Effect of suspended fine sediment on equilibrium local scour depths

UFL/COEL-2002/015

EFFECT OF SUSPENDED FINE SEDIMENT ON EQUILIBRIUM
LOCAL SCOUR DEPTHS

by

Elizabeth Anne Smyre

Thesis

2002

EFFECT OF SUSPENDED FINE SEDIMENT ON EQUILIBRIUM LOCAL
SCOUR DEPTHS

By

ELIZABETH ANNE SMYRE

A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2002

Copyright 2002

by

Elizabeth Anne Smyre

To my family and friends

ACKNOWLEDGMENTS

I would like to extend my gratitude to Dr. Max Sheppard, my advisor and

supervisory committee chair, for his guidance and support on this project. I would also

like to thank Dr. Robert Thieke and Dr. Ashish Mehta for serving on my supervisory

committee.

Several people provided invaluable assistance in the collection of the data used in

this report. I would like to thank Tom Glasser for his work on the Massachusetts scour

tests. I am grateful to Ken Kerr for his guidance on the operation of the RETA; his

assistance was instrumental in designing the tests conducted on the device. I would also

like to acknowledge Vernon Sparkman of the Coastal & Oceanographic Engineering

Laboratory for his assistance in modifying the RETA. Additionally, my sincere

appreciation and thanks go to Dougal Clunie at the University of Auckland for

conducting additional scour tests; the data and assistance he provided were invaluable in

the preparation of this report.

Finally, I would like to thank my parents, family, and friends for their support and

guidance throughout my graduate experience. Their ability to listen, advise, and even to

make me laugh, helped me to realize my goals; I only hope that I can be as helpful to

them in life as they have been to me.

TABLE OF CONTENTS

ACKNOW LEDGMENTS ........................................................................................... iv

LIST OF TABLES .............................................. vii

LIST OF FIGURES ................................................................................................... viii

LIST OF SYMBOLS ................................................................................................... xi

CHAPTER

1 INTRODUCTION .....................................................................................................

Definition of Scour ................................................................................................... 1
The Scour Process..................................................................................................... 2
Current Local Scour Prediction Equations ................................................ ........... 3
Problem Description ............................................................................................... 10

2 BACKGROUND ...................................................................................................... 13

University of Florida Local Scour Tests: USGS-BRD Laboratory ....................... ... 13
Analysis of Suspended Fine Sediment W ater Samples ............................................ 20
Video Data .............................................................................................................. 22
Calculation of Bed Shear Stress.................................................. .......................... 23
Discussion of USGS-BRD Data ........................................ ................................... 24
University of Auckland Scour Tests ........................................................................ 25

3 LITERATURE REVIEW ........................................................................................27

4 TEST APPARATUS AND PROCEDURE ...............................................................33

B background ................................................................................................................... 33
University of Florida Testing Apparatus ................................... ............. ............ 34
Current Testing ....................................................................................................... 35
M modifications to RETA....................................................... ........................... 35
Procedure ......................................................................................................... 36

5 TEST RESULTS AND OVERALL CONCLUSIONS ....................................... ..42

R esu lts........................................................................................................................... 42

D iscussion..................................................................................................................... 45
Conclusions................................................................................................................... 46

6 FU TURE RESEARCH ............................................................................................. 50

APPENDIX

A ROTATING EROSION TEST APPARATUS (RETA) TEST DATA ......................54

B UNIVERSITY OF AUCKLAND SCOUR TESTING ................................................. 79

Introduction................................................................................................................... 79
Facility .......................................................................................................................... 79
Test Param eters............................................................................................................. 79
Procedure .................................................................... .................................................. 80
Results........................................................................................................................... 82
D iscussion..................................................................................................................... 83

LIST OF REFEREN CES...................................................................................................89

BIOGRAPH ICA L SKETCH ........................................................................................ 92

LIST OF TABLES

Table page

2-1. Summary of USGS-BRD scour tests parameters...............................................16

2-2. Test Param eters: Experim ents A, B.................................. ......................................16

2-3. Suspended Fine Sediment Concentration Test Data...........................................21

2-4. Shear stress calculation parameters.................................... .....................................23

2-5. Typical flat bed shear stress values for the local scour tests performed at the USGS-
B RD Laboratory............................................................................................... 24

LIST OF FIGURES

Figure page

2-1. Aerial photo of USGS-BRD Laboratory, Turners Falls, Massachusetts.................. 14

2-2. Schem atic of USGS-BRD flum e.................................................. ........................... 14

2-3. Plot illustrating the effect of suspended fine sediment on scour depth ....................17

2-4. Post-experiment photos of Experiment A and Experiment B .................................18

2-5. Contour plots of Experiment A and Experiment B ............................................19

2-6. Concentration of suspended fine sediment measured during one of the USGS-BRD
scour tests............................................................................................................... 2 1

4-1. University of Florida Rotating Erosion Test Apparatus (RETA).............................35

4-2. Photograph of cylinders used in RETA tests.........................................................37

5-1. Average shear stress lines generated from tests conducted on 0.15 mm sediment
coated acrylic cylinder. ..................................................................................... 43

5-2. Average shear stress lines generated from tests conducted on 0.85 mm sediment
coated acrylic cylinder. ..................................................................................... 44

5-3. Proposed relationship between shear stress and scour depth ...................................49

A-1. RETA Tests 5-8, 0.15 mm Sediment Coated Acrylic Cylinder ..............................54

A-2. RETA Tests 13-16, 0.15 mm Sediment Coated Acrylic Cylinder ..........................55

A-3. RETA Tests 25-28, 0.15 mm Sediment Coated Acrylic Cylinder ..........................56

A-4. RETA Tests 29-32, 0.15 mm Sediment Coated Acrylic Cylinder ..........................57

A-5. RETA Tests 37-40, 0.15 mm Sediment Coated Acrylic Cylinder ..........................58

A-6. All 0.0 g/1 concentration tests conducted on 0.15 mm Sediment Coated Acrylic
C ylinder......................................................................................................... . .59

A-7. All 0.05 g/1 concentration tests conducted on 0.15 mm Sediment Coated Acrylic
C ylinder............................................................................................................ 60

A-8. All 0.5 g/1 concentration tests conducted on 0.15 mm Sediment Coated Acrylic
C ylinder......................................................................................................... . 6 1

A-9. All 1.0 g/l concentration tests conducted on 0.15 mm Sediment Coated Acrylic
C ylinder......................................................................................................... . 62

A-10. RETA Tests 1-4, 0.85 mm Sediment Coated Acrylic Cylinder ............................63

A-i 1. RETA Tests 9-12, 0.85 mm Sediment Coated Acrylic Cylinder ..........................64

A-12. RETA Tests 17-20, 0.85 mm Sediment Coated Acrylic Cylinder ........................65

A-13. RETA Tests 21-24, 0.85 mm Sediment Coated Acrylic Cylinder ........................66

A-14. RETA Tests 33-36, 0.85 mm Sediment Coated Acrylic Cylinder ........................67

A-15. All 0.0 g/1 concentration tests conducted on 0.85 mm Sediment Coated Acrylic
C ylinder......................................................................................................... . .68

A-16. All 0.05 g/1 concentration tests conducted on 0.85 mm Sediment Coated Acrylic
C ylinder......................................................................................................... . 69

A-17. All 0.5 g/1 concentration tests conducted on 0.85 mm Sediment Coated Acrylic
C ylinder......................................................................................................... . .70

A-18. All 1.0 g/l concentration tests conducted on 0.85 mm Sediment Coated Acrylic
C ylinder............................................................................................................71

A-19. RETA Tests 45-48, Smooth Aluminum Cylinder ...........................................72

A-20. RETA Tests 49-52, Smooth Aluminum Cylinder ...........................................73

A-21. RETA Tests 41-44, Rough Aluminum Cylinder.............................................74

A-22. RETA Tests 53-56, Rough Aluminum Cylinder ................................................75

A-23. RETA Tests 57-60, saltwater tests conducted on 0.15 mm Sediment Coated
A crylic C ylinder........................................................................................... .... 76

A-24. RETA Tests 61-64, saltwater tests conducted on 0.85 mm Sediment Coated
A crylic C ylinder............................................................................................... 77

B-1. University of Auckland Scour Tests, 0.95 U.........................................................85

B-2. University of Auckland Scour Tests, 1.1 U........................................................86

B-3. University of Auckland Scour Tests, 1.5 U........................................................87

B-4. University of Auckland Scour Tests, 2.0 Uc........................... .......................88

LIST OF SYMBOLS

b pier diameter

C suspended fine sediment concentration

D5o median bed sediment diameter

ds measured scour depth

dse equilibrium scour depth

Fr Froude number

g coefficient of gravity

ks Nikuradse roughness length

L length of RETA test cylinder

R radius of RETA test cylinder

T torque

U flow velocity

Uc critical velocity

yo flow depth

P dynamic viscosity

v kinematic viscosity

p fluid density

Ps sediment density

S shear stress

Tc critical shear stress

To initial pier shear stress

TU shear stress upstream of pier

Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

EFFECT OF SUSPENDED FINE SEDIMENT ON EQUILIBRIUM
LOCAL SCOUR DEPTHS

By

Elizabeth Anne Smyre

December 2002

Chair: D. Max Sheppard
Department: Civil and Coastal Engineering

Recent clearwater local scour experiments with cohesionless sediments have shown

that the presence of suspended fine sediments can impact equilibrium local scour depths.

Researchers have known for some time that suspended fine sediment can affect shear

stress at the flow boundaries and have observed large drag reductions due to suspended

fine sediment in the flow. Reduced bed shear stress is one possible explanation for the

observed reduction in local scour depths.

This thesis 1) presents the results of an attempt to quantify the effects of suspended

fine sediment on bed shear stress through laboratory experiments, 2) presents local scour

data (obtained by other researchers) that illustrate the reduction in scour depth due to

suspended sediment, and 3) discusses the possible causes for the reduced scour depths.

An understanding of these effects is not only important in scour depth prediction but also

in the proper interpretation of laboratory local scour results. This may also help explain

some of the scatter in reported laboratory and field local scour data.

CHAPTER 1
INTRODUCTION

Definition of Scour

The removal of sediment near a structure located in a flowing water body is

referred to as sediment scour, or simply as scour. Sediment scour can be due to several

mechanisms, and thus it is usually divided into categories or types. Lateral migration of

the channel or the channel thalweg can reduce the bed level at the structure; this is known

as general scour. General degradation of the channel bed can result in a reduced bed

level at the structure. The structure can cause a reduction in flow cross-section and a

corresponding general reduction in bed elevation at the structure (contraction or

constriction scour). Finally, the presence of the structure alters the flow field in the

vicinity of the structure causing an increase in bed stress and a corresponding removal of

sediment near the structure (local scour). The total scour is the sum of these components.

Scour at bridge piers is a major problem. According to Richardson and Davis

(1995), scour around bridge piles and foundations as a result of flooding is the most

common cause of bridge failure. The problems resulting from bridge scour is a

widespread problem and has the potential for tragic results. The potential cost, including

human toll and monetary cost, of bridge failures due to scour damage has highlighted the

need for better scour prediction methods and equations. Under-prediction of scour depth

can lead to costly bridge failure, while over-prediction can result in millions of dollars in

unnecessary construction costs.

Thus much of scour research has been directed at improving scour prediction

equations and methods. Current research is devoted primarily to predicting scour at

specific types of structures, predicting the rate at which scour occurs, predicting scour in

soils other than sand, and the development of 3-D computational models for predicting

scour. Despite continuing research into scour mechanisms, some aspects of the scour

process are still not understood.

The research reported in this thesis is directed at understanding the reasons why the

presence of suspended fine sediment (SFS) reduces equilibrium local scour depths.

Experiments that 1) discovered this effect and 2) attempt to quantify the dependence of

equilibrium scour depth on SFS concentrations and flow velocity are summarized first.

This is followed by experiments conducted as part of this work to measure the effects of

SFS concentrations on the wall (bed) shear stress.

The Scour Process

The depth of scour at a bridge pier or abutment depends on a range of flow,

sediment, and structure parameters. The results of empirical studies assist in reducing the

number of parameters to only those having a primary effect on the local scour processes

(Sheppard et al., 1998). Most research in bridge scour has been devoted to the accurate

prediction of the actual maximum scour depth for a particular bridge structure

arrangement; however, recent studies have also examined the time history of scour.

One of the primary flow mechanisms responsible for local scour is the horizontal

vortex at the base of the leading edge of the structure. This vortex is known as the

"horseshoe vortex" due to its horseshoe like shape when viewed from above. This vortex

increases in size and intensity as the scour hole develops up to a maximum value. The

dissertation of Melville (1975) details the process of scour hole formation. Scour begins

as the velocity around the circumference of the cylinder increases and reaches a

maximum value at approximately 1000 from the upstream edge. At these locations,

indentations begin to form as the shear stress exerted on the bed increases; these

indentations slowly progress to the front of the pier where they meet. Scoured material is

transported downstream by the flow. Though the initial horseshoe vortex is relatively

weak, its intensity increases as more material is eroded, thus causing the vortex to

descend into the hole. The horseshoe vortex derives its energy from the main flow. The

bed shear stress is extremely high near the structure at the base of the scour hole and

decreases radially outward. As sediment is removed from near the structure, the

surrounding sediment avalanches into the hole. The slope of the scour hole is

approximately the submerged angle of repose of the sediment (Melville, 1975). In the

clearwater scour range, scour continues until the shear stress causing the scour is

balanced by the gravitational forces. For the live bed scour regime, scour continues until

the sediment entering the scour hole equals that being transported out.

Current Local Scour Prediction Equations

Scour prediction equations vary greatly in format and in content due to the wide

variety of parameters that affect equilibrium scour depth. The depth of scour depends on

several flow, fluid, sediment, and structure parameters, as shown in the following

functionally dependent equation:

dse =f[p, g, Dso, a, s, o, U, Uc, b, h(pier), C] (Equation 1-1)

where

dse= equilibrium scour depth,

P, Ps = water and sediment densities, respectively,

p = dynamic viscosity of the water (temperature dependent),

g = gravity,

Dso = median sediment diameter,

a= gradation of the sediment,

yo = depth of flow upstream of the structure,

U = depth-average velocity,

Uc= sediment critical depth-average velocity,

b = pier diameter/width normal to the flow,

h(pier) = function that describes the shape and alignment of pier in

relation to the flow, and

C = concentration of suspended fine sediment in the water column.

A dimensional analysis using the variables in the above equation results in a

number of independent (n) groups. Even though the equilibrium scour depth depends to

some degree on all of these groups, researchers at the University of Florida have found

the following groups to dominate:

db fbU b ,ar,h(pier),C (Equation 1-2)
b b 'Uc D,50

The aspect ratio -o relates the depth upstream of the pier to the pier diameter
b

normal to the flow. According to Ettema (1980), the depth of the flow impacts the

formation of the vortices on the upstream side of the pier. Decreasing flow depth

decreases the effect of a surface roller that rotates in the opposite direction to that of the

horseshoe vortex at the bottom of the pier. If the two vortices interfere with each other,

the effect of the surface vortex decreases. Additionally, low values of the aspect ratio

decrease the percentage of the flow that moves through the scour hole (Ettema, 1980).

Laboratory data indicates that if the parameters U and -50 are held constant, dse
U, b

increases rapidly with increasing Y until the ratio reaches 2.5 to 3; dse then remains
b

constant (Pristivelis, 1999).

The velocity ratio determines if scour is occurring in the clearwater or live-bed
U,

velocity range. If the velocity ratio is greater than some value (that depends on the

structure shape) but less than 1, the scour is said to be clearwater scour. If the velocity

ratio is greater than or equal to 1, then the flow velocity is such that sediment motion is

initiated on a flat bed away from the structure; thus live-bed scour conditions exist. The

critical velocity is based on Shields' equations for the critical shear stress and shear

velocity. The Prandtl-Von Karman formula for a fully developed velocity profile can be

used to calculate the depth averaged velocity in terms of shear velocity and bed

roughness (Sleath, 1984). The critical velocity Uc is then calculated from these values.

The dependence of equilibrium scour depth on the ratio for larger values of
Dso

b was not realized until recently. Ettema (1980) concluded that the effect of this ratio
D50

on scour depth was only significant for low ratios. Sheppard and Ontowirjo (1994),

Sheppard (1997), and Sheppard et al. (2002) show that the scour depth dependence

b
extends to much larger values of b. Since there is a lower limit on the size of the
Ds 0

sediment before it becomes cohesive (~0.1 mm), the equilibrium scour depth dependence

b b
on for extremely large values of is due entirely to the dependence on pier width,
D50 D50

b.

Sediment gradation, a, is believed to affect scour depth due to the formation of an

armor layer around the structure; this armor reduces the depth of the scour hole. The

effect is more pronounced in the clearwater scour range; the data of Ettema (1980)

showed a reduction in scour depth as the standard deviation of the particle size

distribution increased.

The physical pier properties, denoted by h(pier), are usually accounted for by a

multiplicative coefficient in the scour equation that is determined empirically.

As stated above, there are many predictive equations for local scour depth in the

literature. These equations differ significantly in their form and in the magnitude of their

predictions. Three of these equations are presented below. The FHWA recommends an

equation developed at Colorado State University in its Hydraulic Engineering Circular

No. 18 (HEC-18) report:

( o0.65
ds, = 2.0yoKK2 _- Fr"o" (Equation 1-3)
\yo)

where

dse = equilibrium scour depth,

yo = flow depth upstream of structure,

K1 = pier nose shape correction factor,

K2 = angle of attack flow correction factor,

b = pier width, and

Fr = Froude number, Fr = -

This equation can be used for both clearwater and live-bed scour conditions and has

limiting values: d- = 2.4 for Fr < 0.8 and d" = 3.0 for Fr > 0.8 for a circular pile
b b

(Richardson and Davis, 1995).

The following equation was developed by Melville (1997) and is based on the

results from laboratory experiments:

dse = KKdKyDKaKs (Equation 1-4)

where

1 U 2
Uc
KI = flow intensity factor = for
U U
v- <1
UC Uc

Kd= sediment-size factor =

1.0

for

0.57 log 2.24 -b
D 50

KyD = flow depth-pier width factor =

Ka = pier-alignment factor, and

Ks = pier-shape factor.

2.4b

245y

4.5y,

b
-<0.7
Yo

for 0.7 < b< 5,
Yo
b
->5
Yo

b
-> 25
Dso

b
< 25
Ds0

The values for Ka and K, are given in Melville and Sutherland (1988).

Sheppard and Ontowirjo (1994) published the original version of the following

equation in 1994. This equation has undergone minor modifications over the years as

more and improved data became available.

In the clearwater scour range 0.47 < < 1.0:

d, Kcsff 2 f3 (Equation 1-5)
b s ^[b U,J D50J

where

U, ln(U/U)2,

S(_b (c3 +c4)
D350 c expc4 (x-Xk)]+C4 exp[-C3(x -xp)]

x =log xpk =log ,

c = 1.0,

c2 = 0.4,

C3 = 2.6,

C4 = 0.45,

c5 = 2.5,

b
S = 44, and
D50 atpeak

U
0.47.
UcO

For live bed scour conditions, the equations become

For 1.0<-< UI
U, U c

d U-Uc b' Ulbp -U
S=Kf ( ULbp= -U U c + c, b f-- (Equation 1-6)
b b Ulbp U UDo D50 -(7- Uc

For f > :
(U UC

=Kc c6tanh c 1- (Equation 1-7)
b [kb}J

where

c6 = 2.2, and
Ulp = velocity at which the peak scour depth occurs in the live bed
scour range (live bed peak velocity).

Note that SFS concentration, C, is not included in these, or to the author's

knowledge, in any of the published scour prediction equations. The dependence of

equilibrium scour depth of SFS concentration was discovered during tests preformed by

Sheppard et al. (2002) at the Conte USGS-BRD Laboratory. The flume used for these

tests was a "flow-through" type flume with the water being supplied by a power plant

reservoir adjacent to the Connecticut River. There was no control on the constituents in

the water, and for some of the tests the SFS concentrations were elevated due to rain

water runoff and/or snow melt. The USGS-BRD flume is similar to other flow-through

flumes used in scour research; thus, research into the effect of SFS has implications for

future scour studies conducted in similar flumes.

Problem Description

Most studies designed to measure scour around bridge pier configurations are

conducted in laboratory situations in which the experiment simulates a proposed bridge

pier design. These situations offer control over most of the parameters affecting local

scour depth. Tests can be designed for a specified bridge pier configuration, bed

sediment size distribution, and flow condition. However, the facilities available to

researchers provide their own limitations. Scour tests conducted in a recirculating flume,

for example, offer greater control of the water properties than those tests conducted in

flow-through type flumes. In the case of flow-through flumes, water properties, such as

suspended fine sediment concentration, cannot be controlled. Depending upon the

structure and flow velocity, scour experiments can last up to several days or weeks.

When a flow-through flume is used, the amount of sediment in the water source can vary

drastically depending upon weather conditions (specifically, the amount of runoff due to

rainfall events or snowmelt) during the duration of the test.

In a previous research project, University of Florida researchers (Sheppard et al.,

2002) conducted a series of clearwater scour tests at the Conte USGS-BRD Laboratory in

Turners Falls, Massachusetts. The facility consisted of a gravity-driven, flow-through

type flume that used an adjacent power plant reservoir as its water source. During several

of the scour tests, the water in the reservoir became turbid due to rainwater runoff and/or

snowmelt upstream of the laboratory. This reduced the equilibrium scour depths, forcing

researchers to repeat the tests. Since this was not expected, there were only few

measurements documenting the conditions during tests with heightened suspended

sediment concentrations.

The purpose of research reported in this thesis was to confirm that the presence of

SFS does affect the measured local scour depth and to delineate possible reasons for the

reduced scour due to SFS and to examine one of these possible causes. One possible

cause is that the presence of SFS causes a reduction in bed shear stress. This would have

a similar effect to reducing the flow velocity. A second possibility for the reduction in

scour depth is that there may be deposition of fine sediment in the scour hole out from the

structure. This could retard the avalanching of sediment into the high bed stress region

near the structure and/or when avalanches occur, they would carry with them the fine,

cohesive, sediment. The mixture of sand and cohesive sediment has, in general, a higher

critical shear stress. Thus the equilibrium scour depth will be less.

This research concentrates on the first possible cause for scour reduction listed

above, namely the reduction in bed shear stress due to the presence of SFS. The

investigation of the SFS problem begins with a review of the clearwater scour data

obtained by Sheppard et al. (2002) in their tests performed at the USGS-BRD Conte

Laboratory in Massachusetts. The data provides insight into the flow, bed, and pier

conditions under which SFS impacts scour depths. Water samples collected during one

of the USGS-BRD scour tests were analyzed as part of this work. The results provide

information about the concentrations of SFS encountered in the USGS-BRD tests. In

addition, the results of local scour tests conducted at the University of Auckland with

various SFS concentrations and flow velocities are reviewed and discussed.

Because the effect of SFS on equilibrium scour depth has not been previously

investigated, the literature review focuses on the research that has been conducted on

drag reduction due to SFS. Laboratory tests were conducted in a Rotating Erosion

12

Testing Apparatus (RETA) in an attempt to quantify the reduction in bed shear stress as a

function of SFS concentration.

The objective of this research is to identify and discuss the most likely mechanisms

responsible for the reduction in scour and to investigate in some detail one of these

possible causes. Once the effects are understood and quantified, they can be incorporated

into scour prediction equations. The task of modifying scour prediction equations is,

however, not part of this research.

CHAPTER 2
BACKGROUND

University of Florida Local Scour Tests: USGS-BRD Laboratory

Between August 1998 and June 2001, scour tests were conducted at the Conte

USGS-BRD Laboratory located in Turners Falls, Massachusetts adjacent to the

Connecticut River (Figure 2-1). The primary purpose of the tests was to extend the

structure-induced local sediment scour database to include data from larger structures

(larger values of ). Three different pile diameters [b= 0.915 m, 0.305 m, and 0.114
D50

m] and three different bed sediment sizes (Dso= 0.22 mm, 0.80 mm, and 2.9 mm) were

used in the investigation. All of these tests were conducted within the clearwater scour

range of velocities (i.e., < 1).

The flume used for these tests measures 6.1 m wide, 6.4 m deep, and 38.4 m long.

The test section within the flume was 6.1 m wide, 9.8 m long, and began 24.4 m

downstream of the flume's entrance. The bed sediment height within the test section was

1.83 m deep. A hydroelectric power plant reservoir (connected to the Connecticut River

upstream of control structures that lower the river's elevation) supplied water for the

flume; water flowed from the reservoir through the flume and was then discharged into

the Connecticut River downstream of the river control structures. The flume itself is

gravity controlled with the elevation of the flume bottom approximately 10 m lower than

that of the reservoir. A sharp-crested weir at the downstream end of the flume controls

water depth and flow discharge within the flume. The advantage of this type of flume is

that pumps are not needed to produce the flow.

Figure 2-1. Aerial photo of USGS-BRD Laboratory, Turners Falls, Massachusetts
(reprinted from Sheppard et al., 2002)

NOrTOSCALI:
All dimensions in meters
I;

Seclion A-A

I F *Ilow Dishorgd
So Connmct icul
Riker

IINI

flow
7---------- /T

rillcr Matrial BOIk ScdincIl Te-A Sdimciii BaIse SCdimnint
',slg I, H-1

Figure 2-2. Schematic of USGS-BRD flume (reprinted from Sheppard et al., 2002).

Flow-through flumes do, however, have disadvantages. There is little or no control

of the incoming water; thus, properties such as water temperature will be that of the water

source (in this case, the power plant reservoir). For these tests, the water temperature of

the Connecticut River varied from slightly above freezing to approximately 260C during

the summer months. In addition, the concentration of SFS within the water supply could

not be controlled. The level of rainwater and snowmelt runoff upstream of the reservoir

governed the concentration of SFS in the water supply. While the temperature of the

water can be accounted for in scour prediction, the effect of SFS had not yet been

identified as a quantity affecting scour depth.

During each of the fourteen scour tests, equilibrium scour depth measurements

were made using both acoustic transponders and video cameras located inside of the

piles. A real time scour depth plot was maintained during each test in order to know

when the scour depth had reached an equilibrium value. Flow velocity, water depth, and

water temperature were measured throughout the tests. A summary of scour test

parameters and results is presented in Table 2-1. Those tests affected by the suspended

fine sediment are not included in the summary, as most of the tests were halted once the

presence of SFS was detected and thus not included in the final data presentation.

In some of the longer duration tests, researchers noted a rather sudden increase in

SFS in the water column and a corresponding change in the rate of scour. In each case,

the rate of scour initially proceeded normally; at a point in the test, the rate of scour

suddenly decreased to zero, and the scour depth remained constant. Figure 2-3 is a

comparison time history plot that illustrates the change in the rate of scour. Experiment

A is clearly affected by the increase in SFS concentration approximately 10 hours into the

test, causing the scour hole depth to level out immediately. Experiment B, which was

with the same structure but at a slightly higher velocity and deeper water depth, was

conducted under ambient SFS conditions. The curve labeled Experiment B is the actual

Experiment B data adjusted to the water depth and flow velocity conditions of

Experiment A using Sheppard et al. (2002) scour prediction equations. Table 2-2

summarizes the parameters for the two experiments.

Table 2-1. Sumary of USGS-BRD scour tests parameters (Seppard et al., 2002).
Test D50 Yo b (m) U Uc Duration y/b b/Dso dse
No. (mm) (m) (m/s) (m/s) (hrs) (m)
1 0.22 1.19 0.114 0.28 0.32 87 10.4 518 0.133
2 0.22 1.20 0.305 0.29 0.32 163 3.9 1386 0.257
3 0.80 1.27 0.915 0.43 0.47 362 1.4 1144 1.112
4 0.80 0.87 0.915 0.38 0.46 143 1.0 1144 0.638
5 0.80 1.27 0.305 0.37 0.47 87 4.2 381 0.416
6 0.80 1.27 0.114 0.38 0.47 42 11.1 143 0.185
7 2.90 1.22 0.915 0.68 0.84 188 1.3 316 1.270
8 2.90 0.56 0.915 0.60 0.76 330 0.6 316 1.058
9 2.90 0.29 0.915 0.56 0.69 448 0.3 316 0.896
10 2.90 0.17 0.915 0.48 0.65 616 0.2 316 0.659
11 2.90 1.90 0.915 0.60 0.93 350 2.1 316 1.004
12 0.22 1.22 0.305 0.31 0.33 256 4.0 1386 0.377
13 0.22 0.18 0.305 0.27 0.27 215 0.6 1386 0.296
14 0.22 1.81 0.915 0.21 0.32 579 2.0 4159 0.787

Table 2-2. Test Parameters: Experiments A, B shown in Figure 2-3 (Sheppard et al.,
2002).
Experiment b (m) D50 (mm) yo (m) U/Uc
A 0.915 0.22 1.22 0.92
B 0.915 0.22 1.8 0.97

The photographs in Figure 2-4 show the differences in the final scour holes for the

two tests. In Experiment A, fine sediment deposition was not apparent adjacent to the

structure; however, there was deposition in the scour hole away from the pier once the

experiment was halted (Sheppard et al., 2002). Contour plots of the scour hole for each

experiment indicate that the scour hole in Experiment A was slightly steeper than the hole

formed in Experiment B, but that the scour hole in Experiment B had a more uniform

shape. In Experiment A, scour only occurred on the upstream side of the pier (Figure 2-

5).

Experiment B (low turbiditv)

Experiment B Adjusted
(adjusted to flow conditions of Experiment A

0 20 40 60 80
Time (hrs)

100 120 140

Figure 2-3. Plot illustrating the effect of suspended fine sediment on scour depth
(reprinted from Sheppard et al., 2002).

At the time of the experiments, the researchers were unaware of the sensitivity of

local scour rates and equilibrium values on SFS concentration; thus, water samples were

only taken during one of the tests that had higher than normal SFS concentrations. For

that test, water samples were taken at 11 different levels in the water column.

E

0.3
o
a
0.2

0

O w ip

Figure 2-4. Post-experiment photos of Experiment A (top) and Experiment B (bottom)
(reprinted from Sheppard et al., 2002).

- 7

Flow -

Flow
-1

I I I
-1 0 1

8.5-

3.5 4 4.5 5

Figure 2-5. Contour plots of Experiment A (top) and Experiment B (bottom) (reprinted
from Sheppard et al., 2002). Flow direction is indicated.

_

Analysis of Suspended Fine Sediment Water Samples

In early 2001, 11 water samples were collected during a scour test with higher than

normal SFS concentrations. The water samples were collected at the following heights

above the sediment bed: 7.6 cm, 15.2 cm, 22.9 cm, 30.5 cm, 38.1 cm, 45.7 cm, 53.3 cm,

61.0 cm, 91.4 cm, 152.4 cm, and 182.9 cm. These samples were gathered in order to

determine the overall SFS concentration and the concentration profile during an actual

scour test.

The procedure used to analyze the samples was to take a pre-measured volume of

the sample and filter it through a vacuum-pump filtration system. The filter paper was

oven-dried and the mass measured prior to the filtration. The water sample was then

filtered. The filter paper with the sediment was then dried overnight and the mass

measured again; the difference in the two measurements was recorded as the mass of the

sediment in the volume. The sediment mass was divided by the volume of the water

sample used in the test to get the concentration in g/1. Table 2-3 shows the results of the

suspended fine sediment concentration tests; the concentration profile is shown in Figure

2-6.

There were errors involved in the concentration measurement process. There was

approximately a ten-month lapse between the time when the water samples were taken

and when they were analyzed. Some of the sediment that settled out during this time

adhered to the walls of the containers, thus causing the measured values to be lower than

the actual value. The pretest drying time for the filters was also not sufficient, again

causing a lower measured concentration.

Measured concentrations ranged from 0 to 0.08 g/l, with the highest concentration

occurring at 45.7 cm above the bed. The concentration profile varied significantly over

the water column. While these measurements were not very precise, they do provide

some insight as to the magnitude of the SFS concentrations in one of the experiments that

was impacted.

Table 2-3. Suspended Fine Sediment Concentration Test Data
Sample Depth Sample Blank Filter w/ Mass Concentration
Number Above Volume Filter Sediment Difference (g/l)
Bed (ft) (mL) Mass (g) (g)
(g)_
1 6 85 1.1378 1.1395 0.0017 0.02
2 5 100 1.1231 1.1171 0 0.00
4 3 100 1.1246 1.1206 0 0.00
5 2 275 1.1246 1.1250 0.0004 0.00
6 1.75 225 1.1306 1.1310 0.0004 0.00
7 1.5 279 1.121 1.1433 0.0223 0.08
8 1.25 200 1.133 1.1320 0 0.00
9 1 100 1.1226 1.1174 0 0.00
10 0.75 100 1.1354 1.1385 0.0031 0.03
11 0.5 100 1.1272 1.1226 0 0.00
12 0.25 209 1.1313 1.1400 0.0087 0.04

Concentration Profile for Massachusetts Scour Test

6
5

3

2

1.75
[]

1.25

1

0.75

0.5

0.25
0.00 0.01 0.02 0.03 0.04 0.05
Concentration (gll)

T

0.06 0.07 '0.08 0.09

Figure 2-6. Concentration of suspended fine sediment measured during one of the USGS-
BRD scour tests.

I I

Video Data

Two small video cameras were mounted on a vertical traversing mechanism inside

of the test pier to monitor the water-sediment interface during the test. The cameras were

situated facing the upstream side of the pier and the scour hole. The camera platform was

made to traverse vertically with an electric motor that was controlled manually. Length

scales were mounted inside of the pier within the visual range of the cameras. Thus the

scour depth could be monitored from inside the pier via the video data. A programmable

control system was used to set the duration of the individual recordings as well as the

intervals between recordings. The control system also switched between the two cameras

at regular intervals (Sheppard et al., 2002).

The video recordings from two of the scour tests were reviewed as part of the

research in order to examine the differences in the scour hole formation between a test

without and one with SFS. In the test with little or no SFS present, the water was very

clear. The horseshoe vortex could be observed by the suspended bed sediment and small

debris particles in the water column. At the end of the test, movement of bed sediment

out of the scour hole was countered with sediment sliding back into the hole; this balance

of sediment transport indicated that the scour hole had reached equilibrium.

There were no major differences in the video from the experiment with SFS

(Experiment A) with the exception of a reduction in sediment movement in the bed.

Also, there was no indication of a sharp change in SFS (turbidity) at or near the time the

scour ceased. It should be pointed out, however, that in the tests with the large pile, the

pile was flooded with water and the cameras housed in waterproof containers. The view

of the external flow and scour hole was somewhat attenuated by the water in the pile. T

he only information obtained from the examination of the video was the reduced

sediment movement that could have been caused by any of the mechanisms outlined

earlier in the thesis.

Calculation of Bed Shear Stress

The equations for the hydraulically rough region of flow (see e.g. Sleath, 1984)

were used to estimate the shear stress values on the flat bed in front of the test pier. Table

2-4 summarizes the information used in the shear stress calculations.

Table 2-4. Shear stress calculation parameters.
Parameter Value
Sediment density, ps 2650 kg/m3
Water density, p 1000 kg/m3
Gravity, g 9.81 m/s2
Water viscosity, v 1.12x10-6 m2/s
Depth-averaged velocity, U Used U given for each test (m/s).
Flow depth, yo Given for each test (m).
Roughness, ks 5D5o for the 0.22,0.8 mm sand; 2.5D50 for the
__2.9 mm sand (m).

The equation for u. in the hydraulically rough flow range is:

U
2.51n 11.0yo

(Equation 2-1)

Solving for zo results in:

(Equation 2-2)

The critical shear stress is calculated by finding the non-dimensional critical shear stress

ratio from the Shields' curve. Thus the critical shear stress is:

zr = (dimensionlesst) (p, p)* g D50. (Equation 2-3)

u k
Finally, the value of the ratio of uk' was checked to ensure that the requirements

for hydraulically rough flow were met. This condition was met for the 2.9 mm and 0.8

mm sediment, but was not met for the 0.22 mm sediment. Similar shear stress values

were calculated for the 0.22 mm sediment using the equations for the transitional flow

region. Table 2-5 lists the average shear stress values calculated for each sediment size.

Table 2-5. Typical flat bed shear stress values for the local scour tests performed at the
USGS-BRD Laboratory.
Sediment Size (mm) Shear Stress (kg/m*s )
0.22 0.16
0.8 0.45
2.9 1.87

Discussion of USGS-BRD Data

A review of the tests conducted at the Conte USGS-BRD Laboratory provides

insight as to the conditions needed for suspended fine sediment to cause a reduction in

local equilibrium scour depth. The majority of the tests affected by the suspended fine

sediment were those with a bed sediment diameter of 0.22 mm. Only one test with a bed

sediment diameter of 0.8 mm was affected by suspended fine sediment.

The tests affected by the suspended sediment were conducted over a range of flow

depths and pier sizes. The time at which the scour rate dropped to near zero varied for

each test, ranging from 20 hours to 200 hours. The point at which the scour rate

decreases sharply is more likely a function of the percentage of the equilibrium scour

depth without SFS.

The video data recorded during a test affected by suspended fine sediment did not

indicate a sharp increase in water turbidity at the time that the scour rate decreased. The

turbidity appeared to be the same throughout the test. This observation suggests that the

SFS has its greatest effect after the scour hole has reached a certain depth where the

effective shear stresses are diminished. In the early stages of scour hole development, the

effective shear stresses near the structure are large, and it is known that the effects of SFS

decrease with increased velocity/turbulence.

There was undoubtedly an increase in turbidity during the tests since otherwise the

tests would not have been started. The increase was, however, more gradual than

originally suspected. The analysis of the video data showed that the change in SFS

concentration was not rapid as believed, but it was not helpful in determining which, if

any, of the proposed scour reduction mechanisms were responsible for the observed

effects.

The measured SFS concentrations indicate that only small concentrations are

needed to cause a dramatic effect on the equilibrium scour depth. The measured

concentrations were less than 0.1 g/1. The dependence of scour depth on SFS

concentration, flow velocity, and possible other parameters is needed before accurate

predictions of scour attenuation can be made.

University of Auckland Scour Tests

In order to confirm that the effect of SFS was not confined to the USGS-BRD tests,

the data collected during a series of scour tests conducted at the University of Auckland

was examined. The tests involved four freshwater/bentonite suspension flows conducted

at four velocities. A reduction in ds occurred in the tests in the clearwater range;

however, the reduction in scour was less than those measured in the USGS-BRD tests.

The tests conducted in the live bed velocity range offered conflicting results. The

clearwater tests do confirm that the final measured scour depth is reduced by the presence

26

of SFS. A complete write-up of the test procedure and a presentation of the final results

are included in Appendix B.

CHAPTER 3
LITERATURE REVIEW

The major thrust of this thesis was to investigate the effects of the suspended fine

sediment (SFS) on the shear stress exerted on the bed by the flow. Previous research

examined the effect of suspended clays on the drag forces exerted on a sediment bed in

an attempt to quantify the drag reduction due to SFS concentrations. Drag reduction is

defined as "the decrease in shear stress in the viscous sublayer with respect to the

apparent shear stress of the logarithmic layer in the upper water column" (Li and Gust,

2000, p.77). Research concerning the effect of drag reduction in open channel flows

began after similar studies, such as that of Toms (1949), measured drag reduction in

turbulent pipe flows with dilute polymer solutions. While several studies have shown

that the addition of suspended sediment within the flow causes a reduction in drag, the

exact mechanisms of drag reduction are not well understood. Several studies have tried

to quantify the effects of SFS concentrations on drag reduction (with concentrations as

high as 9% by volume).

Gust (1976) measured the mean streamwise velocity profiles for three smooth flow

systems: a Lucite and tap water fluid over a smooth bottom, a freshwater flow over a fine

quartz sand bed, and a seawater/clay mineral suspension over a mud bottom. Three

different fluid flows were used in order to determine the drag reduction for the clay

suspension flow at both non-eroding and eroding velocities. The research was designed

to examine whether the universal law of the wall,

U+ = K' lIn y + C, (Equation 3-1)

where

U+= dimensionless velocity-,
U.

u = local mean velocity,

u*= shear velocity,

y = dimensionless wall distance yu*, and
V

C,= integration constant,

that describes the dimensionless velocity profile for hydrodynamically smooth flows, was

applicable in the case of a cohesive suspended sediment flow over a mud bottom. In

addition, the research was to verify the existence of the drag reduction observed in earlier

studies.

Both the Lucite/tap water flow and the freshwater flow considered in the study

fulfilled Newtonian flow expectations by yielding measured velocity profiles that

followed the law of the wall equation. However, all of the seawater/clay suspension

flows indicated deviations from the universal law of the wall. As the seawater flows

were increased to the turbulent flow velocity range, the measured viscous sublayer

thickened to 5 mm above the mud bottom, as compared to a 1 mm thickness for the

Lucite/tap water and freshwater flows. No substantial increase in clay concentration was

measured within the viscous sublayer, thus there was not an increase in the kinematic

viscosity that would indicate compatibility with the law of the wall.

The author presented his results with plots of the dimensionless velocity versus

dimensionless height above the bed. All of the velocity profiles for the Lucite/tap water

and the freshwater systems overlapped, indicating no change in the velocity profile with

increasing Reynolds number. However, the dimensionless velocity profiles of the

seawater/clay suspension flows indicated the presence of a thickened viscous sublayer;

the thickness of the sublayer generally increased with increasing Reynolds number. A

specific relationship between the Reynolds number and the thickness of the viscous

sublayer was not apparent. Thus a correlation between the clay concentrations and the

dimensionless velocity profile could not be obtained. The author assumed that the clay

concentration, the type of clay material, and the shear rate of the flow influenced the final

velocity profile.

Best and Leeder (1993) noted that in previous work, drag reduction occurs when

the near wall turbulence structure is modified, thus linking drag reduction to the

mechanisms that produce turbulence. The authors conducted a series of experiments with

seawater/clay suspension (maximum concentration, 2.2 g/1) flows over a mud bed and

plotted the final velocity profiles for each seawater/clay mixture. The data showed a

decrease in the near bed velocities as the clay concentration increased; a plot of the

dimensionless velocity versus the dimensionless height revealed a thickening of the wall

layer with increasing clay concentration. The authors speculated that a decreased rate in

turbulent bursts within the turbulent boundary layers that leads to a reduced momentum

exchange within the boundary layer is a possible explanation for the reduction in drag.

In an additional series of tests on 0.5 mm and 0.8 mm sand, both plain seawater and

seawater/clay suspensions were allowed to flow over the bed sediment to investigate the

effects of the clay on bed formations. The flow velocity was allowed to increase to the

point of bed form movement initiation. Once the bed forms had reached equilibrium

position, the test was stopped and the bed forms were measured. For the 0.5 mm

sediment bed, the water mixed with clay was found to cause bed forms with much

smaller amplitudes and wavelengths than those generated in clear water even though the

test was run much longer than that of the test with plain seawater. It was believed that

the influence of drag reduction had not allowed the bed to reach a true movement

threshold. However, the tests conducted over the 0.8 mm sediment bed indicated that the

increased concentrations of clay had little to no effect on the bed forms created by the

flow. The sequence of bed forms generated in the seawater/clay suspension flow was

similar to that formed in plain freshwater flows. It was concluded that the thickened wall

layer formed by the clay concentrations had no effect on the course sediment.

Li and Gust (2000) examined the effect that clay suspensions have on the boundary

wall layer. The purpose of the research was to examine drag reduction at various clay

concentrations and flow velocities. In a series of experiments conducted in a

recirculating flume, velocity profiles and shear velocities within the viscous sublayer

were measured. The experiments involved suspended sediment kaolinitee) concentrations

ranging from 0.1 to 8 g/l in both freshwater and seawater. The authors wanted to

examine the effect of higher clay concentrations on drag reduction, as previous studies

had mainly encompassed low clay concentrations (less than 2.2 g/l). Logarithmic

velocity profiles were plotted for each of the flows.

The results of the study indicated that drag reduction increased with an increase in

the suspended clay concentration. The data indicated that the measured shear velocity

decreased as the concentration of the clay suspension increased. Thus the use of the

shear velocity calculated from the velocity profile will result in the overestimation of the

actual bed shear stress. A corresponding thickening of the viscous sublayer was

presumed to be the cause of the drag reduction. The thickening of the wall layer led to a

reduction in bed shear stress. The effects of the suspended sediment were most

prominent at low flow velocities; the effect was weaker as the flow velocity increased.

While the effect of suspended sediment on drag reduction was clearly shown, a

direct relationship between clay concentrations and drag reduction has not been

determined. However, the authors formulated the following empirical relationship:

u- = -72.73 (c +0.107u.*lo+ 0.451 (r2 = 0.760), (Equation 3-2)
U*log P, )

., =-114.93 -c +0.653u.o +0.0410 (r2 =0.903), (Equation 3-3)

where

u*s = measured shear velocity,

u*iog = profile-derived shear velocity,

c= clay concentration, and

p, = density of clay suspension sediment.

These equations show that if the clay concentrations and the velocity profiles are known,

the actual shear velocity, and the corresponding shear stress, can be quickly estimated.

Thus the magnitude of the drag reduction can be determined.

The authors concluded that, based on measurements from flume experiments that

utilized high kaolinite flows, the shear velocity at the top of the viscous sublayer could be

reduced by as much as 70%. The exact amount of the reduction depended on both the

clay concentration and the flow velocity; in general, the drag reduction increases with

increasing clay concentration at a given flow velocity. Also, the drag reduction is higher

at low flow velocity and decreases as the velocity increases for a given suspended clay

concentration. Finally, this thickening of the viscous layer as well as the turbulence

dampening was believed to be the cause of the reduced bed shear stress (Li and Gust,

2000).

In each of these studies, suspended fine sediment within the flow was shown to

increase the depth of the viscous sublayer, thus increasing the drag reduction. While the

effect of the suspended fine sediment itself has been proven, it has not yet been quantified

for varying flows. Further research is needed to quantify the effect of the suspended fine

sediment on the velocity profile and on the shear stress at the bed.

In each of the studies, the concentrations of the suspended fine sediment remained

fairly low, usually lower than 9 g/1. At higher concentrations, the suspended sediment

begins to flocculate; the effect of the flocculated particles is also not known.

CHAPTER 4
TEST APPARATUS AND PROCEDURE

Background

Rotating cylinder devices have been utilized in previous research in order to

measure the erosion rate of selected materials. The operation of the rotating cylinder

device is based on the theory of a rotational viscometer. A rotating viscometer is used to

measure the viscosity of liquids. A cylinder is rotated within a fixed outer cylinder the

annulus of which is filled with the fluid being tested. The viscosity of the fluid is

calculated from the measurements of the torque required to maintain the angular velocity

(Munson et al., 1994).

A rotating cylinder device was developed by Moore and Masch (1962) for the

purpose of measuring the scour resistance of cohesive soils. A cylindrical cohesive soil

sample, supported by a hollow tube, is centered and suspended within a slightly larger

cylinder. The annular gap between the outer and inner cylinder is filled with water. The

outer cylinder is rotated while the inner cylindrical sample remains fixed. A shear stress

is applied to the surface of the sample by the fluid. The torque exerted on the sample is

measured and the average shear stress on the sample surface computed.

Additional research has been conducted with rotating cylinder devices by Rektorik

(1964), Arulanandan et al. (1975), Sargunam et al. (1973), Alizadeh (1974), and Chapius

and Gatien (1986) (Kerr, 2001). Each set of research slightly modified and improved

upon the original device.

University of Florida Testing Apparatus

A Rotating Erosion Test Apparatus (RETA) was developed at the University of

Florida for the purpose of measuring erosion rates of certain types of rock (such as lime

rock, sand stone, coquina, etc.). Even though the RETA was designed to measure erosion

rates as a function of shear stress for erodible rock, it can be used with any sediment that

can support its own weight, such as a stiff clay.

The RETA is very similar to the device developed by Moore and Masch (1962) but

has improved instrumentation for measuring torque. As stated above, the computed shear

stress is the average value over the sample surface; thus the stress can be higher at certain

locations. There may also be variations in surface roughness that cannot be accounted for

using this approach. This problem is not unique to the current RETA. Rohan and

Lefebvre (1991) showed that the shear stress measurements generated by rotating

cylinder devices could be underestimated. The low estimates could be caused by the

centrifugal forces that cause the fluid to move toward the outer cylinder, creating a

helicodial, secondary flow. Fluctuations in the velocity radial components can also lead

to an underestimation of the shear stress when the fluid is in the fully turbulent regime

(Kerr, 2001).

Figure 4-1 is a photograph of the University of Florida RETA. While the RETA is

based on the designs of earlier rotating cylinder devices, several practical modifications

have been incorporated into the new unit. A back slide plate has been added to allow the

test cylinder to be easily lowered into the outer cylinder as well as to allow the sample to

remain attached between tests. A torque cell is used to measure the torque exerted on the

inner cylinder. A variable speed motor is used to rotate the outer cylinder. The

tachometer, which is connected to the motor shaft, is programmed to show the corrected

RPM value based on the gear ratio of the gearing system (Kerr, 2001).

Figure 4-1. University of Florida Rotating Erosion Test Apparatus (RETA).

Current Testing

Modifications to RETA

As stated previously, the RETA was designed to measure the erosion rate as a

function of shear stress for various rock samples. Because rock specimens are highly

resistive to the erosive forces of water, these tests can last for days. Thus the gearing

system was originally designed for long tests at high angular velocities (revolutions per

minute). Even though these angular speeds were suitable for the rock samples (rock tests

required RPM values of approximately 1000 or greater), they were too large for the shear

stresses needed for this work. Therefore, the 3:1 step-up gear system used in the

apparatus was replaced by a 1:1 gear/belt system, allowing the outer cylinder to rotate at

speeds as low as 200 RPM. Since the original tachometer could not be reprogrammed to

show the actual RPM of the cylinder, the tachometer output was divided by three in the

analysis to give the actual RPM value.

In order to eliminate as much noise as possible from the digital readout of the

torque values, an electronic filter was used in accordance to the torque cell

manufacturer's guidelines. The filter eliminates noise as well as allows the torque

readout to stabilize faster once a particular torque value is achieved.

Procedure

In previous tests using the RETA, the test cylinders were core samples taken

directly from rock beds. However, because these tests involved sand beds, a cylinder

could not be created in the same method. For these tests, two acrylic cylinders were

coated with similar sized sediment as was used in the USGS-BRD scour testing: one

cylinder was coated with 0.15 mm sediment and another with 0.85 mm sediment. A two-

part epoxy was used to adhere the sediment onto the cylinder. The cylinders are

approximately 10.2 cm in height and 6 cm in diameter and are closed at both ends. An

attempt was made to coat an additional cylinder with 2.0 mm sediment to simulate the 2.9

mm bed sediment size; however, due to the angularity and the size of the sediment it was

difficult to create a cylinder with an even roughness.

In addition to the sediment-coated acrylic cylinders, two additional test cylinders

were made from aluminum. The cylinders were knurled to simulate two different

roughnesses equivalent to those achieved with the 0.15 mm and 0.85 mm sediment

coated acrylic cylinders. The surface roughness was more uniform for the aluminum

cylinders than for those coated with sand. Figure 4-2 is a photograph of all four cylinders

used in the RETA tests.

Figure 4-2. Photograph of cylinders used in RETA tests. The left two are the aluminum
cylinders; the two cylinders on the right are acrylic cylinders coated with sand.

Tests were performed with each of the cylinders with four different (SFS)

(bentonite/water) concentrations: 0.0 g/l, 0.05 g/l, 0.5 g/l, and 1.0 g/l. The 0.0 g/1 water

tests were performed first to establish the proper RPM values. These RPM values were

then replicated in the subsequent three sets of tests.

Once the test cylinder was placed in the apparatus and the outer cylinder filled with

distilled water, the cylinder was allowed to spin at the lowest possible speed in order to

generate a base torque (and shear stress) value. This shear stress value is compared to the

base shear stress originally calculated for the USGS-BRD tests. After the base torque is

generated, the torque values originally calculated from the USGS-BRD tests representing

increasing magnitudes of the base value were simulated. The RPM values generated

from these torque values are recorded. Following the tests, the clear water was discarded;

however, the cylinder remains on the torque cell for subsequent testing. The initial and

final water temperatures, as well as the final resting torque value, were noted for each test

set.

The first bentonite/water mixture was then poured into the annulus between the

cylinders, and the initial set of tests was repeated. The difference in these tests was that

the RPM values, not the torque values, are the control values. The new torque values,

generated from the same RPM's measured in the 0.0 concentration test, were recorded.

The tests are to indicate if, at the same rotational speed, the shear stress changes with the

addition ofbentonite in the fluid. The tests were designed to simulate velocities

equivalent of ratios between 1 and 4.
UC

The torque values to be simulated in the RETA were calculated from the critical

shear stress values developed in Chapter 2. The torque is calculated from the critical

shear stress using the following equation:

T= 2nR2Ltc (Equation 4-1)

where

R = radius of sample cylinder,

L = length of sample cylinder,

T= torque, and

Tc = critical shear stress.

For each cylinder, a minimum of five readings, representing increasing magnitudes of the

calculated torque value, were recorded.

The test procedure is outlined below:

0.0 g/l Test:

1. Attach the test cylinder onto threaded rod and secure in place with end plate. Place
acrylic lid over the sample and screw top end of threaded rod into torque meter.

2. Fill the plastic insert 1/3 full (approximately 200 mL) with distilled water. Record
the initial water temperature. Place the plastic insert into the acrylic annulus.

3. Lower the test cylinder into annulus, ensuring that the test cylinder is centered and
that water rises just above the top of the cylinder.

4. Add water if necessary.

5. Secure the cylinder lid. Make sure that the torque cell is locked into place on the
slide rail.

6. With the machine on and the cylinder still at rest, tare the torque meter so that it
reads 0.000 N-mm.

7. Make sure that the RPM control dial is set to the lowest speed.

8. With the directional switch pointing to the forward direction, turn on the motor.

9. Once the outer cylinder begins spinning, check that the water level is even; if not,
add more water.

10. Take an initial (base) reading at the lowest RPM.

11. The control values in the 0.0 concentration test are the torque values. Increase the
speed of the outer cylinder slowly until the first torque value is shown on the digital
readout. Record both the actual torque value achieved as well as the RPM.

12. Continue taking torque and RPM measurements at the values for 2*Initial Torque,
3*Initial Torque, etc.

13. Once finished, gradually reduce speed to the lowest possible RPM, then turn off
motor.

14. Record the resting torque value once the cylinder has stopped spinning.

15. Gently slide the test sample out of the acrylic annulus (test sample can remain
attached to the torque cell between tests).

16. Remove the plastic insert and record the final water temperature.

17. Discard water.

18. Carefully dry the test sample.

0.05 g/ 1.0 g/l Tests:

19. Measure 250 mL of distilled water.

20. Measure the appropriate amount of bentonite (record exact amount used), and mix
bentonite with the water.

21. Record the initial water temperature.

22. Fill the plastic insert 1/3 full (approximately 200 mL) with the bentonite/water
mixture. Retain remaining mixture in case additional is needed.

23. Place the plastic insert into the acrylic annulus.

24. Lower test sample into the outer cylinder, securing it in place.

25. Once the lid is tightened, tare the torque readout.

26. Repeat the test procedure from the 0.0 concentration test, this time simulating the
RPM values generated in that test. Record the new torque values achieved at each
RPM speed.

27. Once complete, gradually slow the outer cylinder to the lowest speed, then turn off
the motor.

28. Record the final resting torque value once the cylinder has stopped spinning.

29. Remove the test sample and the plastic insert from the acrylic annulus. Record the
final water temperature.

30. Dispose of the bentonite/water mixture and completely dry the plastic insert prior
to the next test. Carefully dry the test sample.

31. Repeat steps 19-30 for each remaining bentonite/water concentration.

Four sets of RPM versus torque data sets were obtained. The average shear stress

was computed from the measured torque values using Equation 4-1. The results were

plotted as shear stress versus RPM for different SFS concentrations.

As mentioned previously, the studies of Gust (1976), Best and Leeder (1993), and

Li and Gust (2000) incorporated mixtures of seawater and SFS in their research on drag

reduction. In order to see the effect of salinity on the SFS induced shear stress reduction,

one set of tests was conducting using seawater. The seawater was collected from the

Intracoastal Waterway on the east coast of Florida and had a measured salinity of 29.4

parts per thousand. The bentonite/seawater mixture was allowed to stand for ten days

41

prior to the tests in order to facilitate the reactions between the bentonite and salt

particles. The tests then followed the same procedure as was used in the freshwater tests.

The saltwater tests were limited to one on each sediment-coated acrylic cylinder in order

to reduce potential corrosion damage to the RETA.

CHAPTER 5
TEST RESULTS AND OVERALL CONCLUSIONS

Results

Five sets of tests were conducted on each of the two sediment-coated acrylic

cylinders, and two test sets were conducted on each of the two aluminum cylinders. A set

consisted of testing a range of RPM values at each of the four SFS concentrations. The

torques generated in these tests were near the lower limits that can be measured with the

torque cell on the RETA. This resulted in errors in the readings beyond those discussed

earlier. In order to minimize these errors (that are thought to be random in nature) the

tests were repeated for each set of conditions and the results averaged. Plots of the

individual tests are not conclusive, but there is a definite pattern of shear stress reduction

in the plots of the average values.

The tests with the acrylic cylinders (with bonded sand) show a greater reduction in

shear stress than the aluminum cylinders for some reason. There is also a greater shear

stress reduction with SFS concentration for the acrylic cylinder with the 0.15 mm sand

than for the one with the 0.85 mm sand. This is most likely due to the increased

turbulence generated by the larger sand particles and the corresponding reduction in the

viscous sublayer. The results of the average shear stress curves for the two acrylic

cylinders are shown in Figures 5-1 and 5-2. The results of all of the individual cylinder

tests are presented in Appendix A.

0.15mm Acrylic Cylinder- Average Concentration Lines
(Tests with excessive values removed)

1.4

1.2

S 0.8 __

00 . . .
0" .6

-0-0.05 g/l
A- 0.5 g/l
0.2 -X- 1.0g/ -

0
200 250 300 350 400 450 500
RPM

Figure 5-1. Average shear stress lines generated from tests conducted on 0.15 mm sediment coated acrylic cylinder.

0.85mm Acrylic Cylinder- Average Concentration Lines
(All freshwater tests)

2.5

2

S1.5

- 0.5 g/l
-X- 1.0 g/1

0
200 250 300 350 400 450 500
RPM

Figure 5-2. Average shear stress lines generated from tests conducted on 0.85 mm sediment coated acrylic cylinder.

The shear stress plot for the tests conducted with the 0.15 mm sediment coated

acrylic cylinder shows about a ten percent reduction in shear stress with the addition of

SFS. There is an increase in shear stress reduction with increasing RPM, but the data is

not accurate enough to discern an accurate dependence on SFS concentration. The results

of the 0.85 mm sediment coated acrylic cylinder were less conclusive; the plot of the

average concentration curves only shows a 7% shear stress reduction at a suspended

sediment concentration of 0.5 g/1. The curves for 0.05 g/1 and 1.0 g/1 concentrations

overlap the curve for the clear water test.

The results of the tests with the two aluminum cylinders were mixed. For both

cylinders, the first test set indicated a reduction in shear stress, while the second test set

showed an increase in shear stress with increasing SFS concentration. Because of the

conflicting results, an analysis of the average concentration lines was not conducted for

the aluminum cylinders. Why the test results with these cylinders were different from

those with the bonded sand is not clear.

The saltwater tests conducted on the 0.15 mm sediment-coated acrylic cylinder

yielded up to a 7% reduction in shear stress with the addition of SFS. This reduction was

slightly less than what was observed in the freshwater tests. The results of the 0.85 mm

sediment coated cylinder showed only a slight shear stress reduction, less than 5%. This

result was consistent with the fresh water results for the same cylinder. Results from the

saltwater tests are also provided in Appendix A.

Discussion

Based on the results of the RETA tests, the SFS appeared to have a greater effect

on the shear stress of the cylinder coated with the smaller sediment. A reduction in shear

stress up to 10% was noted for the 0.15 mm sediment-coated cylinder. The reason for the

smaller shear stress reduction with the 0.85 mm sediment-coated cylinder is most likely

due to increased turbulence and a reduced viscous sublayer thickness. A similar result

would be expected in a field situation, i.e., less shear stress reduction would be expected

for a bed with large sediment than for one with fine sediment.

The RETA itself could be a source of some error in the tests. Though the gearing

system was modified to operate at lower RPM values, the angular velocity varied with

time at the lowest RPMs.

In spite of these problems, the tests with the RETA did conclusively show that a

reduction in shear stress does occur with the presence of SFS in the water column. It can

therefore be concluded that shear stress reduction is at least one of the possible causes for

the observed reductions in equilibrium local scour depths due to the presence of SFS in

the water column.

Conclusions

The results of the USGS-BRD and University of Auckland scour tests as well as

the shear stress measurements generated in the RETA provide valuable insight as to the

possible causes of the reduction in scour depth due to an increase in SFS. Video data

recorded during the tests at the USGS-BRD laboratory indicated that the turbidity of the

water did not increase suddenly; thus the SFS affected both the upstream flow conditions

as well as the formation of the scour hole. In Gust (1976) and others, an increase in

suspended sediment caused an increase both in drag reduction and in the height of the

viscous sublayer. If this were the case in the scour tests, the velocity upstream of the pier

would be reduced. Thus the velocity and subsequent shear stress within the scour hole is

lower than would be found in a clear water situation. A reduction in shear stress in the

scour hole reduces the sediment removed around the pier.

An additional aspect of the scour depth reduction is the pattern of the initial rate of

scour in comparison with the test not affected by the suspended sediment. The test

affected by the SFS initially follows the same scour rate as the non-affected test; at a

certain time into the test, the scour depth levels off and stops increasing. If the shear

stress was reduced upstream of the pier, the rate of scour would seemingly not be the

same as that of a test not affected by the suspended sediment. Since there is no sudden

increase in turbidity hours into the scour test, the effect of the suspended sediment must

have a gradual influence on the scour hole. The theory of this gradual influence is that

the suspended fine sediment is slowly deposited on the upstream slope of the scour hole.

Initially, this fine sediment acts as a lubricant, allowing bed particles to fall down the

slope of the scour hole slightly faster than normal. The scour hole development initially

proceeds as normal. Once the amount of deposition reaches a certain point, the cohesive

bonds between the fine sediment cause the bed particles to bond together, preventing both

sediment removal and the avalanching of sediment into the scour hole. Thus the scour

depth does not increase.

The results of the RETA tests indicate that the reduction in shear stress with

increasing suspended fine sediment concentration is greater with a smaller bed sediment

size. The shear stress reduction for the 0.15 mm sediment-coated cylinder was higher

than that found for the 0.85 mm sediment-coated cylinder. The tests affected by the

suspended fine sediment were all ones with 0.22 mm bed sediment. Thus one possible

condition of the scour depth reduction is the bed sediment size.

University of Florida researchers are developing a relationship between effective

shear stress in the scour hole and scour depth. This relationship is useful in explaining

the possible effect of the suspended sediment on the scour depth. The shear stress is

normalized by the upstream bed shear stress and the scour depth is normalized by the

equilibrium scour depth (see Figure 5-3). For clearwater scour tests, the predicted
ds

curve pattern shows an initial increase in effective shear stress as the scour process

begins. The non-dimensional shear stress reaches a peak and then begins to decrease. At

some point, possibly when the horseshoe vortex is located entirely within the scour hole,

the slope of the shear stress curve changes sharply and the rate of scour remains small

until the critical shear stress is reached and the scour stops. The location of the "break

point" in this curve varies for flow, sediment, and pier conditions.

It is known that the level of bed shear stress reduction due to the presence of SFS

decreases rapidly with increased velocity and bed roughness. It seems reasonable that the

effects of SFS will be small during the initial stages of local scour, but will increase as

the scour hole develops and the turbulence level decreases. According to the normalized

shear stress versus normalized scour depth plot shown in Figure 5-3, a small reduction in

the shear stress in the latter stages of the scour hole development can reduce the

equilibrium depth significantly.

Another possible mechanism for equilibrium scour depth reduction is as follows.

Some of the SFS is deposited in the scour hole out from the structure due to the reduced

flow velocities in this region. As the surface material in this area avalanches into the

scour region near the structure, the fine sediment acts to increase the critical shear of the

material. As can be seen in Figure 5-3, an increase in tc will decrease the equilibrium

scour depth.

While it is not possible with the limited amount of data to positively identify the

cause of scour depth reduction, the evidence points to the conclusion that the reduction in

shear stress due to SFS is at least one of the causes. A variety of experiments will be

required to determine the relative importance of the various proposed mechanisms.

T,
1;u

"Break Point"

d,

Figure 5-3. Proposed relationship between shear stress and scour depth.

CHAPTER 6
FUTURE RESEARCH

The scour experiments conducted at both the USGS-BRD Laboratory in

Massachusetts and at the University of Auckland have confirmed that an increase in

suspended fine sediment concentration within the flow affects the local equilibrium scour

depth. Because the exact effects of the suspended fine sediment on the local scour depth

appeared to differ with varying flow velocities and sediment diameters, continued

research is needed in order to quantify the change in scour depth for the purposes of scour

prediction.

Clearwater versus live-bed scour: The tests conducted at the University of

Auckland were directed at establishing the relationship between equilibrium scour depth

and SFS concentration and flow velocity. There were, however, problems with the test

procedure that most likely affected the results. Due to the number of tests required and

the time available, the flume with SFS concentrations was allowed to stand overnight

(and even longer periods of time) between tests. It is suspected that this resulted in

deposition of fine sediment throughout the flume, which accumulated as the tests

progressed and impacted the critical shear stress. The initial test results were consistent

with the findings at the USGS-BRD Laboratory but started to deviate as the tests

progressed. In particular, two of the live bed tests actually indicated an increase in scour

depth with the presence of SFS. Additional tests should be conducted to further examine

the effect of SFS in both the clearwater and live bed velocity ranges; however, care

should be take to ensure that the test procedure itself does not cause unintended effects on

the test results.

Bed sediment diameter: Both the tests conducted at the USGS-BRD Laboratory

and at the University of Auckland utilized a similar median grain size sediment. Most of

the research summarized in Chapter 3 indicated an increase in the height of the viscous

sublayer over mud beds. Best and Leeder (1993) examined drag reduction on sand beds

and found that SFS had a greater impact on the bed forms in the finer sands. Thus the

Dso of the bed likely contributes to the effect of the SFS on the local scour depth. Further

testing is required to determine if the effect of the SFS changes with a change in the bed

sediment size.

Suspended fine sediment concentration: During one of the original USGS-BRD

scour tests, a depth-averaged suspended fine sediment concentration of 0.029 g/1 was

measured. Because this concentration was not measured during a test in which the

highest concentrations of suspended fine sediment were recorded, the tests in this paper

incorporated concentrations up to 1.0 g/1. Additional research is needed to determine if

there is a direct correlation between SFS concentration and a change in the measured

scour depth. Scour depth changes may be limited to flows with low suspended fine

sediment concentrations.

Alteration of scour prediction equations: When the effect of suspended fine

sediment on equilibrium is quantified, the current equations used to predict equilibrium

scour depth will need to be modified. The causes of the scour depth reduction point to

possible ways in which this effect can be incorporated into scour prediction equations.

For example, the University of Florida currently uses the following formula for

clearwater scour prediction:

d= KScf f2 (-- 7 3 (-b- (Equation 6-1)

This equation can be modified in one of two ways: first, a constant that quantifies scour

depth reduction can be directly included in the equation. Second, the function of

- can be modified to include the effect, since the reduction in bed shear stress is

similar to a reduction in velocity.

These research topics are necessary in order to quantify the effect of SFS on local

scour depth and to facilitate the alteration of current scour prediction equations to

incorporate the effect of the suspended sediment.

APPENDIX A
ROTATING EROSION TEST APPARATUS (RETA) TEST DATA

0.15mm Acrylic Cylinder Tests 5-8

1.4
.A

1.2 o

1

- 0.8

0.6 -

0.4

-- 0.0 g/1
---0.05 g/I
0.2 ---- A- 0.5 g/l

-X- -1.0 g/l

0
200 250 300 350 400 450 500
RPM

Figure A-1. RETA Tests 5-8, 0.15 mm Sediment Coated Acrylic Cylinder.

0.15mm Acrylic Cylinder Tests 13-16

1.4

1.2

1
-

E 0.8

.b
S

0.6
-

0.4 r

0.2

0
200

Figure A-2. RETA Tests 13-16, 0.15 mm Sediment Coated Acrylic Cylinder.

250 300 350 400 450
RPM

500

0.15mm Acrylic Cylinder Tests 25-28

1.4

1.2

1

E 0.8

S0.6

0.4

0.2

200 250 300 350 400 450
RPM

Figure A-3. RETA Tests 25-28, 0.15 mm Sediment Coated Acrylic Cylinder.

500

0.15mm Acrylic Cylinder Tests 29-32

1.4

1.2

1 ------------__ __------

E 0.8 __

0.6

---- 0.0 g/I

0.2 ---0.05 g/I
- 0.5 g/I
X- 1.O g/I_

0
200 250 300 350 400 450 500
RPM

Figure A-4. RETA Tests 29-32, 0.15 mm Sediment Coated Acrylic Cylinder.

0.15mm Acrylic Cylinder Tests 37-40

1.4

1.2

1

S0.8
S

0.6

0.4

0.2

200 250 300 350 400 450
RPM

Figure A-5. RETA Tests 37-40, 0.15 mm Sediment Coated Acrylic Cylinder.

500

0.15mm Acrylic Cylinder All 0.0 gll Tests

1.4

1.2

E 0.8

0.6

0.4
1--*-Test 5
--Test 13
-h-Test 25
0.2 -X Test 29
-N-Test 37
Test 57 (salt)

0
200 250 300 350 400 450 500
RPM

Figure A-6. All 0.0 g/1 concentration tests conducted on 0.15 mm Sediment Coated Acrylic Cylinder.

0.15mm Acrylic Cylinder All 0.05 g/l Tests

1.4

1.2

1 -

B 0.8

0.6

0 -*-Test 6
-U-Test 14
-.Test 26

0.2 Test 30
-- Test 38
- -Test 58 (salt)

0
200 250 300 350 400 450 500
RPM

Figure A-7. All 0.05 g/1 concentration tests conducted on 0.15 mm Sediment Coated Acrylic Cylinder.

0.15mm Acrylic Cylinder All 0.5 g/l Tests

1.4

1.2

E
z 0.8 -
a

0.6

0.4
0.4 4--Test 7
-U-Test 15
-A-Test 27
0.2 -4- Test 31
-t- Test 39
- -Test 59 (salt)

0
200 250 300 350 400 450 500
RPM

Figure A-8. All 0.5 g/l concentration tests conducted on 0.15 mm Sediment Coated Acrylic Cylinder.

0.15mm Acrylic Cylinder All 1.0 g/l Tests

1.4

1.2

1
-~

E 0.8

0.6

0.4

0.2

0
200

Figure A-9. All 1.0 g/1 concentration tests conducted on 0.15 mm Sediment Coated Acrylic Cylinder.

250 300 350 400 450
RPM

500

0.85mm Acrylic Cylinder Tests 1-4

2.5

2

E 1.5

0.5
0.5 -0-0.0 g/I
0.05 g/I
A- 0.5 g/I
-X- ,1.0 g/I
0
200 250 300 350 400 450 500
RPM

Figure A-10. RETA Tests 1-4, 0.85 mm Sediment Coated Acrylic Cylinder.

0.85mm Acrylic Cylinder Tests 9-12

2.5

2= .5 -----------------f--------------------------------
2

0.5 0

0.5 --0.0 g/I

*- 0.5 g/l
-X- *1.0 g/I

0
200 250 300 350 400 450 500
RPM

Figure A- 1. RETA Tests 9-12, 0.85 mm Sediment Coated Acrylic Cylinder.

0.85mm Acrylic Cylinder Tests 17-20

2.5

2

S-

S1.5

0.5
A

0 -
200

Figure A-12. RETA Tests 17-20, 0.85 mm Sediment Coated Acrylic Cylinder.

250 300 350 400 450
RPM

0.85mm Acrylic Cylinder Tests 21-24

0.50 i

0 -
200

250 300 350 400 450
RPM

Figure A-13. RETA Tests 21-24, 0.85 mm Sediment Coated Acrylic Cylinder.

500

0.85mm Acrylic Cylinder Tests 33-36

2.5

2

r1.5

E

0.5 -",0--0.0 g/I
-- -0.05 g/I
AL- 0.5 g/I
-X- '1.0 g/l

0
200 250 300 350 400 450 500
RPM

Figure A-14. RETA Tests 33-36, 0.85 mm Sediment Coated Acrylic Cylinder.

0.85mm Acrylic Cylinder All 0.0 g/I Tests

2.5

2

1.5

----Tet 1

0.-Test 9
0.5 --Test 17

--Test 21
Test 33
- -Test 61 (salt)

0
0 ,--------------------------------------------

200 250 300 350 400 450 500
RPM

Figure A-15. All 0.0 g/1 concentration tests conducted on 0.85 mm Sediment Coated Acrylic Cylinder.

0.85mm Acrylic Cylinder All 0.05 gll Tests

2.5

2

1.5

a

cn

0.5

0
200

Figure A-16. All 0.05 g/1 concentration tests conducted on 0.85 mm Sediment Coated Acrylic Cylinder.

250 300 350 400 450
RPM

500

0.85mm Acrylic Cylinder All 0.5 g/I Tests

2.5

2

1.5

-*--Test 3
0.5 --Test 11
-Test 19

-*-Test 23
--Test 35
-.--Test 63 (salt)

0
200 250 300 350 400 450 500
RPM

Figure A-17. All 0.5 g/l concentration tests conducted on 0.85 mm Sediment Coated Acrylic Cylinder.

0.85mm Acrylic Cylinder All 1.0 g/l Tests

2.5

2

P 1.5
E

I 1

0.5

0
200

Figure A-18. All 1.0 g/1 concentration tests conducted on 0.85 mm Sediment Coated Acrylic Cylinder.

250 300 350 400 450
RPM

500

Smooth Aluminum Cylinder Tests 45-48

250 300 350 400 450 500
RPM

Figure A-19. RETA Tests 45-48, Smooth Aluminum Cylinder.

1.6

1.4

1.2

S 1

0.8

I 0.6

0.4

0.2

0-
200

Smooth Aluminum Cylinder Tests 49-52

1.6

1.4

1.2

*9 1-

8 0.8
-

0.6

0.4
C-

0.2

0
200

Figure A-20. RETA Tests 49-52, Smooth Aluminum Cylinder.

250 300 350 400 450 500
RPM

550

Rough Aluminum Cylinder Tests 41-44

2.5

2

1.5
E

0

0.5

0
200

250 300 350 400 450 500 550 600
RPM

Figure A-21. RETA Tests 41-44, Rough Aluminum Cylinder.

Rough Aluminum Cylinder Tests 53-56

2.5
A

2 "-0.0 g/I ----
1.5

0.5 g/I
I 1

-X- 1.0 g/I

200 250 300 350 400 450 500 550 600 650
RPM

Figure A-22. RETA Tests 53-56, Rough Aluminum Cylinder.

0.15mm Acrylic Cylinder Tests 57-60
Saltwater Tests

1.4

I. ---- ----- --- -
1.2

E
0.8
a

S0.6

0 .4 _

--*--0.0 g/I
-- --0.05 g/1
0.2 *A- 0.5 g/l
-X- 1.0 g/1

0
200 250 300 350 400 450 500
RPM

Figure A-23. RETA Tests 57-60, saltwater tests conducted on 0.15 mm Sediment Coated Acrylic Cylinder.

0.85mm Acrylic Cylinder Tests 61-64
Saltwater Tests

2.5

X

'r-
E 1.5

--*--o o g/l
0.5 -W-0.05 gil
A- 0.5 g/l
-X- -1.0 g/1

0 ---
200 250 300 350 400 450 500
RPM

Figure A-24. RETA Tests 61-64, saltwater tests conducted on 0.85 mm Sediment Coated Acrylic Cylinder.

APPENDIX B
UNIVERSITY OF AUCKLAND SCOUR TESTING

Introduction

In order to quantify the effect of suspended fine sediment (SFS) concentrations on

local scour in cohesionless sediments, a series of tests were initiated by University of

Florida researchers in a flume in the Hydraulics Laboratory at the University of

Auckland. An undergraduate student completed these tests as part of an Honors Project.

The purpose of these tests was not to provide information regarding the cause of the

scour reduction, simply the magnitude of scour reduction as a function of the suspended

sediment concentration and the normalized flow velocity.

The results are presented here with the permission of Mr. Thomas Macdougal

Clunie, the student who performed the tests. These results are designed to give further

evidence of the effect of SFS on local equilibrium scour.

Facility

The University of Auckland (UA) flume used for these tests is a glass-sided flume

that is approximately 0.440 m wide, 0.44 m deep, and 19 m long. Unlike the USGS-BRD

flume in Massachusetts, the UA flume is a closed flume; all of the water and sediment are

recycled. A 0.15 m diameter supply line controls the discharge, while a tailgate at the

downstream end controls the water level within the flume. Supported by a central pivot,

the slope of the flume can be adjusted using manual screw-jacks located at both ends of

the flume (Ettema, 1980).

Test Parameters

Local scour tests were conducted for four SFS concentrations (0.0 g/l, 0.1 g/l, 0.5

g/l, and 1.0 g/1) and 4 normalized velocities (0.95 U, 1.1 U, 1.5 U, and 2.0U,). The first

velocity ratio is in the clearwater scour range, while the remaining three are in the live

bed scour range. Note that the USGS-BRD Experiments A and B discussed in Chapter 2

were conducted at velocity ratios of 0.92 and 0.97, respectively; thus the clearwater tests

fall within the range of USGS-BRD tests where scour reduction due to SFS was

observed. The sediment concentrations used were designed to approximate the

magnitude of the SFS concentrations measured in the USGS-BRD scour tests. The test

structure was a 50 mm diameter Plexiglas cylinder. The bed sediment had a Dso of 0.24

mm. The water depth was maintained at 170 mm for all of the tests.

Even though the size and type of flume used in the UA tests was very different

from that at the USGS-BRD flume facility, the parameters Y. and U in the UA tests
b U,

were similar to those in the USGS-BRD tests. The bed sediment diameters for the two

tests were also similar (0.22 mm for the USGS-BRD tests and 0.24 mm for the UA tests).

b
However, the UA pier diameter (and therefore the ratio) was only approximately 5%
D50

for that of the USGS-BRD tests. As discussed in Chapter 1, local scour depth has been

found to depend primarily on the non-dimensional quantities o, and b. The UA
b Uc D50

tests were conducted at about the same value of o- as the USGS tests, but the value of
b

b U
- was much smaller. The UA tests covered a range of values including those in
D5o Uv

the USGS tests.

Procedure

A procedure, similar to that used in the USGS-BRD tests, was followed in the UA

tests. Scour depth was measured using a single acoustic transponder that measured depth

approximately 15 mm in front of the circular pier. In order to ensure that the transponder

would not be affected by the addition of SFS, the transponder was tested for a known

depth over the range of suspended sediment concentrations. The transponder recorded

the correct depth each time.

The pre, during, and post-experiment procedure used for these tests is outlined

below:

Pre-experiment:

1. Compact and level the bed in the flume.

2. Thoroughly mix desired concentration of bentonite powder in water. To ensure that
bentonite flocs do not form, pass the mixture through a 60tm sieve.

3. Pour concentrated solution into flume.

4. Start acoustic transponder.

During experiment:

5. Begin experiment, quickly increasing the velocity of the flow to predetermined
value.

6. Measure the scour depth versus time using acoustic transponder.

7. Record water velocity, flow depth, and temperature, as well as observations of the
water turbidity.

8. Gather samples of the water during the experiment to determine actual suspended
fine sediment concentrations.

9. Once the equilibrium scour depth has been reached, stop experiment.

Post-experiment:

10. Note the condition of the bed following the test.

11. Conduct data reduction and analysis; plot time-history of scour depth
measurements.

The tests were conducted in the order of increased SFS concentrations, i.e., all of

the zero concentration tests at the four velocity ratios were conducted first. This was

followed by tests with the smallest SFS concentration at all four velocities, and so on

until the 1.0 g/1 concentration test was completed. Due to time constraints, the flume was

not emptied between scour tests. This appears to have influenced the test results.

Results

The time-history scour plots are compiled into Figures B-1 through B-4. Each plot

represents a comparison of the entire scour tests conducted at a single velocity but at

varying SFS concentrations.

Initial results shown in the 0.95 Uc tests indicate that the presence of suspended fine

sediment did cause up to a 15% reduction in equilibrium scour depth near the end of the

tests. The lowest equilibrium scour depths were recorded at the 0.84 g/l and the 0.1 g/1

concentrations. The test at 0.5 g/l suspended fine sediment yielded a 10% reduction in

equilibrium scour depth.

Three sets of tests were conducted in the live bed range of velocities. The first set,

with a velocity of 1.1 Uc, indicated that as suspended fine sediment was added to the flow,

the measured equilibrium scour depth actually increased; at the end of the allotted time,

almost a 10% increase in the scour depth was shown between the tests with 0.1 g/1 and

0.5 g/l and the test with no suspended fine sediment.

The tests conducted at 1.5 U show the movement of bed forms through the scour

hole that is characteristic of tests conducted in the live bed velocity range. A reduction in

scour depth, which could approach 8% depending upon the position of the bed forms,

was indicated at the end of the allotted test time. However, the average difference in

scour depth between the suspended sediment tests and the tests without the sediment was

only around 5%.

The final test series, conducted at 2.0Uc, shows dramatic bed forms moving through

the scour hole during the test. While the final plot shows that the measured scour depths

overlap for the tests with the suspended fine sediment, it does indicate that the presence

of the suspended fine sediment causes an increase in the scour depth. This is in contrast

with the thinking that the presence of suspended sediment only decreases the equilibrium

scour depth. However, the fact that the flume was not drained between tests may have

affected these results. Additional sediment settling onto the bed between the tests may

actually change the velocity ratio of the 2.0Uc tests by increasing the critical velocity.

Discussion

The four sets of scour tests conducted at UA provide some useful information about

the effect of SFS concentrations on equilibrium local scour depths. In the test conducted

in the clearwater velocity range, reductions in scour depth from 10-15% were measured.

The results from the live bed scour tests were somewhat conflicting, but they do indicate

that the effects of SFS on equilibrium scour depths is reduced significantly with increased

flow velocity. The tests at 1.5Uc indicated a slight reduction (approximately 5%) in scour

depth, while the tests at 1.1 and 2.0Uc actually yielded an increase in the measured scour

depth.

The magnitudes of the scour reduction in the UA tests were lower than those

experienced in the USGS tests (up to 40% in the USGS-BRD tests and only 15-20% in

the UA tests). The suspended sediment in the USGS-BRD tests represented that

occurring naturally from rainwater and snowmelt runoff. The suspended sediment in the

UA tests was bentonite. One explanation for the contradictory finding in the UA tests

that the scour depth increased with the presence of SFS is that fine sediment was

deposited on the bed during the intervals between tests. Small amounts of fine sediment

mixed with cohesionless sediments (sand) have been found to reduce the critical shear

stress compared to that for the sand alone. The thin coating of fine sediment on the sand

84

grains acts as a lubricant and reduces the stresses required to initiate sediment motion. A

U
reduced critical shear stress would mean an increased -, and thus an increased scour
Ud

depth.

0.95 Uc Local Scour Tests

I II II :1111- III .1111
Tirre (sec)
0.0 g/ 1-----0.1 g/l--- 0.2gA -.--.- 0.5 g/l- -0.84 g/1------- 1.0g/1

Figure B-1. University of Auckland Scour Tests, 0.95 Uc.

1.1 Uc Local Scour Tests

O.3E

w- .* .... .. .. .. ....

0.00
0.34
0. 12:11111

0.3:
o.J1

C
0 203C OXO 63SO 83JX -0 X(J

-r-A 0.ec)

0.0 gA - -- 0.1 g/ O.Sg4 ....... 1.0 gn

Figure B-2. University of Auckland Scour Tests, 1.1 Uc.

MILLISECOND	CLASS.METHOD	MESSAGE
0	sobekcm_page_globals.constructor
0	sobekcm_page_globals.constructor	Application State validated or built
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.constructor	Navigation Object created from URI query string
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.display_item	Retrieving item or group information
0	sobekcm_page_globals.get_entire_collection_hierarchy	Retrieving hierarchy information
0	sobekcm_assistant.get_entire_collection_hierarchy
0	cached_data_manager.retrieve_item_aggregation
0	cached_data_manager.retrieve_item_aggregation	Found item aggregation on local cache
0	item_aggregation_builder.get_item_aggregation	Found 'all' item aggregation in cache
0	system.web.ui.page.page_load (ufdc.page_load)
0	sobekcm_page_globals.constructor.on_page_load
0	html_echo_mainwriter.add_style_references	Adding style references to HTML
0	html_echo_mainwriter.add_text_to_page	Reading the text from the file and echoing back to the output stream
53	html_echo_mainwriter.add_text_to_page	Finished reading and writing the file