, and the flow through it is de-
cross section remains fixed regardless of the scribed by the classical laws of hydrodynamics
lateral pressure, and differs from the flow for circular tubes.
through elastic tubes of circular cross section Following our earlier studies1"3 of the flow
in which the cross section increases as the of fluids through collapsible tubes, numerous
lateral pressure increases; it differs from the investigations have been carried out concern-
flow in open channels where there is a liquid- ing such tubes. These include studies involv-
solid interface on part of the boundary and a ing venous return to the chest,4 energy rela-
liquid-gas interface where the surface is in tionships,5 the flow in particular organs,0 and
contact with air; and it differs from a viscous a study of the flow through long, vertical
liquid which, when poured from a container, collapsible tubes by Duomareo et al.,7 in which
falls freely in air and has a liquid-gas inter- he found that there is zero lateral pressure
face extending over the entire surface. In a throughout the length of such a tube. A
collapsible tube there is a liquid-solid inter- comprehensive review of the literature is given
face extending over the entire surface, even by Breeher.*
though the tube is free to change its cross sec- In our earlier studies we were concerned
only with the flow in collapsible tubes that
From the University of Louisville Institute for were in the collapsed state for a short length
Medical Research, Louisville, Ky. of the tube, the remaining part being dis-
Supported in part by grants-in-aid from the Ameri- tended to the point where it closely approxi-
can, Kentucky, Louisville and Jefferson County Heart
Associations.
mated a rigid tube with a circular cross sec-
A preliminary report of this work was given at tion. In our present studies we have been con-
the meeting of the American Physiological Society in cerned with the flow through long tubes col-
Chicago, March 1957. lapsed throughout their entire length; that is,
Received for publication Xovemher 12, 195S. either a vertical collapsible tube, or one in-
.-542 Circulation Rcacarch, Volume VII, Man '''>
FLOW THROUGH COLLAPSIBLE TUBES 343

clined to some angle between slightly above into the upper end of the collapsible tube, Pj,
horizontal and vertical. In mammals the flow through the tube to P 2 where the collapsible tube
through veins that are above heart level is was connected to a rigid tube, and through semi-
rigid tubes to the outflow, Z, located several centi-
generally of this nature; an extreme example meters below P 2 . The flow in milliliters per second
is the flow of blood through the external jugu- through the collapsible tube was changed by vary-
lar vein of the giraffe, when the head is above ing the resistance between the Mariotte bottle and
heart level. In these studies we have been Pj, and was measured by collecting the outflow in
interested in learning the inter-relationships a graduate cylinder in a given time. The collap-
sible tube was placed within a plethysmograph
between the following factors: rate of flow which was connected to a volume recorder con-
(milliliter per second), cross section, mean sisting of a paraffin-coated 120 cm. long glass tube,
velocity (centimeter per second), viscosity, having a radius of 0.15 cm., containing a small
length of tube, resistance, potential energy bubble of water. With zero flow through the
difference between any two cross sections of collapsible tube the position of the water bubble
the tube, and the perimeter of the tube. in the volume recorder was read. A few seconds
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

after a particular flow had been established the


Brecher 4 has suggested that the modified volume recorder was read again. The difference
Poiseuille equation for elliptical tubes, F = between the two readings gave the internal volume
of the collapsible tube for the particular flow.
given by Milne-Thomp- The volume recorder was gently vibrated at the
VL ar+l?
time of reading. Since, as will be shown, the cross
son,8 describes the flow through collapsible section of the collapsible tube was the same at all
tubes. It should be noted that the above points throughout the length of the tube (except
a3b3 . , . .. for a few centimeters at the upper and lower ends)
term, where a is the major semi-axis the cross section of the tube for a particular flow
a?+b2 was calculated by dividing the internal volume
and Z> the minor semi-axis of an ellipse, re- of the tube by its length. Since we knew the
places the radius to the fourth power, r 4 , in perimeter of the collapsible tube and the experi-
mentally determined cross section (for each flow
the classical Poiseuille equation, and that the studied), it was possible to calculate, from the
theclass ic a l Poiseuille equation, and well known equations for the perimeter and cross
circular tubes is replaced b y in the above section of an ellipse, the major and minor axes,
4 and b, of the ellipse having the same cross sec-
elliptical modification. Since there have been tion as that determined experimentally, The ap-
paratus was hinged at the level P 2 , so that the
no experimental studies concerning the ap- collapsible tube and plethysmograph could be
plicability of this equation to the flow through rotated about P 2 and thus inclined to any desired
long collapsible tubes, collapsed throughout angle with the horizontal. In this way the vertical
their entire length, we have attempted to distance between the upper and lower ends of the
ascertain how the equation applies to the flow tube could be changed, thus varying the potential
energy difference between the inflow and the out-
through such tubes; and also to learn how
flow ends of the tube. The potential energy differ-
the flow through these tubes is related to the ence between the two ends is equal to (pghj-pgh2),
classical hydrodynamics based on Bernoulli's where p is the density of the liquid, g the gravita-
theorem. 9 I t will be shown that the flow of tional constant, hx the vertical height of the upper
viscous liquids through such collapsible tubes end of the tube above a fixed reference plane, and
is a special case of Bernoulli's theorem in h2 the vertical height of the outflow end above the
same reference plane. The mean velocity of flow
which the lateral pressure and kinetic energy in centimeters per second through the collapsible
functions become zero, leaving only the po- tube was calculated by dividing the flow in railli-
tential energy and heat energy functions. liters per second by the cross section. The viscosity
of the different solutions studied was measured
METHODS with an Oswald viseosimeter, and expressed as
The apparatus, figure 1, consisted of a Mariotte relative viscosity to 0.9 per cent NaCl solution,
bottle from which 0.9 per cent XaCl solution con- which was taken as 1. More than five different liq-
taining different concentrations of sucrose flowed uids having relative viscosities ranging from 1 to
through heavy-walled, semi-rigid rubber tubing 14 have been studied. The temperature in the
Clear, Bed Label Karo Syrup. plpthysmograph was maintained constant. Pros-
344 HOLT
AWnott BofMe
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

Scale
FIG. 1 Left, Jlodel used in studying flow of liquids through ;i collapsible tube.
PIG. 2 Right. Diagrams of the cross section of a "collapsible" tube through which different
amounts of liquid are flowing: 1, zero flow; S, small flow; 3 and 4, larger flows; 5, flow so
large that the tube has become circular and no longer functions as a collapsible tube, a and
h, major and minor semi-axes of the elliptical cross section.

sures were measured by means of water manome- factors on theflowcross-sectioncurve was studied.
ters or Statham strain gages, and recorded with a All of the studies have been carried out on thin-
Brush electromagnetic oscillograph. The pressure walled Penrose rubber tubing having lengths from
at different points within the collapsible tube was 9.75 to 78.5 cm., and perimeters of 2.0, 3.85 and
measured by passing a small-bore polyethylene 6.0 cm.
tube through a side arm near Pi to different points Two other series of similar experiments were
throughout the length of the collapsible tube. The carried out which differed from the above in that
conformation of the collapsible tube at its upper the plethysinograph was not used; instead an in-
and lower ends differs from that of the rest of dicator was injected instantaneously at P^ and its
the tube, and was studied by photographing the time-concentration curve recorded at P.,. This
ends of the tube. The volume of the collapsible method gave inconsistent results, as shown by the
tube at its ends was calculated by planimetering fact that the outflow calculated from the indicator
the area on the photographs, and in some eases dilution curve frequently differed greatly from the
the measured volume of the tube was corrected by true outflow. This may be explained by the fact,
subtracting the excess volume at the two ends from as shown by Rossi et al.,10 that with streamline
the total. This correction was generally small. flow the indicator dilution technie does not give an
In an experiment all of the factors (viscosity, accurate measure of flow.
length, perimeter, and potential energy difference
between the ends of the tube) were maintained con- RESULTS
stant, and the volume and cross section of the verti-
cal collapsible tube determined for flows ranging In a liquid-filled system such as shown in
from zero to 20 ml./sec. This was carried out with figure .1, when there is zero flow through a 78
the tube inclined to angles of 9, 18, 30, 57 and i n . long collapsible tube inclined to some
00; in this way the potential energy difference, be- angle with the horizontal, the tube has a uni-
tween the ends of the tube was changed. This entire
procedure was carried out for different viscosity form shape extending throughout its length.
liquids, tubes of different length, and tube of differ- Tts cross section is of the shape shown in stage
ent perimeter. Tims the effect of each of the above 1. figure 2. When a small constant flow is
FLOW T l l R O r m i COLLAPSIBLE Tl/BES 345

established through the tube it assumes


throughout its length, except for the upper
3 cm. and the lower 3 cm., the cross section
shown in stage 2. figure 2. At the upper end
the tube decreases in cross section more or
less uniformly from the rigid circular tube
to which it is attached to a point approxi-
mately 3 cm. below the upper end, where it
assumes the cross section shown in stage 2. The
lower 3 cm. of the tube assumes the dilated
hour-glass shape shown in figure 3. This hour-
glass shape is more pronounced with large
flows. With the establishment of a greater
flow, the tube takes on throughout its length,
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

except for the upper and lower 3 cm., the


elliptical cross section shown in stage 3, figure
2. With a still greater flow, the elliptical
shape shown in stage 4 is assumed throughout
its length, except for the upper and lower
ends. If a very large flow is established the
tube becomes circular in cross section through-
out its entire length. It develops a lateral
pressure inside the tube which is greater thau
the pressure outside, and the tube no longer
1'IG. 3. Photograph of lower 10 cm. of a vertical
behaves as a "collapsible" tube but as a rigid
collapsible tube through which a 5.25 viscosity liquid
or elastic circular tube. was flowing at a rate of 14.0 nil./sec. The connection
"Collapsible" Tube Docs Xot Affect Flow. of the outflow end of this collapsible tube to a rigid
The presence of a vertical collapsible tube in glass tube is shown at the bottom.
a system such as that shown in figure 1 causes
no change in the amount of liquid flowing factor the flows became approximately equal.
from the Mariotte bottle system. This is shown Also, when the tube is in the extremely col-
in figure 4/1 where the pressure, Y (figure 1), lapsed state, two small capillary tubes form
was maintained constant and a 78 cm. long, along either side of the collapsible tube and
3.85 perimeter collapsible tube, having a 4.5 serve as more or less rigid tubes in parallel
viscosity liquid flowing through it, was rotated with the collapsible tube. Therefore, with
about V-, changing the flow and the pressure small flows the flow is slightly greater with
head V /',. This was done with the collaps the collapsible tube in the system. It is con-
ible tube in the system and with the collapsible cluded that a perfectly collapsible tube ac-
tube removed from the system; in the latter cepts whatever flow is presented to it without
case the outflow was measured from ZY The affecting the flow through the system.
flows were nearly the same but were always a Lateral Pressure Is Zero at All Cross Sec-
little larger with the collapsible tube in the tions. The lateral pressure at all cross sections
system. The reason for this is that the thin- throughout a vertical collapsible tube, except
walled rubber collapsible tube is not perfectly for a few centimeters at the upper and lower
collapsible, as are veins. Because of its elas- ends, is zero and does not change as the rate
ticity the upper and lower few centimeters of of flow through the tube changes. This is
the tube do not collapse perfectly, and with shown in figure 4B, in which the lateral pres-
small flows the pressure 3 cm. below Py is sub- sure, at a point 3 cm. below the upper end of
atmospheric, as shown in figure 4B. When the collapsible tube, is plotted against the flow
the data in figure 4A were corrected for this when a 7.6 viscosity liquid flowed through a
346 HOLT

40
With Collapsible Tube
x No Collapsible Tube
,30
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

4 8 12 12 4 8 12 16
Flow (ml/sec.) Flow (ml./sec)
FIG. 4. A, Belationship between pressure head, (Y-Pi), and flow with the collapsible tube in
the system and with the tube removed from the system. B, Belationship between lateral pres-
sure and flow. The lateral pressure was measured in the collapsible tube, at a point 3 em.
below the upper end of the tube, taking this point as the zero reference level. C, Kelationshlp
between cross section and flow through a 78 cm. long, 3.85 cm. perimeter collapsible tube
inclined to an angle of 30 for liquids having relative viscosities of 1, 2.9, 5.7 and 12.5.
D, Relationship between cross section and flow of a 12.5 viscosity liquid flowing through a
78.5 cm. long, 3.85 cm. perimeter collapsible tube when inclined to 9, 18, 30, 57 and
90 with the horizontal. E, Relationship between flow and cross section for collapsible tubes of
different lengths. The numbers 9.8, 19.5, 39.5 and 78 cm. are the lengths of the tubes. F,
Relationship between flow and cross section for 78 cm. long collapsible tubes of different
perimeters. The numbers 2.0, 3.85 and 5.8 cm. are the perimeters of the tubes, v is viscosity.

78 cm. long, 3.85 cm. perimeter tube inclined in pressure from 107.8 cm. water below atmos-
90 to the horizontal. In this experiment the pheric to 3 cm. below atmospheric. Further
zero position was taken as the point of pres- increase in flow was associated with a slight
sure measurement in the tube 3 cm. below the increase in pressure, until it reached zero,
upper end of the tube. As the flow increased where it remained for large flows. Measure-
from zero to 2 ml./sec. there was a rapid rise ment of the lateral pressure at all points
PLOW THROUGH COLLAPSIBLE TUBES 347

throughout the length of the vertical tube,


taking the point of pressure measurement in
the tube to be the zero level, gave similar re-
sults to the above, except at the lower 3 cm. of
the tube, where the pressure was always sub-
atmospheric. Similar results were obtained
when the tube was inclined to various angles
with the horizontal instead of being in the
vertical position. The reason that the pres-
sure differed from zero when the flow was
small is that the tube possessed certain elastic
properties and several centimeters of water
pressure was required to maintain the defor-
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

mation of the tube. As a result, with zero


flow through the system two small, more or
less rigid, capillary tubes form along the sides
of the collapsible tube, and because of these
the pressure, Pi, registers the level of the out-
flow from the system where the liquid passes
into the atmosphere. As the flow increases
from zero to 2 ml./sec, the pressure, Pi, rises
rapidly, because the effect of the two small
capillary tubes that form along the lateral
edges of the collapsible tube is largely over-
come ; as the flow increases beyond 2 ml./sec.
there is little change in Pi because the effect of 20 40 60 80
the two small rigid tubes has been overcome Velocity (cnv/tcj
and the tube acts as a collapsible tube with FIG. O. A, Relationship between mean velocity and
the lateral pressure being independent of the flow for the data shown in figure 4C. B, Relationship
rate of flow. between mean velocity and cross section of the tube
Cross Section Same Throughout Tube. As for the data in figure 4C.
described above, the cross section of a collaps-
ible tube through which liquid is flowing at a mean flow time. Although there was consider-
constant rate appears on visual inspection to able scatter of the data in some experiments,
be the same throughout the length of the tube, for a given rate of flow throughout the collap-
except for the upper and lower ends. Meas- sible tube a linear relationship was found
urements, taken by means of calipers through- between the mean flow time and the length of
out the length of a 78 cm. long collapsible tube the collapsible tube traversed by the indicator,
inclined to some angle with the horizontal, indicating that the cross section was uniform
through which liquid was flowing at a con- throughout the tube. Since the cross section
stant rate, showed no measurable difference in is the same throughout the tube and the same
the minor axis of the more or less elliptical amount of liquid is flowing through each cross
cross section shown in stage 3 (fig. 2). Further section at any given time, it follows that the
evidence in support of this was obtained by mean velocity of the flow is the same at all
injecting an indicator at different points cross sections.
throughout the length of the "collapsible" Flow Increase Causes Cross Section In-
tube, when there was a constant flow through crease. In figure iC is shown the relationship
the tube, and recording the indicator time-con- between flow and cross section of the collap-
centration curve at P2, thus measuring the sible tube, when liquid having relative viscos-
348 HOLT
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

FIG. 6. A , Relationship between the elliptical cross tube inclined to different angles with the horizontal.
section function, a
(ill
and flow for liquids of dif- This data was taken from figure 40. p is the spe-
b
cific gravity of the liquid; g is the acceleration due to
ferent viscosities flowing through a 78 cm. long, 3.85 gravity; 7d and It? are the vertical heights of the two
cm. perimeter tube inclined to an angle of 30. This ends of the collapsible tube above a reference plane.
data was taken from figure 4C. B, Relationship be- 1), Relationship between the elliptical cross section
tween viscosity and the elliptical cross section func-
a'b' , for the data shown in figure 4C C, function, -, and the length of the collapsible tube
tion, (f+b
<r+V for tubes of different length. Data from figure
Relationship between the elliptical cross section
4E. E, Relationship between the elliptical cross see-
function, n'b*;, and the product (flow divided by
ar+V tion function, - , and the flow for tubes of dif-
the potential energy difference between the two ends
of the tube) for liquid flowing through a collapsible ferent perimeters. Data from figure 4F.

ities of 1, 2.9, 5.7, and 12.5 flowed through this was done for the data in figure 46', re-
a 78 cm. long, 3.85 cm. perimeter tube inclined sults shown in figure 6A were obtained. The
30 to the horizontal. It will be noted that as relationship is approximately linear for flows
the flow increases the cross section increases greater than 2 ml./sec. with the elliptical
in a curvilinear manner, and that for any function increasing as the flow increases.
particular flow the greater the viscosity the Velocity Increases As Flow Increases. In
larger the cross section. It it is assumed figure oA is shown the relationship between
that the cross section of the tube is of ellip- the flow in milliliters per second and the mean
tical shape, then the function of the elliptical velocity of flow in centimeters per second for
the data in figure 46\ As the flow increases
cross section, -. can be calculated. AYhen the mean velocitv increases in a curvilinear
or + i2
FLOW THROUGH COLLAPSIBLE TUBES 349

manner, and for a given flow the velocity de- Increase in Length Causes Increase in Cross
creases as the relative viscosity increases. Section. The relationship between flow per
Cross Section Increases As Velocity In- unit time and cross section for a 5.7 viscosity
creases. The relationship between mean ve- liquid in 3.85 cm. perimeter tubes of different
locity of flow and the cross section of the lengths, inclined to angles such that the verti-
collapsible tube for the data in figure 4C is cal difference between the ends of the tube
shown in figure 5B. As the velocity increases was always 9.8 cm. (and so having a constant
the cross section increases in a curvilinear potential energy difference between the ends
manlier. of the tube), is shown in figure 42?. As the
Cross Section Directly Belated to Viscosity. flow increases so does the cross section and
When for a particular flow, for the data there is a family of curves which describes
shown in figure 46', the cross section is plotted this relationship for tubes of all lengths. For
against the viscosity, a curvilinear relation- any given flow the cross section is greater
ship is obtained with the cross section in- the longer the tube. For each flow when
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

creasing as the viscosity increases. When the the elliptical function of the cross section,
elliptical function of the cross section, al
a?b3 , is plotted against the length, from
, is plotted against the viscosity for
2 a- + b-
a~ b
the data in figure 42?, a family of curves is
this data the results shown in figure 67J are
obtained as shown in figure QD. As the length
obtained. As the viscosity increases, the
increases the elliptical cross section function
elliptical function of the cross section in-
increases in a curvilinear manner.
creases in a linear fashion for flows greater
Perimeter Increase Causes an Increase in
than 2 ml ./sec.
Gross Section. The relationship between flow
Cross Section Inversely Related to Potential
per unit time and cross section for a 5.6 vis-
Energy Difference Between Upper and Lotoer
cosity liquid in 78 cm. long collapsible tubes
Ends of the Tube. Figure 4D shows the rela-
oi different perimeters, inclined to an angle
tionship between the flow per unit time and
of 9 degrees with the horizontal, and so hav-
cross section of a 78.5 cm. long, 3.85 cm. per-
ing a vertical difference between the two ends
imeter collapsible tube when the potential
of 9.8 cm., is shown in figure 4F. As the
energy difference between the two ends of the
flow increases the cross section increases, and
tube was changed by inclining the tube to
for any particular flow the larger the perim-
angles of 9, 18, 30, 57 and 90 for a
eter the larger the cross section. In figure 67?
liquid having a relative viscosity of 12.5. For
is shown the relationship between the flow
a given flow, as the angle of inclination to
and:r :i the function of the cross section,
the horizontal decreases the cross section in- ab
, for tubes of different perimeters.
creases ; and there is a family of curves which
a" + b-
describes the relationship between flow and
These data are taken from curves in figure
cross section for all angles of inclination.
IF. There is an approximately linear rela-
When the elliptical cross section function,
tionship, and this relationship is the same for
a3b3 tubes of different perimeters. When the po-
, is plotted against the flow divided tential energy difference between the two ends
a- + b" of the tube was increased, by increasing the
by the potential energy difference between the angle of inclination with the horizontal, the
two ends of the tube, F/(h\ p IJ Ibj p g), for cross sections were smaller than those in figure
the data in figure \l) the linear relationship (>/*.', and under these circumstances the ellipti-
obtained is shown in figure (iC. As the frac- cal cross section function, -, was always
2
a- + b
tion, flow divided by the potential energy
larger for the smaller perimeter tubes. The
difference, increases so does the elliptical cross
section function. reason for this is not clear.
350 HOLT

Energy Relationships. On the basis of the in the lateral pressure difference between the
law of conservation of energy, applying Ber- two ends of the tube in addition to an increase
noulli 's theorem9 to the flow of viscous liquids, in the above product. Thus, in collapsible
the sum of the total mechanical energy and tubes, inclined to some angle with the horizon-
heat energy at any cross section of a perfectly tal, there is no lateral pressure gradient be-
collapsible tube must be the same as that at tween any two cross sections, instead there is
any other cross section. This is expressed in a potential energy gradient which is propor-
the equation9: tional to the energy loss along the tube:
AMg (lh-h2) + puiiv, AT p~a2v2 AT = hx P ghi P g
1/2 &M (i^-W) + TL AM ~ T = _ = T P

where AM Gm. of liquid enter a cross section In the above we were concerned with energy
near the upper end of the tube and A M Gm. changes within the collapsible tube extending
simultaneously exit through a cross section from the point farthest upstream where the
near the lower end in the time, AT. pi is tube is collapsed to the farthest point down-
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

the lateral pressure, a^ the cross section, hx stream where the tube is collapsed. In the
the vertical height above an appropriate plane, case of a horizontal collapsible tube, or one
and Vi the velocity of the liquid particles at inclined to some angle below horizontal, at
a cross section near the upper end; while p2, the outflow end of which is applied a subat-
a2, h2, and v% are the corresponding values mospheric pressure, Po, the collapsible tube is
near the lower end of the tube, r is the heat open throughout its entire length except a
energy per unit mass, per unit length, lost by few millimeters of the tube which are col-
the liquid; L is the length of tube between the lapsed at its outflow end.1 The pressure, Plt
two cross sections where the pressure is meas- immediately upstream to the collapsed seg-
ured, and g is the acceleration due to gravity. ment in the open part of the collapsible tube,
Since, as shown earlier, the cross section, lat- is slightly above atmospheric, while the pres-
eral pressure and velocity are the same at all sure at some point within the collapsed
cross sections of the tube, three of the above segment is atmospheric. The pressure differ-
terms drop out of the equation leaving: ence from immediately above to immediately
AMg (hrh,) = L AM below the short collapsed segment is (PiPo).
That is, the flow through such collapsible If the cross sections of the open tube above
tubes is a special ease of Bernoulli's theorem and below the collapsed segment are the same
for the flow of viscous liquids, in which the then the energy difference across the segment
potential energy difference between any two is:
cross sections is equal to the heat energy loss PithXv AT i/=r AT =TL AM.
due to friction as the liquid moves from the A collapsible tube inclined to some angle
upper to the lower cross section. In a collaps- with the horizontal, to which a negative pres-
ible tube having a certain difference in vertical sure is applied on the downstream side of the
height between the two ends of the tube, an collapsible tube, consists of two collapsible
increased flow through the tube causes no segments in series; that is, the long collapsed
change in the lateral pressure but causes an segment extending throughout the length of
increase in the cross section. The increased the inclined tube, and the collapsed segment
energy difference between any two cross sec- at its outflow end. The energy difference,
tions of the collapsible tube, that is associated between the upper end of the long collapsible
with an increased flow through the tube, mani- segment and the point immediately down-
fests itself as an increase in the product: stream to the collapsible segment at the out-
AMg (7*1 - ho). This is in contrast to the flow flow end, is the sum of the energy differences
through a comparable rigid tube in which ease from the inflow to the outflow ends of the
the increase in energy difference between the two collapsed segments.
two ends of the tube, associated with an in- Resistance. The resistance to flow through
crease in flow, manifests itself as an increase a rigid tube system of constant cross sec-
FLOW THROUGH COLLAPSIBLE TUBES 351

tion is generally taken to be the product. The fact that the cross section is the same
/\ t) throughout the length of a long collapsible
-, where Ap is the lateral pressure differ- tube is in contrast to the fact that the cross
F section of a free-falling liquid, poured from
ence between two cross sections along the a beaker, decreases as the distance from the
length of the tube, and F the flow per unit beaker increases. The reason for this differ-
of time. In laminar flow as the Ap increases ence would appear to be the fact that there
the F increases and the resistance remains is little or no frictional resistance at the air-
constant. Since the lateral pressure is zero liquid interface of the free-falling liquid,
at all points in a collapsible tube, inclined to while in the collapsible tube there is frictional
some angle with the horizontal, and does not resistance at the liquid-solid interface at the
change as the flow changes, if we substitute perimeter of the collapsible tube.
the vertiele height (7ii 7i2), above a fixed refer- These studies indicate that in a perfectly
ence plane, of the two cross sections for the collapsible tube inclined to angles between
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

lateral pressure Pi and po, then the resistance horizontal and vertical the lateral pressure
mav be taken to be, R = hj L 9~lh p . This is zero at all points regardless of the flow.
F Our results confirm those obtained by Duo-
resistance has the same dimensions and mean- marco et al.7 in studies on a model, and are in
ing as the conventional resistance. "With zero general agreement with the theoretical and
flow in a collapsible tube, the resistance is experimental results he reported5 for the ve-
infinite, and as the flow increases the resist- nous system of the dog (when held in the
ance decreases. vertical, head-up, position).
In the case of a horizontal collapsible tube, Although the resistance to flow through
collapsed only at its downstream end, the veins in'the circulatory system is generally
PiP.. small when the veins are not collapsed, it
resistance, R, is equal to : As the flow should be pointed out that the resistance is
F high in collapsed veins. The giraffe is an
increases the resistance decreases, and with example; only a relatively small amount of
zero flow the resistance is infinite. Thus, in energy is required to move a given quantity
a collapsed tube system in which the above of blood from the aorta to the head of a
two collapsed segments are in series, as in giraffe, with the head at heart level, however,
a collapsible tube inclined to some angle be- a much larger amount of energy is required
tween vertical and horizontal, and to which to raise the same amount of blood to head
suction is applied at the downstream end of level when the head is in the vertical position
the collapsed segment, the total resistance is because of the large potential energy differ-
the sum of the above resistances. ence between head and heart level. This
DISCUSSION large amount of potential energy is dissipated
Since the rate of flow through a collapsible in the form of heat when the blood passes
tube system, with the tube collapsed through- from the head back to the heart through the
out its length, is unaffected by the presence collapsed veins in the neck where the resist-
of the collapsible tube, it is clear that freely ance is large and the velocity high. It would
collapsible veins running from a part above appear that the collapse of veins in the neck
heart level, such as the head, back to the of the giraffe is a protective mechanism that
heart can exert no siphoning effect on the tends to maintain the pressure in the capil-
flow of blood to the part. Thus, the flow laries of the head above atmospheric, and to
is determined solely by the pressure head, dissipate the large amount of potential energy
from the aorta to the highest point where that the blood possesses at head level on its
the vein is collapsed, and the resistance to return to the heart; in this way the blood
flow along these arterial and capillary chan- returns to the heart without excessive kinetic
nels. energv.
352 HOLT

Attempts to obtain a mathematical equa- -, and the length, L, suggests


function,
tion that will accurately describe, for the a2 + b2
flow of liquid through a collapsible tube, the that this equation is not accurate insofar as
relationships between viscosity, flow per unit length is concerned. As mentioned, this dis-
time, length, perimeter, potential energy crepancy may be caused by "end-effects."
difference between the ends of the tube, and Although some of the data indicates that the
the cross section has not been entirely suc- perimeter of the tube has no effect on the
cessful because of the following: 1. Thin- flow, other data (when the cross section is
walled rubber tubes are not perfectly col- small) suggests that the perimeter does have
lapsible but possess elastic properties which an effect. Further experiments are necessary
cause them to resist deformation. 2. The in order to clarify these questions.
cross section of a collapsible tube may vary
from the extreme of a circle to the other ex- SUMMARY
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

treme of a line. As a result the Poiseuille


Studies of the flow of different viscosity
equation for circular tubes does not apply.
liquids through long "collapsible" tubes
The only other presently-available equation
is that for flow through elliptical tubes, and (Penrose tubing) inclined to some angle be-
it is not satisfactory for all cross sections of tween horizontal and vertical, and collapsed
experimental collapsible tubes because such throughout the length of the tube were car-
tubes are only approximately elliptical for ried out in a model. The flow in such a
certain cross sections. 3. Technical difficulties system is similar to the flow in the external
are encountered in measuring the cross sec- jugular vein of the giraffe when the head is
tion of a collapsible tube. These difficulties above heart level. The inter-relationship be-
are caused by the increase in cross section at tween the following factors was studied: rate
the two ends of the tube, as well as unknown of flow through the tube (milliliter/second),
"end-effects" where the collapsible tube cross section of the tube, mean velocity of flow
joins the rigid tube at its upper and lower (centimeter/second), viscosity of the liquid,
ends. length of the tube, resistance to flow through
The fact that an approximately linear the tube, the potential energy difference be-
relationship is obtained in our experiments, tween the two ends of the tube, and the perim-
with flows ranging between 2 and 16 ml./ eter of the tube. It was shown that, except for
sec, between the elliptical cross section the length of the tube, the flow through such
tubes is described with a reasonable degree
function, 2 _ , and : flow, viscosity, and of accuracy by a modification, for elliptical
a + b~
Flow tubes, of Poiseuille;s law for circular tubes;
suggests that insofar as these and that the flow of viscous liquids through
ll, p (Jh2 p (J
such tubes is a special ease of Bernoulli's
parameters are concerned an elliptical modi-
theorem in which the kinetic energy and lat-
fication of Poiseuille's equation:
eral pressure functions become zero, leaving
n (h. p gh. p g) aV
F = only the potential energy and heat energy
4>/I a* + b* functions. There is no lateral pressure gra-
describes the flow through collapsible tubes dient in the flow of liquids through such
that are collapsed throughout their length. tubes, instead there is only a potential energy
In the case of collapsible tubes that are col- gradient.
lapsed only at the downstream end, to which SUMMARIO IX IXTEKLIXUCA
negative pressure is applied, the above term, Esseva effectuate, in nn modello, studios
(/'i P gh P g) is replaced by the lateral pres- relative al fiuxo de fluidos de differente vis-
sure, (]\-lJ2). The fact that we did not ob- cositate a transverso longe "collabibile"
tain a linear relationship between the elliptical
FLOW THROUGH COLLAPSIBLE TUBES 353

tubos inclinate a angulos inter horizontal e REFERENCES


vertieal e collabite ab tonnino a termino. Le 1. HOLT, J. P.: The collapse factor in the meas-
fluxo in un tal systema es simile al fluxo in urement of venous pressure: The flow of
fluid through collapsible tubes. Am. J.
le externe vena jugular del girafa quando Physiol. 134: 292, 1941.
le eapite del animal es supra le nivello de sii 2. : The effect of positive and negative intra-
eorde. Esseva studiate e relatioues inter le thoracic pressure on peripheral venous pres-
sequente factores: Intensitate del fluxo a sure in man. Am. J. Physiol. 139: 208, 1943.
transverso le tubo (cm3/sec), section trans- 3. :The effect of positive and negative intra-
thoraeic pressure on cardiac output and
verse del tubo, veloeitate medie del fluxo venous pressure in the dog. Am. J. Physiol.
(cm/see), viscostitate del fluido, longor del 142: 594, 1944.
tubo, resistentia al fluxo a transverso le 4. BRECHER, G. A.: Venous Return. New York,
tubo, le differentia de energia potential inter Grune & Stratton, 1956.
le duo terminos del tubo, e le perimetro del 5. DUOMARCO, J. L., ASD RIMINI, R.: Energy
and hydraulic gradients along systemic
Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

tubo. Esseva monstrate que, eon le excep- veins. Am. J. Physiol. 178: 215, 1954.
tion del longor del tubo, le fluxo a transverso 6. RODBARD, S.: Distribution of flow through a
tal tubos pote esser deseribite eon grados pulmonary manifold. Am. Heart J. 51: 106,
adequate de aecuratia per un modification 1956.
(pro tubos elliptic) del lege de Poiseuille 7. DUOMARCO, J., RIMINI, R., AND RECARTE, P.:
La presion de los troncos venosos de torax:
pro tubos circular e que le fluxo de fluidos Apendice. Rev. argent, eardiol. 12: 129,
viscose a transverso tal tubos es un caso 1945.
special del theorema de Bernoulli, eon le S. MILNE-THOMSOX, L. M.: Theoretical Hydro-
functiones del energia cinetic e del pression dynamics. New York, Macmillan, 1950.
lateral deveniente zero de maniera que sol- 9. FERENCE, M., JR., LEMON, H. B., AND STEPH-
ENSON, R. J.: Analytical Experimental
mente le functiones de energia potential e de Physics. Second Revised Edition, Univer-
energia caloric remane. II non existe, in le sity of Chicago Press, 1956, p. 152.
fluxo de fluido a transverso tal tubos, un 10. Rossi, H. H., POWERS, S. H., AND DWORK, B.:
gradiente de pression lateral. II existe sol- Measurement of flow hi* straight tubes by
mente un gradiente de energia potential. means of the dilution technique. Am. J.
Physiol. 173: 103, 1953.
Flow of Liquids Through "Collapsible" Tubes
J. P. HOLT

Circ Res. 1959;7:342-353


Downloaded from http://circres.ahajournals.org/ by guest on November 22, 2017

doi: 10.1161/01.RES.7.3.342
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright 1959 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/7/3/342

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:


http://www.lww.com/reprints

Subscriptions: Information about subscribing to Circulation Research is online at:


http://circres.ahajournals.org//subscriptions/

You might also like