UFL/COEL98/OO1
RHEOLOGY AND RHEOMETRY OF MUD
SLURRY FLOW IN PIPES: A LABORATORY
INVESTIGATION
by
Phinai Jinchai
Jianhua Jiang
and
Ashish J. Mehta
February 1998
UFL/COEL98/001
RHEOLOGY AND RHEOMETRY OF MUD SLURRY FLOW IN
PIPES: A LABORATORY INVESTIGATION
By
Phinai Jinchai
Jianhua Jiang
and
Ashish J. Mehta
Coastal and Oceanographic Engineering Department
University of Florida
Gainesville, FL 32611
February, 1998
SUMMARY
As a response to the need to pump comparatively high density mud slurries though dredged
material discharge pipes, this study examined the question of whether relevant mud properties can
be tested a priori in order to predict slurry transportation in pipes. To that end, laboratory rheometric
experiments were conducted to assess the dependence of slurry flow rate on mud composition and
density. Muds consisted of clays and clay mixtures of varying density. The selected clays included
a kaolinite, an attapulgite and a bentonite. The choice of these clays was based on the need to vary
mud properties widely in terms of their nonNewtonian rheology, as characterized by the three
parameter Sisko powerlaw model for apparent viscosity variation with shear rate. The overall
density range tested was 1,125 kg/m3 to 1,550 kg/m3. Powerlaw parameters (characterizing
ultimate viscosity, consistency and pseudoplasticity) as functions of mud composition and density
were obtained by testing the slurries in two types of rheometers. For low shear rates, a coaxial
cylinder viscometer (CCV) of the Brookfield type generating annular couette flows was used. For
comparatively high shear rates, a specially designed horizontal pipe viscometer (HPV) generating
Poiseuille flow was used. The latter apparatus consisted of a 2.54 cm i.d. and 3.1 m long,
horizontally positioned, PVC pipe through which mud flow was driven by a pistondiaphragm pump.
Rheometric results obtained by combining the CCV and HPV data confirmed the
pseudoplastic (shear thinning) behavior of all slurries. For mud of given composition, the ultimate
viscosity, consistency and the degree of pseudoplasticity generally increased with density. A slurry
cation exchange capacity (CEC,, y) is introduced as a cohesion characterizing parameter dependent
on the weight fractions of clays and water in the slurry. It is shown that the powerlaw parameters
correlate reasonably well with CECSI,, which therefore makes it a convenient measure of the
rheology of slurries composed of pure clays and clay mixtures.
The steadystate slurry transport equation for the Sisko powerlaw fluid is solved numerically
to yield the velocity distribution in the pipe and the corresponding discharge. To test the model,
including the applicability of the derived powerlaw for the slurries, additional tests were carried out
in the HPV at lower pumping pressures than those used to determine the powerlaw coefficients. It
is shown that for most slurries tested the calculated discharge agrees reasonably with that measured.
The approach developed in this study lends itself to further exploration as a method designed
to test the bottom mud to be dredged for its pumping requirements.
Support for this study from the U. S. Army Engineer Waterways Experiment Station,
Vicksburg, MS under contract no. DACW3996M2100 is acknowledged. Dr. T. M. Parchure of
the Coastal and Hydraulics Laboratory was the project manager.
TABLE OF CONTENTS
PAGES
SUMMARY ............................................................... .. ii
LIST OF FIGURES ........................................................... v
LIST OFTABLES .......................................................... ix
1. PROBLEM STATEMENT ................ ............................ 11
1.1 Need for Investigation ................ .......................... 11
1.2 Objective, Tasks and Scope ................ ...................... 13
1.3 Outline of Chapters ...................................... ..... 14
2. SLURRYFLOW IN PIPES ................ ............................ 21
2.1 Rheological Behavior of Materials ................................ 21
2.2 Equations of Motion in Pipes ................ ..................... 22
2.2.1 General Problem .......................................... 22
2.2.2 Poiseuille Flow Problem .................................. 23
3. RHEOLOGICAL MODELS ................ ........................... 31
3.1 Viscous Model ...................................... ..... ....... 31
3.1.1 Flow Type ........................................ .. .... 31
3.1.2 Apparent Viscosity ......................................... 31
3.1.3 End and Slip Effects and Corrections .......................... 34
3.1.4 Slurry Flow Curve ......................................... 37
3.1.5 Some Empirical NonNewtonian Models ........................ 37
3.2 Linear Viscoelastic Models ................ ...................... 38
3.2.1 Boltzmann Superposition Principle ............................ 38
3.2.2 Approximate Relationships Among Linear Viscoelastic Functions .... 39
4. ANALYTICAL SOLUTIONS FOR SLURRY FLOW ......................... 41
4.1 Poiseuille Flow Problems ......................................... 41
4.1.1 Generalized Newtonian Fluid ................................ 41
4.1.2 Bingham Plastic ........................................... 41
4.1.3 PowerLaw Fluid .......................................... 44
4.1.4 Maxwell Model ........................................... 44
4.2 Initial Value Problem and Periodic Motion ............................ 44
4.2.1 Initial Value Problem ....................................... 45
4.2.2 Periodic Motion ....................................... 410
5. EXPERIMENTAL SETUP, PROCEDURES AND MATERIALS ............... 51
5.1 Coaxial Cylinder Viscometer ................ ..................... 51
5.2 Horizontal Pipe Viscometer ........................................ 52
5.3 Sediments and Slurries .......................................... 52
5.3.1 Sediment and Fluid Properties .............................. 52
iii
5.3.2 Mud Slurries .............................................. 55
6. EXPERIMENTAL RESULTS ............................................ 61
6.1 Rheometric Results ............................................ .. 61
6.1.1 PowerLawfor Mud Flow .................................. 61
6.1.2 Calculation of PowerLaw Coefficients ......................... 63
6.2 PowerLaw Parameters ........................................... 64
6.3 Slurry Discharge with Sisko Model ................................. 610
6.4 Calculation of Slurry Discharge ................ .................. 611
7. INTERPRETATION OF RESULTS AND CONCLUSIONS .................... 71
7.1 PowerLaw Correlations with Slurry CEC ............................ 71
7.2 Concluding Comments ................ .......................... 71
8. REFERENCES ................................................... 81
APPENDICES
A SLURRY VISCOSITY DATA .................................... .... A1
B DEPENDENCE OF POWERLAW PARAMETERS ON DENSITY ............. B1
LIST OF FIGURES
FIGURE PAGE
1.1 Potential for entrainment and spreading of contaminated mud at intake and discharge
points during dredging operation ....................................... 11
1.2 Schematic drawing of Dry DREdge (after Parchure and Sturdivant, 1997) ......... 12
2.1 Poiseuille flow definitions ............................................... 24
3.1 Typical pressuredrop/flow rate relationships for slurry flow in pipes ............. 32
3.2 Plot of lnt, versus InF. Dashed line is tangent to the curve through a data point .... 32
3.3 Equivalent "extra length" due to end effects ................... ............ . 36
3.4 Length and pressure components in two pipes, where Ll and Ls are the fully
developed flow sections for longer and shorter pipes, respectively ................ 36
3.5 Slurry flow curve for nonNewtonian fluid. Dashed line is tangential extrapolation
to obtain the yield stress ............................................. 37
3.6 Applied strain in the stress relaxation test ................................ . 312
3.7 Responses to stress relaxation for Maxwell and Newtonian fluids and an elastic solid.
Note that for a Newtonian fluid the instantaneous stress is theoretically infinite. (This
is not indicated in the plot, however, for simplicity of depiction of the curves for all
three responses.) ....................................... .............. 312
4.1 Velocity profile for a generalized Newtonian fluid with Q=0.003 m 3/s and R=0.1
m ................ ......................................... ....... 42
4.2 Schematic of velocity profile for a Bingham plastic ........................... 42
4.3 Velocity profile for a Bingham fluid, with Q=0.003 m3/s and R=0.1 m, for slurry
density p=1000, 1099, 1198 and 1314 kg/m3 (from right to left) ................. 43
4.4 Velocity profiles for a powerlaw fluid, with Q=0.003 m 3/s, R=0. m, for flow index
n=0.5, 1.0, 2.0 and 3.0 (from left to right) ................................... 45
4.5 Velocity profiles during the generation of Poiseuille flow of a weakly elastic fluid
with S,=0.04, and S2=0.02, for t,=0.1, 0.18, and 0.34 and (from left to right).
Broken curves are the corresponding Newtonian profiles ....................... 49
4.6 Velocity profiles during the generation of Poiseuille flow of a general viscoelastic
fluid with S =0.04, and S2=0.04, for t1 =0.15, , 0.85 and 0.5 (from left to right) ... 49
4.7 Curves showing the variation of mean square velocity profiles with the parameters
(c3,1, 0 2)=(4,1), (4,2) and (4,3), (from left to right). Solid line corresponds to a
Newtonian fluid ...................................................... 411
4.8 Illustrative curves for Vo, V1, V2 as functions of r for Tio = 10 poises, =0.1 poise,
'1=0.6 s, X2 =0.1 s, e=0.25, p=1,000 kg/m3, w=l s1 , 0=200, R=0.25 cm. Broken
line is the velocity profile for Newtonian fluid giving rise to the same rate as Vo .... 415
4.9 Theoretical (I,A) curves for various values of o (s1), ro=10 poises, g0=l poises
=0.06 s, X2=0.01 s and e=0.25 ........................................ 415
4.10 Theoretical (S,D) curves for a Newtonian fluid (broken line) and a viscoelastic fluid
with viscosity function for a 1.75% aqueous solution of polyacrylamide and with
X 3=2.75 s, X4=0.3 s, o=l s1, R=0.25 cm and e=0.25 (solid line) ............... 416
5.1 Schematic drawing of a coaxial cylinder viscometer ........................... 51
5.2 Brookfield viscometer with an attached spindle shearing a clay slurry ............. 51
5.3 Schematic drawing of experimental setup for the horizontal pipe viscometer (HPV)
............... ............... .................. ............... 52
5.4 Photograph of HPV setup ............................................... 53
6.1 Comparison between pseudoplastic (shearthinning) and Newtonian flow curves. The
nature of the shearthinning curve is such that while at low shear rates its viscosity
is higher than the constant value for the Newtonian case (line), with increasing shear
rate the pseudoplastic curve becomes asymptotically parallel to the Newtonian line,
hence the pseudoplastic viscosity approaches that of the Newtonian case .......... 61
6.2 Excess apparent viscosity as a function of shear rate for kaolinite slurry no. 3 ....... 67
6.3a. Variation of g, with density of kaolinite slurries ............................. .69
6.3b. Variation of c with density of kaolinite slurries ............................... 69
6.3c. Variation of n with density of kaolinite slurries ............................. 610
6.4 Computed velocity profiles and corresponding discharges for slurry no. 1. Line is
numerical solution using Sisko model; dots represent analytic Newtonian solution .. 612
6.5 Computed velocity profiles and discharges for slurry no. 24 ................... 613
vi
7.1 Variation of p with slurry CEC (CEC, ,U) for all slurries ...................... 72
7.2 Variation of logc with slurry CEC (CECsI) for all slurries ..................... 73
7.3 Variation of n with slurry CEC (CECs IY) for all slurries ....................... 73
A1 Viscosity data for slurry no. 1 ......................................
A2 Viscosity data for slurry no. 2
A3
A4
A5
A6
A7
A8
A9
A10
A11
A12
A13
A14
A15
A16
A17
A18
A19
A20
A21
A22
A23
A24
A25
A26
A27
A28
A29
A30
A31
A32
A33
A34
A35
A36
A37
A38
A39
Viscosity data for slurry no. 3
Viscosity data for slurry no. 4
Viscosity data for slurry no. 5
Viscosity data for slurry no. 6
Viscosity data for slurry no. 8
Viscosity data for slurry no. 9
Viscosity data for slurry no. 10
Viscosity data for slurry no. 11
Viscosity data for slurry no. 12
Viscosity data for slurry no. 13
Viscosity data for slurry no. 14
Viscosity data for slurry no. 15
Viscosity data for slurry no. 16
Viscosity data for slurry no. 17
Viscosity data for slurry no. 18
Viscosity data for slurry no. 19
Viscosity data for slurry no. 20
Viscosity data for slurry no. 21
Viscosity data for slurry no. 22
Viscosity data for slurry no. 23
Viscosity data for slurry no. 24
Viscosity data for slurry no. 25
Viscosity data for slurry no. 26
Viscosity data for slurry no. 27
Viscosity data for slurry no. 28
Viscosity data for slurry no. 29
Viscosity data for slurry no. 30
Viscosity data for slurry no. 31
Viscosity data for slurry no. 32
Viscosity data for slurry no. 33
Viscosity data for slurry no. 34
Viscosity data for slurry no. 35
Viscosity data for slurry no. 36
Viscosity data for slurry no. 37
Viscosity data for slurry no. 38
Viscosity data for slurry no. 39
Viscosity data for slurry no. 40
................................. ......... A 1
.......................................... A1
......................................... A 1
......................................... A 2
........ .... ..................... ......... A 2
............... ..... ...................... A 3
.................................. .... .... A 3
....... ... ................................ A 3
................................... ....... A 3
............................... ..... ...... A 4
.................................... ...... A 4
.......................................... A 4
.......................................... A 4
..................................... ..... A 5
................................... ....... A 5
....................... .................. .. A 5
.................................... ...... A 5
................... ...... ................. A 6
................................ ... ....... A 6
.......................................... A 6
.......................................... A 6
.......................................... A 7
......... ......... ...... ................. A 7
.......................................... A 7
................................... ....... A 7
............. .. ........................... A 8
................................ ......... A 8
.......................................... A 8
.................................... ...... A 8
........................ ................. A 9
.................................... ...... A 9
.......................................... A 9
......................................... A 9
......................................... A 10
......................................... A 10
......................................... A 10
......................................... A 10
......................................... A 11
A40 Viscosity data for slurry no. 41 ........................................ A11
A41 Viscosity data for slurry no. 42 ......................................... A11
A42 Viscosity data for slurry no. 43 ....... .......... ....................... A11
B1 Variations of powerlaw coefficients with density for kaolinite slurries. Top: p_;
middle: c; bottom: n ................................................... B1
B2 Variations of powerlaw coefficients with density for 75% kaolinite + 25%
attapulgite slurries. Top: p,; middle: c; bottom: n ............................ B2
B3 Variations of powerlaw coefficients with density for 50% kaolinite + 50%
attapulgite slurries. Top: pL ; middle: c; bottom: n ............................ B3
B4 Variations of powerlaw coefficients with density for 25% kaolinite +75% attapulgite
slurries. Top: Lt; middle: c; bottom: n ............... ................... .B4
B5 Variations of powerlaw coefficients with density for attapulgite slurries. Top: [p;
m iddle: c; bottom : n ............... ...................................... B5
B6 Variations of powerlaw coefficients with density for 90% kaolinite + 10% bentonite
slurries. Top: gL ; middle: c; bottom: n ..................................... B6
B7 Variations of powerlaw coefficients with density for 65% kaolinite + 25%
attapulgite + 10% bentonite slurries. Top: lIm; middle: c; bottom: n .............. B7
B8 Variations of powerlaw coefficients with density for 40% kaolinite + 50%
attapulgite + 10% bentonite slurries. Top: pl; middle: c; bottom: n .............. B8
B9 Variations of powerlaw coefficients with density for 15% kaolinite + 75%
attapulgite + 10% bentonite slurries. Top: g_; middle: c; bottom: n .............. B9
B10 Variations of powerlaw coefficients with density for 90% attapulgite + 10%
kaolinite slurries. Top: p,; middle: c; bottom: n ............................. B10
LIST OF TABLES
TABLE PAGE
5.1 Chemical composition of kaolinite ........................................ 53
5.2 Chemical composition ofbentonite ........................................ 54
5.3 Chemical composition of attapulgite (palygorskite) ........................... 54
5.4 Chemical composition of water ........................................... 54
5.5 Size distribution ofkaolinite ............................................. 55
5.6 Size distribution ofbentonite ............................................. 56
5.7 Size distribution of attapulgite .......................................... .. 56
5.8 Properties of mud slurries tested ....................................... 57
6.1 Pressure drop, discharge, shear rate and wall stress data from HPV tests ........... 65
6.2 Sisko model coefficients and HPV flow Reynolds number ...................... 67
6.3 Low pressure HPV test parameters for selected slurries ....................... 611
1. PROBLEM STATEMENT
1.1 Need for Investigation
As coastal navigation channels are deepened by dredging due to requirements for greater draft
vessels calling at ports, the need to maintain these channels against sedimentation has also risen.
Disposal of dredged material has become problematic in many urbanized estuaries where the number
of sites at which the material can be discharged has been restricted for ecological reasons (Marine
Board, 1985). Added to this is the problem of contaminated bottom sediments, which can pose a
threat to habitats both at the intake point and at the discharge point, as a result of the potential for
the dispersal of sediment and associated pore fluid into the ambient waters. Referring to Fig. 1.1, it
is a common practice to cut and loosen bottom mud at the intake point mechanically or by high
pressure water jets in preparation for the suction of the diluted slurry and its transport to the outtake
point. In this process the water content of mud may increase from, say, 20% for a compacted bed to
as much as 400% in a diluted slurry (Parchure and Sturdivant, 1997). It is therefore evident that this
procedure can exacerbate contaminant dispersal at both ends, especially where strong currents or
waves are present. The question therefore arises whether it is feasible to transport the slurry at in situ
bottom density, as this would be a "cleaner" operation and therefore would be highly desirable for
areas where sediment at toxic "hot spots" needs to be removed and transported as safely as possible.
Although this seems like an obvious solution, the reason for diluting the mud at the intake point in
the first place is that it is often difficult to transport undiluted, relatively dense bottom mud without
a very highpressure pump.
Discharge
Intake
JI / Entrainment and Spreading
ud
Fig. 1.1 Potential for entrainment and spreading of contaminated mud at intake and discharge points
during dredging operation.
To address the above issue, at the U. S. Army Engineer Waterways Experiment Station in
Vicksburg, Mississippi, a collaborative research program was undertaken by WES and DRE
Technologies of Brentwood, Tennessee, to design, construct and operate a dredge which can
transport bottom mud at in situ density. Concurrently, a laboratory experimental program was carried
out at WES as part of the same investigation to explore the effect of mud composition and density
on its transportability (Parchure and Sturdivant, 1997).
The Dry DREdge (Fig. 1.2) had a specially designed sealed clashell mounted on a rigid,
extendable boom. The open clamshell could be hydraulically driven into the bed sediment at a low
speed, thus minimizing sediment disturbance, resuspension and spreading. The clamshell could be
closed hydraulically and sealed, trapping a plug of sediment at its in situ water content. The sediment
could then be deposited in the hopper of a positive displacement pump which could deliver it
through a pipeline to the disposal site. Based on the results of prototype tests conducted in a shallow
lake with a muddy bottom it was found that this dredge was useful for the following operating
conditions: 1) shallow water depth (less than about 4.5 m), 2) small quantity of dredging, low output
rate (less than about 30 m3/hr), 3) sediment with a high proportion of fines and low amount of sand,
4) short pumping distance, and 5) no significant wave or current action (Parchure and Sturdivant,
1997).
SHIS DRAWIO AND THE INFORIAION IT CONTAINS. IHNCUWD 1HE PRINCIPAL OF DESIGN IS PATENT PENDING
THE PROPERTY OF ORE ENIRONIENTM. SERVICES, RENTWOOO TENNESSEE. AND IS SUBUTIED
1H E PRESS AGREEMENT TAT IT Mr NOT BE REPRODUCED, COPED. OR O, MSE V.J. 2O E RONMENTAL
DISPOSED OF. DOECRY OR WIRECRY, AND UL NOT BE USED AS A BASIS FOR AHUFACMURE Iom 11 392 cM I ONM
OF EMOIJENT OR APPARATUS WIOUT WIM PBIOSSION OF ORE ENWRM0NENTAL SERVICES 
.. FIRST OBTAIN AND SPECIC AS TO EACH CASE. ONE FICURE1 BRENTMOOD, mEEE
Fig. 1.2 Schematic drawing of Dry DREdge (after Parchure and Sturdivant, 1997).
The main objective of the laboratory tests at WES was to ascertain the validity of the
assumption that a highpressure positive displacement pump is capable of transporting dredged
sediments through a discharge pipe at near in situ water content. A 12.7 cm diameter, 274 m long
steel pipe was laid out for this purpose, and various mixtures in water of two clays: a kaolinite and
a bentonite, and 0.4 mm diameter sand were transported with the help of a pistontype concrete
pump. One noteworthy conclusion based on twelve tests carried out was that it was possible to
transport a dense slurry with a specific gravity as high as 1.75. Also, in general, it was found easier
to transport clays alone, than clay mixed with sand.
In general, because bottom mud composition and density can vary widely, the pumping
requirements can also vary accordingly. It is therefore a matter of considerable engineering interest
to have the knowhow to determine a priori what the pumping requirements will be in a given
situation, in order to design the dredging operation and its execution. Since the transport of a slurry
of given composition and density depends on the theological behavior of the slurry (Heywood, 1991;
Wasp et al., 1977), by determining this behavior of mud to be dredged in a rheometer should make
it feasible to ascertain: 1) if the bottom mud can or cannot be transported without dilution, and 2)
what the rate of discharge of transportable will be. This laboratory study was therefore concerned
with a twostep procedure, namely: 1) relating mud discharge in a pipe to mud rheology, and 2)
relating mud rheology characterizing coefficients to mud composition and density. Accordingly, the
study objective, tasks and scope were accordingly as follows.
1.2 Objective, Tasks and Scope
The objective of this study was to correlate mud discharge in a pipe with mud composition
and density through mud rheology, as a basis for developing a predictive tool for assessing pumping
requirements for transportation of relatively dense dredged bottom mud. The associated tasks were
as follows:
1. To select muds of widely varying composition.
2. For mud of given composition and density, to determine its theological behavior.
3. To pump each mud of given composition and density through a horizontal pipe, and measure the
discharge and pressure loss.
4. For mud of given composition, to determine the relationship between discharge, density and
rheology characterizing parameters.
5. To explore the feasibility of using the above relationship as a predictive tool for assessing the
transportability of a given bottom mud.
The experimental scope of the work was defined by the choice of muds selected and the
facility used for experiments. Since this work was of an exploratory nature, it was decided to select
pure clays and their mixtures in water at different densities as mud slurries. The theological behavior
of each slurry was determined by combining the data from two types of rheometers: a coaxial
cylinder viscometer (CCV) and a horizontal pipe viscometer (HPV). A Brookfield (model LVT)
viscometer was used as CCV for testing the slurries at low shear rates. For relatively high shear rates,
an HPV was designed specially. This particular type of viscometer was chosen because it also served
as the arrangement used to measure slurry discharge at different pump pressures. The benchtop
apparatus consisted of a 2.54 cm i.d. and 3.1 long PVC pipe through which mud was made to flow
due to pressure applied by a diaphragm pump. The objectives of the study were met entirely by data
obtained using these two types of viscometers, as described in the subsequent chapters.
1.3 Outline of Chapters
In Chapter 2, the slurry transport problem is formulated in terms of Poiseuille flow in
horizontal pipes. Chapter 3 covers theological models for slurry flow, and in Chapter 4 several
analytic solutions of the Poiseuille flow problem are explored, as well as the initial value problem
and periodic motion. The experimental setup, materials used and test procedures are described in
Chapter 5. Results of rheometry along with theological simulations are presented in Chapter 6. These
are further interpreted with reference to the overall objective of the study in Chapter 7. Finally,
relevant references are given in Chapter 8.
2. SLURRY FLOW IN PIPES
2.1 Rheological Behavior of Materials
In this chapter we refer to the works by Huilgol (1975) and Darby (1976). This chapter briefly
describes the theoretical basis for the rheologial behavior of materials, such as mud, in pipes. For
high density flow in pipes, the classic Newtonian theological model cannot describe flow behavior
appropriately, because here the elastic as well as viscous characteristics of the mixture of water and
soil are also very important. The general relationship between an applied stress and the resulting
deformation of a material is a unique function of the material. This relation may be represented
symbolically by the general constitutive equation
P(,t ij,,..)= )e(y,ij" .....) (2.1)
where Tyi, "t *ij, *, y ij, **" are the shear stress, strain and their first and higher order derivatives
with respect to time. Generally, a real material has nonlinear characteristics, and it is difficult to
accurately construct its theological model. In practice, we therefore make use of some simple
empirical models for nonNewtonian behavior, e.g., the Bingham plastic model, powerlaw model,
generalized Newtonian fluid model, etc., and also linear theological models, e.g., Maxwell model,
to express the theological properties approximately.
The constitutive Eq. (2.1) implies that a unique property of viscoelastic fluids, such as mud
slurries, in contrast with purely viscous fluids, is the timedependence of material response. If a
material undergoes a largeamplitude or continuous deformation, the coordinate positions of a given
material point (with reference to a fixed origin) will not be "fixed". As a result, any measure of
deformation based on infinitesimal displacements of fixed coordinate positions will be meaningless,
since this would not always refer to the same material element. Thus, in order to describe the
deformation of a viscoelastic material under these circumstances, it is necessary to follow a given
material element or point with time as it moves from place to place in the system. One way in which
we might follow the deformation of a material element as it moves through a system is by means of
a reference frame which is defined by a set of base vectors that move and deform with the material.
The position vectors in this frame are referred to as convective coordinates. These coordinates follow
the path line of material particle, defined by
S
i=xi V j(tt'),tt']dt' (2.2)
0
where xi is the absolute coordinate; V, is the absolute velocity; , is the position of a particle at time
ts, with O0ssao and s is the time history of a particle. Under convective coordinates, the strain tensor yi
becomes
Y ij(xk,t,s)= gij(xk) Cj(xk,t,s) =gij(x) ( i. gm(k) (2.3)
ax, a^^ )^ gm (Xk ) (2.3)
Note that yi is called the covariant convected strain tensor, where cij is referred to as the Cauchy
deformation tensor and gy is called the metric tensor. ci andgi are defined as
cij(xk,t,S)= m n k) (2.4)
axi ax g(Xk) (2.4)
8x, O ixn
aX m a (2.5)
ax. ax
where xi and x, are the absolute Cartesian and curve coordinates, respectively.
In order to determine the flow behavior of a given material in a pipe, the theological
equations of the material, which relate the force at all points in the material to the local deformation,
must be solved simultaneously with the equations of conservation of slurry mass and momentum.
We will derive the general equations of motion in next section.
2.2 Equations of Motion in Pipes
2.2.1 General Problem
The general equations of motion in a pipe in cylindrical coordinates (z, r, 0) are as
follows:
Incompressible continuity equation:
1arV+ aVe O= (2.6)
r ar r A azz
Momentum equations:
rcomponent:
P aVr + Vr V V ar V2 aVr ap I r, 1 Oa re 'ee (2.7)
+V ar v v = +pg+ +r r (2.7)
at r r r 6 r z r r r r 6 r z
0component:
aVe aVe V v dv rV ,V aV 1 r+p 1 e r2 r 1 O z
+V_ + +Vr r +0 +i
at r, ra r Z az r a r ar rA z
zcomponent:
aVz aVz Va9Vz avz
a+V + +V z
at r r r e zz
B zp 1 arz 1 a te at
az r ar r oe az
where TzZ, Tr, TOO are the normal stresses which contain the elastic effect; Tz, Tre, rez are shear
stresses; Vz, Vr, Veare the velocity components in z, r and 0 directions, gz g, goare the
corresponding gravity acceleration components; t is the time; p is the pressure and p is the density
of mixture of water and sediment.
2.2.2 Poiseuille Flow Problem
In this section we will describe the steady, isothermal, axial and laminar flow of an
imcompressible fluid in a pipe (Fig. 2.1), known as Poiseuille flow. It is assumed that the flow is
symmetric and that the axial (z) velocity component is the only nonzero component. These
conditions can only be satisfied if
Vr=V6=O' VZ*f(O)
(2.10)
Under the above conditions, the continuity Eq. (2.6) reduces to
(2.11)
oV
a=
so that Vz f(z), i.e., Vz = V(r) only. The rate of strain tensor therefore becomes
0 dV/dr
,i = dVIdr 0
0 0
0
0
0
0 1 0
= Y 1 0 0
0 0 0
(2.8)
(2.9)
(2.12)
so that the shear stress tensor has, at most, the following nonzero components
t.. = t ;r
rz
0
TO
rz
rr 0
0 6oe
The three component momentum equations are simplified as
zcomponent:
ap 1
 pg r(rz)
9z r 9r
rcomponent:
Lp g=1a (rTr) 
ar r ar r
0component:
1 ap
r a P
Po PL
I L L
1 z
Fig. 2.1 Poiseuille flow definitions.
(2.13)
(2.14)
(2.15)
(2.16)
V(r)
I
Considering the fact that the velocity varies only in the r direction, all internal stresses which depend
on flow deformation must also be functions only of r. Using this condition, differentiating Eq. (2.15)
with respect to z leads to
a ap a p (2.17)
az[ ar ar[ az
In other words, the pressure gradient (ap/9z) is independent of r. So we can calculate the pressure
gradient along the pipe as 9p/az =(Ppo)/L. Considering Eq. (2.14), the left hand side is independent
of r, whereas the right side is a function only of r. This can be true only if both are equal to a
constant (say, D):
ap 1
pgz=o= (rr,) (2.18)
az r dr
Integrating Eq. (2.18) leads to
r@ C1
T =+ (2.19)
r2 r
The integration constant, C1, must be zero, since otherwise an infinite stress would be predicted at
the center (r=0). Considering the stress exerted by the fluid on the pipe wall
iR 20 RD
S= (2.20)
w 27R 2
then Eq. (2.19) becomes
rH r
r = t (2.21)
2 Rw
which is the final form of motion equation for Poiseuille flow. It is valid for either laminar
(Newtonian or nonNewtonian) or turbulent flow.
3. RHEOLOGICAL MODELS
3.1 Viscous Model
3.1.1 Flow Type
In this chapter we refer to the works by Alfrey and Doty (1945), Williams and Ferry (1953),
Ferry (1970), Darby (1976) and Mehta (1996). In terms of flow properties for homogeneous, non
settling slurries considered here under steady state in pipes, the flow type, the apparent viscosity and
the flow curve (j versus 1) must be obtained. The experiment data required include the pressure
drop dp over the fully developed flow length L and the volumetric flow rate Q (or mean velocity
Vm).
From the plot of ln(Ap/L) versus InQ or InVm shown in Fig. 3.1, where n is the powerlaw
index, the flow type can be determined. It can be laminar or turbulent, Newtonian or Non
Newtonian, etc.
3.1.2 Apparent Viscosity
For a nonNewtonian fluid, the ratio of shear stress and shear rate is not constant, i.e.,
doubling the shear stress will not result in twice the shear rate, or vice versa. Thus the viscosity is
not independent of the shear rate. We therefore define a function called the apparent viscosity as:
T) (3.1)
Now, from Eq. (2.21) it is evident that measurement of the pressure gradient 0c provides a direct
means of determination of the shear stress at any point in the pipe. The calculation of the apparent
viscosity of the fluid also requires the determination of the shear rate at some point in the pipe. An
expression for the shear rate can be derived by considering the following relation for the volumetric
flow rate, Q:
R
Q=f27rV,(r)dr (3.2)
0
Integrating Eq. (3.2) by parts, with the condition that Vz=0, at r=R, leads to
R R( dVR
Q= fr2dVz=t7 for J dr (3.3)
0 0
lnQ or InVm
Fig. 3.1 Typical pressuredrop/flow rate relationships for
slurry flow in pipes.
Inr
Fig. 3.2 Plot of Int, versus InP. Dashed line is tangent to the curve through a data point.
Equation (3.3) can now be used to change the variables from r to T (for a given wall shear
stress Tand R) to give
Q =R (3.4)
T "3
W 0
By taking the derivative of Eq. (3.4) with respect to Tw, we obtain
d(rw3) =4"w2'w (3.5)
d'w
where F=4Q/cR 3 and is the shear rate at the wall. Solving Eq. (3.5) for ,w leads to
Tw Ar' 3r
YW4 d.cw +ri (3.6)
If we let n '=dln(Tw)/dlnF, Eq. (3.6) can also be written as
w 3n/]l (3.7)
Based on Eq. (3.7), the apparent viscosity takes the form
T Tw 4n'/
Sl(*)== (3.8)
1, r 3n '+1
From the laboratory data (measurements of discharge and pressure drop), In', can be plotted
against Inr, as shown in Fig 3.2. For that purpose, the wall shear stress, tw, is calculated from Eq.
(2.20). Values of the coefficient, n are obtained manually from tangents drawn to the curve, as
shown by the example in Fig.3.2.
If necessary, end and slip corrections can be applied to correct for the measured values of the
pressure drop, Ap. These corrections are described next.
3.1.3 End and Slip Effects and Corrections
A. End Effects:
A major error which may arise in pipe flow measurements is due to end effects. Near the
entrance and exit regions of the pipe, the velocity profile is not constant along the pipe but is in a
state of transition between the flow configurations outside and inside the pipe, and the pressure
gradient is not constant over these regions. For viscoelastic fluids, there is an additional end pressure
drop due to elastic or normal stress effects. Therefore, if the measurement of pressure drop Ap is not
carried out within the fully developed flow section, a correction for Ap becomes necessary.
Entrance effects can be corrected for experimentally in various ways. One approach is to
determine an equivalent "extra length" (Le) of the pipe that would have to be added to the actual
length if the total measured ,p were that for an entirely fully developed flow region. This can be
done as follows.
Consider the total pressure gradient due to friction in fully developed flow in the pipe (of
length L), plus an extra pressure drop due to entrance effects which would be equivalent to friction
in fully developed flow over an additional length Le:
A g= (3.9)
L+Le R
If the pipe is horizontal (g=0) and noting that zw is a unique function of F as shown in Eq. (3.6), Eq.
(3.9) can be rearranged to give:
Ap = 2t1~ L =f()( L (3.10)
R R R RR
Hence, if several pipes of different L/R ratios are used, and ip is plotted against L/R for the same
value of r in each pipe, the plot should be linear if the flow becomes fully developed within each
pipe, and the intercept at Ap=0 determines Le (Fig.3.3). The intercept on the Ap axis at LUR=0 is the
pressure drop (Ape) due to the combined end effects. Since a different value of Le would be obtained
for each value of F, LjR can be empirically correlated with r.
An alternate procedure involves the use of two pipes of the same diameter, operating at the
same flow rate (Q or F). Using subscript S for the shorter pipe and L for the longer, the various
lengths and pressure components are defined in Fig. 3.4. Care must be taken in choosing the pipe
lengths so that errors in pressure measurement are not unduly compounded by taking differences of
large numbers. Assume the pressure gradient in the fully developed flow sections of the longer and
shorter pipes are identical, i.e., 4L,='s=c1. Then following relationships are satisfied
A,= Ape +(L,L,) =Ap, +LL'4
Aps= Ae+(LsLe) =Ape+Ls'( (3.12)
Subtracting Eq. (3.12) from Eq. (3.11), the true pressure gradient in the fully developed flow section
reads
^ APLA Ps
= PLP(3.13)
LLLS
B. Slip Effect
An error in the measurement of Q can arise from an apparent slip between the fluid and the
solid wall. This effect is actually due to the general inhomogeneity of the fluid near the wall.
However, the extent of the region affected is often very small, so that the effect may be accounted
for by assuming an effective slip velocity (us) superimposed upon the fluid in the pipe, and
modifying Eq. (3.4), i.e.,
Q=usR2 +R 2d (3.14)
w 0
or
r + w fr2ld (3.15)
4Tw R zw40
w0
where p =ujrw is a slip coefficient. This coefficient can be evaluated as follows:
1. Using various pipes of the same length but different radii, plot r/4 versus tw for each
pipe. If P=0, these curves should coincide. If not, the curves will be distinct, in which case proceed
as follows:
2. At constant Tw, plot r/4Tw versus 1/R from the above curves. This plot should be linear
with a slope =P;
(3.11)
3. Repeat step 2 for various values of Zw, and then plot P versus T.
The appropriate value of F to use in evaluating ,w is then a "corrected" value corresponding
to no slip:
4pT
Inoslip ( slip)measured R (3.16)
4Ps
r=const.
Short pipe, radius R
) Q
L, Ls
I Long pipe, radius R
 Q
LI LL
Fig. 3.4 Length and pressure components in
two pipes, where LL'andLs' are the fully
developed flow sections for longer and shorter
pipes, respectively.
3.1.4 Slurry Flow Curve
After completing the above calculations, the slurry flow curves characterizing the theological
behavior of the slurry can be drawn, i.e., plots of rc, and rl versus (Fig. 3.5). The yield stress of
the mud, is obtained by extrapolating the curve of T, versus Y,. Next we will attempt to
determine the empirical relationships between t, and Il versus jw.
Fig. 3.5 Slurry flow curve for nonNewtonian fluid. Dashed line is tangential extrapolation to obtain
the yield stress.
3.1.5 Some Empirical NonNewtonian Models
The most successful attempts at describing the steady rate of shearstress behavior of non
Newtonian fluids have been largely empirical. The following represents some of the more common
empirical models which have been used to represent the various classes of experimentally observed
nonNewtonian behavior.
versus y,
r, versus y,
I
A. Bingham Plastic:
Given T and j both positive, this model is
r=Ty+pl for Tt>y (3.17)
j=0 for T
This is a twoparameter model, with as the yield stress and p as the plastic viscosity. The apparent
viscosity function thus becomes, rl =p(T/), for rz and rl.oo, for T
B. PowerLaw Fluid:
This is described as
t=pj," (3.19)
This is also a twoparameter model, with n as the flow index, t as the consistency. The apparent
viscosity for this model is rI =,t nl.
C. Ellis Model
In this model, the apparent viscosity is obtained as
S r (3.20)
The three parameters in Eq. (3.20) are pt, T1/2 and a. Here T1/2 is the value of : at which rI =0.5p,
and a is an empirical constant.
3.2 Linear Viscoelastic Models
3.2.1 Boltzmann Superposition Principle
For any linear viscoelastic system, the constitutive equation has the general form
d d2 dN d d2 dm
Po*PI P2 dpN d +q1+2 + ...... *+Md (3.21)
dt dt2 dt dt2 dt
38
where the parameters po, Pl......, pN and q0, q, ......, qM are material properties and is the jth
dtJ
derivative. According to the Boltzmann superposition principle for a linear system, the total effect
will be a linear combination of the separate contributions:
T(t)= V, i=E 1r(tt)VY(ti) (3.22)
i=0 i=0
where r is a relaxation function. When the system is continuous, the sum can be replaced by an
integral:
T(t)= j(tt)(t')dt' = ft(s)(ts)ds (3.23)
0 0
Therefore, in the convective coordidate system the stress tensor Tij takes the form
t I a m
T ij(xkt)= f tt (zt)dt' (3.24)
Under a steady state, =const., so that we obtain the apparent viscosity from Eq. (3.23) as
S=T = f(s)ds (3.25)
Y 0
which implies that the motion of any general linear viscoelastic fluid falls into the category of
"viscometric flow" at steady state.
3.2.2 Approximate Relationships Among Linear Viscoelastic Functions
A. Dynamic Experiments:
In dynamic experiments to determine the specific type of viscoelastic behavior of a material
it is found that, if the viscoelastic behavior is linear, the strain also alternates sinusoidally but is out
of phase with the stress. This can be shown from the constitutive equation (3.21) as follows:
Let
Y =Yosin(ot)
where Yo is the maximum strain amplitude and oi is the angular frequency of oscillation. Then
S=OYo0cos(t) (3.27)
Substituting Eq. (3.27) into Eq. (3.23) we have
00
Tr(t)=f(s)owycos[w(ts)]ds
o (3.28)
=Yo[w f(s)sin(ws)ds]sin(wt)+yo[k fo(s)cos(ws)ds]cos(wt)
0 0
It is evident that the term with sin((t) is in phase with Yo and the term with cos(ot) is 900 out of
phase. In other words, T is periodic in 0 but out of phase with y to a degree depending on the relative
magnitudes of these two terms. The quantities in brackets are functions of frequency but not of
elapsed time, so that Eq. (3.28) can be conveniently written as
T =Yo[G'sin(t) +G"cos(ot)] (3.29)
thereby defining two frequencydependent functions the shear storage modulus G' and the shear
loss modulus G". It is instructive to write Eq. (3.29) in an alternative form displaying the amplitude t0
of the stress and the phase angle 6 between stress and strain. From trigonometric relations,
T =:oSin(Ot+8) =ToCos6sinot+trosin6cosGt (3.30)
Comparison of Eqs. (3.27) and (3.28) shows that
G'=(roYo)cos6
G" = 0)Sin8 (3.31)
G" =(rdYo)sin6
Thus it is evident that each periodic, or dynamic, measurement at a given frequency provides
simultaneously two independent quantities, G' and G".
310
(3.26)
B. Relaxation Function:
To better illustrate the relaxation function, let us consider a Maxwell fluid response to a stress
relaxation test in simple shear, in which a shear strain of magnitude Yo is suddenly applied at time
zero, and is then held constant (Fig. 3.6). The Maxwell model has the form
r+, t=LY, (3.32)
where X= p/G is the relaxation time and p is the Newtonian viscosity. In the stress relaxation test,
the stress response will be (Fig. 3.7)
T(t) =GYoe (3.33)
where G is the constant shear modulus. It is seen that the initial stress response, (To GYo as t0 +)is
purely elastic. It then decays exponentially with time, reaching 37% [=(1/e)x 100] of its initial value
Gy0 at the time t=X. Thus, X is a characteristic time constant of the material, representative of the
time scale for stress relaxation, and j(t)=Ge t/" is called the relaxation function. Note also in Fig.
3.7 the corresponding responses of a Newtonian fluid and an elastic solid.
For a generalized linear viscoelastic model defined in Eq. (3.21), the appropriate relaxation
function corresponding to an infinite series of relaxation times is
*(t)= Gi e (3.34)
i=1
If the spectrum of relaxation times in continuous rather than discrete, Eq. (3.34) can be written as
l(t) =F(X)e ^tAd (3.35)
0
Let us set F(X)=H(X)/X and s=t. Then, the relaxation function for continuous linear viscoelastic
system takes the form
1 (s)= e^d = H(X)e (se)dln (3.36)
0 0
0 0,;
311
Fig. 3.6 AI
T
y(t)=y0
t=O t
strain in the stress relaxation test.
t=O X t
Fig. 3.7 Responses to stress relaxation for Maxwell and Newtonian fluids and
an elastic solid. Note that for a Newtonian fluid the instantaneous stress is
theoretically infinite. (This is not indicated in the plot, however, for
simplicity of depiction of the curves for all three responses.)
where H(X) is a distribution function of relaxation times X. Theoretically, H can be obtained from
G'or G". In practice, the functional forms of G' or G" are so complicated that no attempt is made
to represent them by analytical expressions. For this reason, a variety of approximation methods have
been developed for performing such calculations.
Eq. (3.36) shows that H is multiplied by the kernel function e (s), which varies from 0 at
312
X=0 to 1 when Xoo. If the latter were approximated by a step function varying from 0 to 1 at X=s,
we would have
(3.37)
V~s) ad H(Inn
Ins
and the integral would not be grossly different Eq. (3.36). Moreover, by differentiating Eq. (3.37)
with respect to the limit in Ins, we obtain
Sd*(s) =H )
dlns
(3.38)
Hence the relaxation spectrum at X=s is obtainable in the first approximation as the negative slope
of the relaxation modulus. This is called Alfrey's rule.
Williams and Ferry (1953) provided two formulas to calculate the relaxation spectrum from
the storage modulus, depending on whether m, the negative slope of H against ; on a double
logarithmic plot, is greater or less than 1. If m
H(X) =AG'dlogG'/dlogo 11 =;,
(3.39)
A sin(m7i/2)
m7c/2
(3.40)
If, on the other hand, l
H(A)=AG'(2dlogG'/dlogo) I /,,=
(3.41)
A sin(m7/2)
~7(1 m/2)
313
where
where
(3.42)
The calculation is carried out in two stages. First A is set equal to unity, and a preliminary
calculation is made with each point at a given value of ( yielding a value of H at X=1/o. From the
tentative graph of H versus o, the value of m is measured at each point, and the appropriate
correction factor A is applied. Then, through the calculated values of H and ,, one can construct an
empirical formula for H.
314
4. ANALYTICAL SOLUTIONS FOR SLURRY FLOW
4.1 Poiseuille Flow Problems
4.1.1 Generalized Newtonian Fluid
In this chapter we refer to the works by Oldroyd (1950, 1958), Jones and Waters (1967),
Waters and King (1970, 1971) and Huilgol (1975). As stated earlier, any linear viscoelastic system
which falls in the category of viscometric flows is known as a generalized Newtonian fluid. The
apparent viscosity of such a fluid is ir. For convenience, we will use p to represent the fluid viscosity,
so that the theological model is r =i. By using the boundary condition
Vz=0 at r=R (4.1)
The solution for the velocity profile is
V= r 2 (4.2)
The corresponding volumetric flow rate is
Q 3 (4.3)
4p
and the ratio between V, and mean velocity V, is
V =21( r2] (4.4)
V. R
At steady state Vm is constant, and the velocity profile only depends on R. A sample of velocity
profile in the pipe is shown in Fig. 4.1.
4.1.2 Bingham Plastic
For a material that conforms to the Bingham plastic model, the theological formulas are given
in Eqs. (3.17) and (3.18). If Ir,< Ty, the material will behave like a rigid solid. Therefore, from the
pipe centerline to the point at which  r, = Ty, the material moves as a solid plug", as shown in Fig.
4.2. Solving Eqs. (2.21), (3.17) and (3.18) leads to:
0.5
r/R 0
0.5 1 1.5
v/V.
Fig. 4.1 Velocity profile for a generalized Newtonian fluid with Q=0.003
m3/s and R=0.1 m.
Fig. 4.2 Schematic of velocity profile for a Bingham plastic.
for ro< rR, where ro=(:y/:w)R:
and for rro:
VZ= 1 r 21_ ' 1
2
plug 2 V
The corresponding volumetric flow rate is:
Q=Qfluid QpIug =n 3 1_ y 1'14
4g Z 3r
Examining of above solution indicates that the velocity profile depends on Try, V, R and Q. A sample
of velocity profile is shown in Fig. 4.3.
1
0.5
I 0 
0.5
0.5 1 1.5 2
V/V.
Fig. 4.3 Velocity profile for a Bingham fluid, with Q=0.003m3/s, R=0. lm, for slurry density p=1000,
1099, 1198 and 1314 kg/m3 (from right to left).
(4.5)
(4.6)
(4.7)
4.1.3 PowerLaw Fluid
The theological equation of state for a powerlaw fluid takes the form (3.19). Combining Eqs.
(2.21) and (3.19) leads to the following solutions:
Velocity profile:
V= n ) (1) () (4.8)
z n+1
Volumetric flow rate:
Q= n 'l 7R 3 (4.9)
Ratio between V, and mean velocity Vm:
V =_ 3n+i) 1 r[(I) (4.10)
Therefore, when Vm is constant, the velocity profile only depends on n and R. Examples of velocity
profiles in the pipe is shown in Fig. 4.4.
4.1.4 Maxwell Model
From Eq. (3.32), the theological formula for a Maxwell fluid is of the form
L +d = d z (4.11)
z G dt dr
For steady flow, d,/dt=0, so that in this case the Maxwell fluid is a viscometric flow, as noted
earlier. The solution is identical to that given in Section 4.1.1.
4.2 Initial Value Problem and Periodic Motion
0 0.5 1 1.5 2 2.5
V/.,
Fig. 4.4 Velocity profiles for a powerlaw fluid, with Q=0.003m3/s, R=0. m, for flow index
n=0.5, 1.0, 2.0 and 3.0 (from left to right).
Thus far, we have mentioned some analytical solutions of viscometric flows at steady state.
However, in so doing we lack knowledge of the dynamic behavior of viscoelastic fluids in the
unsteady state, e. g., the initial value problem and periodic motion which commonly exist in real
situations. Therrfore, in this section we will examine the initial value problem and periodic motion
in pipes for a linear viscoelastic material.
To solve initial value problem, one is required to consider the integral model for finite linear
viscoelasticity, i.e., Eq. (2.24), in which the only unknown quantity is the relaxation function. As an
illustration, we will us consider the socalled "start up" problem. Here the fluid, which is initially
at rest, is set in motion by a suddenly applied pressure gradient. The transitional velocity field as well
as the final velocity field are of interest in the analysis of this flow.
Next, the motion of a viscoelastic fluid under a periodic pressure gradient will also be
examined. Here we will concentrate on the results obtained for the theory of finitely linear
viscoelasticity for flow due to a pulsating pressure gradient superposed on a mean pressure gradient.
4.2.1 Initial Value Problem
Consider the axial flow of a finitely linear viscoelastic fluid in the cylindrical coordinates.
The known path lines are given by
Vr=Ve=O, Vz=Vz(r,t) (4.12)
From Eq. (2.12) it is seen that, for Poiseuille flow, the only nonzero strain tensors Yij are
rz =zr=i =dVzdr. Hence from Eq. (3.24) the stress tensor reduces to
/ a '2 a1 a 12 aVz
rij= f(tt) a+ dt (4.13)
axi 9ax ax a9x, ar
Therefore, the only nonzero stress components are rz, trz and the normal stress rr,, where the first
two components, using the initial problem condition, i.e., =0, for t<0, take the form
s av at
Trz zrf= (tt') dt' = i(s)V(rts)ds (4.14)
oo 0
Then, the zcomponent of the equation of motion, (2.9), combined with the Poiseuille flow
properties, Eq. (4.12) and (4.14), reduces to
atv
0
where v2 a2 1 a and ,=const. is the pressure gradient as in Eq (3.18). The boundary conditions
are ar2 r r
Vz(R,t) =0, (0,t) =0 (4.16)
Let W denote the Laplace transform of V:
W(r,wo) = Vz(r,t)e "'dt (4.17)
Then, by using Eq. (3.36) and (4.17), Eq. (4.15) becomes
Then, by using Eq. (3.36) and (4.17), Eq. (4.15) becomes
a2_W +1 aW q 2W= L
+q2W
2_ p
0
which is subject to the conditions
W(0,) =0
a (O,r )=0
ar
A solution of Eqs. (4.18) and (4.20) is
Let the distribution function H(,) be assumed to be
H(X)= =tx 8(XL) 1O )Ll)28(,XX
'XI
where 6 is the Dirac delta function. It is assumed that 1 >A2>0, where 1 is the relaxation time and 12
is called the retardation time. Using Eq. (4.19) and (4.22) we have
2 p)(1 +W)1)
11o(1 +6)2)
(4.23)
where
(4.18)
(4.19)
W(R,o) =0,
(4.20)
W(r,o) = 
1 J(iqr)
Jo(iqR))
(4.21)
(4.22)
If we set
r t lot
r1 t ,
R 'pR 2
V 
8rio
S a a= 1,2,
pR 2
pR2
i0o
_ o(1+S10)
1(1 +S2)
The complex inverse integral of Eq. (4.21) is given by
1 C+il
Vz(rtl)=i f W,(ri,o)exp(ot)do,
ci.
8Vm Jo(iq,rl)
W(r,o)= 1 jo(iql)
02 0 Jo1
Through singularity analysis of W,(r1,o)exp(ot1), and using Eq. (4.24), the solution becomes
n G,,(t,)
2 "n~l
(4.26)
1=( r2)8E exp
2V, z. J(Z,)Zn3
where
G(tP=cos ntl 1+Z2,2(S22Ss) in ntI
2S Pn 2S,
(4.27)
Zn=iql, an=' +S2Zn2,
n=(1 +S2Zn2)4SZ,,2]12,
s=0, S Sn
2S1
For a Newtonian fluid, S =S2 =0 and thus
(4.29)
Note that Jo, J1 are the Bessel functions of the first kind and order 0 and 1, respectively. It can be
shown that a weakly elastic fluid behaves like a Newtonian fluid as the velocity increases from zero
to its final steady state value gradually (Fig. 4.5). However for a general viscoelastic fluid, its
velocity at the center oscillates about the Newtonian value before reaching it (Fig. 4.6).
(4.24)
(4.25)
(4.28)
Vz(r",tl) =(1r2)8E Jr1Zn) exp(Z 2t )
Vm z, J1(Zn)Z,3
0.2 i I /
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
V/2 V
Fig. 4.5 Velocity profiles during the generation of Poiseuille flow of a weakly elastic fluid with
Si=0.04, and S2=0.02, for t,=0.1, 0.18, 0.34 and (from left to right). Broken curves are the
corresponding Newtonian profiles.
1I
0 0.5 1
v/2V.
1.5 2
Fig. 4.6 Velocity profiles during the generation of Poiseuille flow of a general viscoelastic fluid with
S,=0.4 and S2=0.04, for t,=0.15, , 0.85 and 0.5 (from left to right).
.t I
4.2.2 Periodic Motion
To examine flow behavior under a periodic pressure gradient, let us assume that the pressure
and velocity fields are given by
p =Re[zAe 'r],
(4.30)
Now, the equation of motion for the axial flow of an incompressible fluid [Eq. (4.15)] becomes
d2f 1 df +k 2f Ak2,
dr2 r dr ipa
01
k2= H(Xi) dX
1+i) )
The solution of Eq. (4.31), satisfying the conditions f(R)=0 andf(O)
A [ Jo(kr)
f(r) = i IR
ipw Jo(kR)
(4.31)
(4.32)
where JA is the Bessel function of the first kind of order 0.
For small frequencies w, that is, small k, one can approxiamate Eq. (4.32) and obtain Vz(r,t)
Vz(r,t) A (R 2r 2)c(t+8)
IT I
(4.33)
where r1 *(() is the zero shear rate complex viscosity, written as r *(W)= I r exp(i6). Thus, for a
slowly varying pressure gradient the flow is equivalent to that of a Newtonian fluid, in which the
velocity field is parabolic at each instant, but leading the applied pressure gradient by a phase angle
6.
When Ikrl is very large (except near the centerline of the pipe, where r=0), Jones and
Walters (1967) obtained an asymptotic expansion forJo(kr) in powers of (kr)1
(4.34)
R)r Rl i(Rr) (Rr)(9R+7r) ik(Rr)
ip\ r) 8kRr 128k2R2 2
410
Vz(r,t) =Re[f(r)e 't]
Near the centerline, for large values of kr, Eq. (4.32) yields
A
V(r,t)= sinot
po
(4.35)
This is a noteworthy result, indicating that at large frequencies the velocity at the centerline of the
pipe lags the applied pressure gradient by it/2 in phase. The mean square velocity Vm2 is given by
V2= 1 [(ReVz)2+(I V)2] =
2
2p2l B 1 iS S(16B7S) eiKS
2WW22 BS 8B(BS)K 128B 2(BS)2K2
(4.36)
where K=(o/vO)12K, v0=r/p, B=(o/vo)R and S=(o/lv)12(Rr). For a fluid with the distribution
function H(X) given by Eq. (4.22), the plot of 2p2 2Vm2/A 2 versus S (Fig. 4.7) shows the following:
CI
S0.8
at.
3
1A A
S=(o/Jo)m(Rr)
Fig. 4.7 Curves showing the variation of mean square velocity profiles with the parameters (A,1,
~)X2)=(4,1), (4,2) and (4,3), (from left to right). Solid line corresponds to a Newtonian fluid.
411
(1) the mean square velocity has a higher value than for the corresponding Newtonian value;
(2) this higher value occurs closer to the wall than in the Newtonian fluid;
(3) the thickness of the boundary layer in which the rapid velocity fluctuations occur is
reduced from that for the Newtonian fluid.
For flow under a small periodic pressure gradient superposed on a constant mean value, the
above problem becomes considerably more complicated. If the pressure gradient dp/dz is given by
dp (1+ee i) (4.37)
dz
The equation of motion, (4.15), becomes
avz la
p =(1 +eei)+(rz) (4.38)
at r ar
In view of the form of the pressure gradient producing the flow, one can write the axial
velocity and the shear rate as
Vz(rt) =Vo(r) +eV,(r)e +2[V2(r)+V 1)(r)e +V (2)r)e 'i] (4.39)
0 (r V2)e +Y2 (r)e ] (4.40)
where terms of flow under 3 and higher order have gradient ignored. Note that V and correspond to thegiven by
case of flow under a constant pressure gradient ,. The shear stress zrz is given by
Tr=11 (YO)%O +fpl(jo',s)y(s)ds +ffp2(iO';si S2)XY(sl)Y(s2)dsdS2 (4.41)
0 00
where yo=dVodr and we let
412
(s) = f K(r,to)d yos (4.42)
or{
0
as the oscillatory shear part, correct to O(e2). Here, (p,, cP2 are called memory functions. Note that
these kernels obey the consistency relations
0 00 2 df2
d 1 do2
It is impossible to determine the kernel functions qp, p2. However, qualitative predictions may
made by making plausible assumptions concerning the unknown functions. Barnes et al. (1971)
obtained following approximation relation
(1+o2 32)
I= +I (4.44)
(1 +2,X42)
where
I=100( 1 (4.45)
Q and Qs are the mean flow rates with and without the pressure fluctuation, respectively, and
R
Q f(Vo+e2V2)21rrdr (4.46)
0
Eq. (4.44) implies that a nonzero o will give rise to a constant magnification of the (Il,D) curve by
a factor depending on the frequency of the fluctuation and the values of X3, 3X4
Barnes et al. (1971) also employed the Oldroyd (1958) model through a numerical method
to arrive at the following conclusions (Figs. 4.8 and 4.9):
(1) the mean flow rate Q is not affected if the viscosity is constant;
413
R
(2) let I denote the percentage fluctuation due to the E2 in Eq. (4.46) over fVo2irdr. Then
for low values of 0, I is positive and for high value 1, I is negative; o
(3) there is also a strong indication of a "resonance" effect where a large percentage increase
in mean flow rate can be expected for a given o. This could be useful in increasing the mean flow
rate Q for a given mean pressure gradient D.
In order to determine any practical significance of above conclusion (3), Barnes et al. (1971)
examined the energy required to produce a given mean flow rate. The mean energy Ep required to
produce the flow in the case of a fluctuating pressure gradient is given by
Ep=27rc [ Vo+e +V2 ]dr (4.47)
The mean energy Es required to produce the same flow rate in the case of a constant pressure
gradient is given by
R
Ep=21 fs fr(Vo +e2V2)dr (4.48)
0
where 0s is the constant pressure gradient required to produce a flow rate Q. Fig. 4.10 shows the
theoretical curves of S against (D, where S is the percentage increase in the energy
E
5=100 1) (4.49)
It can be observed that for low flow rates, S is lower in the case of the viscoelastic fluids, but the
opposite is the case at higher flow rates. Of considerable interest is the observation that under some
conditions, S can take fairly large negative values, which implies that a significant saving of energy
by the use of a fluctuating pressure gradient is possible. This may be due to the retardation of the
development of a fully developed, steady flow, boundary layer as a result of periodic flow reversal
and associated interruption of the developmental process due to momentum transfer. The behavior
at high flow rates is also noteworthy, indicating that energy can be wasted in pulsatile flow in this
range.
414
Fig. 4.8 Illustrative curves for V0, V1, V2 as functions of r for rb=10 poises, Np =0.1
poise, X1=0.6 s, 2=0.1 s, e=0.25, p=l,000 kg/m3, o=l s1, 0=200, R=0.25 cm.
Broken line is the velocity profile for Newtonian fluid giving rise to same rate as V0.
40
30 Inelastic
1 20
(0=10
10 0
10 20
< (Pa)
Fig. 4.9 Theoretical (I, D) curves for various values of o s', o=10 poises, po=1
poises X1=0.06 s, X2=0.01 s and e=0.25.
415
400 c (pa)
Fig. 4.10 Theoretical (S, c ) curves for a Newtonian fluid (broken line) and
a viscoelastic fluid with viscosity function for a 1.75% aqueous solution of
polyacrylamide and with X3=2.75 s, 4=0.3 s, j=1 s, R=0.25 cm and e0.25
(solid line).
416
5. EXPERIMENTAL SETUP, PROCEDURES AND MATERIALS
5.1 Coaxial Cylinder Viscometer
The coaxial cylinder viscometer (CCV) (Fig. 5.1) used was of the Brookfield (model LVT)
type (Fig. 5.2). The general procedure for using the CCV involves rotating a metallic bob (a right
circular cylinder) or a spindle at a selected rate in a beaker containing the mud of known density. In
the present case the spindle could rotate at fixed speeds, giving a shear rate range of 0.063 to 20.4
Hz. The torque generated in mud by the rotation of the spindle was recorded from a readout meter.
The shear stress, which is proportional to the torque, was calculated directly from the torque using
a formula supplied by the maker (Brookfield Dial Viscometer, 1981).
Outer Torque Inner Cylinder
Cylinder (Bob)
Fig. 5.1 Schematic drawing of
a coaxial cylinder viscometer.
Fig. 5.2 Brookfield
viscometer with an
attached spindle
shearing a clay slurry.
The Brookfield viscometer is actually equipped with a series of spindles. The spindle
required to shear a particular slurry depends on the density and viscosity of the slurry. Use of these
spindles along with the charts provided results in values of the apparent viscosity, p, which is
inherently corrected for endeffects in the viscometer (Brookfield Dial Viscometer, 1981).
5.2 Horizontal Pipe Viscometer
The horizontal pipe viscometer (HPV) constructed at the Coastal Engineering Laboratory of
the University of Florida is shown schematically in Fig. 5.3, and a photographic view is given in Fig.
5.4. The 3.1 m long, 2.54 cm i.d. PVC pipe was clamped on to the a workbench, with one end
attached to a pistondiaphragm pump and the other end open, with a bucket receptacle to collect the
slurry. The ARO model 6661A3344C nonmetallic doublediaphragm pump was operated at a
nominal pressure of 40 psi (276 kPa). Pressurized air required at the pump inlet was supplied by a
compressed air line observed in Fig. 5.4. Over the central 2.46 m length of the pipe the pressure drop
was measured by two flushdiaphragm gage pressure sensors. The positions A and B of these sensors
were selected to be 32 cm from the pipe ends in order to minimize endeffects. Note that endeffects
tend to become significant when the ratio of pipe length to diameter exceeds about 100. In the
present case this ratio was 3.1/0.0254 = 122, i.e., well over 100. The pressure readings at A and B
were recorded by a PC using Global Lab software. Mud slurry could be fed through a hopper above
the pump and connected to it. Any slip effects that might have occurred during slurry flow were
likely to have been small compared to the high velocities that were generated during the tests, and
therefore were not explored further.
Hopper
Pump Pipe
A B
Bucket
I CD
Pressure Drop
Fig. 5.3 Schematic drawing of experimental setup for the horizontal pipe
viscometer (HPV).
5.3 Sediments and Slurries
5.3.1 Sediment and Fluid Properties
Three types of commercially available clays: a kaolinite, a bentonite, and an attapulgite,
which together covered a wide range of cohesive soil properties, were selected. Kaolinite (pulverized
kaolin), a light beigecolored powder, was obtained from the EPK Division of Feldspar Corporation
in Edgar, Florida. Its Cation Exchange Capacity (CEC), as given by the supplier, was 5.26.5
milliequivalents per 100 grams. Its granular density was 2,630 kg/m 3. Bentonite was obtained from
the American Colloid Company in Arlington Heights, Illinois. It was a sodium montmorillonite
(commercial name Volclay) of a light gray color. Its CEC was 105 milliequivalents per 100 grams,
and its granular density was 2,760 kg/m 3. Attapulgite, of greenishwhite color, was obtained from
Floridin Company in Quincy, Florida. Also called palygorskite, its CEC was 28 milliequivalents per
100 grams, and its granular density was 2,300 kg/m3. Tables 5.1 through 5.3 respectively give the
chemical compositions of the three clays (provided by the suppliers).
MI
Fig. 5.4 Photograph of HPV setup.
Table 5.1: Chemical composition of kaolinite
Chemical % Chemical %
SiO, 46.5 MgO 0.16
A1203 37.62 Na2O 0.02
Fe2O3 0.51 KO 0.40
TiO2 0.36 SO3 0.21
P205 0.19 V20s < 0.001
CaO 0.25
Table 5.2: Chemical composition of bentonite
Chemical % Chemical %
SiO, 63.02 A1203 21.08
Fe203 3.25 FeO 0.35
MgO 2.67 Na2O & K20 2.57
CaO 0.65 H20 5.64
Trace Elements 0.72
Table 5.3: Chemical composition of attapulgite (palygorskite)
Chemical % Chemical %
SiO2 55.2 A1203 9.67
Na20 0.10 K20 0.10
Fe203 2.32 FeO 0.19
MgO 8.92 CaO 1.65
HO 10.03 NHO 9.48
Table 5.4 gives the results of chemical analysis of the tap water used to prepare mud, whose
pH value was 8 and conductivity 0.284 millimhos. This analysis was conducted at the Material
Science Department of the University of Florida (Feng, 1992). The procedure was as follows: firstly,
an element survey of both the tap water and doubledistilled water was performed, which determined
the ions in tap water. Secondly, standard solutions of these ions contained in the tap water were
made, and the tap water was analyzed against the standard solutions to determine the concentrations
of the ions by an emission spectrometer (Plasma II).
Table 5.4: Chemical composition of water
Chemical Concentration
(ppm)
Si 11.4
Al 1.2
Fe 0.2
Ca 24.4
Mg 16.2
Na 9.6
Total Salts 278
The particle size distributions of kaolinite, bentonite and attapulgite are given in Tables 5.5,
5.6 and 5.7, respectively. The procedure for determination was: firstly, a particular suspension was
prepared at about 0.5% by weight concentration, and run for at least 15 minutes in a sonic
dismembrater (Fisher, model 300) to breakdown any agglomerates. Secondly, the suspension was
analyzed in a particle size distribution analyzer (Horiba, model CAPA 700), and allowed to gradually
settle down to the bottom. Particle concentration and fall velocities were determined with an Xray,
which could be converted to Stokes equivalent diameters. The median particle sizes of kaolinite,
bentonite and attapulgite were 1.10 pm, 1.01 prn and 0.86 pu, respectively.
Table 5.5: Size distribution of kaolinite
Frequency distribution
(%)
0.0
0.0
2.9
4.0
2.6
4.1
4.0
6.0
5.7
6.2
5.5
6.2
5.8
5.0
10.4
11.2
13.6
6.8
Cumulative frequency distribution
(%)
0.0
0.0
2.9
6.9
9.5
13.6
17.6
23.6
29.3
35.5
41.0
47.2
53.0
58.0
68.4
79.6
93.2
100.0
5.3.2 Mud Slurries
Mud slurries of different densities were prepared by thoroughly mixing the selected dry clays
and clay mixtures with tap water at the ambient temperature, allowing these mixtures to stand for
a minimum of 24 hours before testing them in the CCV and the HPV. Mud composition, density
and water content are given in Table 5.8. Also given, in the last column, is the CEC of the slurry,
calculated as follows:
CECsiuny= fko1 CECkoino += puigitCECatMpujgite +fb.toiteCECbenton
(5.1)
where f represents the weight fraction of the subscripted sediment, and subscripted CEC are the
Diameter
(tm)
5.00<
5.003.20
3.203.00
3.002.80
2.802.60
2.602.40
2.402.20
2.202.00
2.001.80
1.801.60
1.601.40
1.401.20
1.201.00
1.000.80
0.800.60
0.600.40
0.400.20
0.200.00
corresponding cation exchange capacities. Note that given water as the weight fraction of water in
the slurry, we have: fkaoiite attapulgite bentonitewater = 1. The CEC values (in milliequivalents per
100 g) were selected to be: 6 (nominal) for kaolinite, 28 for attapulgite and 105 for bentonite. The
weight fractions depend on the composition of each slurry given in Table 5.8.
Table 5.6: Size distribution of bentonite
Frequency distribution Cumulative fr
(%)
5.9
1.9
2.3
2.5
3.0
3.0
4.9
5.3
8.1
4.5
9.3
9.1
equency distribution
(%)
5.9
7.8
10.1
12.6
15.6
18.6
23.5
28.8
36.9
41.4
50.7
59.8
71.2
82.4
93.3
100.0
Table 5.7: Size distribution of attapulgite
Diameter Frequency distribution Cumulative frequency distribution
(Pm) (%) (%)
2.00< 11.8 11.8
4.1 15.9
4.9 20.8
5.3 26.1
5.6 31.7
5.8 37.5
17.4 54.9
25.5 80.4
12.3 92.7
6.1 98.8
1.2 100.0
Diameter
(pm)
3.00<
3.002.80
2.802.60
2.602.40
2.402.20
2.202.00
2.001.80
1.801.60
1.601.40
1.401.20
1.201.00
1.000.80
0.800.60
0.600.40
0.400.20
0.200.00
2.001.80
1.801.60
1.601.40
1.401.20
1.201.00
1.000.80
0.800.60
0.600.40
0.400.20
0.200.00
...
Slurry No.
Table 5.8: Properties of mud slurries tested
Sediment Density Water Content
(kg/m3)
100%K 1,250 210
100%K 1,300 167
100%K 1,350 139
100%K 1,400 117
100%K 1,450 100
100%K 1,500 86
100%K 1,550 75
75%K+25%A 1,243 210
75%K+25%A 1,291 169
75%K+25%A 1,339 139
75%K+25%A 1,387 117
50%K+50%A 1,236 210
50%K+50%A 1,283 169
50%K+50%A 1,306 153
50%K+50%A 1,329 139
25%K+75%A 1,175 289
25%K+75%A 1,200 253
25%K+75%A 1,225 215
25%K+75%A 1,250 189
100%A 1,125 409
100%A 1,150 333
100%A 1,175 280
100%A 1,200 239
90%K+10%B 1,200 273
90%K+10%B 1,250 211
90%K+10%B 1,300 169
90%K+10%B 1,350 140
CEC,,Su
(meq/100 g)
1.94
2.25
2.51
2.84
3.00
3.23
3.44
3.71
4.28
4.80
5.29
5.48
6.33
6.72
7.10
5.79
6.38
7.14
7.77
5.50
6.46
7.38
8.26
4.26
5.12
5.90
6.63
28 65%K+25%A+10%B 1,225 231 6.47
29 65%K+25%A+10%B 1,250 204 7.04
30 65%K+25%A+10%B 1,275 182 7.59
31 65%K+25%A+10%B 1,300 163 8.12
32 40%K+50%A+10%B 1,175 299 6.74
33 40%K+50%A+10%B 1,200 257 7.54
34 40%K+50%A+10%B 1,225 224 8.31
35 40%K+50%A+10%B 1,250 197 9.05
36 15%K+75%A+10%B 1,125 423 6.20
37 15%K+75%A+10%B 1,150 345 7.28
38 15%K+75%A+10%B 1,175 290 8.31
39 15%K+75%A+10%B 1,200 248 9.30
40 90%A+10%B 1,125 415 6.93
41 90%A+10%B 1,150 339 8.13
42 90%A+10%B 1,175 284 9.29
43 90%A+10%B 1,200 244 10.39
From Table 5.8 we note that the density range covered was from a low 1,125 kg/m3 to a high
1,550 kg/m3. The water content varied from a high 423% to a low 75%. Finally, the CEC,,,, values
ranged from 1.94 meq/100g for a kaolinite slurry (no. 1) to 10.39 meq/100g for a slurry (no. 43)
composed of attapulgite and bentonite.
6. EXPERIMENTAL RESULTS
6.1 Rheometric Results
6.1.1 PowerLaw for Mud Flow
Previous work on flocculated bottom muds in the coastal environment has established their
the pseudoplastic (shear thinning) flow behavior (e.g., Parker and Kirby, 1982). Subsequent work
by, among others, Feng (1992) has revealed that the wellknown Sisko (1958) powerlaw provides
a reasonable fit to the measured decrease in apparent viscosity, p, with increasing shear rate, j, a
behavior that is consistent with the pseudoplastic flow curve (Fig. 6.1). Note that for convenience
we will use the symbol [ for the apparent viscosity in place of r! used for example in Chapter 3.
Pseudoplastic shear
thinning flow curve
Newtonian flow curve
Fig. 6.1 Comparison between pseudoplastic (shearthinning) and
Newtonian flow curves. The nature of the shearthinning curve
is such that while at low shear rates the viscosity is higher than
the constant value for the Newtonian case (straight line), with
increasing shear rate the pseudoplastic curve becomes
asymptotically parallel to the Newtonian line, hence the
pseudoplastic viscosity approaches that of the Newtonian case.
With reference to the Sisko model, we begin by noting that general powerlaw equations that
predict the shape of the curves representing the variation of viscosity with shear rate typically need
at least four parameters. One such relation is the Cross (1965) equation given by
(cl )p (6.1)
where go and I_ refer to the asymptotic values of the viscosity at very low and very high shear
rates, respectively, c, is a constant parameter having dimensions of time, p is a dimensionless
constant, and I is the apparent viscosity.
It is generally found that << go, hence the above equation can be simplified as
o = (cl') (6.2)
It 1 .
which can be rewritten as
I9 = F+ + (6.3)
(c1j')P
or
I = g +cY"1 (6.4)
Equation (6.4) is the Sisko model, where ._ is the constant ultimate viscosity at the limit of high
(theoretically infinite) shear rate, c is a measure of the consistency of the material, and n is a
parameter which indicates whether the material is shearthinning or shearthickening, that is, when
n > 1 the material exhibits shearthickening, and n < 1 denotes a shearthinning behavior. When n
= 1 the behavior is Newtonian, with a with a constant viscosity equal to gp + c. Note also that when
Ip=0, Eq. (6.4) becomes consistent with the powerlaw given by Eq. (3.19). Note that we are
conveniently using the symbol c denoting consistency in place of p used in Eq. (3.19).
It is important to recognize that the coefficients of Eq. (6.4) must be derived from
measurements conducted under a laminar flow. The laminar limit for Newtonian slurries is given
by the wellknown Reynolds number criterion:
Re VD < 2100 (6.5)
9
For a powerlaw [Eq. (3.19)] slurry, the limit is given by (Ryan and Johnson, 1959):
Re' = VD =< 4(n +2)n4)
8V n1 3n+ n 3n+1 (6.6)
D ) 4n)
In a strict sense, condition (6.6) is applicable to the Sisko powerlaw fluid only when = 0. It can
perhaps be used to represent the flow of a Sisko fluid in general in an approximate way in cases
where p_ is small, e.g., for many of the slurries in the present investigation, for which the value of p_
was found to be close to that of water.
6.1.2 Calculation of PowerLaw Coefficients
To solve for the three Sisko parameters, p., c and n, the least squares method can be used
for fitting the curves obtained from Eq. (6.4) to the experimental data on the apparent viscosity, t,
as a function of the shear rate, j, obtained from the measured relationship between stress (T) versus
j, such as shown qualitatively in Fig. 6.1. For this least squares method it is required that the
viscosity difference between the model [Eq. (6.4)] and data, D, be minimized, that is,
N
D= > (Pip)2= minimum (6.7)
i=1
or
N
D=L (PipCc~'1)2= minimum (6.8)
i=1
where Pi is the mud viscosity obtained from the experiment, and N is the number of data points.
Setting
_D D aD
0 =0; =0; D 0 (6.9)
a8Lp an ac
from Eq. (6.8) we obtain the following by differentiation:
N
S(pipc'n1) =0 (6.10)
i=1
N
E {"n1(i,c n1)} =0 (6.11)
i=1
and
N
S{cn1 logY(,ij~c"n1 )} =0 (6.12)
i=1
In this way, gp, c and n can be determined by solving Eqs. (6.10), (6.11) and (6.12). A requirement
for the determination of these coefficients is that each slurry be tested over a comparatively wide
range of the shear rate j, so that the lowshear rate nonNewtonian and high shear rate Newtonian
behaviors are identified.
6.2 PowerLaw Parameters
All 43 slurries noted in Table 5.8 were tested in the CCV and the HPV; the CCV for data at
low shear rates, and the HPV for high shear rates. Slurry no. 7 did not flow through the HPV due to
insufficient pump pressure. For this slurry therefore the Sisko powerlaw coefficients were not
calculated. The overall rage of shear rates covered in the CCV was 0.063 Hz to 20.4 Hz, whereas
the in the HPV they were considerably higher, in the range of 150.7 to 1094.5 Hz. Note that in the
CCV the shear rate is an independent parameter which is inputted, whereas in the HPV it depends
on the pipe diameter, length, pressure drop and slurry rheology. As discussed further in Section 6.3,
at the high shear rates in the HPV the behavior of the slurry was close to Newtonian, hence the shear
rate, j, and the corresponding shear stress, Tw, both at the pipe wall, could be calculated from the
following Newtonian flow equations:
8V
S (6.13)
D
S A (6.14)
S 4L
here V is the mean flow velocity in the pipe, D = 2R is the pipe diameter, R is the pipe radius and
L is the distance over which the pressure drop, Ap, occurs. Then g =w/j For each slurry the
measured pressure drop, Ap, the measured discharge, Q, and the calculated wall stress, ',, are given
in Table 6.1. Note that while in the CCV each sample was tested only once after the correct spindle
was selected, in the HPV each sample was tested three times. The reported pressure reading and the
discharge (obtained by timing the rate of flow of the slurry out of the pipe, weighing the mass
accumulated in the bucket placed to receive the slurry, converting this weight to volume knowing
the density and dividing the volume by the measured time) are means of the three measurements.
Table 6.1: Pressure drop, discharge, shear rate and wall stress data from HPV tests
Slurry no. Pressure drop, Ap Discharge, Q Shear rate, j Wall stress, Tw
(Pa) (m3/s) (Hz) (Pa)
1 76,065.9 0.00150 931.8 196.3
2 54,452.4 0.00140 869.7 140.6
3 150,503.0 0.00136 844.9 388.5
4 192,067.5 0.00130 807.6 495.8
5 205,348.3 0.00124 770.3 530.1
6 144,997.7 0.00086 534.3 374.3
7 N.D.a N.D. N.D. N. D.
8 119,050.5 0.00144 898.7 307.3
9b 177,736.9 0.00129 803.4 458.8
10 312,957.7 0.00076 472.1 807.8
11 54,381.6 0.00024 150.7 140.4
12 93,544.0 0.00149 927.7 241.5
13 85,376.3 0.00123 766.2 220.4
14 78,436.3 0.00104 647.1 202.5
15 73,784.9 0.00064 399.7 190.5
16 95,384.3 0.00158 980.9 226.2
17 137,272.7 0.00141 874.8 354.3
18 168,589.8 0.00121 754.2 435.2
19 179,701.1 0.00117 728.2 463.9
20 75,907.4 0.00156 907.0 195.9
21 106,010.3 0.00141 873.2 273.6
22 128,008.6 0.00130 809.2 330.4
23 167,889.2 0.00123 764.1 433.4
25 73,888.0 0.00153 953.5 190.7
26 95,130.6 0.00141 872.9 245.6
27 126,692.1 0.00130 810.1 327.0
28 95,659.2 0.00166 1030.1 246.9
29 113,064.7 0.00157 974.4 291.9
30 115,010.4 0.00145 898.7 296.9
31 249,492.7 0.00134 829.6 644.0
32 58,314.3 0.00161 1002.2 150.5
33 33,372.3 0.00150 931.8 86.1
34 55,968.5 0.00146 907.9 144.5
35 96,128.4 0.00130 809.2 248.1
36 71,254.7 0.00157 976.1 183.9
37 57,862.2 0.00144 895.5 149.4
38 54,352.8 0.00153 952.2 140.3
39 121,193.6 0.00131 811.6 312.8
40 82,942.3 0.00156 968.8 214.1
41 45,127.2 0.00147 911.3 116.5
42 134,873.8 0.00141 878.0 348.1
43 160,813.2 0.00084 520.6 415.1
aNo data obtained as slurry did not flow at this high density.
bFor the particular Brookfield spindle which had to be used for this slurry, the torque reading was found to be below the
minimal value at which the measurements were considered to be reliable by the maker of the viscometer.
An example of the Sisko relationship [Eq. (6.4)] based on the combined CCV and HPV data
is shown in Fig. 5.5, in which the eight points within the lower shear rate range were obtained by the
CCV, and the single value at the higher shear rate from the HPV. Bestfit coefficients p_, c and n
for all the slurries obtained in the same way are listed in Table 6.2, which also gives the
characteristic HPV flow Reynolds number, Re, calculated according to Eq. (6.5). Thus, in other
words, the Newtonian Reynolds number is given as the characteristic value in lieu of Re' from Eq.
(6.6) because the flow regime in the HPV was practically Newtonian (see Section 6.3). All plots of
excess apparent viscosity, p, as a function of shear rate are given in Appendix A.
As seen in Table 6.2 from the range of Reynolds numbers experienced, all tests were carried
under nonturbulent conditions, as required for the theological analysis. The powerlaw coefficients
0.00176 1094.5
209.2
24 81,047.4
show considerable variability with slurry composition and density. Note that the lowest value of
, was chosen to be 0.001 Pa, the viscosity of water. In other words, in the least squares analysis p_
was not allowed to have values lower than the viscosity of water.
^,o3
10.
0101
0
S10
1.
0. 1
U 10
10
Kaolinite, density = 1350 kg/m3
10 102 1
Shear rate (Hz)
Fig. 6.2 Excess apparent viscosity as a function of shear rate
for kaolinite slurry no. 3.
Table 6.2: Sisko model coefficients and HPV flow Reynolds number
Slurry No. Io, c n Ref
(Pa.s)
1 0.1904 1.99 0.308 223
2 0.0024 3.79 0.520 282
3 0.2813 9.75 0.403 100
4 0.0011 6.81 0.610 74
5 0.4633 8.19 0.448 66
6 0.5837 14.46 0.228 46
7 N.C." N.C. N.C. N.C.
8 0.0015 2.43 0.581 132
9 0.4983 8.02 0.296 73
10 1.6880 9.98 0.003 15
p. = 0.2813 Pa.s
c = 9.75
n = 0.403
0.7513
0.2380
0.2723
0.2965
0.3140
0.2357
0.3585
0.5632
0.6242
0.0019
0.3070
0.3628
0.5604
0.1899
0.1608
0.1530
0.2432
0.1544
0.2012
0.2179
0.4926
0.0970
0.0013
0.0719
0.1473
0.0012
0.0566
0.0912
0.2990
0.1787
41.31
4.31
13.21
21.67
86.17
3.14
9.61
22.07
34.01
1.27
3.36
13.55
21.38
2.97
9.43
15.86
47.92
9.46
18.29
23.12
66.70
8.05
12.71
11.37
40.29
2.31
6.55
9.23
23.96
4.20
0.083
0.230
0.022
0.116
0.047
0.223
0.212
0.124
0.205
0.559
0.068
0.148
0.236
0.152
0.199
0.289
0.149
0.318
0.239
0.216
0.188
0.274
0.297
0.280
0.174
0.538
0.398
0.251
0.160
0.318
9
130
45
45
31
185
105
65
58
229
129
94
65
277
240
163
109
213
164
140
56
317
488
282
133
236
249
307
102
199
0.0950
0.3570
0.5263
15.24 0.100
17.95
104.00
0.095
0.048
" Not calculated, because slurry did not flow.
In Fig. 6.3, pi, c and n are plotted as functions of kaolinite slurry density (for slurry nos. 1
through 6; slurry no. 7 with a density of 1,550kg/m3 did not flow in the HPV). Analogous data for
all ten mud types tested are given in Appendix B. Observe in Fig. 6.3 that gt shows an overall
increasing trend with increasing density. The curve seems to suggest that this increase may not be
monotonic, although it should be recognized that the number of data points is too few to arrive at
a firm conclusion in this respect. In any event, it is logical to expect [lt to increase with density. The
consistency, c, is seen to increase linearly with density. This trend can also be expected as it implies
that for a given shear rate (and holding n invariant), slurry viscosity increases with density. Finally,
n seems to increase at first with density, go through a maximum, and then decrease. Since n<1
throughout, over the entire density range the slurry behavior is seen to be pseudoplastic.
Reviewing the data for the other nine muds in Appendix B we note that the trends are
qualitatively akin to those for kaolinite, especially with regard to consistent pseudoplastic behavior.
The nonmonotonic effects of density on the coefficients in many cases may reflect the complex
physical and physicochemical interactions between the particles and the pore fluid at different shear
rates.
Kaolinite
0.8
0.6
S0.4
2 0.2
0 *
190n 12M3 14 19 00 1800
Kaolinite
Fig. 6. ariion of with density of kolinite sl es.
15
S10
5
0
1200 1300 1400 1500 1600
Density (kg/m^3)
Fig. 6.3b Variation of c with density of kaolinite slurries.
Kaolinite
0.8
0.6
c 0.4
0.2
0
1200 1300 1400 1500 1600
Density (kg/m^3)
Fig. 6.3c Variation of n with density of kaolinite slurries.
6.3 Slurry Discharge with Sisko Model
Given horizontal velocity of the slurry v(r), we have the shear rate =av/ar, and the steady
state momentum equation for pipe flow is
r
P = , (6.15)
Rw
where Tw is the wall shear stress. For expressing the apparent viscosity, V, in terms of the shear rate,
the Sisko model is given by Eq. (6.4). Next, we let
= 1 (6.16)
R
Then, combining Eqs. (6.4), (6.15) and (6.16) we obtain
C+ + (f1)r = 0 (6.17)
R Rn a T u /
which must satisfy the noslip boundary condition at the wall, i.e., v(rl=0) = 0. Equation (6.17) can
next be written in the finite difference form as
610
(6.18)
r+C ri+ Vi in V1 1i+1 +Ti
+ + 2 1 = 0
R Rs Ad A 2 n iti
Based on Eq. (6.18) the following iterative relation was used:
m+1 A
i+1 Vi + AI
STl i+1I
1 ,
2 w
n1
m
C vi+1Vi
R R All
where v,+1, (i = 1,2, ..... N) is the layer velocity, N is the total number of layers into which the
distance from the wall to the centerline is divided, and m is the iteration index. Equation (6.19) was
solved with the initial condition v1=0 at the wall. The chosen criterion for convergence was
IViml V mi' < 106
Finally, the discharge, Q, is obtained from the summation
Q=2 ( ViVi rdr
i=1 2
(6.20)
(6.21)
6.4 Calculation of Slurry Discharge
In order to test the applicability of the Eq. (6.21) for numerical determination of the
discharge, Q, several slurries were pumped again through the HPV. The data are given in Table 6.3.
These tests were deliberately carried out at pressures lower than those used to obtain the data points
for determining the powerlaw relationships. (Compare the pressure drops in Table 6.3 with the
corresponding ones in Table 6.1).
Table 6.3: Low pressure HPV test parameters for selected slurries
Slurry no. Pressure drop, Ap Measured Shear rate, Wall shear stress, T, Computed
(Pa) discharge, Q (Hz) (Pa) discharge, Q
(m3/s) (m3/s)
1 68,465.5 0.00135 836.6 176.7 0.00133
12 82,678.5 0.00136 844.9 213.4 0.00128
16 67,697.0 0.00128 794.5 174.8 0.00108
611
(6.19)
28 67,464.2 0.00133 825.4 174.2 0.00093
32 73,927.5 0.00129 800.9 189.2 0.00200
40 74,920.0 0.00137 850.0 193.4 0.00136
Using the powerlaw coefficients for these slurries form Table 6.2, Eqs. (6.19) and (6.21)
were solved along with the convergence criterion of Eq. (6.20). The number of layers, N, into which
the pipe radius was divided was 20. As an example, the calculated result for slurry no. 1 is shown
in Fig. 6.4, which plots the computed velocity profile and gives the corresponding discharge.
Observe that this discharge agrees well with that measured in Table 6.2. Also plotted is the velocity
profile assuming the slurry to be Newtonian [Eq. (4.1)], and the corresponding discharge is
calculated from Eq. (4.2). It is seen that the Newtonian assumption is reasonable at the high shear
rate (836.6 Hz) at which the data were obtained, and justifies the Newtonian assumption for
calculating the viscosity from HPV for determining the powerlaw coefficients in Table 6.2. In fact,
for each slurry the data point for calculating the powerlaw coefficients was obtained at a higher
shear rate (931.8 Hz) than in Table 6.3, thus making it even more acceptable to assume Newtonian
flow. Another illustrative plot (for slurry no. 24) is given in Fig. 6.5. Measured and computed
discharges for all slurries tested are given in Table 6.3. The degree of agreement varies, and can be
shown to be sensitive to the powerlaw approximation of the theological data, i.e., to the extent to
which the powerlaw fits the measured data from the viscometers.
1
QNewtonian = 0.00136 m3/s
0.5
QSisko = 0.00133 m3/s *
 
0.5
0 1 2 3 4 5 6
Velocity (m/s)
Fig. 6.4 Computed velocity profiles and corresponding discharges
for slurry no. 1. Line is numerical solution using Sisko model; dots
represent analytic Newtonian solution.
612
24 71,668.8
0.00133 823.2
185.0 0.00155
7
QNewtonian = 0.00156 m3/S
0 ___ _
0.5
1
0 2 4 6 8
Velocity (m/s)
Fig. 6.5 Computed velocity profiles and discharges for slurry
no. 24.
613
7. INTERPRETATION OF RESULTS AND CONCLUSIONS
7.1 PowerLaw Correlations with Slurry CEC
The slurry CEC (CECSu,), as defined by Eq. (5.1), potentially lends itself as a measure of
slurry cohesion, hence its rheology, at least to the extent to which cohesion and rheology are likely
to be physicochemically related. In Figs. 7.1, 7.2 and 7.3, data from Table 6.2 have been used to
plot the powerlaw coefficients pt, (logarithm of) c and n against CECsy given in Table 5.8.
(Slurry nos. 7 and 9 have been omitted for reasons cited in Table 6.2.) Observe that while there is
considerable data scatter, not all of which is likely to be "random", correlations indeed seem to exist
in the mean (lines).
We observe in Fig. 7.1 that in the mean pt increases with CECSuy, which can be expected
since greater cohesion would imply greater interparticle interaction, hence viscosity. Similarly, in
Fig. 7.2 the consistency, c, is seen to increase with CEC, ,J, which is consistent with the trend in Fig.
7.1, given that consistency can be expected to vary directly with viscosity. Finally, in Fig. 7.3, n is
seen to decrease with CECy. Note that since n = 1 would mean a Newtonian fluid, and for a shear
thinning material n < 1, the observed trend of variation of n with CECsIu, implies increasingly non
Newtonian, shearthinning behavior of the slurries with increasing cohesion.
The mean trend lines in Figs. 7.1, 7.2 and 7.3 respectively correspond to the following
relations:
p, = 0.0125CECsiu + 0.181 (7.1)
logc = 0.125CECsly + 0.27 (7.2)
n = 0.047CECslu + 0.50 (7.3)
7.2 Concluding Comments
The experiments and analyses presented in the previous chapters essentially highlight a
method which may be explored further in future for assessing pumping requirements. A drawback
of course is that without knowing the rheology of a given mud its transportation characteristics
cannot be determined. Secondly, the use of CEC as a measure of theological behavior of slurry cannot
be extended to sediments that are not clayey. Given these two limitations, it will be necessary to: 1)
examine a wide range of natural muds for their theological behavior, and 2) develop correlations
between rheology characterizing parameters and readily determinable parameters perhaps including,
but not limited to, CEC. Finally it should be added that slurries of densities higher than those tested
must be studied, using pumps which can supply higher pressures, in order to fulfill the need to
quantify the understanding of transportability of high density muds at in situ densities. In any event,
the following procedure, developed as part of this study, can serve as a guide for future efforts in this
regard.
1. For the site to be dredged determine the required pipe discharge.
2. Collect the bottom mud sample to be discharged.
3. With mud rheology known, backcalculate the pressure drop required to achieve the discharge
for a pipe of known dimensions. This can be done by matching the required discharge with that
calculated from Eqs. (6.19) and (6.21), for a given (calibrated) value of Ap.
4. This value of Ap should be considered to be the minimum pressure drop required for the pump
to be selected.
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Slurry CEC
Fig. 7.1 Variation of pL with slurry CEC (CEC,,S) for all slurries.
*4
4
2.5
1.5
1
0.5
0
0 2 4 6 8 10
Slurry CEC
Fig. 7.2 Variation of logc with slurry CEC (CECsIur) for all slurries.
0.8
0.6
0.4
0.2
0
0.2
0.4
Slurry CEC
Fig. 7.3 Variation of n with slurry CEC (CEC,uy) for all slurries.
4 *
4.
0 2 4. 6 * 8 10 1
*
8. REFERENCES
Alfrey, T., and Doty, P., 1945. The method of specifying the properties of viscoelastic materials,
Journal of Applied Physics, 16, 700713.
Barnes, T. N. G., Townsend, P., and Walters, K., 1971. On pulsatile flow of nonNewtonian liquids,
Rheologica Acta, 10, 517527.
Brookfield Dial Viscometer, 1981. Operating Manual, Brookfield Engineering Laboratories,
Stoughton, MA.
Cross, M. M., 1965. Rheology of nonNetonian fluids: a new flow equation for pseudoplastic
systems, Journal of Colloidal Science, 20, 417437.
Darby, R., 1976. Viscoelastic Fluids, Marcel Dekker, New York.
Feng, J., 1992. Laboratory experiments on cohesive soil bed fluidization by water waves, M. S.
Thesis, University of Florida, Gainesville, FL.
Ferry, J. D., 1970. Viscoelastic Properties of Polymers, Wiley, New York.
Heywood, N. I., 1991. Rheological characterisation of nonsettling slurries, In: Slurry Handling
Design of SolidLiquid Systems, N. P. Brown and N. I. Heywood (eds.), Elsevier, Amsterdam, 5387.
Huilgol, R. R., 1975. Continuum Mechanics of Viscoelastic Liquids, Wiley, New York.
Jones, J. R., and Walters, T. S., 1967. Flow of elasticoviscous liquids in channels under the
influence of a periodic pressure gradient, Rheologica Acta, 6, Part 1, 240245; Part 2, 330338.
Marine Board, 1985. Dredging Coastal Ports: An Assessment of the Issues, National Research
Council, Washington, DC.
Mehta, A. J., 1996. Interaction between fluid mud and water waves, In: Environmental Hydraulics,
V. P. Singh and W. H. Hager W H (eds.), Kluwer, Dordrecht, The Netherlands, 153187.
Oldroyd, J. G., 1950. On the formulation of theological equations of state, Proceedings of the Royal
Society (London), A200, 523541.
Oldroyd, J. G., 1958. NonNewtonian effects in steady motion of some idealized elastoviscous
liquids, Proceedings of the Royal Society (London), A245, 278297.
Parchure, T. M., and Sturdivant, C. N., 1997. Development of a portable innovative contaminated
sediment dredge. Final Report CPARCHL972, Construction Productivity Research Program, U.
S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Parker, W. R., and Kirby, W. R., 1982. Time dependent properties of cohesive sedment relevant to
sedimentation management European experience, In: Estuarine Comparisons, V. S. Kennedy (ed.),
Academic Press, New York, 573589.
Ryan, N. W., and Johnson, M. M., 1959. Transition from laminar to turbulent flow in pipes,
American Institute of Chemical Engineers Journal, 5,433435.
Sisko, A. W., 1958. The flow of lubricating greases, Industrial Engineering Chemistry, 50, 1789
1792.
Wasp, E. J., Kenny, J. P., and Gandhi, R. L., 1977. SolidLiquid Flow Slurry pipeline
Transportation. Trans Tech Publications, San Francisco.
Waters, N. D., and King, M. J., 1970. Unsteady flow of a elasticviscous liquid, Rheologica Acta,
9, 345355.
Waters, N. D., and King, M. J., 1971. The unsteady flow of an elasticoviscous liquid in a straight
pipe of circular cross section, Journal of Physics D : Applied Physics., 4, 204211.
Williams, M. L., and Ferry, J. D., 1953. Second approximation calculations of mechanical and
electrical relaxation and retardation distributions, Journal of Polymer Science, 11, 169175.
APPENDIX A: SLURRY VISCOSITY DATA
2
10
10.
0
a
i
10
0
I
a
x
W n2
10 L3
Kaolinite, density = 1250 kg/m^3
2 0 2
10 100 10
Shear rate (Hz)
Fig. Ai Viscosity data for slurry no. 1.
S Kaolinite, density = 1350 kg/m*3
10
2 Mu (inf)= 0.2813 Pa.s
10 c = 9.75
10,
0
0
Col
a
10
x
0 10
2
10" 10 10
Shear rate (Hz)
Fig. A3 Viscosity data for slurry no. 3.
102
.10
0
100
ao
w
x10"
u I
10.2
Kaolinite, density =1300 kg/m^3
Mu (inf) = 0.0024 Pa.s
c = 3.79
Sn = 0.520
102 10 10
Shear rate (Hz)
Fig. A2 Viscosity data for slurry no. 2.
2
10
0
10
a
x
u
Kaolinite, density = 1400 kg/m^3
S Mu (inf)= 0.0011 Pa.s
c=6.81
S n = 0.610
10
2 0 2
10 10 10
Shear rate (Hz)
Fig. A4 Viscosity data for slurry no. 4.
Mu (inf)= 0.1904 Pa.s
c=1.99
n = 0.308
i I i I
I i I
102
.101
0
01
x10
'U
x10
Kaolinite, density = 1450 kg/m^3
2 0 2
10 100 102
Shear rate (Hz)
Fig. A5 Viscosity data for slurry no. 5.
Kaolinite, density= 1500 kglm^3
, I U
0
10
W01
o
l0
o
0
iW 4,1
1021L
10" 10 102
Shear rate (Hz)
Fig. A6 Viscosity data for slurry no. 6.
A2
Mu (inf) = 0.4633 Pa.s
c=8.19
S 0.448
xKV
Mu (inf)= 0.5837 Pa.s
c= 14.46
n = 0.228
* \)...
I i I
I i I i
102
10
.101
S100
010
oIl
x10
I
102I
75% K + 25% A, density = 1243 kg/m^3
2 0 i
2 0 2
10 10 10
Shear rate (Hz)
Fig. A7 Viscosity data for slurry no. 8.
10
102
W100
to
UI
0 2
10
X10
w
75% K + 25% A, density = 1339 kg/m^3
S 2 0 2
10 10 10 10
Shear rate (Hz)
Fig. A9 Viscosity data for slurry no. 10.
2
10
10
0
to
U10
x
WIii
10.2
102
12
10
0 2
10 10 1
Shear rate (Hz)
Fig. A8 Viscosity data for slurry no. 9.
75% K + 25% A, density = 1387 kg/m^3
102
'10
Q.
102
100
100
w
UJ10
104 ,
2 0 2 4
10 10 10 10
Shear rate (Hz)
Fig. A10 Viscosity data for slurry no. 11.
A3
Mu (inf)= 0.0015 Pa.s
c = 2.43
n = 0.581
Mu (inf) = 0.4983 Pa.s
c = 8.02
n = 0.296
Mu (inf)= 1.688 Pa.s
c=9.98
n = 0.003
Mu (inf)= 0.7513 Pa.s
c = 41.31
n = 0.083
75% K + 25% A, density = 1291 kg/mA3
50% K + 50% A, density = 1236 kg/m^3
10 11
102 100 10 104
Shear rate (Hz)
Fig. A 1 Viscosity data for slurry no. 12.
50% K + 50% A, density = 1306 kg/m^3
10 1 1
10 10 102 104
Shear rate (Hz)
Fig. A13 Viscosity data for slurry no. 14.
50% K + 50% A, density = 1283 kg/m^3
2
.10
0.
o10
2
,10
X
WJl
I0 L
2 0 2 4
10 10 10 10
Shear rate (Hz)
Fig. A12 Viscosity data for slurry no. 13.
4
10
2
.102
0
0
S10
UJ
50% K + 50% A, density = 1329 kg/m^3
10 I
2 0 2 4
10 10 10 10
Shear rate (Hz)
Fig. A14 Viscosity data for slurry no. 15.
A4
10
m
o
'10
o
0.
010
xa
X
Mu (inf) = 0.2723 Pa.s
c= 13.21
n = 0.022
.X
10
610
0.
810
0
oX1
N,
n,
Nu
Mu (inf) = 0.2965 Pa.s
c = 21,67
n = 0.116
I (
25% K + 75% A, density = 1175 kg/m^3
10
2 0 2 4
102 10 102 104
Shear rate (Hz)
Fig. A15 Viscosity data for slurry no. 16.
25% K + 75% A, density = 1225 kg/m^3
10
102
Z.
0
010
'9
(I
S10.2
w
1n n4
2 0 2 4
102 100 102 104
Shear rate (Hz)
Fig. A17 Viscosity data for slurry no. 18.
25% K + 75% A, density = 1200 kg/m^3
10
2 Mu (inf)= 0.3585 Pa.s
10 c = 9.61
10 X
0
0
'0 X
2
10
10 10 10 10
Shear rate (Hz)
Fig. A16 Viscosity data for slurry no. 17.
25% K + 75% A, density = 1250 kg/m^3
2 0 2 4
102 100 102 104
Shear rate (Hz)
Fig. A18 Viscosity data for slurry no. 19.
A5
Mu (inf)= 0.5632 Pa.s
c = 22.07
n = 0.124
i I i I
2 Attapulgite, density= 1125 kg/mA3
10 
101
aO.
(0
I100
w.
310
102
2
10 10 10 10
Shear rate (Hz)
Fig. A19 Viscosity data for slurry no. 20.
10
.10
0.
10
0 2
x10
8 .2~
10I
Attapulgite, density = 1175 kg/m^3
10.2 100 102 10
Shear rate (Hz)
Fig. A21 Viscosity data for slurry no. 22.
Attapulgite, density = 1150 kg/m^3
10 10 102 104
Shear rate (Hz)
Fig. A20 Viscosity data for slurry no. 21.
10
2
V
'10
a.
I,
0 0
0.10
lO
x 10U
Attapulgite, density = 1200 kg/m^3
i41
1041 ..I
102 100 102 10
Shear rate (Hz)
Fig. A22 Viscosity data for slurry no. 23.
A6
Mu (inf)= 0.0019 Pa.s
c=1.27
n = 0.559
Mu (inf)= 0.3628 Pa.s
c= 13.55
n = 0.148
Mu (inf) = 0.5604 Pa.s
c = 21.38
n = 0.236
f i I i
I I I
10
2
?102
a.
0 0
0 10
0 
x100
w
rl
0 o
90% K + 10% B, density = 1200 kg/m^3
102 10o 10 104
Shear rate (Hz)
Fig. A23 Viscosity data for slurry no. 24.
90% K + 10% B, density = 1300 kg/mA3
10
90% K + 10% B, density = 1250 kg/m^3
10 1
2 0 2 4
10 10 10 10
Shear rate (Hz)
Fig. A24 Viscosity data for slurry no. 25.
90% K + 10% B, density = 1350 kg/m^3
10 1 I
2
010
o
0
10
x
LU
uJ
10 I
102 100 10 104
Shear rate (Hz)
Fig. A25 Viscosity data for slurry no. 26.
102
0 2 4
10 10 102 10
Shear rate (Hz)
Fig. A26 Viscosity data for slurry no. 27.
Mu (inf)= 0.1899 Pa.s
c = 2.97
n = 0.152
Mu (inf)= 0.2432 Pa.s
c = 47.92
n= 0.149
S I I
I I
S65% K + 25% A + 10% B, density = 1225 kg/m^3
10
SMu (inf)= 0.1544 Pa.s
"102 c=9.46
n = 0.318
lo
(0
0
10
2
10
2 0 2 4
10 10 10 10
Shear rate (Hz)
Fig. A27 Viscosity data for slurry no. 28.
65% K + 25% A +10% B, density = 1275 kg/m^3
10
2 Mu (inf)= 0.2179 Pa.s
I10 c = 23.12
0. n = 0.216
., m1 0
*10
S101
X X
10
2 0
w 10'
10 10 10 104
Shear rate (Hz)
Fig. A29 Viscosity data for slurry no. 30.
S65% K + 25% A + 10% B, density = 1250 kg/m^3
n = 0.239
_10
2
10
ti
2 0
x
102 100 102 104
Shear rate (Hz)
Fig. A28 Viscosity data for slurry no. 29.
S65% K + 25% A + 10% B, density = 1300 kg/m^3
10
Mu (inf)= 0.4926 Pa.s
c = 66.7
a 102 n = 0.188
10
w 2
10
0 2 0 2 4
10 10 10 10
Shear rate (Hz)
Fig. A30 Viscosity data for slurry no. 31.
A8
40% K + 50% A + 10% B, density = 1175 kg/mA3
10
2 Mu (inf)= 0.097 Pa.s
X10 c = 8.05
e n = 0.274
0
10
2
10
102 10 102 104
Shear rate (Hz)
Fig. A31 Viscosity data for slurry no. 32.
40% K + 50% A + 10% B, density = 1225 kg/m^3
10
,2 ~ Mu (inf) = 0.0719 Pa.s
102 c = 11.37
n = 0.280
10
>0
w 1
10,
2
102
o22
102 10 102 104
Shear rate (Hz)
Fig. A33 Viscosity data for slurry no. 34.
40% K + 50% A +10% B, density = 1200 kg/m^3
10
2 Mu (inf)= 0.0013 Pa.s
10c = 12.71
n = 0.297
101
10
0
2
102
102 10 102 10
Shear rate (Hz)
Fig. A32 Viscosity data for slurry no. 33.
10
u
2
o
10
x
LU
40% K + 50% A + 10% B, density = 1250 kg/m^3
Mu (inf) = 0.1473 Pa.s
c = 40.29
n = 0.174
2
1021
102 10 102 10
Shear rate (Hz)
Fig. A34 Viscosity data for slurry no. 35.
A9
S15% K + 75% A + 10% B, density = 1125 kg/m^3
10 I 1
i101
0
Q.
100
100
>
15% K + 75% A + 10% B, density = 1150 kg/m^3
Sn _________________
IU I
610
(I
0
010
0
X10
1021
10 10 10 10
Shear rate (Hz)
Fig. A35 Viscosity data for slurry no. 36.
15% K + 75% A + 10% B, density = 1175 kg/m^3
10
,1 2 x Mu (inf) = 0.0912 Pa.s
10 m c = 9.23
n = 0.251
0
0
2
10 2
2 0 2
1 2 0 24
10.2 100 102 104
Shear rate (Hz)
Fig. A37 Viscosity data for slurry no. 38.
12
2 0 2 4
10 10 10 10
Shear rate (Hz)
Fig. A36 Viscosity data for slurry no. 37.
15% K + 75% A + 10% B, density = 1200 kg/m^3
10
2 Mu (inf) = 0.299 Pa.s
'10 c c=23.96
Sn = 0.160
=10X
2
0
0 2 0 2
> U
210
102 10 10 10
Shear rate (Hz)
Fig. A38 Viscosity data for slurry no. 39.
A10
Mu (inf)= 0.0012 Pa.s
c = 2.31
m n = 0.538
SX
Mu (inf)= 0.0566 Pa.s
c = 6.55
n = 0.398
I
90% A +10% B, density = 1125 kg/m^3
10
Mu (inf)= 0.1787 Pa.s
10 c = 4.120
S.n n=0.318
0X
e
0
= =
2
10
>
S10
10 10 102 10
Shear rate (Hz)
Fig. A39 Viscosity data for slurry no. 40.
90% A + 10% B, density = 1175 kg/m^3
210
0o
a
r0
010
'o
Q)
x10
u'
10
2 0 2 4
10 10 10 104
Shear rate (Hz)
Fig. A41 Viscosity data for slurry no. 42.
90% A + 10% B, density = 1150 kg/m^3
10
Mu (inf)= 0.095 Pa.s
c 2 c= 15.24
(L n= 0.100
00
o 2
110
4
10
2 0 2 4
10 10 10 10
Shear rate (Hz)
Fig. A40 Viscosity data for slurry no. 41.
90% A + 10% B, density = 1200 kg/m^3
lunl
1 0 2 1 1 2
10 10 10 10
Shear rate (Hz)
Fig. A42 Viscosity data for slurry no. 43.
A11
10
0
o
10
0
x
UJ
Mu (inf)= 0.357 Pa.s
c= 17.95
n = 0.095
Mu (inf)= 0.5263 Pa.s
c=104
n = 0.048
I
IU I
APPENDIX B: DEPENDENCE OF POWERLAW PARAMETERS ON DENSITY
Kaolinite
20
15
o 10
5
0
1200 1300 1400 1500 1600
Density (kg/m^3)
Kaolinite
0.8
0.6
c 0.4
0.2
0
1200 1300
Density (kg/m^3)
Fig. B1 Variations of powerlaw coefficients with density for
kaolinite slurries. Top: [lt; middle: c; bottom: n.
1400
1500
Kaolinite
0.8
o 0.6
0.4
0.2
0 
1200 1300 1400 1500 1600
Density (kg/m^3)
1600
75%K +25%A
0.8
0.6
0.4
r
0.2
Density (kg/m^3)
Fig. B2 Variations of powerlaw coefficients with density for
75% kaolinite + 25% attapulgite slurries. Top: pL; middle: c;
bottom: n.
B2
75%K + 25%A
2
1.5
S* 1
0.5
0
0.52 4 0Density (kg m3)0
Density (kg/m^3)
75%K + 25%A
50
40
30
20
10
0^ ^I
1200 1250 1300 1350 1400
Density (kg/m^3)
Fig. B3 Variations of powerlaw coefficients with density for
50% kaolinite + 50% attapulgite slurries. Top: Vp; middle: c;
bottom: n.
50%K + 50%A
0.4
" 0.3 *
S0.2
02 .1
0
1200 1250 1300 1350
Density (kg/m^3)
50%K + 50%A
100
50
so
0
12O0 1250 1300 150
50
Density (kg/m^3)
25%K + 75%A
40
30
o 20
10
0 1
1160 1180 1200 1220 1240 1260
Density (kg/m^3)
25%K + 75%A
0.4
0.2
0
0 1180 1200 122T 1260
0.4
Density (kg/m^3)
Fig. B4 Variations of powerlaw coefficients with density for
25% kaolinite + 75% attapulgite slurries. Top: p.; middle: c;
bottom: n.
25%K + 75%A
0.8
S0.6
0.4
S 0.2
0 
1160 1180 1200 1220 1240 1260
Density (kg/m^3)
Attapulgite
0.6 
0.4
0.2
0.21 140 60 1 1 20
0.4
Density (kg/mA3)
Fig. B5 Variations of powerlaw coefficients with density for
attapulgite slurries. Top: pt; middle: c; bottom: n.
Attapulgite
0.6
S0.4 
 0.2
0
1120 1140 1160 1180 1200 1220
Density (kg/m^3)
Attapulgite
25
20
15
10
5
0
1120 1140 1160 1180 1200 1220
Density (kg/m^3)
Fig. B6 Variations of powerlaw coefficients with density for
90% kaolinite + 10% bentonite slurries. Top: Vp; middle: c;
bottom: n.
B6
90%K + 10%B
0.3
. 0.2
i 0.1
0
1150 1200 1250 1300 1350 1400
Density (kg/m^3)
90%K + 10%B
60
40
20
0
1150 1200 1250 1300 1350 1400
Density (kg/m^3)
Fig. B7 Variations of powerlaw coefficients with density for
65% kaolinite + 25% attapulgite + 10% bentonite slurries.
Top: p ,; middle: c; bottom: n.
65%K+25%A+10%B
0.6
 0.4
I 0.2 *
0
1220 1240 1260 1280 1300 1320
Density (kg/m^3)
65%K+25%A+10%B
80
60
o 40
20
0 
1220 1240 1260 1280 1300 1320
Density (kg/m^3)
40%K+50%A+10%B
50
40
30
20
10
0
1160 1180 1200 1220 1240 1260
Density (kg/m^3)
40%K+50%A+10%B
0.4
0.3
c 0.2
0.1
0
1160 1180 1200 1220
Density (kg/m^3)
1240 1260
Fig. B8 Variations of powerlaw coefficients with density for
40% kaolinite + 50% attapulgite + 10% bentonite slurries.
Top: g_; middle: c; bottom: n.
B8
40%K+50%A+10%B
0.2
C 0.15 
. 0.1
0.05
0 +
0 , ,,
1160 1180 1200 1220 1240 1260
Density (kg/m^3)
15%K+75%A+10%B
30
20
10
0
1120
1140 1160 1180
Density (kg/m^3)
Fig. B9 Variations of powerlaw coefficients with density for
15% kaolinite + 75% attapulgite + 10% bentonite slurries.
Top: p_; middle: c; bottom: n.
B9
1200 1990
90%A + 10%B
0.4
0.3
c 0.2
0.1 
0
1120 1140 1160 1180 1200 1220
Density (kg/m^3)
Fig. B10 Variations of powerlaw coefficients with density for
90% attapulgite + 10% kaolinite slurries. Top: p,; middle: c;
bottom: n.
B10
90%A + 10%B
0.6
 0.4
0.2 
0
1120 1140 1160 1180 1200 1220
Density (kg/m^3)
