SOCIETY OF PETROLEUM ENGINEERS OF AIME
6200 North Central Expressway =R SPE 16$5
Dallas, Texas KJZ?06
THIS IS A PREPRINT --- SUBJECT TO CORRECTION
Flow of Viscoelastic Fluids Through Porous Media
By
Narendranath Y. Gaitonde and.Stanley Middleman,*
U. of Rochester, Rochester, N- Y.
0 Copyright 1966
American Institute of Mining, Nletrdlrrrgical and Petroleum Engineers, Inc.
This paper was prepared for the Society of Petroleum Engineers Symposium un Mechanics of
RtreologicallyComplex Fluids) to be held in Houston> Tex.? DeCW 15-16~ 1%6* pe~ission to cOPY
restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abst
should contain conspicuous acknowledgment of where and by whom the paper is presented. Publicati
elsewhere after publication in the JOURNAL OF PETROLEUM TECHNOLOGY or the SOCIETY OF PETROLEUM
ENGINEERS JOURNAL is usually granted upon request to the Editor of the appropriate journal provid
agreement to give proper credit is made.
Discussion of this paper is invited. Three copies of any discussion should be sent to the
Society of Petroleum Engineers office. Such discussion may be presented at the atxrve-citedmeeti
and, with the paper, may be considered for publication in or~eof the two SPE magazines.
INTRODUCTION and in more detail by Christopher and by
Gaitonde.7 polyisobutylene [pIB L-1OO, Enja
This conmrunicationpresents an extension Co.] solutions in toluene were run at 25.0°C
of an earlier study by Christopher and Ptiwer-lawparameters were obtained frum
Middlenlan5of the applicability of a non- capillary viscometry performed vver a range
Newtonian generalization of the Blake-Kozeny shear rates corresponding to the shear rates
equation to the laminar flow of non-Newtonian achieved in the porous medium. Table 1 show
polymer solutions through porous media. The the solution properties, including the zero-
earlier work, done with dilute aqueous solu- shear viscosities of the solutions and the
tions of carboxy?nethylcellulose,indicated that viscosity average molecular weight of PIB L-
the model put forth correlated data with an 100,
average error of 18 per cent and a standard
deviation of the friction factor of 0.21 over a CORRELATION OF DATA
range of three orders of magnitude in a modi-
fied Reynolds number. Fig. 1 shows the friction factor-Reynol
number correlation of the data obtained usin
A survey of the literature available at tubes packed with glass spheres of a narrow
that time8 indicated some evidence of visco- size range. The data are correlated with an
elastic effects, in contrast with the conclu- average error of 9.8 per cent and a standard
sions drawn by Christopher and Middleman. Thus deviation of 0.14 by
it was decided to extend the earlier study by
examining a more elastic fluid, polyisobutylene f = l/Re, . . . . . . . . . . . . [
dissolved in toluene. In addition, data were
obtained for tubes packed with sand, and for where
tubes packed with binary mixtures of glass &?D C3D
spheres. We still fail to see viscoelastic f = 2 “ “ “ “ ““”* [
effects. .LG2 (1 -c) “
The basic experimental techniques, and the and
methods of reducing and correlating the data, D G2-n Pn-l
,are as described by Christopher and Middlemm5 Re = . .... ● .[
*Present address: E. I. dupont de Nemours lSOH (1 -c)
I Wi& i t D l
2 FLOW OF VISCOEL4STIC FLUIDS THROUGH POROUS MEDIA SPE-1685
The factor H is given as comparable and even higher Ellis numbers
indicates that the Ellis number i~ not really
H= the appropriate viscoelastic parameter with
* (9 + 3/n)n (150 k .)(~+’? which anomalous behavior can be correlated.
. . . . . . . . . . . . . . . . . . . [4] Astarital cumnents thatqo/Tl/2 is not
necessarily a viscoelastic parsmeter in
These resuits are consistent with those re- contrast with the arguments or Bird.$
ported by Christopher.
FLOW THROUGH NONUNIFORM PACKING
As a test for viscoelastic effects, these
data were replotted in Fig. 2 as f Re vs Having thus established the applicability
VoTI/Dp. This latter group plays the role of of the power-law model to flow through a
a viscoelastic parameter. T1 is a molecular packing composed of uniform spheres, it was
relaxation time, given by Bueche3 as decided to investigate somewhat more complex
media. Fig. 3 shows results for flow through
[51 beds packed with sand of fairly uniform grain
T1=~ ~...*******
c3T size. AS in the other cases, the permeability
of the sand was first determined by using a
The viscosity average molecular weight Mq was Newtonian fluid. The experimental results in-
used in place of M in calculating TI.6 Values dicate the applicability of the proposed model
of TI are included in Table 1. to flow of non-Newtonian fluids through beds of
sand.
~/V. is a characteristic time for the
acceleration induced in the fluid due to its As a final test it was decided tw prepare
passage around and between spheres. When the a bed composed of a ‘cinarymixture of spheres,
relaxat~.untilfie
for the fluid is long in In addition to providing a non-uniform par-bicl
comparison with the time scale of flow dis- size distribution, binary sphere mixtures allow
turbances, one might expect to see viscoelastic one to extend the data to porosity values of c
phenomena. In this case the viscoelastic as low as 0.27.
phenomena would be observed as a failure of a
purely viscous flow model to successfully TWO beds were prepared, one with a sphere
correlate the data. Fig. 2 ~eveals no such diameter ratio of 3:1 and the other with a
failure for values of V. 11~ up to 1.2. ratio of 6:1..In both beds, tests with ~
Newtonian fluids indicated that the Blake-
Sadowski8~y il:trtiduced a similar ljarswiet,erKozeny equation held if ~ was re~>lacedby th;
into his study of this problem. He used the volume-average particle size, < ~.’v. Fig.3
Ellis number: shows res’~ltsfor non-Newtonian fluids in
V. q. binary sphere mixtures. when k i.s measured
. [6] with a Newtonian fluid, and when d ~ 7V is
51=
JlpT1/2 ‘ ‘ “ “ “ “ ‘“ “ used for the particle diameter, the mtidified
Blake-Kozeny equation is adequate to describe
where TI pin an Ellis model parameter which the flow of non-Newtonian fluids through binary
/ sphere mixtures.
represents the shear stress at which the
apparent viscosity falls to one-half of its
zero-shear value ~o. Dunleavy6 has shown, in NOMENCLATUTIE
a study of the shear behavior of solutions of
polyisobutylene, that c = concentration of polymer solution,
$/100 cc
no
T1=o.78—” T 1/2 “ “ “ “ “ “ ‘
● ●
[71 ~ s particle dianeter, cm
= Ellis number [Eq. 6]
f= friction factor [Eq. 2]
Hence our groupVoT1/Dpis very nearly equal to G ‘ mass velocity, gram-cm2 -see-l
the Ellis number. Yet Sadowski observes, for H= non-Newtonian bed factor [Eq. h]
El > 0.1, significant deviations between his dynes-seen -cm-~-n
theory and experimental results. Sadowski’s k = bed permeability, sq cm
theory is based upon the Ellis model rather than K = power law parsmetert T = K “(Y)nS
the power law model, but i.tis nevertheless dynes-seen -cm-2
a purely viscous theory. N= molecular weight of polymer
viscosity average molecular weight
Sandowskl’s results may indeed be due to ‘!!: power law parameter
viscoelasticity, perhaps arising from a pecu- AP/L pressure drop per unit length, dynes/cc
■
liarity of the fluid for which this effect was R = gas constant = 8.3Lx L07 g -sq cm see-2
observed, Natrosol 250-H. If this is the case, g -mole-l ‘K “1
10S
15 –
,.O
u
? 10
0000000
f lo~
,.8 05L__—_—J 02 0.4 06 00
r, VOI Dp
10 12 14
Fig, 2 - Data of Fig. 1 Replotted aa a Test
I I L1 for V+scoelastlc Effects
10*
,@ 10-’ 10-4 10-’
Re
Fig.l- Test of Modified Blake-Kozeny Equation
for Polyieobutylene Solutioqs
,.6
10’
0
,.-s ,.- B 10”* ,.- s
fee
Fig, 3 - Teet of Modified Blake-Kozeny Equation
for Polyisobutylene Solutions and Nonuniform Beds
~PF.lm5
“.9 .---,
NARENDRANATH Y. GAITONDE and STANLEY MIDDLEMAN
. ..— — .—-_— .. —--
T= absolute temperature,°K 3. Bueche, F.: Jour. Chem. Phys. [1954] ~,
V. = superficial velocity in bed, cm/sec 1570.
shear rate, sec ‘1
Y= 4. Christopher, R. H.: MS Thesifi,U. Of
E = void fraction Rochester, Rochester, N. Y. [19651.
~o = zero-shear viscosity, poise 5. Christopher, R. H, and Middleman, S.:
P= fluid density, gram/cc I.E.C. Fund [2.9651~, 422.
TS shear stress, dynes~sq cm 6. Dunleavy,
J. E.: MS TheSiS,u. of
TI = molecular relaxation the [Eq. 5], sec Rochester, Rochester, N. Y. [19651.
T L/2 = shear stress at %hick viscosity falls to 7. Gaitonde, N. Y.: MS Thesis, U. Gf
1/2 To, dynes/sq cm Rochester, Rochester, N. Y. [1966].
8. Sadowskf, T“.J.: PhD Thesis, U. of
REFERENCES Wisconsin, Madison, Wis. [1963].
9* Sadowski, T. J.: Trms., SOC. RheologY
1. Astarita, G.: Cdn. J. Chem. Eng. [19661 [1965] ~, 251.
44, 59.
2. ~rd, R. B.: Cdn. J. Chem. Eng. [1965]
Q 16I..
Table 1
PIB L-1OO in toluene at ‘2j.0°C,Ml = 7.3 x 105
T1
c K n no -1
wt. q) Dynes-5ecn poise Sec
* Sq,cm.
8.25 61.0 0.554 10.5 0. 004-(
8.7 75.0 0.55 12.2 0.0052
q.o 95.2 0.52 13.7 0,0056
y.~ 286.0 0,46 15.0 0.0061
9.9 112.0 0.51 23.8 0.0086
