Small-Scale Sounding Tests on Soft Saturated Cohesive Soils

I

I

I

I

,I

I

I

I

I

,J,?t,J(

TU

Delft

I

Technische Universiteit Delft

Faculteit der Civiele Techniek Vakgroep Waterbouwkunde

I

I

I

Small-Scale Sounding Tests

on Soft Saturated Cohesive Soils

I

Thijs van Kessel

report no. 7-96

,

I

I

I

,~i

'

TU Delft

I

Ph.D. student, Hydromechanics Section, Department of Civil Engineering, Delft University

of Technology, P.O. Box 5048, 2600 GA, the Netherlands. Tel. +31 15 278 40 70;Fax +31 152785975; E-mail: T.vanKessel@ct.tudelft.nl

I

I

I

I

I

I

Abstract

I

Knowledgeof the yield strength of subaqueous sediment beds is essential to calculate their sensitivity to erosion and failure. Also for dredging operations yield strength is a parameter of the utmost importance. This report describes measurements of the yield strength profiles of cohesive sediment samples with a sounding instrument developed by the Hydromechanics Laboratory of this university. From the forces acting on a geometry slowly penetrating into a sediment sample the strength profile is calculated. During some experiments also pore and total pressures were measured. The influences of penetration geometry and traversing speed are discussed. The method is shown to be suitable to measure bed strength profiles fairly accurately. Several sediment types with varying consolidation times were tested.

I

I

I

I

I

I

I

I

Contents

1 Introduction 2

I

2 Sounding instrument 5

I

3 Calculation of strength

3.1 Drained and undrained shear strength .

3.2 Shear stress at the shaft . . . . 3.3 Normal vertical stress difference between top and bottom 3.4 Sum of contributions 3.5 Wall effects . . . . 8 8 9 10 13 13

I

4 Experimental results and discussion

4.1 Reproducibility... 4.2 Influence of geometry ... 4.3 Influence of column diameter 4.4 Influence of traversing speed . 4.5 Negative skin friction ...

4.6 Yield strength at sediment-water interface 4.7 Influence of sediment type . . .

4.8 Influence of consolidation time . 4.9 Influence of effectivestress. . . .

4.10 Comparison between balance and pressure transducer results 4.11 Bulk density profiles . . . .

16 16 16 17 19 22 24

25

26

26

28

29

I

5 Conclusions and recommendations 33

I

Bibliography 35 List of Figures 37 List of Tables 38

I

1

I

I

I

I

Chapter

1

Introd uction

I

I

The strength development of freshly deposited sediment beds is important in several respects.

Ifthe strength development is fast, the sediment is unlikely to be re-eroded and will become a permanent part of the subaqueous soil.Itis therefore essential to know the time needed for strength increase compared to the time scale of the fluctuations of the largest forces causing shear stresses at the bed and in the bed. These forces are generally tidal forces and wave-induced forces caused by storms, but in some areas in the world also earthquakes should be taken into account.

Strength of freshly deposited sediment is also important for determining navigable depth,

which is a decisiveparameter for dredging operations. Navigable depth is currently most often defined in terms of bulk density,but a criterion based on strength would be more refined and accurate., as the link between density and strength is not unique. It is strongly related to factors such as sediment type, stress history and pore fluid composition.

Strength development of sediments can partly be explained by consolidation. As the sedi-ment particles settie under the influenceof gravity,pore water is expelled and the volumetrie sediment concentration increases. As aresult, the permeability will gradually decrease and the consolidation process slows down. At the end of this compaction process, all sediment particles support each other, and pore water flow stops, resulting in a hydrostatic pore pres-sure. For a sediment bed consisting of sand this process is relatively fast, as the permeability of sand is high. However,for cohesive sediment beds, which have a low permeability, the consolidation process is relatively slow. For more information about consolidation theory,the reader is referred to the literature (Terzaghi and Peck 1967;Gibson, England, and Hussey 1967;Toorman 1996; Lambe and Whitman 1979).

As a result of the weight of the overlying sediment layers,sediment particles are subjected to an effective vertical stress. They can now resist shear stresses up to a value in the order of the effective vertical stress by partiele interlocking. For cohesive sediments, however, ad-ditional mechanisms for strength development exist. Cohesive sediment particles are small

(about 1 /-Lm), and therefore have a large specific area (Mitchell 1976;De Wit 1992). They mainly consist of day, and have a layered,plate-like structure. Their surfaces are charged,

resulting in electrostatic attractive and repulsive forces. Also Van der Waals attractive forces aresignificant for particles of this size. Except day particles, also organic material is present

I

I

I

I

2

I

I

CHAPTER 1. INTRODUCTION 3

I

in cohesive sediment. The flora and fauna present in it produce, depending on their life cycle,

adhesives like biopolymers which may significantly contribute to sediment strength. Thus by 'cementation' or 'aging' the cohesive sediment may show a gain in strength independent of the vertical effective stress, explaining strength or cohesion of cohesive sediments at zero effective stress. In the upper layer of a freshly deposited eohesive sediment bed, where the vertical effective stress is smal1, the internal cohesive strength of the sediment (generally about 1-100 Pa) may be the major contribution to bed strength. The upper layer is the most relevant for hydraulic engineering problems sueh as erosion and navigable depth.

Measurement of bed strength is not straightforward. The method to be used depends in the fust place on the order of magnitude of the bed strength. Sediment layers that are buried or have been buried generally have a strength in the order of a few kPa and more. They can be tested with classieal geotechnical methods, for example with a triaxial test. Weaker beds can not be tested with this method, as it is intended for much higher stresses. Measuring such small strengths means vacillating at the interface between geotechnics and rheology. Another disadvantage of the triaxial test is that the strength of the sample as a whole is obtained, and not the strength profile within the sample.

Measurement of bed strength is also possible with a vane test, during which the torque needed to displace a vane immersed in the sediment is measured. This torque T can easily be related to strength c via T =7rc(d2h/2

+

d3/6), where d is the overall vane width and h is

the vane immersion depth. By increasing the immersion depth, a strength profile is obtained. However, to measure bed strengths of about 1 Pa accurately, air bearings are essential. This makes the instrument expensive.

Another possibility is the cone penetration test. In this test a fall cone is dropped from a certain height above the sediment bed, and its penetration depth is measured. From this depth the bed strength can be calculated. Zreik, Ladd, and Germaine (1995) analyzed cone penetration in soft marine muds extensively. However, no strength profile is obtained and the shear rates during this short test (approximately 1 s) are high.

Other methods are statie stability tests on inclined planes (Coussot and Boyer 1995) and methods based on statie equilibrium of an immersed body (Nguyen and Boger 1992). With the latter method it is assumed that the buoyant weight of the body is carried by the vertieal component of the yield stress Ty acting over the body surface. For spheres, for example, this leads to a eriterion for incipient mot ion of Ya

=

Ty/gd(ps - p) =2/37r ::::::0.212, where Ya is the gravity yield group, Ps and p the density of the sphere and surrounding material, respectively, and d the sphere diameter. However, onee the mot ion of the body ceases, neither the normal nor the shear stress distributions on the surface are known. Pressure may not be hydrostatic and so the buoyant weight of the sphere may not be relevant; the shear stress acting on the surface may not equal the yield stress everywhere (Chhabra and Uhlherr 1988). Experimental values for Ya in the range of 0.04-0.2 have been reported. When the force exerted on a body moving through a visco-plastic fluid is measured, another problem arises. The flow field around the body can be very complex, as the object may be enclosed by a fluid envelope within which no shearing occurs. The size of this envelope depends on the fluid yield strength and the velocity of the object relative to the fluid and at present cannot be predieted accurately.

Whatever method is used, sampling should be earefully performed. In situ testing is to be

I

I

I

I

I

I

I

I

CHAPTER 1. INTRODUCTION 4

I

preferred, as even ideal sampling without any disturbance affects the external stresses acting on the sample, which may lead to changes in strength. Bad sampling causes the sample to be sheared and the stress history of the material will-at least partially-be lost. In this case the remoulded strength is measured, which is not representative of the in situ strength.

In this report, a method-based on the principle of sounding-is described to measure

in situ the strength profile of a cohesive sediment bed consolidated in the laboratory in a settling column by recording the force exerted on a small rod slowly sinking into the bed. It provides a simple and accurate way of measuring bed strength and is easily achievable in the laboratory. The only instruments needed are an accurate balance to measure the force and a traversing system to control the movement of the sample column with respect to the balance. Penetration geometries are easily made.

In §2 of this report the experimental set-up is presented and explained. In §3the necessary elaborations to obtain bed strength from the measured forceare presented. In §4experimental results are presented and discussed. The sounding instrument has been tested on consolidated beds consisting of both an artificial mud (China clay) and a natural mud (Caland channel mud). In §5 conclusions are drawn and some recommendations are made.

I

I

I

I

I

I

I

I

I

I

I

Chapter

2

Sounding instrument

I

I

The experimental set-up used for the sounding tests consists of a sensitive balance to measure

the force exerted on a geometry penetrating into a submerged sediment layer,and an

auto-matic traversing unit to control the motion of the cylinder containing the sediment layer with respect to the penetration geometry and the balance. A sketch of the experimental set-up is presented in Figure 2.1. The balance used is a Sartorius Research R 200 D, which has a weighing capacity of 205 g and an accuracy of 0.3 mg (3 JLN). The accuracy for weights up to 42 g is 0.05 mg. With this balance accurate measurements of yield stresses as low as 1

Pa are possible with penetration geometries with a surface of only a few cm2. The traversing

unit used is an Elmo motion control unit equipped with a stepping electric motor, resulting

in a velocity range of 0.01 - 10 mm/s and an accuracy of vertical position of 0.1mmo Both

traversing unit and balance are controlled with a PC.

Two plates, a rod and a disk were used as penetration geometries. A sketch of these

geometries is shown in Figure 2.2, their dimensions are given in Table 2.1. By using different geometries, an estimate could be obtained of the relative contributions of the shaft surface and the under-surface of the geometries to the bearing capacity.

The tip of one plate was equipped with a pore and a total pressure transducer (Druck

DPCR 81, used with a Druck DPI 260 pressure indicator, range 0-75 mbar). From the

pressure recordings the effectivestress profile in the sediment bed can be evaluated.

Measurements were performed by slowly immersing the geometry into the sediment and recording the balance output and, if measured, the total and pore pressures at the tip of the geometry. The diameter of the columns used was 8 cm. Traversing speeds were in the order of 0.1 mm/s for most experiments. In some experiments the traversing speed was varied in order to investigate its influence on the force response. At the start of each experiment the geometry was positioned in the cylinder with its tip above the interface between water and

sediment. The top of the geometry was also immersed into the water in order to minimize

the change in water level, now only resulting from the immersion of the suspension wire. The experiments were stopped when the tip of the geometry reached the bottom of the column. In some experiments also weight changes during the subsequent traversing in reversed direction were recorded. After each experiment the bed concentration profile was measured with a conductivity probe.

I

I

I

I

5

I

I

CHAPTER

2.

SOUNDING INSTRUMENT

6

I

-balance

-

--

-/1

tr ~ -water

I

I

sediment aversing unit motor

I

Figure 2.1: Experimental set-up

I

Sediments tested extensively with the sounding instrument are China clay,an artificial clay mainly consisting of kaolinite, and Caland channel mud, a natural mud dredged from the harbour of Rotterdam. The properties of these sediments are described in De Wit (1995). Initial concentrations prior to sedimentationj consolidation ranged between 60 and 275 kg m-3 by mass; the columns were filled with sediment suspension up to a height of 22-30

cm. China clay suspensions were prepared by adding tap water with 0.5% NaCI to the dry clay powder and subsequent mixing. Caland channel mud was available as wet, but highly concentrated dredged material; suspensions were prepared by adding water also originating from the Caland channel to the concentrated slurry and subsequent mixing.

During one experiment also mud sampled at Nieuw-Statenzijl in the Dollard estuary was tested. At present its composition is unknown. It was tested without any mixing after sampling. The sample was allowed to consolidate for 6 months.

I

I

I

I

CHAPTER 2. SOUNDING INSTRUMENT

I

I

I

7 L D

-

--D

I

rod B plate disk

Figure 2.2: Sketch of geometries used

I

I

Table 2.1: Dimensions of geometries used (Figure 2.2); plate 2 is fitted with total and pore

pressure transducers

---~~~--~~~----~~----~~---geometry D (mm) B (mm) L (mm) M (g)

I

rod 5 150 25.26 plate 1 1 20 150 19.22 plate 2 10 21 disk 20 150 84.8 3 7.71

I

I

I

I

I

Chapter

3

Calculation of strength

I

_{In this section the calculation of the strength profile from the weight changes as a function} of the vertical coordinate z recorded with the balance are discussed. Weight changes can be attributed to three effects:

1. Changes in shear stress at the shaft of the geometry.

2. Changes in the difference between vertical normal stress at the bottom and the top of the geometry.

3. The increase in water level in the column resulting from the gradual immersion of the suspension wire.

The weight change contribution of the wire .ó.Ww resulting from an upward vertical movement of the column .ó.z can be easily calculated from

I

I

7r 2 (

1)

-.ó.Ww =4_dwPwg.ó.z 1

+

2

(t)

-1

(3

.

1)

I

where dw is the diameter of the wire, de the diameter of the column and Pw the density ofwater, which may contain some'dissolved salt. The weight change .ó.Ww is generally small compared to the other two contributions. These will be discussed in two followingsubsections. First the definition of shear strength will be discussed,and the difference between undrained and drained shear strength.

I

3.1

Drained and undrained shear strength

I

The shear strength, yield strength or yield stress of a material is defined as the smallest shear stress at which the material fails, i.e. the stress at which interpartiele bonds are broken up. A failure criterion that is often adopted because of its simplicity is the Mom-Coulomb criterion,

in which the shear strength Ty of the material isassumed to be dependent on its cohesion c,

normal stress (Jn and an angle of internal friction <p via

Ty =C

+

(Jntan<p

(3

.

2)

8

I

I

CHAPTER 3. CALCULATION OF STRENG TH 9

I

The cohesion of materials that mainly consist of sand is low and may generally be neglected

compared to the contribution of internal friction. However, for cohesive materials such as

clay, cohesion is important by definition. Under saturated and undrained conditions, which

are likely to occur at short time-scales because of the lowpermeability of sediment layers that consist of clay,an increase in normal stress does not result in an increase in yield strength,

as the normal stress increase is absorbed by the pore water and not by the grain structure.

Thus

<p

=

0 for saturated clays under undrained conditions, resulting inTy

=

c. The material

behaves purely plastic in this case. Under drained conditions however,when the pore water

pressures are allowed to dissipate during a measurement,an approach in terms of total stress is not allowed any more. In this case a failure criterion in terms of effective stress is more appropriate, changing the Mohr-Coulomb criterion into Ty =

d

+

O"~ tan</>',where primes refer to effective stress. Now

<p'

f.

0 also in the case of cohesive sediments. In the following section an undrained condition during the sounding test is assumed. Also yield strength

measurements with a vane, which can be used to compare with the sounding experiments,

usually take place under undrained conditions. Total and pore water pressure measurements

during a sounding test can validate the total stress approach. If necessary, an analysis in

terms of effectivestress then is possible also.

In order to asses whether a test is drained or undrained, the Fourier number Fo=

Dt/l

2

has to be known, where D is the pore water diffusion coefficient (related to permeability),

l the bed height and t time. If Fo

«

1 then the test is undrained, if Fo

»

1 then the test is drained. If Fo ~ 1 then partially drained conditions prevail. A simple rule of thumb is that the test is undrained if the test time is much shorter than the consolidation time of the sample, otherwise it is drained or at least partially drained.

I

I

I

I

3.2

Shear stress at the shaft

I

Let, at a certain time t during an experiment, a portion .D..z of the geometry be immersed in the sediment bed and a portion L -.D..z in the overlayingwater, where L is the length of the

geometry and

z

=0 the interface between sediment and water. The force .D..Ws exerted at

the shaft of the prismatic geometry is then given by

I

.D..Ws =fot.z OsO!Ty(Z)

(1

+

Ac ~s

AJ

dz

where Ac and As are the crossectional areas of the cylinder and shaft, respectively,Os is the perimeter length of the shaft, Ty(Z) is the local bed strength or yield strength and O!the shaft

adhesion factor accounting for wall slip. For an undrained test the yield strengths equals the cohesion of the sediment:

(3.3)

ïv =C,

whereas for a slow,drained test the yield stress is given by:

(3.4)

Ty =c'

+

K

O"v,eff tan

<p'

,

(3.5)

I

where O"v,eff is the vertical effective normal stress and K the ratio between horizontal and

vertical effectivenormal stress: K =(O"h,eff/O"v,eff).

I

CHAPTER 3. CALCULATION OF STRENG TH

10

I

The value of a may range from 0.3 to 1;for very weak cohesive soil samples it is likely to be close to unity (poulos 1981). By repeating a sounding test several times with geometries with different shaft areas, the value of a can be estimated. Unless stated otherwise, it is assumed to be unity throughout this report.

In (3.3) the viscous shear stress exerted by water on the shaft has been neglected. For the typical values of L).z

=

0.05 m, Ty

=

10 Pa, L

=

0.15 m, traversing speed Ut

=

0.1 uuuf«,

de

=

0.08 m and ds

=

0.05 m, the force resulting from this shear stress amounts to 2 nN,

whereas the force calculated with (3.3) amounts to 10 J.LN, 5000 times larger.

The sign ofL).Ws needs some special attention. If the geometry is moving downward,the shear stress will act upward only if the settling or consolidation velocity is smaller than the traversing speed. Otherwise the shear stress may act downward, which is known as 'negative skin friction'. If experiments are started with the tip of the geometry above the interface between sediment and water, and the traversing speed is kept constant, negative skin friction is impossible, as the interface will never be reached as long as the traversing speed is smaller than the consolidation speed. However,if the geometry is halted at a certain position inside the sediment bed, negative skin friction may develop. In this case, the value of the shear stress is uncertain and may lie between plus and minus the yield strength. If the geometry is moving upward the shear stress will always act downward.

I

I

I

3.3

Normal vertical stress difference between top and bottom

I

Normal vertical stresses act on the horizontal surfaces at the top and the bottom of the geometry. If the traversing speedUt is sufficiently low, dynamic forces will be small compared

to static forces. At Ut = 0.1 mm/s, for example, the dynamic pressure is 10-5 Pa, at least 5

orders of magnitude smaller than the yield strength of the materials tested.

Thus, at the top of the geometry,which is always in the water layer overlying the sediment bed, a hydrostatic pressure can be assumed. At the tip, however, also the presence of the bed material contributes to the normal stresses acting on the tip surface. At low traversing speed, these forces will equal the bearing capacity of the footing.

Using plasticity theory, Prandtl derived an exact expression for the bearing capacity qf

of a strip footing of infinite length on the surface of a semi-infinite, isotropie weightless soil for the undrained condition (4J= 0), given byqf

=

(2

+

7r)c::::::5.14c, where qf is the bearing capacity expressed in Pa (Figure 3.1). For the drained condition (4J=1= 0) it is necessary to consider a surcharge pressure qoacting on the soil. The solut ion for this case, due to Prandtl and Reissner, is (Lambe and Whitman 1979),

cexp(7rtan4J)tan2

(i

7r+~4J)-1

qf = ---_,__-~-'---~

tan 4J

+

qoexp (7rtan 4J)tan2

(i

7r+~4J)

(3.6)

I

I

I

which for 4J~ 0 also results inqf =(2

+

7r)c.

However,an additional term must be added to take into account the component ofbearing capacity due to the self-weight of the soil, which can only be approximated by numerical or graphical means. If the footing is not located on the surface of the soil,but at a depth

z

below

(a) (l) (e)

--(4) p,.ndII Doe-___ nd ...,_ H1M_nd Rol... JÓIIy Y-

-Caqual M.,tIoI vnié Skompton,Y_uin.

Buisrnan ...Gi_

TerucN

I

CHAPTER

3.

CALCULATION

OF

STRENG

TH

I

q

I

I

11

Figure 3.1: Assumed failure patterns under deep foundations

the surface, a third term has to be taken into account representing the surcharge pressure. For drained conditions (</J

#

0), the bearing capacity will also be influenced by </J.Based on these considerations, Terzaghi developed a general expression for the bearing capacity of the soil under a shallow footing given by (Lambe and Whitman 1979),

where

B

is the breadth of the footing, 'Y the specific volumetrie weight given in this specific case by 'Y = P - Pw9 with P the bulk density of the soil and z the vertical coordinate. Ny,

Ne and Nq are dimensionless bearing capacity factors depending only on <iJ and representing contributions resulting from the self-weightof the soil,the component c of shear strength and surcharge pressure respectively, and S"p Seand Sq are dimensionless shape factors depending only on the footing geometry used. It should be borne in mind that the superposition of the components of bearing capacity is theoretically incorrect for plast ie materials; however,any resulting error is considered to be on the safe side, thus underestimating the actual bearing capacity. Values for these shape factors for several geometries are given in Table 3.1. For high values of </Jthey are considered to give conservative values for qf' Alternative proposals for shape factors have been presented by De Beer (1970) and Hansen (1968).

Bearing capacity factors based on the Prandtl-Reissner solution are given by

I

I

I

I

_N

q =exp (71" tan </J)tan2

(45

0

+

~</J)

N. _ Nq-1

e - tan</J

(3.7)

(3.8)

(3.9)

I

The most widelMeyerhof(1963)y: used values for the factor N-y are those obtained by Hansen (1968)and

Ny = 1.80(Nq - 1)tan </J

I

I

CHAPTER

3.

CALCULATION

OF

STRENG

TH 12

I

Table 3.1: Values for shape factors in (2) proposed by Terzaghi and Peck (1967)

I

geometry s"'( Sc Sq strip 1 1 1 square 0.8 1.2 1 circle 0.6 1.2 1 rectangle 1- 0.2(BjL) 1 + 0.2(BjL) 1 N"'( =(Nq - 1) tan (IA</» (3.11)

I

bearing capacity theory, which will lead to errors if compressibility effects are important (VanIt is important to notice that the effect of soil compressibility is not incorporated in the

den Berg 1994). An analysis based on the cavity expansion theory is then more appropriate. Throughout this report the sediment bed is considered to be incompressible. For saturated sediments (no air!) under undrained conditions this is a very good approximation. However, for slow, drained tests sediment particles may get more closely packed under the weight of the penetrating geometry and compressibility effects may become important, depending on the surface/volume ratio of the geometry.

I

U ndrained

test

I

In the case of an undrained test, the duration of which is too short for dissipation of the excess pore pressures generated by the penetration of the geometry into the sediment bed, an analysis in terms of total stress is allowed. Now (</> =0), N"'(=0, Nq = 1 and Ne =5.14.

This is in agreement with the original Prandtl solut ion and leads to

qf =eSeNe

+

foz "(dz (3.12)

I

According to Skempton (1951), the shape factorratio of the footing. Skempton proposed for undrained conditionsSc is also dependent on the depthjbreadth

I

Sc = (1 + 0.2~) (1 + 0.2~) if ~

<

2.5

Sc =1.5 (1 +0.2~) if ~

>

2.5.

where

B

<

L

with

L

the length ofthe footing. The values proposed by Skempton are generally accepted and give as a simple relation between undrained shear strength and bearing capacity

qf =ge

+ J~

"(dz for shallow circular footings and (zjB)

>

2.5.

(3.13)

I

I

CHAPTER 3. CALCULATION OF STRENG TH 13

I

Drained test

I

In the case of a slow,drained test, during which dissipation of excesspore pressures is allowed,

an analysis in terms of effectivestress is necessary. Equation (3.7) may then be rewritten as: qf =~')'Bs"{N"{

+

c'seNe

+

sqNq foz ')'dz (3.14)

I

Values for N"{, Ne and Nq are dependent on

cP'

according to (3.10), (3.9) and (3.8). The value for

cP'

followsfrom the experiments, so an iteration procedure to determine the strength profile may be necessary.

The assumption of incompressibility is doubtful in the drained case. The freshly deposited,

uncompacted grain skeleton is easily compressed by the penetrating geometry if pore water is allowed to flowfreely and no excess pore pressures develop. The sediment below the geometry

may consolidate under the extra load of the geometry,and its strength will increase. This

strength is no longer representative for the strength of the original, undisturbed sample.

I

Partially drained test

If excess pore pressures are only partially dissipated, (3.14) changes into:

(3.15) where aw,exc is the excess pore pressure and aeff the effective stress, which can be derived from the experimental pore and total pressure data.

I

3.4

Sum of contributions

I

Now that all factors contributing to the weight change of a geometry penetrating into a submerged sediment layer are quantified, it is possible to calculate the strength of this layer. From the pore pressure data it can be assessed which analysis has to be applied. For sediments

mainly consisting of clay the undrained analysis is generally the most suitable. The weight

change of the geometry ~ Wexp simply is the sum of the factors considered:

(3.16)

I

where As is the base area of the geometry and ~Ww is given by (3.1), ~Ws by (3.3) and qf

by (3.12), (3.14) or (3.15). If evaluation of (3.16) does not lead to an explicit expression for Ty, it must be obtained iteratively.

3.5

Wall effects

I

The diameter of the settling columns used for the sounding test should be sufficiently large to avoid wall effects. A first prerequisite is that the increase in fluid level because of the immersion of the geometry should be small compared to the total fluid volume in de cylinder,

which for rods results in the criterion d~~z/d?;_hw

«

1,where ~z is the length ofthe immersed

I

CHAPTER 3. CALCULATION OF STRENGTH 14

I

part of the rod and hw is the height of the Huid in the cylinder. The ratio should be preferably

smaller than 0.01, which in the limiting case ~z = hw results in dsjde

<

0.1.

From a geotechnical point of view this criterion is certainly sufficient, as can be concluded from an analysis of the size of pressure bulbs under and next to a geometry in a sediment

bed (Smith 1982). An important result from this analysis is that the determination of the

shear strength of a material within two geometry diameters from the bottom of the cylinder is inaccurate, as the pressure distribution then will be influenced by the bottom.

Atapattu et al. (1986) studied the effect of cylindrical walls on the creeping terminal

velocity Vd of spheres in viscoplastic media and presented the following correlation for a wall factor

f

= vdjv, where v is the terminal fall velocity for an infinite cylinder diameter:

I

f

=1 d (d)

I

-<

de

-

de crit = 1 -1.7

(t -

(t)crit)

_{~ >}

_(~)crit

f

(3.17) (~) crit = 0.056

+

3.32YG 0.0091 ~ YG ~ 0.053

I

where d is the sphere diameter and YG = Tyjgd(ps - Pw) is the gravity yield group. From

(3.17) it can be concluded that for fluids with a yield strength of 20 Pa, df d; must be smaller than 0.2 if an iron sphere is used. For objects that are supported by a wire to control their fall velocity, djde can even be larger as the gravity yield group effectively increases. Thus this analysis leads to a criterion for the maximal value ofdj de that is less strict than the criterion based on Huid level increase. Numerically calculated velocity distributions for creeping mot ion of a sphere in a Bingham viscoplastic medium are reproduced in Figure 3.2.

I

I

I

I

I

I

CHAPTER 3. CALCULATION OF STRENGTH 15

I

I

I

1·5,...---'---r---r---, n=0·1

I

[rochet et al (45) Beris et al (36) u o Q;0·5 >

I

..

:: " e N

...

10 Radial Position (- )

I

Figure 3Bingham medium (Beris et al. 1985).2: Numerically calculated velocitydistributions forcreeping mot ion of a sphere in a

I

I

I

I

I

I

Chapter

4

Experimental

results and discussion

I

I

In this section experimental results are presentedand discussed. First the reproducibility of the measurements isdealt with, then the influenceof the penetration geometry used and the influence ofthe sedimentation column diameter, traversing speed,sediment type and sediment consolidation time. The yield stress is calculated from the weight change as described in the previous section. Unless otherwise indicated, undrained behaviour has been assumed.

4

.

1

Reproducibility

I

I

In Figure 4.1 three test on identically prepared beds are presented. A plate geometry has been used for these tests. The beds were prepared from sedimentation of a suspension of China day in tap water (with 0.5% NaCI) with a concentration of 275 kg m-3 and an initial

height of 0.07 m. The consolidation time was approximately 3 days. Agreement is fair, the measurements are reproducible within an error band of approximately 2 Pa for the present experiments. If one bed is tested several times, reproducibility is only good if the geometry penetrates at a position in the bed that is not disturbed by the previous experiment(s), as bed strength is negatively affected by disturbance of the sample.

4.2

Influence of geometry

I

In Figure 4.2 experimental results for the geometries discussed in §2 are shown. The beds were prepared from sedimentation of a suspension of China day in tap water (with 0.5% NaCI) with a concentration of 275 kg m-3 and an initial height of 0.23ID. The consolidation

time was approximately 10 days.

It is clear that with the bearing capacity theory a good overlap between results is obtained, even without adapting the coefficients for shaft adhesionCl! and cone penetration Ne, which

were set at the theoretically predicted values 1.0 and 5.14, respectively.The error band is not larger than for duplicate measurements with the same geometry,thus the testing method is found to be insensitive to geometrical variations.

I

16

I

CHAPTER4. EXPERIMENT AL RESULTS AND DISCUSSION 17

I

0 expo 1 0.005 _expo2

+-expo 3 -G-0.01

I

S

_"-" 0.015_0.02

_+-.

N ':of.. 0.025 '~_~, "k 0.03 ~

'-t-I

0.035 ,,-i'" 0.04 0 2 4 6 8 10 12 14 Ty (Pa)

I

Figure 4.1: Reproducibility of results

I

The disk is only suitable to determine the yield strength at the water-sediment interface properly, as its resistance is so high that it gets stuck befare it reaches the bottom of the sediment bed in all but the weakest samples. Moreover, its nonprismatic geometry makes exact calculations difficult at

z

greater than about 0.5 cm.

Results for the plate equipped with pressure transducers is not shown in Figure 4.2. As the volume of this geometry is not very small with respect to the bed volume, deviation from the results obtained with the other geometries is to be expected. Also wall effects may become important. However,this geometry is primarily intended to compare and relate pressure data with strength. Whether this is the strength of an undisturbed or partially disturbed bed is not our first interest.

I

4.3

Influence of column diameter

I

The influence of the column diameter was investigated by comparing the results obtained with the standard settling column with a diameter of 8 cm to results obtained with a 14 cm diameter column. Sediment beds were prepared in an identical way from sedimentation of a suspension of China clay in tap water (with 0.5% NaCI) with a concentration of 275 kg m-3

and an initial height of 0.17 m. The consolidation time was approximately 7 days.

As was to be expected from the criteria described in §3.5,no significant difference is found for the plate geometry used. Wall effects may therefore be neglected for this geometry. For the ot her geometries with approximately the same dimensions this will also be true, except for the plate equipped with pressure transducers, as was explained in the previous section.

I

I

CHA

P

TER

4.

EX

P

ERlMENTAL RESULTS

AND

DISCUSSI

O

N

I

₀ 0.02 0.04 .-El _0.06

----

_~ 0.08 0.1 0.12 0 rod ~ plate

+-disk

-G--I

I

10 20

25

5

15

I

Figure 4.2: Yield strength profiles fTy

(Pa)

or different geometries

I

_0.01

o ~---~---~---,---~---~

0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 L-- ...__ ....__ ..._ _._ _'

o

column -beaker -- -\

--~.

I

-

-

-

... , ..._,

_,

'-

_-

---

-,

-

-;:.:~

---_

...

I

5 10

15

20

25

Ty

(Pa)

Figure 4.3: Influence of column diameter

I

I

I

CHAPTER

4.

EXPERIMENTAL RESULTS AND DISCUSSION

19

I

-0.05 0.001

mm/s ~

0.002

mm/s

+-0 0.02

mm/s

-G--0.001

mm/s

0.05_0.1

mm/s

_mm/s

·

x..

·

-A-. 0.2

mm/s

,,*-I

..--.. 1

mm/s

+-8

0.05

mm/s

-+-_.. ~ 0.1

I

0.15 0 10 20 30 40 50 60 70 -~W (g)

I

Figure 4.4: Influence of traversing speed on weight changes for China clay

4.4

Influence of traversing

speed

I

Traversing speed is an essential parameter in the sounding process. If it is very high, the viscous contribution to weight changes may be important (§3.2), if it is very small, the test may be drained or partially drained instead of undrained. The lat ter effect can be assessed with pore pressure measurements.

In Figure 4.4 the influence of the traversing speedUt on the forces acting on the geometry

during a sounding test is shown. Traversing speeds ranged from 10-3 to 10 mm S-1. The

beds were prepared from sedimentation of a suspension of China clay in tap water (with 0.5% NaCI) with a concentration of 275 kg m-3 and an initial height of 0.23 m. The consolidation

time was approximately 7 days. The plate geometry equipped with pressure transducers was used for these experiments in order to be able to discern between drained and undrained tests.

From Figure 4.4 is clear that the force decreases with increasing traversing speed at Ut

<

0.02

mm/s.

In the range 0.1

<

Ut =1

mm/s

the forceis virtually independent ofthe traversing

speed, whereas at Ut

>

1

mm/s

the force slightly decreases with traversing speed again.

The decrease for Ut

<

0.02

mm/s

can be explained by the transit ion from drained to

undrained behaviour. This can be concluded from Figure 4.5, where the excess pore pressure is shown for sounding tests at several traversing speeds. The excess pore pressurePexc is found by subtracting the hydrostatic pressurePhydr = Pw9z from the measured pore pressure at the tip of the geometry. The interface between sediment bed and overlaying water is defined as reference levelz =0,wherePhydr

=

Pexc

=

O. From Figure 4.5 it is clear that the pore pressure is nearly hydrostatic at Ut

=

10-3

mm/s

,

the sounding test is therefore drained. In this case

the yield strength is calculated from the force with the effective stress analysis presented in

§3.3. At Ut

>

0.1

romls

the excess pore pressure is independent oftraversing speed and the

I

I

I

I

CHAPTER

4.

EXPERIMENTAL RESULTS AND DISCUSSION

20

I

-0.05 0.001 mm/s ~ 0.002 mmfs

+-0 0.02 mmfs0.05 mmfs -G--

.*

..

.

.

0.1 mmfs -A-. 0.2 mmfs

",*-I

..- 1 mmfs ~-El 0.05 2 mmfs

-+

-..__. N

. *-~:-x.

_{S_ ~.}

.

_'

~

_{-,_.} 0.1 ~--~ . X A.'A. [!J ~X.... ...

4-I

I:]-EJG-_...~ X~,-'-0 .&... 0.15 --"8

*

-0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 aexc (kPa)

I

Figure 4.5: Excess pore pressure as a function of traversing speed for China clay

I

sounding test may be assumed to be fully undrained for the samples considered. An analysis in terms of total stress (§3.3) is now allowed.

The much higher strength in the drained experiment compared to the undrained

experi-ments is caused by consolidation effects in the bed under the load at the tip of the geometry.

In the drained case excess pore pressures are allowed to dissipate, which causes an increase in effective stress, as is illustrated in Figure 4.6. Thus in the drained case the sample is much

affected by the measurement and no

in situ

strength is obtained. At relatively high traversing

speeds, the geometry reaches the bottom of the cylinder, whereas it gets stuck at very low

traversing speeds, when the weight of the geometry is fully supported by the sediment bed

and the wire connecting the geometry with the balance is unloaded. For Caland channel

mud approximately the same behaviour was observed (Figure 4.7),although the bed is much

stronger and more cohesive.

In Figure 4.8 the influence of the traversing speed is presented for China clay tested with

the plate geometry without pressure transducers, which is much smaller and therefore less

disturbing. The range in traversing speed is limited between 10-2 and 1 mm/s. Results are

virtually speed-independent betweenUt =0.1 and 1 mm/s. The slightly increased penetration

resistance at Ut =1 mm/s may be explained by the viscous contribution which is not negligible

anymore. Therefore the optimal traversing speed for the samples used in this study is 1 mm/s, as it allows fast tests without compromise.

I

...

I

I

I

GHAPTER

4.

EXPERIMENT AL RESULTS AND DISGUSSION

I

_-0.05 0.001

mmfs ~

0.002 mmZs

+-0 0.05 mmZs0.02 mmZs -B--.~ ....

I

0.001 mm/s 0.1 mmZs -8-. 0.2 mmZs

-*-1 mmZs

~-s-

0.05 2

mm/s

-+---

~

I

0.1 0.15

o

0.5 1 1.5 2 2.5 3

3

.

5

21 4

I

·

Figure 4.6: Effective vertical stress during an experiment with China clayl1v,O (kPa)

I

-0.04 -0.02 0 ~ 0.02

8

--

~ 0.04 0.06 0.08 0.1 0 2.5 . 10-4

mm/s ~

0.001

mm/s

+-0.005

mm/s

-B--0.02

mmfs

·x

·

·

·

0.5

mm/s

-8-.

I

I

50 60 70 10 20 30 40 80

-.6.W

(g)

Figure 4.7: Infiuence of traversing speed on weight changes for Caland channel mud

I

I

I

CHAPTER

4.

EXPERIMENT AL RESULTS AND DISCUSSION

22

I

I

0 0.02 0.04 ..-0.06

S

--

~ 0.08 0.1 0.12 0 0.01

rnmfs

+-0.03 mm/s

+-0.1 rnmls -B--0.3 mm/s .~.... 1

rnmfs

-8-.

I

I

4 -DoW (g)

Figure 4.8: Influence of traversing speed on weight changes for China clay; focussed on traversing speeds of 0.01-1 rnm/s

1 2 3 5 6 7 8

4.5

Negative skin friction

I

The effect of negative skin friction can easily be studied by reversing the direction of the traversing speed. The force acting on the shaft of the geometry now reverses in direction. If the contribution of the force at the tip is small, which is the case for the plate geometry, the sum of forces for the upward and downward movement should approximately be zero. This is shown in Figure 4.9, where the force response to traversing in both directions is presented. The deviation from zero represents two times the tip resistance, as it acts in the same direct ion for both traversing directions. In this way tip resistance and shaft resistance can be determined separately.

Skin friction at the shaft can also be used to obtain a time-scale for partiele settling and stress relaxation. Just prior to the moment movement of the geometry into the sediment is stopped, the shear stress will act in the upward direction. However, because of stress relaxation and partiele sedimentat ion, the shear stress at the shaft will decrease and may even become negative, depending on the ratio between shaft and tip area. This is illustrated in Figure 4.10, where the weight changes after cessation of motion at t

=

1000 s (z

=

0.10 m) of a plate inside a bed of China clay are displayed. The response is fitted with the function

~W

=

a

+

bexp (-c(t - to)), where a

=

1.93 g,b

=

-3.587 g, c

=

0.0020 s-1 and to

=

1000 s.

I

I

I

I

CHAPTER 4. EXPERIMENT AL RESULTS AND DISCUSSION

o

~W

(g)

Figure 4.9: Skin friction, 'up' and 'down' refer to the mot ion of the plate geometry relative to the settling column

1500

t (s)

Figure 4.10: Relaxation of stress at shaft after cessation of mot ion at

t

=1000 s

I

-0.02 -0.01 0 0.01 ..--. _0.02

S

--

~ 0.03 0.04 0.05 0.06 0.07

I

I

I

I

,

-4 -3 -2 -1 bil

--

_S

0 <l 1 2 3 4 0

I

I

I

I

23 up --down ---up

+

down --- -~ /

.

.

/ ' : I ,. I I •. I ' . I I : / / / / /

"

" ",'/ ,," " / I / / / / / ,I " ,/

·

.

·

.

·

:

·

.

.

.

..

"

-2 -1 1 2 experiment exponential fit \

,

,\

\

_...

')0.

'\

,

________________________________ )o.~ _ 500 1000 2000 2500 3000

I

CHAPTER 4. EXPERIMENTAL RESULTS AND DISCUSSION 24

I

Table 4.1:Yield strength at sediment-water interface for China clay geometry Ty (Pa) rod

5.3

rod

5.9

I

diskdisk

3.7

3.6

average

4.6

CS rheometry

,...,4

I

4.6

Yield strength

at sediment-water

interface

I

Measurement of the yield strength at the interface between sediment and water is important as it determines the true cohesion of the material at zero effective stress. It is also essential if sounding test results are to be compared with rheological measurements. However, the accuracy of the sounding test at the sediment-water interface is less than at a lower level inside the sediment bed, owing to the followingeffects:

I

• The yield strength at the interface is small for freshly deposited layers, especially for China clay. Absolute errors of 1 Pa, in the order of magnitude, then are significant.

• The density of the bed at the interface, which has to be known to be able to calculate yield strength from weight changes, is difiicult to determine with a conductivity probe.

• Surface tension forces are not accounted for.

I

In Table 4.1 the interface yield strength of China clay after 26 days consolidation from a suspension with an initial concentration of 275 kg m-3and an initial height of 0.22 m is

presented in duplicate for the rod and disk geometries. The plate geometry is unsuitable for interface measurements, as its tip area is very small. Agreement between the duplicate experiments is satisfactorily. However,a significant deviation between plate and disk geome-tries can be observed, which may be a consequence of the last two points itemized above. Agreement between these results and controlled stress (CS) rheological measurements is sur-prisingly good. The results for the disk geometry are more reliable than for the rod geometry,

as the tip area of the former is much larger.

It is desirable to compare sounding test results with rheological measurements, not only at the interface but also inside the bed where the effective stress is not equal to zero. With rheological measurements the yield strength ismeasured directly,whereas with the sounding test the yield stress has to be calculated from the measured forces from bearing capacity theory. However,strength profiles in sediment beds can not easily be measured with standard rheological instruments, as sample volumes are often small and sample injection into the instrument causes structural damage and stress conditions unequal to the in situ situation. Until nowthis comparison has not yet been made.

I

I

I

CHAPTER

4.

EXPERIMENT AL RESULTS AND DISCUSSION

25

I

o ~..,.

.

:

:

'+

,

0.01

u

"

:r

~ ~ : \ ~ +\ 0.02 ~\

+\

: ~ ~

"

0.03

1-

:

+\ ~

~

.

.f.

_\ 0.04 - : -k

~

-+

o

,

I

,

_o

-I I China expo 0 Caland expo

₊

-N-St. expo 0 China fit Caland fit

-,

N-St. fit

,

,

o

,

_,

0 'IJ ct·

-I

0.05 - 0

+

+

-I

o

I I I 0.06 L.--..;._ __ ----JL.-- ----L ----L

___~

o

00

100

100

Ty (Pa)

Figure 4.11: Influence of sediment type and consolidation time 200

Table 4.2: Consolidation times (Figure 4.11)

China day 1 week

Caland channel mud Nieuw Statenzijl mud

2 weeks 7 months

I

4.7

Influence of sediment type

I

Apart from China day and Caland channel mud, results of which were shown in previous

sections,mud from Nieuw Statenzijl, a small harbour in the Dollard estuary in the Netherlands has also been tested. A comparison between these three muds is presented in Figure 4.11. Apart from sediment type, the stress histories and consolidation times of the samples were different. It is therefore incorrect to attribute all differences to sample composition. In Table 4.2 the consolidation times are summarized. Strength very significantly increases from China day via Caland channel mud to Nieuw Statenzijl mud.

The strength profile can be roughly represented with Ty = a

+

bz, a being the cohesion intercept and b the slope of the line. The positive correlation between depth and strength is very significant,and can be attributed to the increase in effective stress with depth as a result of the weight of the overlying sediment. In §4.9 this effectisdiscussedindetail.

I

I

I

I

CHAPTER

4.

EXPERIMENT AL RESULTS AND DISCUSSION

20

Ty (Pa)

Figure 4.12: Dependency of yield strength on consolidation time for China clay

I

0.12 0.1 0.08

S

'--' 0.06 ~ 0.04 0.02 0 0

I

I

I

26 3 days ~ 6days

+-8 days -G--10 days

·

x

··

·

.

~x.

.

~~EJ

...

·

XB

"*

_{+---)i;...__}

X -G_

_-s

~

+-

-

-

..,

A "___ L.:..r' .... 5 10

15

25

30 35 40

4.8

Influence of consolidation

time

One of the features of the measuring method under consideration is that it can be used to measure strength development in consolidating beds. Results are presented in Figure 4.12, where the strength profiles for China clay after 3, 6, 8 and 10 days of consolidation are shown. From this Figure it is clear that for a China clay sediment bed with a height of approximately 0.10 m, the consolidation phase ends after about one week, after which the height of the sediment bed remains constant. Evaluation of strength increase is also possible on much short er timescales,as illustrated in Figure 4.13. A significant differenceexists between strength at 1 hour after the start of sedimentation/ consolidation and strength after 3 hours. The strength increase appears to be most pronounced near the bottom of the cylinder. The characters 'a' and 'b' refer to the two columns used. Reproducibility is good, except for the lowest 2 cm after 6 days of consolidation.

For Caland channel mud the same experiments have been carried out (Figure 4.14). Again the results from the sounding test appear sensible and reproducibility is good. Bed strengths of less than 1 Pa can be measured accurately with the plate geometry.

I

I

I

4.9

Influence of effective stress

In §4.7 the roughly linear increase of bed strength with depth was illustrated (Figure 4.11) and explained with the effective stress concept. lnitial effective stress for consolidated beds can be related with z byO"eff =

J;(p -

Pw)g dz. The density profile can be measured with a

conductivity probe as described in §4.11. The initial effective stress should not be confused

I

I

CHAPTER 4. EXPERlMENTAL RESULTS AND DISCUSSION 27

I

0.05 1 hr, a ~ 1 hr. b

-+-3 hr. a-8- -3 hr. b ·K··· 5hr. a -8-. 5 hr. b

"*-6 days a

+-6 days b

-+-I

0.04 0.03 0.02

I

0.01 100

I

Figure 4.13: Dependency of yield strength on consolidation time for China clay, focussed onTy (Pa) changes within hours

I

I

0.06 0.05 0.04

S

_0.03 -...-N 0.02 0.01 0 0.1 2 hr, a-2hr. b ---27 hr. a - - --... 27 hr. b . ....:..'....".., '\... _{~. 6 days b -}6 days a - ---_-

-.

\

"

"

...:.. I " ~ \ \ \. 1\ ":.'

'

,"

<

.

:

.:: \\

\

ï\

~.' ''-Ï_ .' ' :1-".:, v, '.,~,.\ ".\

.

I

1 10 100

I

Ty (Pa)

Figure 4.14: Dependency yield strength on consolidation time for Caland channel mud

I

CHAPTER

4.

EXPERIMENT AL RESULTS AND DISCUSSION

28

I

I

with the effective stress acting at the tip of a geometry during penetration. For undrained tests the latter will generally be smaller than the former, as excess pore pressures are generated by the penetration of the geometry. For drained tests, however, the lat ter tends to be larger than the former, as excess pore pressures are allowed to dissipate and consolidation effects under the weight of the geometry may occur.

In Figure 4.15 the yield strength of Caland channel mud is shown as a function of initial vertical effective stress I7v

,o.

A continuous increase in yield strength with effective stress can be observed. The dependency of the yield strength of China day on z and thus I7vO_, is less pronounced than for Caland channel mud(Figure 4.11). This may be explained by the differences in the ratio between vertical and horizontal effective stress. From Mohr's cirde it followsthat the maximal shear stress within a material equals half of the difference between the major and minor principle stresses, being the vertical and horizontal stresses in the case ofaxisymmetric consolidation columns. If it is conjectured that consolidation proceeds as long as the gravity-induced stresses in the bed exceed the bed strength, the maximal shear stress in a fully consolidated bed will approximately equal the yield strength. For very weak beds like China clay the bed will be supported by the column walls for the major part and the differencebetween major and minor principle stress will be small. A theoretical maximum for the yield stress is found if the horizontal stress is zero. The yield stress then equals half the vertical effective stress. This maximum is only valid for freshly deposited layers which have not been exposed to higher effectivestresses in the past than the present effective stress. Also cohesion effects-which are responsible for strength at zero effectivestress-are not taken into account in this argument. For (partially) drained conditions the yield strength may exceed the theoretica! maximum, as the penetration of a geometry then results in a vertical effective stress higher than the initial vertical effective stress.

A clear, positive trend exists between effective stress and yield strength for freshly con-solidated sediment bed. However, for overconcon-solidated beds, for which the effective stress somewhere in the past has been higher than the actual effective stress, the highest effec-tive stress in its stress history may show a better correlation with the yield strength than the actual effective stress. Overconsolidation is caused byerosion, for example, as sediment that carried the weight of overlying layers becomes exposed. To what extent memoryeffects are important, is not investigated in this report, but there are no practical impediments to investigate this with the sounding instrument.

I

I

I

I

I

4.10

Comparison between balance and pressure transducer

re-sults

I

In the previous sections, the yield strength was calculated from the weight changes measured with the balanee. However, another possibility exists. Yield strength can also be calculated from the effectivestress acting on the tip of the geometry used in the sounding test. The effec-tive stress-the difference between total and pore pressures-can be measured with the plate geometry equipped with pressure transducers. From (3.12) one can estimate the undrained yield strength with Ty

=

C=l7eff/ScNc.

In Figure 4.16 the strength calculated in this wayiscompared with the strength calculated

I

CHAPTER 4. EXPERIMENT AL RESULTS AND DISCUSSION

29

I

0 20 40 "";d 60 c, .__.. =,

_.,

₈₀ b 100 120 140 0

I

I

I

50 100 150 200 250 300 350 400 450 500 Ty (Pa)

Figure 4.15: Yield strength versus initial vertical effectivestress for Caland channel mud

I

from weight changes for several traversing speeds. Caland channel mud was used as the bed material. Qualitative agreement between the two totally independent methods is excellent,

quantitative agreement is fair. Differencesare attributed mainly to the much lower accuracy of the pressure transducers compared to the accuracy of the balance. Compare Figure 4.16 with Figure 4.7,which is based on the same experimental data.

4.11

Bulk density profiles

I

In order to calculate strength profiles, the bulk density of the bed has to be known. The average bulk density of the bed is easily calculated from the initial height of the suspension

lu, the initial suspension concentration C; by mass and the actual height h:

Pav =[(hdh)Cd Pc]{Pc - Pw)

+

Pw, (4.1)

I

where Pc is the dry density of the clay particle, usually around 2,700 kg m-3. However,the

bulk density varies in the vertical. Lower layers are generally more compacted than upper layers, resulting in a higher bulk density. If accurate calculations of the yield strength are desirable, this should be taken into account.

Bulk density profilescan be measured with a conductivity probe as described in Chapter 2. Examples ofthese profilesare shown in Figures 4.17en 4.18 for China clay and Caland channel mud respectively. Sediment beds in columns 1, 2 and 3 were prepared in an identical way. The trend of increasing density with increasing depth can be clearly observed. Measurements in the upper en lower few millimeters of the bed are inaccurate, as the sensor then is only partly immersed in the bed or is influenced by the bottom. The bulk density of Caland

I

I

CHAPTER

4.

EXPERIMENT AL RESULTS AND DISCUSSION

30

I

I

-0.02 -0.01 0 0.01 0.02 ... S _0.03

--

~ 0.04 0.05 0.06 0.07 0.08 0

I

--'i-::- _ 0'

-+-~~_-_-::.~*_

---_

---:~--+

--

-

- ==::~---

I!::t" --~-~~--,., ... []

+

I

50 100 150 200 250 300 350 400 450 500 Ty (Pa)

Figure 4.16: Comparison between yield strength calculated from weight changes and pore and total pressures for Caland channel mud

I

channel mud is markedly lower than that of China clay. Because of the stronger interpartiele bonds (cohesion) the consolidation process stops at a higher void ratio and lower density at the same effective stress. This is in agreement with yield strength measurements, as shown

in Figure 4.11, for example.

From the density and yield strength profiles a plot of the yield strength versus the density can be constructed. An example is shown in Figure 4.19, where a clear, positive correlation

between density and yield strength can be observed. On the basis of this nice correlation,

one may wonder if it is possible to calculate yield strength directly from the easily measured

density profile without the hassle of measuring the force on a rod or the torque on a vane.

Although there often is a positive correlation between yield strength and density, no

unique relation between these two parameters exist. Different sediment types with the same

density may have a totally different yield strength, as a comparison between China clay and

Caland channel mud illustrates. But even if only one sediment type is considered, the relation between density and yield strength is not unique. By disturbance of a freshly consolidated

sediment bed excess pore pressures are generated, as the loosely packed particles tend to get

more closely packed. However,as the permeability of cohesivesediment beds is low,initially

no flow of water occurs and thus no bulk density changes are observed. Opposite to this, the disturbance does affect the yield strength. This shows that even for one sediment type at a

certain density, the bed strength may vary.

I

I

I

I

CHAPTER

4.

EXPERIMENT AL RESULTS AND DISCUSSION

I

₀ column 1 ~ 0.02 column 2

+-, column 3 -B- -0.04

8

0.06 ..._.. ~ 0.08

I

I

0.1 0.12 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500 P (kg m-3)

Figure 4.17: Density profile for China clay

I

0 0.01

-+

El

column 1 ~

+

G1

column 2

+-I

_[4

column 3 -B--0.02

+"

" ""'*

rJ \

Q

0.03 ~

I

8

_{..._.. 0.04} \;

\-

Q,!SI,

\

~

,

, 0.05 ~" "El, ", _'El 0.06

I

0.07 0.08 1000 1050 1100 1150 1200 1250 1300 p (kg m-3)

Figure 4.18: Density profiles for Caland channel mud

I

I

I

I

CHAPTER

4.

EXPERIMENTAL RESULTS AND DISCUSSION

I

I

I

40 35 30 25 '";d c, 20

--

~." 15 10 5 0 0

I

I

Chinaclay ~ Caland channel mud

+-32

100 150 200 250 300 350 400 450 500

P - Pw (kg m-3)

Figure 4.19: Yield strength versus density profiles for freshly deposited China clay and Caland channel mud

I

I

I

I

f

+

+

I

f

I

t

+

I

,,+

/ -I' 50

I

I

I

Chapter

5

Conclusions and recommendations

I

The sounding test described provides a simple and fairly accurate method to measure

in situ

yield strength and yield strength profiles of sediment beds. Yield strengths down to 1 Pa

ean be measured fairly accurate. A good agreement is obtained between plate, rod and disk

geometries. The method can be used for example to measure the evolution of yield strength

with time during the consolidation process.

If the penetration geometry is equipped with pore and total pressure transducers, then

valuable information on quantities such as effectivestress and excess pore pressure is obtai

-ned in addition to the weight changes of the geometry. The transit ion between drained and

undrained behaviour can be clearly derived from excesspore pressures, and from the effective stresses acting on the tip of the geometry during penetration the bed strength can be

ealcu-lated independently from the weight changes. Qualitative agreement between these methods

is good, quantitative agreement is only fair because of the lower accuracy of the pressure

sensor.

The results of the sounding test should also be eompared with vane test results. In one

case comparison of the cohesion intercept at zero effeetivestress determined with the sounding

test with yield strength determined with controlled stress rheometry showed satisfactory

agreement. Additional comparisons are still to be made.

A clear,approximately linear relationship between vertical effective stress-which is easily

calculated from the density profile of the consolidated bed-and shear strength is observed.

However,a priori estimation of the yield strength from the effective stress is not possible in

general, beeause additionally the ratio between horizontal and vertical effective stress has to

be known. Also the internal cohesion of the particles is important; especially at loweffective stress levelsit may be dominant. Cohesion may increase by aging. Last but not least sediment beds can have a much higher strength than to be expected from their actual state of effective

stress, which may be caused by their effectivestress history. For example, if a sediment layer

is re-exposed after being buried in the past, it has been exposed to much higher effective

stresses than the actual stress.

The geometry equipped with pressure transducers used in this study was quite bulky

com-pared to the dimensions of the settling columns. However,if a more compact instrument is

designed,preferably also equipped with a miniature acoustic probe to measure bed

concen-I

I

I

I

I

33

I

I

CHAPTER 5. CONCLUSIONS AND RECOMMENDATIONS

34

I

trations, a very powerful instrument would be obtained. The combined measurement of force,

pore pressure, total pressure and concentration as a function of the vertical coordinate, makes a thorough analysis of bed strength and strength evolution during consolidation possible.

I

I

I

I

I

I

I

I