|
O
^ \
I
EFFECT OF BED SHEAR STRESS ON THE EROSIONAL
CHARACTERISTICS OF KAOLINITE
By
T.M. PARCHURE
A THESIS PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
1980
ACKNOWLEDGEMENTS
My deepest gratitude goes to Dr. Ashish J. Mehta, Associate Professor,
Department of Coastal and Oceanographic Engineering, who has been my
advisor and chairman of the committee, for his valuable guidance,
encouragement, and support during the course of studies. I thank
Dr. J. L. Eades, Professor, Department of Geology, and Dr. B. A. Benedict,
Professor, Department of Civil Engineering, for serving on my supervisory
committee. I have to thank Dr. E. Partheniades, Professor, Department of
Engineering Sciences, for his advice regarding measurement of bed density.
I am grateful to the personnel at the Coastal Engineering Laboratory,
Mr. George Jones in particular, for their excellent cooperation and for
extending every possible help for a successful completion of the experi-
mental work. I am very thankful to Adele Koehler for careful typing of
the manuscript and to Lillean Pieter for the excellent drafting work.
I would like to express gratitude to my loving wife Aparna and my
mother Indira for their continued encouragement in my endeavor and for
accepting to bear the hardships caused during my long stay away from them.
The present study was conducted as a part of the research project
entitled "Deposition of Fine Sediments in Turbulent Flows" supported by
the National Science Foundation under Grant Number GK-31259. Support
was also received partially from the Environmental Protection Agency
under Grant Number R806684010. This support from both the agencies is
sincerely acknowledged.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS
LIST OF TABLES
LIST OF FIGURES
NOMENCLATURE .
ABSTRACT .
CHAPTER
INTRODUCTION . . . . .
FINE SEDIMENTS . . . . .
2.1 General Description . . . .
2.2 Properties of Fine Sediments . . .
2.3 Parameters Influencing the Properties of Cohesive
Sediments . . . . .
2.4 Processes for Deposited Beds . . .
2.5 Clay-Water System . . . .
PREVIOUS LABORATORY STUDIES . . .
3.1 General Review . . .
3.2 Review of Literature on Erosion . ...
3.3 Review of Literature Pertinent to the Present
Study . . . . .
3.4 Shear Strength of Clay . . .
3.5 Shear Strength and Bed Density of Clay . .
PRESENT INVESTIGATION . . . .
4.1 Objective . . . . .
4.2 Material . . . . .
4.3 Apparatus for Erosion Tests . . .
4.4 Experimental Procedure . . . .
4.5
V TEST
5.1
5.2
5.3
5.4
5.5
Apparatus for Measurements of Bed Density .
RESULTS AND ANALYSIS . . .
Effects of Parameters in Steps I, II and III
Multiple Steps of T .
Discretized Sinusoidal Velocity Variation
Correlation Plots of C30 and C60 Values .
Analysis of Data . . . .
I
II
III
IV
74
Page
ii
v
vi
xi
xiii
1
5
5
7
18
26
31
37
37
39
45
54
62
68
68
71
CHAPTER Page
VI DISCUSSION AND PROPOSED TEST PROCEDURE ........ 132
6.1 Bed Density and Other Soil Parameters Correlated
to the Shear Strength . . .... 132
6.2 Critical Shear Stress . . .... 139
6.3 Proposed Test Procedure . . .... 143
6.4 Illustrative Example . . . .... 151
VII SUMMARY AND CONCLUSIONS . . . .... 164
7.1 Summary of Literature Review ........... 164
7.2 Conclusions of the Present Study . ... 165
REFERENCES . . . . ... ......... 169
BIOGRAPHICAL SKETCH . . . . ... ..... 174
LIST OF TABLES
Specific surface area and liquid limit for typical
clays
Summary of selected studies on cohesive soil erosion
Experimental conditions for tests conducted with a
multiple steps of Te
V
Table
1
2
3
Page
LIST OF FIGURES
Figure Page
1 Inter-particle forces on clay minerals and the clay
micelle 14
2 Rheological models 33
3 Concentration versus time plot obtained in erosion of
placed bed: Partheniades (1962), Expt. Series-I 48
4 Concentration versus time plot obtained in erosion of
remolded bed: Partheniades (1962), Expt. Series-II 49
5 Concentration versus time plot obtained in erosion of
deposited bed: Partheniades (1962), Expt. Series-III 50
6 Relationship between rate of erosion and average bed
shear stress: Partheniades (1962), 50
7 Concentration versus time plot: Krone (1962) 52
8 Concentration as a function of bed shear stress:
Krone (1962) 52
9 Results of a 500 hour long erosion test: Krone (1962) 53
10 Concentration versus time plots for Series I obtained
by Lee (1979) 55
11 Concentration versus time plots for Series II obtained
by Lee (1979) 55
12 Lee's (1979) data for Series I re-plotted to indicate
variation of suspension concentration as a function of
time and bed shear stress 56
13 Lee's (1979) data for Series II re-plotted to indicate
variation of suspension concentration as a function of
time and bed shear stress 57
14 Concentration versus time data obtained by Yeh (1979)
for erosion of Kaolinite 58
15 Concentration as a function of bed shear stress obtained
by Yeh 59
Figure
16 Schematic diagrams showing shear strength of cohesive
soil related to other parameters
17 Relationship between shear strength and bed density
observed by Owen (1970)
18 Bed density profiles: Owen (1970)
19 Definition sketch for notations used to describe experi-
mental conditions
20 Size gradation of Kaolinite used for the experiments
21 Photograph:
22 Photograph:
the ring
23 Photograph:
24 Photograph:
the ring
25 Photograph:
26 Photograph:
27 Photograph:
of sediment
28 Photograph:
29 Operational
and channel
30 Correlation
channel and
The rotating channel facility
Close view of the annular channel and
The motor controllers
The electric motors for the channel and
Millipore filter apparatus assembly
Device for measurement of bed density
Equipment for determining the concentration
suspensions
Sampling bottles
speeds and controller meter readings for ring
at different bed shear stresses
between r.p.m. and meter reading for the
the ring
31 Schematic drawings of apparatus developed for measurement
of bed density
32 Effect of parameters in Step I on suspension concentration
33 Suspension concentration versus time for Expt. 3
34 Suspension concentration versus time for Expt. 4
35 Suspension concentration versus time for Expt. 5
36 Effect of Step II parameters on suspension concentration
Page
Figure
37 Effect of Step III parameters on suspension concentration
38 Representation of a linearly varying bed shear stress by
two different discretized time step functions
39 Suspension concentration versus time for Expt. 7
Suspension concentration versus time for
Suspension concentration versus time for
Suspension concentration versus time for
Suspension concentration versus time for
Suspension concentration versus time for
Suspension concentration versus time for
Suspension concentration versus time for
Time-step function for bed shear stress
Suspension concentration versus time for
Suspension concentration versus time for
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Page
93
97
98
99
100
101
102
103
104
105
106
107
108
50 Variation
stress as
51 Variation
stress as
52 Variation
stress as
53 Variation
stress as
54 Variation
stress as
55 Variation
stress as
56 Variation
stress as
57 Variation
stress as
of suspension
a function of
of suspension
a function of
of suspension
a function of
of suspension
a function of
of suspension
a function of
of suspension
a function of
of suspension
a function of
of suspension
a function of
concentration with bed shear
time for Expt. 8
concentration with
time for Expt. 9
concentration with
time for Expt. 10
concentration with
time for Expt. 11
concentration with
time for Expt. 12
concentration with
time for Expt. 13
concentration with
time for Expt. 14
concentration with
time for Expt. 15
bed shear
bed shear
bed shear
bed shear
bed shear
bed shear
bed shear
58 Effect of shear stress variation on suspension
concentration
viii
110
Figure Page
59 Comparison of suspension concentration obtained under
two different discretized time step functions 119
60 Variation of suspension concentration with bed shear
stress, all data for Td = 24 hours 120
61 Variation of suspension concentration with bed shear
stress, a-i data for Td = 40 hours 121
62 Suspension concentration versus bed shear stress
(Expt. 7) 124
63 Variation of suspension concentration as a function of
bed shear stress (Expt. 9) 125
64 Variation of suspension concentration as a function of
bed shear stress (Expt. 11) 126
65 Variation of suspension concentration with bed shear
stress for two different discretized time step functions 127
66 Cr versus Tr at 10 minutes 128
67 Cr versus tr at 20 minutes 129
68 Comparison of Cr versus T for two different bed density
structures
69 (AC)ex plotted against the corresponding values of (AT)ex 131
70 Example of erosion test result and critical shear stress
for erosion as a function of dry density of mud surface
given by Thron and Parsons (1980) 135
71 Parameters influencing bed density and plasticity of soil 138
72 A typical test result reported by Espey (1963) 140
73 Notations for critical shear stress 142
74 Erosion rate versus time for different values of bed
shear stress 144
75 Definition sketch for various parameters 145
76 Explanatory sketch for EQ type profile of c-t curve 147
77 Explanatory sketch for ER type profile of c-t curve 148
78 Suspension concentration during deposition under the
bed shear stress of 0.05 N/m 153
Figure Paoe
79 Suspension concentration during deposition under the
bed shear stress of 0.015 N/m2 154
80 Suspension concentration during deposition under zero
bed shear stress 155
81 Variation of bed density with depth for three different
conditions of flow deposited beds 156
82 Suspension concentration versus time for Expt. 17 157
83 Suspension concentration versus time for Expt. 18 158
84 Suspension concentration versus time for Expt. 19 159
85 Variation of suspension concentration with bed shear for
different flow deposited beds 160
86 Suspension concentration versus bed shear stress for
different flow deposited beds 161
87 Variation of erosion rate as a function of bed shear
stress for different flow deposited beds 162
88 Shear strength of bed as a function of depth 163
NOMENCLATURE
C = Ratio of the consecutive suspension concentrations, e.g.
C2 C
S2 etc.
C C2
AC = Excess (suspension) concentration, e.g. C2 C1.
C2 C1
(AC)ex = Normalized excess concentration, e.g. C
C1
Co = Suspension concentration at the end of initial mixing.
C30 = Suspension concentration at the end of 30 minutes after change
of bed shear stress.
p = Density of bed.
Tm = Bed shear stress for initial mixing.
Tm = Duration of initial mixing.
Td = Bed shear stress for deposition.
Td = Duration of deposition plus consolidation.
Te = Bed shear stress for erosion (varied as a function of time).
Ts = Time step for Te, i.e. duration over which different magni-
tudes of Te prevailed.
r = Ratio of the consecutive values of bed shear stress, e.g.
r T T
e2 e3
etc.
Te Te
el e2
AT = Excess shear stress, e.g. T Te
22 -
(AT)ex = Normalized excess shear stress, e.g.
e1
Abstract of Thesis Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Master of Science
EFFECT OF BED SHEAR STRESS ON THE EROSIONAL
CHARACTERISTICS OF KAOLINITE
By
T.M. Parchure
December 1980
Chairman: Dr. Ashish J. Mehta
Major Department: Coastal and Oceanographic Engineering
The degree of resistance to the erosion of a cohesive sediment bed
under an applied shear stress is controlled by the physico-chemical
properties of the sediment and the fluid, as well as by the depth-
variation of the bed shear strength characterized by the inter-particle
bonds. Previous attempts to correlate the erodibility of cohesive
sediment beds with the shear strength determined by such devices as a
penetrometer or a viscometer, or with soil indices, have not been suc-
cessful. The objective of this study was to evolve a test procedure
for conducting laboratory experiments to obtain the "layer by layer"
shear strength of deposited beds of various density structures.
Experiments were conducted using kaolinite with salt water of 35 ppt
concentration as the eroding fluid. All the tests were carried out in
an annular rotating channel apparatus. New techniques developed for
measurement of the density structure of deposited beds have been de-
scribed. An illustrative example outlining the procedure for the
determination of the depth-variation of the bed shear strength is
given.
Chairman
xiii
CHAPTER I
INTRODUCTION
Study of the properties of transportation and deposition of sedi-
ments has engaged the attention of several research workers. The
motive behind these studies has primarily been to assist in the design
of dams and irrigation canals, water treatment and sewage disposal
works, navigation channels, docks and harbors, etc. Special phenomena
such as mud-banks along the shorelines, density flows at docks, floc-
culation and deposition of sediments due to mixing of salt water and
fresh water in estuaries are associated with the fine sediments. Among
various ways of classifying the sediments based on their geological
origin, chemical properties, physical properties, etc., one of the
classifications has been to treat the fine sediments of various composi-
tions as a class in itself due to their special properties which differ
considerably from the other sediments, and hence they need to be studied
separately.
The aspects related to fine sediments in some of the engineering
projects are illustrated by case studies. In the Mersey Estuary,
England, because of the abundance of fine cohesive sediment, attempts
to increase the depths in one channel by removing 2.3 million m3 of
material each year failed utterly. The channel in fact became shallower
after five years of intensive dredging than it was before. Another
example is Savannah harbor in the U.S.A., where in spite of continuous
dredging over thirty years, the siltation rate has almost doubled and
displaced the major deposition zone 20 km upstream along the river to
an inconvenient location. Construction of a dam across a river trans-
porting fine sediments in suspension results in the deposition of
sediments immediately upstream of the structure and may, as in the case
of the Aswan Dam, Egypt, deprive farmers downstream of their annual
replenishment of fertile sediment during the flood season. Density
currents created by the presence of fine sediments in sea water cause
excessive siltation of navigation basins as experienced at the Tilbury
tidal basin on the Thames River, England.
Cohesive sediments also form an important consideration in the
nuclear power generation planning and disposal of radioactive waste.
Research has shown that the transuranic elements such as plutonium and
americium are reconcentrated strongly by marine sediments. Their
presence in sediment has focused attention on routes by which contaminated
sediment might present a route of exposure to man especially in the
longer term in the form of airborne dust in a respirable form or uptake
through crops grown in soils reclaimed from areas of contaminated
sediments.
In the more recent years the study of fine sediments has assumed
an important place in the context of pollution control. The transport
and ultimate fate of contaminants is a complex process involving
physical, chemical, and biological aspects, all of which play an
important role. Sediments in a way are pollutants themselves since
they increase turbidity. Their deposition may pose serious engineering
and environmental problems. However, the more important aspect is their
property to adsorb other pollutants very effectively and transport them
along. The content of heavy metals adsorbed to the sediments is found
to depend on the grain size of the sediment. The smaller the grain
size, the higher is the contamination with heavy metals. The sedi-
ments less than 16 microns in size are likely to have very high con-
taminants. The bulk of the pollutants may be carried on the sediments
rather than the water. Hence, the dispersal of pollutants cannot be
dissociated from the dispersal of sediments. Study of dispersal of
sediments alone, however, cannot be expected to provide the required
information related to pollutants since the chemical, biological, and
physical aspects involving oxidation, reduction, exchange of elements,
decay, etc. considerably change the properties of pollutants.
An understanding of the depositional and resuspension character-
istics of the fine sediments would therefore be beneficial in obtaining
better solutions to engineering problems and in exercising a more
effective pollution control. Over the last about twenty years, studies
have been carried out to investigate the erosional and depositional
properties of fine sediments. The influence of different parameters
connected with the sediment and the eroding fluid on the character-
istics of erosion and deposition has also been studied. These have
been described in Chapter III. The experimental work reported in
this thesis was carried out to study the resuspension of flow deposited
bed of Kaolinite under varying shear stress and to study the effect of
bed structure on the process of resuspension. Saline water with 35
parts per thousand concentration was used as the eroding fluid.
An understanding of the physical processes associated with the
movement of fine, cohesive sediments is clearly essential for obtaining
improved engineering solutions to estuarine problems. The phenomena of
-4-
fine sediment transport, deposition, bed formation and consolidation,
and bed resuspension are rather complex, and in the estuarine environ-
ment they are inter-linked in a cyclical manner within time-scales
imposed typically by the astronomical tides. Investigations of these
phenomena under laboratory scales is an important first step towards
an elucidation of the mechanics of the transport processes in the
prototype, since it is possible to isolate and control the important
governing parameters in laboratory tests. In that context, this in-
vestigation is concerned with studying the characteristics of resus-
pension of flocculated cohesive sediment beds. Under an applied bed
shear stress, the surficial erodibility of such a bed is contingent
upon the structure of the bed, as defined by the inter-particle bond
strength of the floc network. Inasmuch as this network is formed
under a given set of conditions specified by floc deposition and con-
solidation of the settling suspension at the bed, the magnitude and
the duration of the applied shear stress during bed formation are
important governing parameters for the subsequent process of resus-
pension. Hence the investigation of the erosion of beds formed under
a variable shear stress is emphasized in this study. The overall
objective of this study was an attempt to establish a laboratory test
procedure for specifying the "layer by layer" erodibility of a deposited
bed in terms of parameters) involving the critical shear stress for
the erosion of a particular layer. The observed variation of the
erodibility of a given bed with depth has been discussed with reference
to the depth-variation of the floc shear strength and the bulk
density.
CHAPTER II
FINE SEDIMENTS
2.1 General Description
Fine sediments, commonly called clays or muds,are a product of
weathering or hydro-thermal action on the rock and other soil on
earth's surface. Although the maximum size of particles in the clay
grade is defined somewhat differently in different disciplines, the
general tendency has been to classify sediments finer than two microns
as clays. Mixtures of clays and silt are usually called muds. The
classification of fine-grained soil as either a silt or a clay is not
merely on the basis of particle size but rather on the plasticity or
non-plasticity of the material. Clay soil is plastic over a range of
water content; that is, the soil can be remolded or deformed without
causing cracking, breaking, or change of volume, and will retain the
remolded shape. The clays are frequently cohesive. When dried, a clay
soil possesses very high resistance to crushing. A silt soil possesses
little or no plasticity and when dried has little or no strength.
The basic differences between the elementary particles of non-
interacting coarse minerals and the interacting fine minerals such as
Kaoline could be briefly described as follows:
1. The non-interacting particles have no electric charge. Hence, they
interact only hydrodynamically without any inter-particle attraction.
-5-
The fine sediments are interacting particles which attract or repel
each other due to the presence of electric charge. Hence, the non-
interacting particles remain separate from each other, whereas the
interacting particles can form flocs under suitable environment.
2. If a small sample of moist silt is shaken easily but rapidly, water
will appear on the surface but disappear when shaking stops. This
phenomenon is referred to as dilatancy. The non-interacting par-
ticles show dilatancy with high concentration of sediment, whereas
a sample of moist clay when shaken similarly does not show wetting
of the surface.
3. The interacting particles exhibit elastic or plastic properties.
4. The non-interacting particles are usually unsaturated, whereas the
flocs of interacting particles are saturated with water molecules.
5. With a low concentration of sediment, the suspension of non-
interacting sediment is close to a Newtonian fluid, where the
deformation is linearly proportional to shear stress. The sus-
pension of interacting particles is non-Newtonian in behavior.
6. The erosional, transport, and depositional characteristics of
non-interacting sediments are based mainly on the physical proper-
ties such as size, specific gravity, compaction, etc. The proper-
ties of the fluid such as salinity, pH, temperature have no sub-
stantial effect. On the other hand, the fluid properties have a
substantial effect on the formation of flocs and hence on the
erosional and depositional properties of interacting particles.
7. Fine sediments have a relatively much higher compressibility than
the coarser non-interacting sediments.
-7-
8. Surface forces are dominant in respect to fine sediments, whereas
gravitational forces predominate in the case of non-interacting
particles. Due to the tendency of fine sediments to attach to each
other due to surface forces, fine sediments are also sometimes
referred to as cohesive sediments and the other sediments as non-
cohesive sediments.
9. Fine sediments are transported in the form of the suspended load or
wash load, whereas the coarser sediments are predominantly trans-
planted as bed load.
2.2 Properties of Fine Sediments
Several parameters affect the properties of clay materials, par-
ticularly the following:
1. Clay mineral composition.
2. Non-clay mineral composition.
3. Organic matter.
4. Exchangeable ions and soluble salts.
5. Texture, i.e. the particle size distribution of the constituent
particles, the shape of the particles, their orientation in space
relative to each other, and the forces tending to bind the par-
ticles together.
In the context of clays, it is necessary to distinguish between
material structure and property anisotropy. In general, "anisotropy"
refers to the material structure and/or properties which do not exhibit
the same characteristics and/or properties in every direction. The
material structure anisotropy relates primarily to the anisotropy of
fabric which would influence development of interparticle force rela-
tionships. Property anisotropy refers to strength, compressibility,
permeability, conductivity, and other mechanical properties which are
not equal in all directions, i.e. the material property demonstrated
is a function of the sample tested. The external constraint anisotropy
refers particularly to the applied stresses and boundary constraints.
While considering properties of fine sediments, the anisotropy needs
to be taken into account.
2.2.1 Size, Range, and Definition
The maximum size of particles in the clay size grade is defined
differently in different disciplines. In geology, the tendency has been
to follow the Wentworth Scale to define the clay grade as materials
finer than about 4 microns. In soil investigations, the tendency is
to use 2 microns as the upper limit of the clay size grade. Although
there is no sharp universal boundary between the particle size of clay
minerals and non-clay minerals, in argillaceous materials, a large
number of analyses have shown that there is a general tendency for the
clay minerals to be concentrated in a size less than 2 microns (Grim,
1968).
Clays contain varying percentages of clay-grade material and
therefore varying relative amounts of non-clay-mineral and clay-mineral
constituents. Clays almost always contain some non-clay mineral
material coarser than the clay grade, although the amount may be very
small. In many materials called clays the clay grade and the clay-
mineral constituents make up considerably less than half the total.
In such materials the non-clay is frequently not much coarser than the
maximum for the clay grade, and the clay mineral fraction may be par-
ticularly potent in causing plasticity. In general, fine grained
materials have been called clays so long as they have distinct plasti-
city and insufficient amounts of coarser material to warrant the
appellations "silt" or "sand." If particle size anslyses are made, the
term clay would be reserved for a material in which the clay grade
dominates. However, names have been and are applied most frequently
on the basis of appearance and bulk properties of the material.
The expression clay material is used for any fine-grained, earthy,
argillaceous, natural material. Clay material includes clays, shales,
and argillites. It would also include soils, if such materials were
argillaceous and had appreciable contents of clay-size-grade material.
Clay particles are usually within a range of diameter smaller than
0.002 mm, but larger than the molecular size (10-6 mm).
For the civil engineer, the fine sediments eroded from the earth's
crust are of interest and the term clay is primarily a particle size
term. For the chemical engineer the interest includes synthetic and
other materials in their fine form, rather than the natural clay
minerals.
There are three levels of first order fabric recognition of clay
structure. These are categorized on the basis of the degree of magni-
fication required for a proper observation of the fabric pattern.
1. Macroscopic: The fabric units are distinguishable by the naked
eye. They consist of an aggregation of clay particles called peds.
2. Microscopic: The fabric units are visually observed under the light
microscope. For clays, single particles are not distinguishable
-10-
at this level of viewing. The fabric units identified in the
microscopic range consist of several particles or groups of par-
ticles defined as clusters, also sometimes called flocs.
3. Ultra-microscopic: The fabric units are visually observed in the
ultra-microscopic level using electron microscopy (either trans-
mission or scanning electron microscopy). Single or individual
clay particles can be distinguished at this level.
2.2.2 Surface Area
Colloidal properties such as plasticity and adsorption of molecules
arise from the large surface area associated with a small mass. The
surface forces are dominant in respect to fine sediments and the in-
fluence of gravitational forces is small. The clay minerals are plate-
shaped or tabular because the layer-lattice structure results in strong
bonding along two axes but weak bonding between layers. The clay par-
ticle thickness depends upon the magnitude of forces of attraction
between the layers. The variation in specific surface area of different
minerals is primarily due to different thicknesses of the plate-shaped
particles. Variation in the other two dimensions of clay particles is
related to the degree of crystallinity of the clay minerals. A well
crystallized Kaolinite has large particles. If it is poorly crystalline,
the Kaolinite particles may not be larger than those of montmorillonite.
The true meaning of the particle size becomes more evident in terms
of its specific area. For example, a single sphere 1 cm in diameter
has a surface area of 3.14 cm2. The same volume in terms of one micron (1 u)
diameter spheres has a surface area of 10,000 times greater and for one
-11-
millimicron diameter spheres the surface area is 10 million times as
great.
The specific surface area of different clay minerals is as
follows:
Montmorillonite: 800 m2/g Chlorite: 80 m2/g
Clay Mica: 80 m2/g Kaolinite: 15 m2/g
Kaolinites show the most uniform crystals, often hexagonal plates
with a typical diameter of 0.3 to 0.5 p and a thickness of 0.05 to 2 u.
Montmorillonite particles are thin plates typically around 30 A thick
and 0.1 to 1 p in diameter. Illite particles are plates with a typical
thickness of 300 A.
Surface area is one of the most important properties of fine sedi-
ments. Most of the differences between clay minerals in properties such
as water retention, plasticity, or cohesion can be explained by the
differences in the surface areas of particles. This explains the high
swelling and high liquid limit of montmorillonite. Liquid limit being
closely related to the surface area, it is considered as a measure
of the nature of the surface as well as the area.
2.2.3 Shape
The shape of a particle is also an important factor in determining
the specific area. A spherical particle has the least surface area,
and a disc or plate shaped particle has the greatest surface area. A
sphere of 1 cm3 has an area of 4.836 cm2, while an equi-volume disc
one micron thick has 20,000 cm2 surface area. Typical plate shaped
-12-
particles may have surface areas as much as ten times as great as the
spherical particles of equal volumes.
2.2.4 Electric Charge
Substitution of one ion for another in the clay crystal lattice
and imperfections at the surface, especially at the edges, lead to
negative charges on clay particles. Cations from the pore water are
attracted to the particles (and anions repelled) to maintain electro-
neutrality. These are the exchangeable cations and their number is
the cation-exchange capacity (i.e. the amount of negative charge per
unit weight or per unit surface area) of the clay. This is usually
expressed as milliequivalents per gram (me/g) or per 100 g (me/100 g).
The force with which ions are held at the surface varies with the nature
of the charge. The amount of charge for different clay minerals is
given below:
Kaolinite: 5 to 15 me/100 g
Clay Mica and Chlorite: 20 to 40 me/100 g
Montmorillonite: 80 to 100 me/100 g
Vermiculite: 100 to 150 me/100 g
The kind and number of exchangeable cations have an important in-
fluence on the behavior of soils, e.g. monovalent cations such as
sodium increase the activity of the clay, its swelling, etc.
2.2.5 Inter-Particle Forces
The behavior of clay particles is controlled more by the surface
forces than by the gravity forces. Thus it can be shown that the
-13-
average electrochemical force exerted on one clay particle is of the
order of one million times greater than the average weight of the
particle.
The inter-particle forces are both repulsive and attractive in
nature. The most important repulsive forces generated by the electrical
charges are
i) Repulsion caused by the negatively charged particle faces.
ii) Repulsion of adsorbed positively charged cations.
iii) Osmotic pressure resulting from the high concentration of cations
near the surface of the particles in the pore water.
The attractive forces binding the particles together in clay
minerals are the following:
i) Forces due to the attraction of the mass of one clay mineral
particle and the mass of another.
ii) Inter-molecular forces resulting from the nearness of one par-
ticle to another with the overlap of fields of force of molecules
in the surface layers of adjacent particles.
iii) Electrostatic forces due to changes in the lattice resulting
from unbalanced substitution within the lattice, broken bonds
on the edges of the lattice, and the attractive force of cation
ions adsorbed on the clay-mineral surfaces.
iv) Chemical cementation between particles by various compounds.
v) Cation bonds: Forces exerted by cations attracting and trying
to neutralize negatively charged particles (Fig. 1.A).
vi) Water dipole linkage is the bonding action of adsorbed polar
molecules (Fig. 1.B). Oriented water molecules between two clay-
mineral surfaces may form a bridge of considerable strength if
-14-
+ And Indicate Electric Charge
(+ () (W+)
.--- Clay Particle
(A) Cation Bond
E:]
t+-
---- Water Dipole
E:l Clay F
Particle
(B) Water Dipole Linkage
S---o Clay Particle
-- Water Dipole
-- Cation
1?
(C) Dipole Cation Dipole
@0
a @
E G
H-
_-b Diffused
Double Layer
(D) The Clay Micelle
Fig. 1 Inter-particle forces on clay minerals and clay micelle
I- -
_1
I- -
i_ -- 1-
I -I
G a
-15-
only a few molecules thick, and of practically no strength if
more than a few molecules thick. Similarly, adsorbed polar
organic molecules could serve as a bond between clay-mineral
particles.
vii) Dipole-cation-dipole linkage (Fig. 1.C).
viii) Hydrogen bond occurs when an atom of hydrogen is strongly attracted
by two other atoms.
ix) van der Waals forces are secondary valance forces of an electro-
chemical nature. They are generated by the mutual influence of
the motion of electrons of the atoms and they are always attrac-
tive. These forces acting between all units are property of
the matter and are independent of the chemical characteristics
of water. Although other attractive forces of electrical nature
may exist, the van der Waals forces are the main cause for bond-
ing together of clay particles. An important characteristic of
these forces is that they decay very rapidly with distance and
hence the particles must come very close to each other so that
the forces can be effective.
2.2.6 Flocculation
The bonding of clay particles when they are brought together is
known as flocculation. Both collision and cohesion are essential to
flocculation. Cohesion is believed to result from the predominance of
attractive forces on the surface of clay particles. Collision of par-
ticles may be caused by the Brownian motion of the suspended particles,
by internal shear of water, and by the differential settling velocities
-16-
of the flocs. Brownian motion is the erratic movement of small sus-
pended particles caused by the thermal agitation of the suspending
medium, which enables particles to come in contact with each other.
The inter-particle forces have been described in the previous
paragraph. Whatever the origin of the surface electric charges, any
such charged particle in an ion containing water will attract ions of
opposite charges, called "counter-ions,' to compensate its own electric
charges. At the same time, the ions tend to diffuse away from the sur-
face because of their thermal activity since such a diffusion takes
place from a zone of high concentration to a zone of lower concentration
in a way analogous to the diffusion of the molecules of the air in the
atmosphere. Thus a clay particle idealized by a thin rectangular plate
will be surrounded on either side by a diffused layer of counter-ions
whose positions will be determined by the balance of the electrostatic
attraction and their thermal activity. This layer is known as a
"double layer" and it plays a dominant role in the mechanical properties
of suspended clays and clay deposits. The system of clay particle and
double layer is electrically neutral and is known as "clay micelle."
Figurel.D shows a simplified, schematic diagram of a clay micelle.
Any form of agitation, including Brownian motion,will eventually
cause two particles to approach each other sufficiently close for their
double layers to interact. This interaction causes changes in the
distribution of the cations in the double layer of both particles. The
result will be determined by the potentials of the van der Waals force
and the electric forces. Sufficiently far from the particle, the
repulsive forces may dominate, whereas closer to the surface the net
effect would be attractive which causes flocculation of particles.
-17-
It has long been observed that gentle stirring promotes floccula-
tion. This is due to the velocity gradients which are induced in the
liquid causing relative motion of the particles. Such velocity
gradient-controlled flocculation is called orthokinetic flocculation.
A simple theory of flocculation kinetics can be derived for a uniform
liquid shear field, giving a constant velocity gradient. Such constant
velocity gradients are difficult to achieve in practice; the closest
experimental form has been in the annular gap between coaxial rotating
cylinders, also known as Couette apparatus. Consequently, the theory
has to be extended to velocity gradients created in turbulent flow con-
ditions. The orthokinetic rate of flocculation has a high dependence
on initial particle size, is linearly dependent on velocity gradient,
and it is independent of temperature.
While considering Brownian motion, it is useful to consider one
particle (the collector) as stationary, and to calculate the diffusion
rate caused by Brownian movement of other particles to this collector.
Because particles become attached to the collector, and are therefore
removed from the suspension, a concentration gradient is formed radially
outwards from the collector. This diffusion-controlled flocculation is
called perikinetic flocculation. Temperature and viscosity effects are
significant under perikinetic flocculation, and the rate of flocculation
is independent of particle size. In the benthic boundary layer, ortho-
kinetic (i.e. velocity gradient-controlled) flocculation has a much
greater influence on the frequency of collision between particles than
perikinetic (i.e. diffusion-controlled) flocculation (Williams, 1980).
The effect of flocculation is formation of aggregates. Repeated
inter-particle collisions in a turbulent flow field are predominantly
-18-
due to the internal shearing of the suspending water. Aggregates formed
by fluid shearing are denser and stronger than those formed by the
Brownian motion or by differential settling velocities. The aggregates
can be dispersed by local high shearing rates and may re-aggregate later.
The settling velocities of aggregates and the probability of their
sticking to the bed are determined by the size, density, and shear
strength. Mineral particles cohering in a cluster with uniform porosity
are called primary aggregates or flocs. At lower shearing rates,
primary aggregates would collide with each other and bond to form
first-order aggregates. At still lower shearing rates, the first order
aggregates would collide and bond with each other to form weaker, less-
dense second-order aggregates, and so on. Each higher-order aggregate
would include water in the new pore volume formed, and because shear
stress can be transmitted only through inter-aggregate contacts, the
higher order aggregates are weaker. A procedure for determining the
order of aggregation has been given by Krone (1976).
2.3 Parameters Influencing the Properties of Cohesive Sediments
An attempt wasmade by the Committee on Tidal Hydraulics (1960) to
identify the soil parameters which affect the process of shoaling in
estuaries. However, the literature review contained in the report is
limited. Based on the findings of subsequent research, Paaswell (1973)
reviewed the causes and mechanism of cohesive soil erosion and identified the
following parameters used in evaluating the erosion of cohesive beds:
-19-
Character Parameter
Physical Soil type (clay mineral)
Percentage of clay ,
Liquid and plastic limits, and activity
Specific gravity
Physico-chemical Base exchange capacity
Sodium absorption ratio
Pore fluid quality
Pore fluid environment
Mechanical properties Shear strength (surface and body)
Cohesion
Thixotropy
Swelling and shrinkage properties
Conditions of environment Weathering (wet-dry)
Freezing and thawing
Prestress history
In addition to the above, Alizadeh (1974) has included the parame-
ters bulk density, water content, effective stress, time (aging),
testing methods, and soil preparation methods.
Kandiah (1974) mentioned that the erodibility of cohesive soil is
controlled by the mineralogical, chemical, and environmental factors.
The mineralogical factors include the type and amount of clay mineral
present in the soil while the chemical factors include the total salt
concentration, sodium absorption ratio, and pH of the pore fluid. The
influence of these two factors is mutually independent and they are the
"key parameters" of erosion since their effect is far more pronounced
than other factors.
Attempts have been made to correlate the critical shear stress to
various parameters affecting soil erosion. For example, Smerdon and
Beasly (1959) presented the following equation between critical shear
stress ( c) and plasticity index (PI):
Tc = 0.0017 (PI)0.84 (2.3.1)
-20-
Dunn (1959) found that not only the plasticity index but vane
shear strength (S ) needs to be included in the expression as follows:
Tc = 0.0098 + 0.00049 (Su + 180) tan (30 + 1.73 PI) (2.3.2)
Carlson and Enger (1963) suggested the following relationship:
Tc = -0.017 + 0.000181(PI) + 0.000186(v)
+ 0.00268(K) + 0.000465(LL) (2.3.3)
where
v = sample density
K = phi-skewness of the grain size distribution
LL = liquid limit
Sargunam (1973) presented the following expression related to the
composition of pore fluid C, and sodium adsorption ratio (SAR)
Tc = C1 + (C2 n log SAR) log C (2.3.4)
where
C = pore fluid concentration
C1, C2, and n = constants which vary with the type and amount
of clay minerals
For describing the influence of various parameters the following
classification appears more appropriate:
a) Hydrodynamic factors.
b) Properties of sediment.
c) Properties of bed.
d) Properties of pore fluid and eroding fluid.
-21-
a) Hydrodynamic factors (bed shear stress). These are principally
embodied in the instantaneous bed shear stress and its frequency dis-
tribution, as specified by the flow characteristics, including the
surface roughness of the bed-fluid interface. The studies reported
later in this thesis have indicated that the concentration of suspen-
sion resulting from the erosion of bed is not only a function of the
applied shear stress but also the shear stress previously acting on
the bed.
b) Properties of sediment (composition, shape, size gradation,
organic matter, cation exchange capacity, moisture content). Fine
sediments include interacting particles such as clays, as well as non-
interacting fine particles such as silt. The differences in the
physical properties of these have already been described. The clays may
be composed of different clay minerals each of which has its own physi-
cal properties such as shape, size, surface area, liquid limit, etc.
Since these properties in turn influence the erosional properties, the
basic composition of fine sediments in terms of individual clay
minerals or their mixtures, and the clays alone or mixture of clay and
silt has a considerable influence on the erosional properties.
Sediment composition is specified by the clay mineral, its weight
fraction, and the amount and type of organic matter. Recent attempts
at the University of California, Davis, have been directed toward
characterizing the composition of clays through measurements of the
dielectric constant at selected frequencies. A dielectricc dispersion
parameter" is derived from these measurements. Each clay tested seemed
to have a characteristic value of this parameter (Alizadeh, 1974;
Arulanandan et al., 1973). The dielectric constant is a measure of
-22-
the ability of clay to store electrical potential energy under the in-
fluence of an electric field. The dielectric constant for a soil
sample is defined as
Scd (2.3.5)
VA
where
c = capacitance
d = length of specimen
A = cross-sectional area
ev = dielectric constant of vacuum
= 8.85 x 10-14 farad/cm
The dielectric constant of a dry silicate mineral is 4, and of water
about 80.
Alizadeh (1974) has defined the magnitude of dielectric dispersion
(AE') as the total amount of decrease in the measured dielectric con-
stant. The dielectric dispersion depends mainly on the type and amount
of clay; the other factors such as pore fluid composition, water con-
tent, particle orientation, etc. have a secondary effect. It has been
used as a quantitative index for soil characterization. Measurements
have shown that for 10 percent Kaolinite and 21.2 percent water content,
AE0 has a magnitude 7.5, whereas for 60 percent Kaolinite and 30.4
percent water content, AE' is 18.
The shape and size of individual clay particles have an influence
on their surface area. Since the surface forces predominate in respect
to cohesive sediments, the colloidal properties such as plasticity and
adsorption of molecules are governed by the surface area. Natural
sediments usually have a wide range of particle sizes. The effect of
-23-
such heterodisperse sediments is to increase the probability of colli-
sion of primary particles. Also, natural sediments typically are not
uniform in their shape. Since Brownian motion is rotary in character,
the largest dimension is appropriate for the collision diameter,
whereas the mean diameter is applicable to the diffusion constant.
Consequently, anisodimensional particles have a relatively large colli-
sion diameter combined with a relatively large diffusion constant, thus
enhancing the collision rate. After flocculation has progressed, how-
ever, the flocs tend to be more nearly spherical, and the effect of
anisotropy on collision rate is small.
Organic matter has an effect on the properties of flocculation of
sediment and hence it affects the erosional properties of sediment.
Kandiah (1974) found that organic matter strengthens the soil aggre-
gates against slaking. Studies on the erodibility of 30 percent
illitic soil showed that the critical shear stress for erosion increased
from 1.7 N/m2 to 4.0 N/m2 when the organic content was increased from
0 to about 4 percent.
It has been mentioned earlier that the type of clay is one of the
factors determining the erosional characteristics. The Cation Exchange
Capacity is a property of the soil which can be used to indicate the
type while associating it with the corresponding critical shear stress
for erosion. Cations from the pore fluid are attracted by the nega-
tive charge on clay particles and the anions are repelled in order to
maintain electro-neutrality. The number of these exchangeable cations
is known as the cation exchange capacity, which is usually expressed
as milliequivalent per gram (me/g). The kind and number of exchangeable
cations have an important influence on the behavior of soils. For
-24-
instance, monovalent cations such as sodium increase the activity
of clay and its swelling. The predominant exchangeable ions in a soil
are usually Na Ca+, and Mg The rate of cation exchange reaction
varies with the clay mineral, the concentration of the cations, and with
the nature and concentration of the anions. In general, the reaction
for Kaolinite is most rapid, slower for Illite, and still slower for
Montmorillonite. The CEC is independent of the ionic strength of the
solution with which the clay is.in contact or the physical structure
of clay. However, it is a function of pH of the fluid in contact.
Kandiah (1974) has shown that the critical shear stress of cohesive
soil increased from 1.3 N/m2 to about 2.7 N/m2 with a change in CEC
from 2 me/100 g to 34 me/100 g. These results were with an average
sodium absorption ratio of 2.5. A decrease in critical shear stress
was observed for the same range of CEC values when the average sodium
absorption ratio was 48.0 (Fig. 4).
c) Properties of the bed (moisture content, density). Attempts
have been made to study the effect of moisture content in the bed
which can be compared with standard soil indices such as plastic limit
and liquid limit. Lutz (1934) noted that soils with high plasticity
and low plasticity had different resistance to erosion. The low plas-
ticity soil was more erosive than the high plasticity soil. Fukuda
(1978) has shown that an increase in suspension concentration can be
expected as the water content of the sediment increases, with the stress
being held constant. A very small change in the water content of the
sediment may give a large increase in the concentration of suspension. For
instance, for a 12 percent increase in water content in his tests, the
concentration of suspended sediment increased from 50 to 3000 mg/l.
-25-
Owen (1970) reported that the shear strength variation within
the bed could be correlated satisfactorily with the variation of den-
sity. The shear strength was observed to increase rapidly with density.
The method of formation of bed namely, remolded, deposited, or com-
pacted with external force, have an effect on the density structure
of the bed and hence on the erosional property of the soil.
d) Properties of the pore fluid and the eroding fluid (salt
concentration, sodium adsorption ratio, pH, temperature). In the case
of a fully saturated soil which has been equilibrated with the eroding
fluid over a sufficiently long time, the pore fluid and the eroding
fluid are expected to have the same properties. However, in case they
have different properties, osmotic pressure gradient is formed which
changes the properties of the pore fluid and this in turn may change
the erosional properties of the bed. The soil having an aggregated
structure resists erosion more effectively than a soil having a dis-
persed structure. This is because the net force between the particles
in an aggregated structure is attractive while in the latter it is
repulsive. Since the pore fluid has a substantial influence on the
structure, it influences erosion. An increase in the salt concentra-
tions of the pore fluid usually increases resistance to erosion.
Sodium Adsorption Ratio (SAR) is defined as
SAR = Na (2.3.6)
[- (Ca+ + Mg )]'1
The concentrations of individual ions are in milliequivalents per
liter.
The SAR is used as an index to characterize the pore fluid and
eroding fluid in terms of the relative strength of the Na, Ca, and
-26-
Mg ions. Fluids with the same salt concentration can have different
SAR. An increase in SAR decreases cohesion, lowers the critical shear
stress, and hence increases the erosion rate (Kandiah, 1974). Alizadeh
(1974) has presented experimental data to this effect.
Kandiah (1974) also studied the effect of pH and concluded that
"pH influences the interparticle cohesion which strongly affects
aggregation and dispersion of clay soil properties" (p. 118).
The effect of water temperature on erosion rate has been studied
by Grissinger (1966), Christensen and Das (1973), Randkivi and Hutchison
(1974), and Gularte (1978). It is observed that the temperature had a
significant effect on the strength of inter-particle cohesion and the
rate of erosion increased significantly with increasing temperature.
Kandiah (1974) showed from particle-by-particle surface erosion
of a remolded illite soil that the critical shear stress for erosion
varies as
Tc = 1.8 x 10-5 exp[4100/T] (2.3.7)
where T is the absolute temperature in the Kelvin scale. The critical
shear stress dropped from 3.6 to 0.8 N/m2 over the temperature change
from 9.50C to 420C.
2.4 Processes for Deposited Beds
The deposited beds are distinguished from the placed or remolded
beds because of the fact that they are formed by the process of settling
of the sediment in suspension. The settling may take place under
quiescent conditions or under a low shear stress which permits
-27-
deposition of the suspended material. The settled material may undergo
consolidation and form a bed. The aspects related to the settling
and consolidation which are the primary processes in formation of
deposited beds are briefly described below.
2.4.1 Settling
The effect of Brownian motion on a suspension of sediment under ap-
parent quiescent conditions is to induce inter-particle contacts. This
may result in bonding of particles through the process of flocculation
(Einstein and Krone, 1962). The mechanics and importance of floccula-
tion have already been described. Hence only the parameters which
influence settling velocity of the flocs are briefly described here.
Owen (1970) conducted detailed study of the settling velocities of
an estuary mud and noted the following conclusions:
i) Suspended sediment concentration: The median settling velocity
increases with concentration up to a value between 4 and 20 grams
per liter, depending on the salinity, at which hindered settling
begins. For a concentration higher than this, the settling
velocity reduces.
ii) Salinity: The median settling velocity increases with salinity
(except during hindered settling) up to a value between 28 and
43 grams per liter depending on the concentration. For higher
values of salinity, the settling velocity decreases. It may
however be noted here that certain clays such as Kaolinite
flocculate even in distilled water.
iii) Depth: The effect of depth of settling is fairly complex, the
settling velocity reducing to a minimum at a depth of 1 meter
-28-
and then increasing with depth to reach its terminal value at
about 2 meters. For a fixed depth of settling, there is an
absolute maximum value of settling velocity, which is attained at
a fixed salinity and concentration.
iv) Temperature: The effect of temperature is not very clear. It
is largely limited to the effect temperature changes have on the
viscosity of suspension. There appears to be a slight tendency
to increased flocculation at higher temperatures, generally
accompanied by slightly greater median settling velocities.
However, at temperatures above 150C, the settling velocities of
flocs formed in low salinity suspensions decreases.
Effect of other parameters is as follows:
v) pH: High pH contributes to dispersion, whereas low pH enhances
flocculation.
vi) Organic matter: Usually flocculation is promoted by the
organic matter.
vii) Dissolved chemicals: Only those chemicals which enter in some
way into the physico-chemical reaction with soil can probably
have an effect on flocculation and settling.
2.4.2 Consolidation
Consolidation is the term used to refer to that portion of the
compressibility of a soil that is essentially inelastic, i.e. its volume
changes under load. Since the pore water and the soil grains in a
saturated system are relatively incompressible, the volume change ob-
served under load is the result of the expulsion of water from the
interstices between soil grains.
-29-
Most soils regain only a fraction of the volume lost during con-
solidation. This results from the fact that in order to undergo volume
change, the soil particles are displaced relative to one another to
assume a more closely packed condition, and consequently greater
density. In sands these movements are not reversible, and in most
clayey soils, they are only slightly reversible. For sand, the volume
decrease is proportional to the logarithm of the pressure. For clay,
the relation is not linear although at higher pressures, it is approxi-
mately so.
In a process of continuous sedimentation in water, the soil at any
depth is being consolidated under the influence of the weight of the
soil above it. Since new material is continually being added, a gradient
tending to cause the pore water to flow upward out of the system exists
at all points within the stratum. If the material being deposited is
sand, consolidation takes place at a rapid rate since the spaces between
particles are relatively large and the water can escape readily. Fur-
thermore, the sand particles are essentially inert and are not greatly
influenced by adjacent particles, and the initial position of each grain
within the mass is relatively stable. Very little volume change can
take place in a sand deposit except in certain circumstances as a
result of outside influences such as mechanical vibration or shock.
Therefore, sands can be considered to be virtually completely consoli-
dated at all times even when the accumulation of sediment is rapid.
Consolidation of clayey deposit proceeds at a comparatively much
slower rate. The total pore space in a clay mass is large but is com-
posed of a multitude of small channels between the individual particles.
The flow of water in the channels is restricted by their small size
-30-
and also by the affinity of the particle surfaces for water, which in
effect reduces even further the dimensions of the available flow
channels. Also, unlike sand grains, clay particles, due to their
shape and the interactions of their force systems, do not tend to fall
into stable positions. As a result, the upper portion of a clay de-
posit is very porous and contains a large percentage of water. The
actual porosity at the clay surface can vary considerably, depending
upon the amount and type of clay mineral present, and on the factors
that affect the interparticle forces, but it is always large in com-
parison with that of sand.
Soil concentrations of the order of 10 to 20 percent by weight may
be expected in newly formed clayey deposits. It is likely that at about
this concentration, a continuous, interdependent network of soil par-
ticles is formed. This condition has been referred to as the "hindered
settling." It might as well serve as a useful dividing line between
the processes of sedimentation and consolidation. At lesser concen-
trations, settling occurs as the individual particles of flocs inde-
pendently move downward through water. At higher concentrations, where
the units can no longer move independently, the downward movement is
accomplished by water moving up through the soil voids.
In continually accumulating deposits, hydraulic gradients indica-
tive of incomplete consolidation are present throughout the clay layer.
If the accumulation is relatively rapid, the degree of consolidation
at any depth is slight and it will be found that the density of the
clay deposit will be virtually independent of the depth. When sedi-
mentation ceases, the thickness of the layer will continue to decrease
for some time probably for many years, until the water pressures
-31-
induced by the weight of the sediment have been dissipated (Committee
on Tidal Hydraulics, 1960).
The properties of clays in respect to settling and consolidation
are very important in connection with the studies for their depositional
and erosional properties.
2.5 Clay-Water System
Water which can be held by the clay system is grouped into two
categories, namely low-temperature water, which can be driven off by
heating to about 100C to 150C, and the OH lattice water which is lost
at temperatures above about 3000C. The nature of low-temperature water
and the factors that control its characteristics are of great impor-
tance, since they largely determine the plastic, bonding, compaction,
suspension, and other properties of clay materials, which in turn con-
trol their behavior under the given flow field.
Water, though neutral, has its oxygen and hydrogen atoms spaced
in such a manner that the center of gravity of the positive and nega-
tive electrical charges do not coincide. The resulting molecule has a
positive charge acting at one end and a negative charge acting at the
opposite end. Water molecules are thus considered polar molecules.
Because of the net positive charge of the cations, they attract negative
charges. The negative tips of water molecules are attracted and held
to the cation, which in turn is held by the clay particle due to the
negative charge on its surface. The resulting effect is that water
becomes bonded to the clay. Additional water molecules are also at-
tracted to the clay particle because of a chain-like arrangement of
negative ends to positive ends of molecules.
-32-
The term diffused double layer has already been described. With
the water and clay molecules in contact with each other, it is believed
that immediately surrounding the clay particle, there'is a thin, very
tightly held layer of water, perhaps 1 x 10-6 mm (10 A) thick, and a
second, more mobile, diffused layer extends beyond the first layer to
the limit of attraction. The molecular movement occurs continually
in both the layers. The water which is held in the diffused double
layer is frequently termed adsorbed water or oriented water to dif-
ferentiate it from normal pore water which is not oriented.
The plasticity possessed by clay soils is attributed to the water
which is attracted and held by the clay particles. Experiments per-
formed with clay using non-polar liquid in place of water do not in-
dicate plasticity and the particles act similar to those of a coarse
grained sandy soil (McCarthy, 1977).
A dispersion of particles subject only to hydrodynamic interaction
will exhibit Newtonian flow characteristics, i.e. the shear stress and
shear strain have a linear relationship no matter how small the mag-
nitude. A clay-water suspension of high concentration on the other
hand shows properties of a non-Newtonian fluid as shown in Fig. 2 which
is a plot of equilibrium shear stress (T) versus shear rate (v). Curve
A represents Newtonian behavior. Very dilute cohesive suspensions may
exhibit this flow behavior, particularly if they are weakly flocculated.
Curve D describes Bingham plastic behavior. A Bingham fluid will not
flow at all until the yield stress is exceeded. This implies that the
soil structure fails at some critical stress Ty and for stresses in
excess of this the dispersion flows in a quasi-Newtonian manner.
Curves B and C show more realistic rheograms where there is a gradual
(D) Bingham Plastic
(C) Plastic
(B) Pseudo-Plastic
'(A) Newtonian
SHEARING RATE
Fig. 2 Rheological models
(V)
-34-
reduction in the contribution of structure to apparent viscosity as the
shear rate is increased. These indicate pseudo-plastic and plastic
behavior, respectively, also known as shear thinning.' The intercept
T shown in Fig. 2 is the lower yield stress, whereas TB is known as
the upper Bingham yield stress which is a measure of the work done in
disrupting the floc structure in cohesive suspensions.
The content of water in a clay soil determines the consistency in
the remolded state. The criteria to determine the various states of
consistency are known as Atterberg Limits. At high water content, the
soil-water mixture possesses the properties of a liquid; at lesser water
content the volume of the mixture is decreased and the material exhibits
the properties of a plastic; at still lesser water content, the mixture
behaves as a semi-solid and finally as a solid. The water content
indicating the division between the liquid and plastic state is called
the liquid limit. The division between the plastic and semi-solid
state is the plastic limit. The water content at the division between
the semi-solid and the solid state is the shrinkage limit. All these
three limits are expressed as percentage of water with respect to the
weight of solids. Below the shrinkage limit, there is little or no
change in volume as water content varies. However, above the shrinkage
limit, the change in total volume of soil-water mixture is related to
the change in water content. The plasticity index is the numerical
difference between the liquid limit and the plastic limit, and indicates
the range of water content through which the soil remains plastic. It
is necessary to know both the liquid limit and the plasticity index for
a proper evaluation of the plasticity properties of soil.
-35-
The liquid limit expresses the overall effect of the inter-particle
forces within the clay mass and this soil index varies with clay
mineralogy and with the associated cation. The following data given
in Table 1 were presented by the Committee on Tidal Hydraulics (1960).
Table 1. Specific surface area and liquid limit for typical clays.
Primary Specific
Clay Associated Surface Liquid
Mineral Cation Area Limit
Montmorillonite Na 847 710
Montmorillonite H 768 490
Illite (< 2 p) H 79.8 100
Kaolinite H 13.1 53
Kaolinite (< 5 v) H 26.1 110
Many clay soils exhibit the property of rheotropy at water con-
tents above the liquid limit, and also to a lesser degree at water
contents in the plastic range. Rheotropy is the change to a more
fluid consistency on stirring or disturbance. When the disturbance
has ceased, the system reverts to its less fluid or more rigid condi-
tion. This is often called thixotropy, although the strict definition
of thixotropy is a reversible, isothermal sol-gel transformation. A
sol, by definition, has no yield value, while a gel has rigidity. The
change in clay-water systems is generally from a system with higher
yield value to one with a lower yield value. A sol may be considered
as a colloidal dispersion. This restricts sols to liquid-like behavior.
-36-
When hardening of the sol occurs, a gel is formed. This requires a
change of state from a semi-liquid substance (sol) to a semi-solid
(gel).
Rheotropy of clay soils can be measured by a vane shear or at
higher water contents by means of a viscometer.
The property of thixotropy (or rest-hardening) has been explained
either by changes in particle rearrangement and inter-particle forces,
or by changes in adsorbed water. On stirring, the particles and
fabric units are rearranged and the bonds between particles and units
are broken. Also, the structure of the adsorbed water is broken up
and the clay mass will be more susceptible to deformation under self-
weight. After deformation, the clay fabric will seek a status of
minimum energy with maximum attraction between particles and fabric
units. The adsorbed water also regains its quasi-crystalline form to
give the system sufficient rigidity to have a yield value. There are
several factors which contribute to the regaining of part or all of the
strength. These are original structure, activity of the clay minerals,
and the degree of disturbance. The activity is a characteristic
parameter of the electrochemical action of the colloids and is defined
as the ratio of plasticity index and clay fraction less than two
microns.
CHAPTER III
PREVIOUS LABORATORY STUDIES
3.1 General Review
Over the past two decades, considerable laboratory work has been
carried out on cohesive sediments. In order to get an idea about the
variety of different ways in which the research work has been carried
out, some of the topics under which the literature could be classified
are given below along with a typical reference on the same as an
illustration.
a) Sediment used:
Clay mineral alone: Kaolinite: Christensen and Das (1973)
Mixture of clay minerals: Yolo Loam: Arulanandan et al. (1975)
Mixture of clay and silt: Grundite: Gularte (1978)
Natural sediments: Brisbane Mud: Thorn and Parsons (1980)
Fernandina Mud: Yeh (1979)
San Francisco Bay Mud: Partheniades (1962)
b) Fluid used:
Salt water: Partheniades (1962)
Distilled water: Mehta and Partheniades (1979)
Fresh water: Fukuda (1978)
c) Type of bed:
Remolded: Gularte (1977)
-37-
-38-
Deposited from suspension: Yeh (1979)
Compacted: Christensen and Das (1973)
d) Characterizing indices:
Dielectric dispersion: Alizadeh (1974)
Sodium adsorption ratio: Kandiah (1974)
Cation exchange capacity: Kandiah (1974)
Chemical and electrical parameters: Arulanandan et al. (1973)
e) Basic parameters:
Bed density and salinity: Owen (1977)
Temperature: Gularte (1978)
pH: Kandiah (1974)
Pore fluid and eroding fluid: Arulanandan et al. (1975)
Water quality (pH, conductivity): Migniot (1968)
f) Microstructure studies:
Kaolin: McConnachie (1974)
Marine sediments: Bowels (1969)
g) Other studies:
Hydrodynamic aspects: Turbulent drag reduction: Gust (1976)
Colloidal dispersion: Zeichner and Schowalter
(1977)
Attempts have been made from time to time in the past to review
the information available in respect to cohesive sediments. These are
listed below in chronological order:
1960: The Committee on Tidal Hydraulics, U.S. Army Corps of Engineers
conducted literature review to study soil as a factor in shoaling
processes (Committee on Tidal Hydraulics, 1960).
-39-
1964: The present knowledge on the behavior of fine sediments in
estuaries was summarized by Partheniades (1964).
1968: Task Committee of ASCE on Erosion of Cohesive Materials prepared
a report on literature review (Task Committee, 1968).
1973: State of the art paper on causes and mechanisms of cohesive soil
erosion was presented by Paaswell (Paaswell, 1973).
3.2 Review of Literature on Erosion
Since the scope of the present study deals with erosion of Kaolinite,
a brief account of the work carried out in the past on erosion of co-
hesive sediments is given here. Effect of various parameters and pro-
cesses influencing the erosional properties of cohesive sediments has
been studied by several research workers. Important findings of these
studies have already been given in Chapter II.
The earlier experiments conducted to study the erosion of cohesive
sediments were oriented to obtain solution to a specific engineering
problem such as the model studies performed by the Tennessee Valley
Authority for the Fontana Project (1953) and the Fort Patrick Henry
Project (1960).
Most of the work to understand the basic physical processes related
to cohesive sediments has been carried out during the past two decades.
Smerdon and Beasley (1959) applied the tractive force theory to the
stability of open channels in cohesive soils. Dunn (1959) used a sub-
merged jet to determine the tractive resistance of cohesive soils and
correlated it with plasticity index. Masch et al. (1965) conducted
studies on remolded cohesive sediments using a rotating cylinder
-40-
apparatus and found that the critical shear stress was related to water
content and vane shear strength. Flume study of natural soils conducted
by Lyle and Smerdon (1965) showed that the critical tractive force
correlated to void ratio, cation exchange capacity, and plasticity
index. Grissinger (1966) found that the erosion rate of soil decreases
with increasing clay content and decreasing void ratio and temperature.
Findings of various research workers related to determining the
effect of various parameters and characterization indices on erosion
have already been described in Chapter II. Paaswell (1973) summarized
selected studies on cohesive soil erosion. The same are reproduced
from his paper along with addition of subsequent investigations. It
may be noted that the mode of formation of bed in the laboratory equip-
ment is an important factor. This can be done in the following 3 dif-
ferent ways.
a) Placed or remolded or uniform bed: Formed by mixing the sediment
thoroughly with required water content and placed evenly in the
apparatus so as to have uniform density without any external com-
paction.
b) Flow deposited or flocculated or stratified bed: Formed by allowing
a sediment suspension with high concentration under a low flow
velocity which would permit most of the material to slowly deposit
on the bed. When the flow velocity is zero, the term deposited bed
is used. Such a bed is flocculated and has density stratification
over depth.
c) Compacted bed: Sediment with low moisture content is compacted
with external pressure.
Table 2. Summary of selected studies on cohesive soil erosion.
Investigator Mode of Placement of Sample Mode of Measurement of Erodibility
Lutz (1934)
Peele (1937)
Anderson (1951)
Dunn (1959)
Smerdon and Beasley (1959)
Laflen and Beasley (1960)
Flaxman (1962)
Moore and Masch (1962)
Abdel-Rahman (1964)
Comparison of physical tests with
erosive properties of natural soils
In-place topsoils
In-place topsoils
Remolded, subjected to jet
Slightly recompacted natural soil,
top leveled
Remolded at unspecified percentage
of water, then saturated
Natural soils
Remolded and natural (trimmed) jet
Remolded in duct
Use of qualitative physicochemical
analyses
Soil loss and runoff tables
Correlation of erodibility with
shear measurements
Jet to produce erosion; visual
measures
Visual observation of bed movement
Visual correlation or erosion with
calculated inactive stress
Correlation of permeability and
unconfined compressive strength
with natural erosion (channel
measures)
Measurement of scour depth and
weight loss
Visual; measurement of erosion
depth
Table 2. Continued
Investigator Mode of Placement of Sample Mode of Measurement of Erodibility
Partheniades (1965)
Grissinger (1966)
Masch, Espey, and Moore
(1965)
Mirtskhulava (1966)
Liou (1967)
Liou (1970)
Arulanandan et al. (1973)
Christensen and Das
(1973)
Grissinger (1973)
Sargunam et al. (1973)
Alizadeh (1974)
Remolded natural deposited (salt
water) in duct
Remolded in channel
Unspecified but trimmed as hollow
cylinder
Remolded in flume
Remolded in flume
Remolded in flume
Molded in ring
Remolded in tube
Natural samples remolded in channel
Remolded (compacted) in rotating
cylinder test apparatus
Remolded (compacted) in rotating
cylinder test apparatus
Measurement of suspended sediment
concentrating with time
Rate of erosion by weighing
Weight loss versus rotating shear;
visual correlated with shear
Weight of floc loss
Point-gauge measurement of erosion
depth
Weight comparison
Weight comparison
Rate of erosion by weighing
Weight loss of sample
Weight loss of sample
Table 2. Continued
Investigator Mode of Placement of Sample Mode of Measurement of Erodibility
Kandiah (1974)
Raudkivi and Hutchison
(1974)
Thorn and Parsons (July 1977)
Owen (Nov. 1977)
Gularte et al. (1977)
Gularte (1978)
Remolded (compacted) in a rotating
cylinder test apparatus
Remolded in a recirculating refri-
gerated water tunnel
Deposited in a flume
Deposited in a flume
Remolded in a
gerated water
Remolded in a
gerated water
recirculating refri-
tunnel
recirculating refri-
tunnel
Weight loss of sample
Weight loss of sample before and
after test
Withdrawal of samples, filtering
and weighing or use of photo-
absorptiometer to determine sus-
pension concentration
Withdrawal of samples, filtering
and weighing or use of photo-
absorptiometer to determine sus-
pension concentration
Measurement of suspension con-
centration by using laser-photocell
system
Measurement of suspension con-
centration by using laser-photocell
system
Fukuda (1978)
Deposited Measurement of suspension con-
centration by using laser-photocell
system
Table 2. Continued
Investigator Mode of Placement of Sample Mode of Measurement of Erodibility
Lee (1979) Deposited Measurement of suspension concen-
tration by filtering and weighing
Mehta and Partheniades Deposited Measurement of suspension concen-
(1979) tration by filtering and weighing
Yeh (1979) Deposited Measurement of suspension concen-
tration by filtering and weighing
Thorn and Parsons (1980) Deposited Measurement of suspension concen-
tration by using photo-absorptio-
meter
-45-
Types (a) and (b) mentioned above are usually used in flumes,
whereas type (c) is adopted in rotating cylinder type apparatus.
3.3 Review of Literature Pertinent to the Present Study
It may be noticed from 3.2 above that the literature even on the
erosion aspects of fine sediments is quite extensive. Comparative
study of the results obtained by various research workers may pose
some problem due to the experimental technique involved in these studies.
Details of some of these are given below.
a) Type and size of apparatus used:
Christensen and Das (1973): Rotating circular cylinder: 2.5 cm
dia, 10 cm long
Gularte (1978): Water tunnel: 5.5 m long, 2.0 m wide, 1.5 m high
Yeh (1979): Circular flume: 20 cm wide, 1.5 m centerline dia
Krone (1962): Straight steel flume: 0.9 m wide, 30.5 m long
Partheniades (1962): Straight steel flume: 0.3 m wide, 0.45 m
deep, 18 m long
Owen (1977): Straight flume: 0.3 m wide, 0.2 m deep, 17.6 m long
Thorn and Parsons (1980): Straight flume: 0.3 m wide, 0.2 m
deep, 17.6 m long
b) Method of reproducing shear stress:
Rotating the apparatus in the case of circular cylinders.
Flow of fluid in the case of straight flumes and water tunnels.
Rotation of ring alone or both ring and channel in the case of
circular flumes.
-46-
c) Measurement of erosion:
By photo-electric cell, filtration, and weighing weight loss of
sample before and after test.
Although the importance of several parameters in influencing the
erosion of cohesive sediments has now been established, data in respect
to these are not available for each study. The variation in respect
to fluid used, sediment used, and type of bed have already been men-
tioned under paragraph 3.1. Also there exists considerable variation
in duration of test from 1 minute (Espey, 1963) to 500 hours (Krone,
1962). Sampling time for measurement of concentration is substantially
different from one test to another. Hence, some of the observations
made at small intervals of concentration-time history are not avail-
able in respect to studies where observation of concentration was made
only at long time intervals. Visual observations of erosion permitted
by the transparency of apparatus help in a more realistic interpreta-
tion of data than in the case of an opaque apparatus where erosional
characteristics are indirectly inferred from the observations available.
It is therefore necessary to exercise caution while comparing the
results of various studies.
In view of the implications mentioned above, results of only those
studies where the size of the apparatus was comparable in order of
magnitude, the bed was of a deposited type, and the emphasis was on
the bed shear stress and the bed density are primarily considered here.
Partheniades (1962) conducted erosion tests on San Fransisco Bay
mud using a straight flume. He tested two types of beds, viz. i) placed
bed at natural density and water content and ii) flow deposited bed.
The shear strength of the flow deposited bed was 1/136 to 1/14th of the
-47-
strength of the placed bed. The entire experimental work consisting
of 32 runs was divided into three series.
Series I: Tests on placed bed by changing the flow'velocity by small
positive and negative increment. Results are given in
Fig. 3 (runs 1 to 13 only).
Series II: The same bed was used as for series I except that it was
remolded after the upper surfacing was removed. Results
are given in Fig. 4.
Series III: Tests on flow deposited bed. Results of these tests are
given in Fig. 5.
Comparison of the test results of series I and series III are given
in Fig. 6 in terms of rate of erosion.
Important conclusions drawn by Partheniades (1962) from his erosion
tests were as follows:
i) The rates of erosion were independent of concentration.
ii) The erosion rates for the flocculated bed changed abruptly several
times. This change was proven to be caused by changes of the
bed properties.
iii) The eroded surface did not cause any measurable increase of the
frictional resistance of the bed.
iv) The minimum shear stress to start erosion was about 0.05 N/m2
for both the placed and the flocculated bed, although they had
different densities.
v) Erosion rates for both the beds were of the same order of
magnitude.
vi) The overall resistance to erosion of a cohesive bed is inde-
pendent of the macroscopic shear strength of the bed.
E 12-
z
0 -
I- 10-5 -
I-
u 8-4 O
z
o w
o >
z 6-3 o
u .J CONCENTRATION
S uJ ------ AVERAGE FLOW VELOCITY--
o w .
o w
TIME AFTER START OF SERIES I hrs
Fig. 3 Con tr n er run ruin in run of rup n bed: Parthen run(1
Expt. Series I 13
200 400 600 800 1000 1200 1400
TIME AFTER START OF SERIES I hrs
Fig. 3 Concentration versus time plot obtained in erosion of placed bed: Partheniades (1962),
Expt. Series I
-49-
E -
10
z
o uw
u
0 < -
r- 4
8 6- 0
w
S2- rn run run CONCENTRATION
w _
S------ VELOCITY
00
0 200 400 600 800
TIME AFTER STARTOF SERIES -I,hrs
E
8 -
w*
6-
z
( ) 4
c?.
C )
run 22
0 20 40 60 80
TIME AFTER START OF RUN, hrs
Fig. 4 Concentration versus time plot obtained in erosion of remolded
bed: Partheniades (1962), Expt. Series II
-50-
Concentration versus
Partheniades (1962),.
200 400 600 800
TIME AFTER START OF SERIES-M, hrs
time plot obtained in erosion of deposited bed:
Expt. Series III
O 002 0.04 006
AVERAGE BOTTOM SHEAR STRESS (bs/fl2)
Fig. 6 Relationship between rate of erosion and average bed shear stress:
Partheniades (1962)
Fig. 5
-51-
vii) The erosion rates strongly depend on the average shear stress.
For both dense beds, the erosion rates increase very rapidly for
shear stress greater than 0.478 N/m2 for series I and 1.34 N/m2
for series II.
viii) The observed independence of erosion rates from the macroscopic
shear strength of the clay and the fact that clay gets eroded
at shear stresses which are infinitesimal compared to its
strength suggest that the mechanism of failure of clay particles
by surface erosion is basically different than the mechanism of
failure of clay particles in the interior of the clay mass, when
subjected to shear stresses.
Krone (1962) conducted studies on San Fransisco Bay mud in order
to relate transport and deposition processes to properties of the
sediment. Erosion tests were carried out on flow deposited beds. The
concentration-time data of erosion tests plotted on log-log coordinates
are given in Fig. 7. Concentration as a function of bed shear stress is
given in Fig. 8. Results of a 500 hour long erosion test are given
in Fig. 9. An arithmetic coordinate plot of these data would pro-
duce a curve with a steadily decreasing slope, suggesting an approach
to an equilibrium or steady state concentration. However, the con-
tinuing straight line plot on log-log coordinates as shown in the
figure discourages this suggestion. Krone (1962) found that "the
log-log erosion curves and their slopes were difficult to explain." He
presented a qualitative hypothesis based on interchange between sus-
pended flocs and the bed and on the dependence of erosion on time and
weakly, if at all, on shear (p. 86).
-52-
I10,". I 1 1I .1 I I I I I I I ---- -- I /
..l
~.\,
oo0
TIME AFTER VELOCITY INCREASE, rin
loo0
10000
Fig. 7 Concentration versus time plot: Krone (1962)
SHEAR ON SEDIMENT BED Ir,), dynes/sq nc
Fig. 8 Concentration as a function of bed shear stress: Krone (1962)
BED NO. 7
S, 20 0
-. 0
D07
0 0 v 0.34
*
' ' '
" '"
L
O L ~ .. ...." .I I. | ''1.I ..'---- -- ,.
t '
Z
W
S10
I-
0
0
SC 316 MARE ISLAND STRAIT SEDIMENT U-
St VELOCITY, 1.14 ft/sec
A)
j I
o : OPTICAL DENSITY
0 SUSPENDED SOLIDS
w
a1.
0.1
0.01 0.1 1.0 10 100 1000
(
/ TIME AFTER VELOCITY CHANGE, hr
Fig. 9 Results of 1 500 hour long erosion test: Krone (1962)
I.-- ----- -- '' ,- tJI-- .
-54-
Lee (1979) conducted studies on resuspension and deposition of
Lake Erie sediments, using a circular channel apparatus with an
annular ring similar to the apparatus used for the present studies.
However, in the case of apparatus used by Lee, the shear stress was
produced by rotating the ring alone. The results of concentration-
time curves for series I and II of tests as presented by him are given
in Figs. 10 and 11. The same data have been re-plotted in the form of
concentration under a time-variant bed shear stress (Figs. 12 and 13).
Lee found that the entrainment rate was a strong function of shear stress,
water content, and mineralogy.
Concentration-time data obtained by Yeh (1979) for different
values of bed shear stress are given in Fig. 14, and concentration as a
function of bed shear stress given in Fig. 15.
3.4 Shear Strength of Clay
The classical Coulomb's equation for shear strength of soils is
s = c + p tan ( (3.4.1)
where
s = shear strength
c = cohesion
: = angle of internal friction
p = pressure normal to the failure plane
Since c and are found to depend on the loading rate and drainage
condition, the modified Coulomb's equation is given as follows:
s = ce + o tan e (3.4.2)
where
-55-
-s
SI
00JO
U
100
Fig. 10 Concentration
Lee (1979)
scole
90
70
50-
6 owl
run 3 Tw=6.0
-*- --- .--- run 2 T,=4.3
SERIES I
- --o run I Tw=3.2 dyne/cm2
versus time plots for Series I obtained by
100 200
/ (mi;n
Fig. 11 Concentration versus time plot for Series II obtained by
Lee (1979)
1.0 I I
Reference Lee (1979) Series I 0.60
E
z
0
Bed Shear Stress in N/m2 0.43
z
U
z
S0.5- T AT (AT)ex
0.32
z 0.11 0.34
o 0.43
(n 0.17 0.39
z -0.60
D 0.32
0 I I I i I I -I I
0 2 4 6 8
TIME (hrs)
Fig. 12 Lee's (1979) data re-plotted to indicate variation of suspension concentration as a
function of time and bed shear stress
2 4 6 8 10 12 14
TIME (hrs)
Fig. 13 Lee's data re-plotted to indicate variation of suspension concentration as a
time and bed shear stress
function of
I, I;. r
0 U
0
-58-
- A LV.J I 11/ItI
C'A
z 8
0
S6
z
0
4
2-
0I- IL
0 40 80 120 160 200
TIME (Hours)
Fig. 14 Concentration versus time data obtained by Yeh (1979) for
erosion of kaolinite
-59-
KD = Kaolinite in Distilled Water
z
0
S20-
cr
z
LuJ
C-)
Z
U
O
0
10
00
OO 0.2 0.4
Tb BED SHEAR STRESS (N/m2)
Fig. 15 Concentration as a function of bed shear stress obtained by
Yeh (1979)
-60-
ce = true cohesion, being only a function of the void ratio
of the material (Fig. 16(a))
(e = true angle of internal friction, practically independent
of the void ratio
a = the effective pressure normal to shear plane
For normally consolidated clays, the following expression is
used:
s = Pc tan d (3.4.3)
where
P = consolidation pressure for 100 percent consolidation
6d = angle of drained shear resistance
The magnitude of Pd can be much greater than
the increase of cohesion with increase of consolidation pressure (Fig.
16(a)).
The shearing strength that a clay deposit possesses is related to
the type of clay mineral and the water content but the more important
factor is the stress history, that is, the effective stress or con-
solidation pressure to which the soil has been subjected previously
(McCarthy, 1977). The pore water drainage occurring due to shearing
deformation results in change of shear strength of clay. Three condi-
tions could be considered to be related to the shear strength of soil,
viz. i) unconsolidated-undrained (U-U). (Both time and drainage
which are necessary for consolidation are not permitted.)
ii) consolidated-undrained (C-U). (Consolidation permitted, but
no drainage or volume change permitted during shearing.)
iii) consolidated-drained (C-D). (Drainage and volume change are
permitted.)
-61-
Pe =True Angle of Internal Friction
IPd = Angle of Drained Shear
Resist
(a)
ance
RPfiro nr'. :
Partheniades (1962)
EFFECTIVE NORMAL STRESS
(b)
Reference:
McCarthy (1977)
NORMAL STRESS 0-
C- D = Consolidated, Drained Soil
C-U = Consolidated, Undrained Soil
U-U = Unconsolidated, Undrained Soil
(c)
Reference:
McCarthy (1977)
% WATER CONTENT
-J
Fig. 16 Schematic diagrams showing shear strength of cohesive soil
related to other parameters
~1-
Ce
T
m m I
1- 01
\3'
-62-
A qualitative comparison of shear strength results for these conditions
is shown in Fig. 16(b). Shearing of U-U type is completed relatively
quickly because the prevention of volume change results in development
of excess pore pressures and consequent reduction of the shear strength.
Shearing of C-D type takes place very slowly since both drainage and
the volume change are permitted during shearing.
The shear strength of clay is essentially composed of two com-
ponents: i) physical component due to frictional resistance and inter-
locking between particles, and ii) physico-chemical components due to
the inter-particle attractive and repulsive forces.
The shear strength of clay soil improves with consolidation, pro-
vided that time is available for permitting the necessary pore water
drainage to take place. In effect, consolidation results in decreasing
the water content of the clay with a subsequent increase in shear
strength (Fig. 16(c);McCarthy, 1977, p. 234).
3.5 Shear Strength and Bed Density of Clay
Attempts have been made by several research workers to measure
the bed density and shear strength of clays and to establish correla-
tions between the two parameters.
Ariathurai and Kandiah (1979) have developed an electrical method
to measure in situ sediment densities. Dayal et al. (1980) have developed
a method for obtaining in situ soil strength by use of low velocity
projectile penetration technique. Gularte (1978) measured shear
strength with a fall cone device. He also used a modified viscometer
for this purpose.
-63-
Krone (1962) measured shear strength of the bed by using a screen
penetrometer for consolidation time of 8, 24, 46, 72, 97, 120, 146, 168,
240, and 312 hours, and also measured densities. He'concluded that the
ultimate density appears to be independent of total depth, i.e. con-
solidation occurs independently of the weight of material above a
consolidating layer.
Partheniades (1962) used a conventional vane shear test apparatus
as well as a simple penetration test device for measurement of the shear
strength of soil in connection with the erosion tests conducted on a
flocculated bed and arrived at the following important conclusion
(p. 108): "The observed independence of erosion rates from the macro-
scopic shear strength of the clay and the fact that clay gets eroded
at shear stresses which are infinitesimal compared to its strength
suggests that the mechanism of failure of clay particles by surface
erosion is basically different than the mechanism of failure of clay
particles in the interior of the clay mass, when subjected to shear
stresses." He also observed that the erosion resistance of the floc-
culated bed seemed to increase with depth, and attributed this to the
heterogeneity of the bed which was deposited from a suspension of high
initial concentration containing a wide range of particles from clay
size to fine sand.
Owen (1970) measured the shear strength of the surface layers of
the bed by using a Brookfield viscometer. After measuring the shear
strength of the top layer, that portion of the bed was allowed to spill
slowly. The shear strength of the next layer thus exposed was again
determined by using the viscometer. Samples of the bed were taken
simultaneously and a correlation of shear strength with bed density was
-64-
established (Fig. 17). The tests were carried out in a perspex settling
column 10 meters high and 99 mm internal diameter. The variation of
the density at various depths for different consolidation times is
given in Fig. 18. He concluded that the shear strength variation with-
in the bed could be correlated satisfactorily with the variation of
density and that the shear strength increased rapidly with density.
It has been mentioned earlier that consolidation results in de-
creasing the water content of the clay with a subsequent increase in
shear strength (McCarthy, 1977, p. 234). However, the following
interesting observations have been made regarding the erosion rate as
a function of moisture content:
i) Partheniades (1962, p. 54) noted that "in spite of the lower
overall strength of the bed and its higher water content, the
erosion rates of series II were lower than the corresponding
rates of series I."
ii) Christensen and Das (1973, p. 13) noted the following: "It is
generally assumed that under similar conditions, the rate of
erosion will decrease with increasing density. However the evi-
dence in previous studies has not been conclusive. For this
phase of the laboratory investigation, saturated soil samples
were prepared at varying densities and moisture content and sub-
jected to a constant hydraulic tractive stress. Because the soil
samples were saturated, the density decreases with increasing
moisture content. The duration of the test and the temperature
of the water were kept constant for each type of soil. The
laboratory test results exhibit a sharp decrease in erosion with
increasing moisture content."
-65-
CONCEN-
10 0 ___TRATION SALINfIT7Y fEr
17 mg/1 g/1 m
90 0 16 290 32 9 10 06
15 520 17 0 9-39
SO 17 475 17 8 6 96
i0a 17280 16-7 4 64
7.0 _____ 6 705 2 7 9 73
7-0
SI 7 2d 4 6 9 74
0 0 4 392 a 8 9-76
6.0 10 272 16 8 0 02
a 6 666 33 3 9 72
+_ 6 10 6 974
5-0
4.0 _
3.0 x -
2 -5
SHEAR 2-0
STRENGTH x
N/m2
1.0
09 -
06
0-5 -
0. I 4
60 90 100 150 200 250 300 350 400
DENSITY g/1
Fig. 17 Relationship between shear strength and bed density
observed by Owen (1970)
(A) AFTER 250 MINUTES (B) AFTER 500 MINUTES
10 -I- 1-0-
-0 \
09 09-... -- 09 ..
08 -- ----- ---- ----- 08- -
07 ---- -- ---- --- 0 -- --
06 --- -- 06 .-
0 5 _- -- --- - -- -F g.'l
03 ------- ----- 3 SUS PEND 0
2 -- -- -- -- 02 CONCENIO
01 --- --- 01 SLINIIY:
S 173 g/1
o 0 020 06 08 1.0 I I 14 16 1820 22 24 26 2.8 30 0 0-2 04 06 08 10 12 I1 16 18 20 22 2i 26 28 30
100
08 I -- 7
06 ---- ----- 0 1--06
07 - -- -~- -- 07-----------------------6 4
o 2 03 --
04 --- -- 0 4- -
0 -- ---- -- ------ ----- o0 ------_-_-_-
0.1 _IF 10
0 0
0 02 04 0608 10 12 1 1 1.6 1 8 20 22 24 26 2-8 30 0 02 04 06 08 10 12 1 16 18 20 22 2, 26 28 30
Wp DENSITY/ EAN DENSITY
CALCULATED DENSITY PROFILES FOR VARIOUS BED THICKNESSES
Fig. 18 Bed density profiles: Owen (1970)
-67-
iii) Owen (Nov. 1977, p. 11) conducted studies on erosion of Avonmouth
mud and concluded the following: "In terms of mean shear stress
the onset of continuous erosion is almost simultaneous for mud
beds of different density, but the rate of erosion is greater for
mud beds of lower density."
iv) Thorn and Parsons (July 1977, p. 8) studied properties of Grange-
mouth mud and made the following observation: "There does not
seem to be any strong relationship between bed shear stress and
surface density, although further tests would be needed to estab-
lish this with confidence. This result is rather surprising as
it would seem likely that the thicker or denser the mud the more
resistant it should be to erosion. The surface density at
equilibrium was divided by mean bed density to give a relative
density but this did not give any stronger relationship with bed
shear stress. This is an interesting result because an earlier
investigation of Avonmouth mud showed that both equilibrium sur-
face density and relative equilibrium surface density were
linearly related to bed shear stress."
v) Arulanandan et al. (1980) studied the effect of changing the
density structure of bed by remolding the soil and found that
remolding generally decreased both the critical shear stress and
the rate of change of erosion rate. They also found that the
salt concentration of eroding fluid influenced the erosion of
remolded soil samples. A decrease in salt concentration of eroding
fluid decreased the critical shear stress and increased the rate
of change of erosion rate.
CHAPTER IV
PRESENT INVESTIGATION
4.1 Objective
The parameters and processes influencing the behavior of fine sedi-
ments in contact with water have been described in Chapter II. Also,
the results of important investigations carried out to study the ero-
sional properties of fine sediments have been presented in Chapter III.
It is clear from the presented information that a range of physical and
chemical parameters are necessary for characterizing the properties of
the sediment bed as well as the properties of the eroding and the pore
fluids. When a given sediment is equilibrated with the eroding fluid
over a sufficiently long time, the pore fluid and the eroding fluid
have the same properties. The erosion process is then predominantly
governed by the following parameters:
i) The structure of the bed in terms of its floc shear strength varia-
tion over the depth, which is a function of the type of bed, viz.
placed, deposited, or compacted.
ii) The bed shear stress Te which causes erosion when it has a magni-
tude greater than the critical shear stress for erosion.
The moisture content is an important parameter in the case of the
placed bed and the compacted bed. In the case of the deposited bed,
which is of interest in this study, the process of bed formation is
important, involving the following parameters (Fig. 19):
-68-
Deposition
ITm
H--Step I
Mixing
r- Tm --
Step I and Step I1
=Pre-erosion stress history
STm= Bed Shear Stress for Initial Mixing
Td = Bed Shear Stress for Deposition and
Consolidation of Sediment
Tel Te2, = Bed Shear Stress for Erosion
Step U
and Consolidation
-Step III ----
Erosion
+Ts _+TS H_
etc.
Tm = Duration of Initial Mixing
Td = Duration of Deposition and
Consolidation
Ts = Duration of Time Step
Fig. 19 Definition sketch for notations used to describe experimental conditions
Te3
Te I
Td I
-70-
a) The process of bed formation in the case of deposited bed starts
with an initial concentration of suspension, Co. Under the labora-
tory conditions, the sediment and the eroding fluid are mixed under
a comparatively high shear stress in order to obtain a suspension
with C as its uniform concentration throughout the depth of fluid.
b) The shear stress Tm under which the initial mixing of the sediment
and the eroding fluid are carried out.
c) The duration of mixing, T With a sufficiently long duration of
Tm, the maximum size of the flocs in suspension is controlled by
the balance between the local shear stress and the floc shear
strength.
d) The bed shear stress Td which is sufficiently small in its magnitude
so as to permit deposition of most of the material in suspension.
e) Duration of the total time for deposition plus bed consolidation,
Td, which influences the density of the bed.
If Co is kept constant, the erosion of bed will depend upon the
following two important processes.
Formation of bed: influenced by Tm, Tm, Td, and Td'
Erosion of bed: influenced by Te which may vary in its magnitude and
duration.
Attempts made by previous research workers to directly measure the
shear strength of the bed or to correlate it to the bed density have
not been satisfactory. The overall objective of the present study was
to develop a laboratory test procedure which would enable the deter-
mination of the variation of the shear strength of a deposited bed
over the depth. This was accomplished by increasing the applied bed
shear stress in small increments of selected short time periods.
-71-
Different types of bed structures were formed by using different com-
binations of Tm, Tm, Td, and Td. Concentration of suspended sediment
resulting from the different values of Te was measured as a function of
time.
The term resuspensionn" is usually used in the case of erosion of
a flow deposited bed. Although all the experiments reported under the
present study were for the deposited beds, the terms erosion and resus-
pension are considered to be synonymous.
4.2 Material
Commercially available Kaolinite was used as sediment in the studies.
Size gradation curve for the material is given in Fig. 20. The median
diameter was 1.4 microns. Ninety-five percent of the material was within
the size range of 1 to 7 microns. Seventy-four percent of the material
was finer than 2 microns. The maximum size was 15 microns. This size
distribution was obtained by using "Sedigraph" Particle Size Analyzer.
Before using Kaolinite for conducting tests, it was kept submerged
under the eroding fluid for a period of three months for the purpose of
equilibration.
The pore fluid and the eroding fluid was identical in these studies.
The fluid was prepared by dissolving commercial salt in tap water and
was adjusted to have a concentration of 35 parts per 1000 by weight.
The pH of the eroding fluid was 7.6.
100
z
w
w
<0 50
w
i-"
50
100 50 10 5 I 0.4
EQUIVALENT SPHERICAL DIAMETER, LLm
Fiq. 20 Si7p qradation of Kaolinite tisd for thp experiments
EQUIVALENT SPHERICAL DIAMETER, JLnm
Fiq,. 2(? Si7p qrarlation of Kaolinite nlspd for the experiments
-73-
4.3 Apparatus for Erosion Tests
A system of rotating circular channel and ring was used for
conducting experiments. The annular channel was 20 cm wide, 46 cm deep,
and had a mean radius of 76 cm. The channel was made of 9.5 mm thick
fiberglass. Four windows with a transparent plexiglass were provided
on the channel to permit visual observations. The channel was supported
on a rigid steel frame. An annular ring made of 6 mm thick plexiglass,
having the same mean radius as the channel, was provided within the
channel. The width of ring was smaller by 6 mm than the width of
channel. The ring could be positioned at any required height within
the channel and it could be freely rotated while in contact with the
water surface. Taps were provided on the vertical outer wall of the
channel for obtaining samples of suspension from the channel. Details
of the apparatus assembly are given by Mehta (1973).
Accessary equipment consisted of Millipore Filtering Apparatus,
an oven, and a Mettler balance having 0.05 mg precision. For deter-
mining the concentration of sediment in suspension, the following
procedure was adopted:
i) Obtain a sample of the eroding fluid having sediment in suspension
through the tap provided on the channel and collect it in a
sampling bottle.
ii) Measure volume of the sample.
iii) Filter the sample through pre-weighed Millipore filter paper
discs with 0.45 micron pore diameter.
iv) Wash the salt using distilled water.
v) Dry the filter paper discs in the oven at 600C.
-74-
vi) Weigh the filter papers containing sediment.
vii) Obtain concentration by calculating the weight of sediment from
the difference in the weight of the filter paper with and without
the sediment, and dividing it by the volume of the sample.
A mercury thermometer was used to measure water temperature.
Photographs of the apparatus and accessory equipment are given
in Figs. 21 to 28.
The rotating channel facility was previously calibrated for measure-
ment of bed shear stress, details of which are given by Mehta (1973).
The required bed shear stress could be attained by adjusting the speeds
of rotation of the ring and the channel. Calibration curves used for
this purpose are given in Figs. 29 and 30. The fing and the channel are
rotated in directions opposite to each other in order to minimize the
effects of the radial secondary currents (Mehta, 1973).
4.4 Experimental Procedure
The experimental procedure consisted of the following three
parts:
i) Formation of bed in the rotating channel: Kaolinite equili-
brated with the eroding fluid for a period of 3 months was put in the
channel. The quantities of Kaolinite and the eroding fluid (which was
saline water with 35 parts per 1000 concentration) were adjusted in
order to have a sediment suspension of the order of 40 parts per 1000
concentration by weight when fully mixed (C ). The bed was formed in
the channel by initial mixing and allowing the sediment to deposit on
the channel bottom. Figure 19 schematically indicates the procedure
-75-
Fig. 21 The rotating channel facility
Fig. 22 Close view of the annular channel and the ring
Fig. 22 Close view of the annular channel and the ring
-76-
Fig. 23 The motor controllers
Fig. 23 The motor controllers
,kI
*-- .i- ,a $ '
..o'-. o?
Fig. 24 The electric
motors for the channel and the ring
-77-
Fig. 25 Millipore filter apparatus assembly
Fig. 26 Device for measurement of bed density
-78-
c ~-I
MIU
Fig. 27 Equipment for determining concentration
of sediment suspensions
Fig. 28 Sampling bottles
RING CONTROLLER
15 20
30
35
40
RING SPEED (RPM)
Fig. 29 Operational speeds and controller meter readings for
bed shear stresses
ring and channel at different
30 2
2
2
m
r
5
0
o
r
r
m
15
m
-3r
5 0
z
oZ
METER READING
30- Ring
Sco
- 20 -
10-
I0
5 10 15
REVOLUTIONS PER MINUTE (rpm)
Fiq. 30 Correlation hbtween r.p.m. and mnter reading for the channel and the ring
-81-
followed and the notation used. Initial mixing was carried out at a
shear stress Tm over a mixing duration of Tm. The bed shear stress was
then reduced to Td to permit settlement of the suspended sediment and
form a bed. The time for settling plus consolidation is indicated by
Td in the sketch.
All the tests reported in this study were conducted with a total
depth of 30.5 cm in the channel which was kept constant. The quantity
of individual sample withdrawn from the rotating channel was of the
order 20 c.c. out of a total volume of 300 liters in the channel. With
a suspension concentration of say 2 percent, the amount of sediment
withdrawn at each sample was about 0.4 grams out of a total of 12 kg
of sediment in the channel. The quantity of water and sediment with-
drawn from the channel were small enough not to have any measurable
effect on the fundamental processes taking place during the experiment.
A large quantity of the eroding fluid having properties identical to
that in the channel was kept in stock. This was used to replenish the
eroding fluid by adding small quantities from time to time over the
duration of experiment in order to keep a constant depth. Sediment was
replenished at the mixing stage of the next experiment. The sediment
and the eroding fluid was the same throughout the study. All the tests
were conducted on the deposited bed. No remolding or compaction with
external force was carried out.
ii) Erosion of bed: The bed shear stress (re) varying in its
duration and magnitude was applied by rotating the ring and the channel
in accordance with the calibration curves. The time-step function of
bed shear stress is shown schematically in Fig. 19. For any one experi-
ment, the duration of time step (Ts) was kept constant (such as 30 min,
-82-
60 min, 90 min) and only the magnitude of shear stress was varied.
The values of re were obtained by discretization procedure of a linearly
increasing shear stress or the one equivalent to a sihusoidal velocity
variation, etc. If T and T are the two consecutive magnitudes of
e e
bed shear stress, then excess shear stress is given by
AT = T (4.4.1)
e2 e
The normalized excess shear stress was defined as
T -T
e e
ex T
For example, (Ar)ex = 0.2 represents magnitude of T which is 20 percent
ex e2
greater than the magnitude of T and so forth. Different values of
e1
AT as well as (AT)ex were selected for variation of bed shear stress.
iii) Data analysis: Data collection consisted of obtaining samples
of suspension at pre-determined time intervals after every change of
the bed shear stress. The sampling time used was 1, 2, 3, 5, 10, 15,
20, and 30 minutes in the case of Ts = 30 min. For Ts of longer dura-
tion such as 60 and 90 minutes, additional samples were taken at every
10 minutes after the first 8 samples were collected in 30 minutes.
In order to study the variation of suspension concentration over the
water depth, samples were taken from two locations, viz. Tap A located
125 mm above the channel bottom and Tap B located 225 mm above the
channel bottom.
The concentration of sediment in suspension for each sample was
determined by following the procedure described earlier under section
4.3. The basic data consisted of plotting of a concentration versus
-83-
time graph for each experiment. Further analysis of these data was
used to study the variation in suspension concentration as a function
of bed shear stress and for computation of erosion rates.
Although the room housing the rotating channel assembly was air-
conditioned, facilities to maintain a constant temperature of water in
the channel were not available. Hence, typically a change in the water
temperature of the order of 2 to 30C took place over the duration of
the experiment.
4.5 Apparatus for Measurements of Bed Density
During the course of the present study, two different apparatus
were developed for the following measurements:
(a) Measurement of bed density for sediment deposited under
quiescent conditions (Td = 0). The apparatus developed for this pur-
pose consisted of a 30 cm high, 15 cm dia. polyvinyl chloride cylinder
provided with a bottom plate. Ten plastic tubes of various heights
ranging from 0.95 to 6.35 cm, all having a 0.95 cm inner diameter,
were glued to the bottom plate (Fig. 31). The cylinder was made in
two pieces, the bottom cylinder being 7.5 cm high (photograph in Fig.
26). After placing the 22.5 cm tall piece of cylinder on the bottom
cylinder, the circumferencial joint was sealed with a tape to make it
water-tight. The cylinder was then filled with a sediment suspension
of known concentration. The sediment was allowed to deposit under
quiescent conditions (rd = 0) for the required consolidation time
(Td = 24 hrs, 40 hrs, etc.). The supernatant water was siphoned out,
and the top cylinder was removed after peeling the tape off. All the
-84-
I--Top Cylinder 15cm dia.2 2 I
Plastic Tubes of various heights,
0.95 cm dia. glued to the
bottom plate
Bottom
Cylinder
15cmdia.-
In fl II 1
Bottom Plate
SKETCH OF APPARATUS I
-2 cm dia plastic tube
-15 cm dia. plexiglass cylinder
2.5 cm dia metal tube
-- Annular space for mixture of alcohol
and dry ice
Porcelein
Dish -
Piston with Screw Rod
SKETCH OF APPARATUS TI
Fig. 31 Apparatus developed for measurement of density as a function
of depth for deposited beds
5 cm
225 cm
1
T
15cm
_L
-85-
sediment outside of the tubes was removed. Entire sediment from each
tube was taken out in porcelain dishes by using a hypodermic syringe
with repeated washing by small quantities of distilled water. The weight
of the sediment was determined after evaporating water in an oven at
500C temperature. From the height and diameter, the volume of sedi-
ment in each tube was calculated and by knowing the weight, the density
of sediment in each tube was calculated. Further calculations were
made as follows:
Let L1 and L2 be the heights of two adjacent tubes 1 and 2 with a small
change in heights (of the order of 0.3 cm).
Let p, and P2 be the densities of sediment in each tube, calculated as
above.
Let V1 and V2 be the volumes of sediment in each tube.
Let L2 < L1 and hence V2 < V1.
It was assumed that the bottom sediment of height L2 in tube L1 had the
same density as that of tube 1, viz. pI. The reason p2 is not equal to
p1 is the fact that the sediment contained in the upper portion of tube
L1, viz. in the incremental height (L1 L2), has a different density
(Ap)1-2 which was calculated as follows:
A "pV1 P2V 2 (4.5.1)
plV1 P2V2
(Ap)1-2 V- V (4.5.1)
(b) Measurement of bed density for sediment deposited in the
rotating channel under a low bed shear stress (td = 0.015 N/m2,
0.05 N/m2, etc.). The apparatus consisted of a 2.5 cm dia. metal tube
15 cm high placed concentric in a 15 cm dia., 15 cm high circular
plexiglass cylinder having a sealed bottom. At the center of bottom
-86-
plate 2.5 cm dia. hole was provided in the plate to match with the con-
centric metal pipe, thus leaving the bottom only for the annular space
between the metal tube and the plexiglass canister. After the bed was
formed in the rotating channel under the required conditions of Td and
Td, a transparent plastic pipe of about 2 cm dia. was placed vertically
through the sediment bed over the bottom of rotating channel. The
plexiglass canister was then lowered vertically so as to insert the
plastic tube through the metal tube of the apparatus. The annular
space around the metal tube was filled with commercial grade denatured
alcohol and dry ice was added in pieces to the alcohol. In less than
about 30 minutes this resulted in freezing of the suspension inside
the plastic tube which was then removed and placed horizontally covered
with ice cubes in order to keep it frozen. A piston which could be
activated by a threaded rod was used to push about 4 mm length of
frozen sediment projecting outside the plastic tube. A metal plate
held vertically in contact with the projected portion quickly melted
the frozen sediment which was collected in a porcelain dish. Next, a
5 mm portion of the sediment was then pushed out and the process was
repeated. The density of each 5 mm thick layer could be determined
by knowing the volume and the weight. The freezing resulted in swelling
of the sample and thus increasing the height of the sediment bed in
the tube. The total thickness of frozen sediment was therefore divided
into ten equal parts and the density of each layer was measured which
was taken to be corresponding to the ten parts of the thickness of the
original depth of the bed.
|
RSS
HIDE
