Local sediment scour at large circular piles

UFL/COEL-99/016

LOCAL SEDIMENT SCOUR AT LARGE CIRCULAR
PILES

By

ATHANASIOS PRITSIVELIS

Thesis

1999

LOCAL SEDIMENT SCOUR AT LARGE CIRCULAR PILES

By

ATHANASIOS PRITSIVELIS

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

1999

ACKNOWLEDGMENTS

I would like to express my deepest gratitude to my advisor and committee

chairman Dr. D. Max Sheppard. His continuous support and his insight have permitted

this study to become a reality. I would like also to thank Dr. Mufeed Odeh, P.E., head of

the Engineering Section at the USGS Laboratory part of my committee, for his support

during my 14-month stay in Massachusetts. My thanks also go out toward Dr. Robert

Dean and Dr. Robert Thieke for serving on my supervisory committee.

John Noreika, Stephen Walk, Phil Rokasah and the other members at the USGS-

BRD Laboratory in Massachusetts also deserve my thanks for putting up with me and

helping me on the practical parts of the experiments. Also deserving my deepest

appreciation are Tom Glasser and Edward Albada for helping me, sometimes working 15

hours, with the experiments.

Sidney Schofield, Jim Joiner, Vernon Sparkman, Chuck Broward, and the other

members of the Coastal and Oceanographic Engineering Laboratory also deserve my

thanks for their assistance, as well as Chris Jette for answering all my questions about

transducers fast and efficient.

I am also grateful to J.S. Jones, P.E. with the Federal Highway Administration and

Shawn McLemore, P.E., and Rick Renna, P.E., with the Florida Department of

Transportation for their financial support for this research and their interest and helpful

suggestions.

Many others deserve a hearty thank you. My fellow students, including Ed, Roberto,

Guillermo, Hugo, Tom, Bill, and Erika, for the help on classes. Additional gratitude goes

to Becky, Helen, Laura, and so many other people at the department.

Finally, I want to express my deepest gratitude to my parents, George and Vasiliki,

and my sister Helen for their patience and lifelong support.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ........................................................ ... ......................... ii

LIST OF TABLES ....................................................................................................... vi

LIST OF FIGURES..................................................................................................... vii

KEY TO SYMBOLS ........................................................................................................ xii

A B STR A C T ...................................................................................................................... xv

CHAPTERS

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

2 BACKGROUND AND LITERATURE REVIEW.................................. .......... .7

Description of the Flow Field Around a Pier.......................................... ...............7
Scour Formulation for Steady Flow....................................................................... 9
Effect of the Various Parameters .............................................................................11
Effect of Aspect Ratio yo/b .............................................. ................................ 11
Effect of Velocity Ratio U/Uc .................................................................. 13
Effect of Sediment to Pier Size Dso/b................................................................17
Effect of Sediment Gradation ............................................................................ 19
Effect of Pier Properties (Shape and Alignment) ....................................... ...21

3 EXPERIMENTAL APPROACH.........................................................................24

Facilities and Instrumentation....................................................................................24
Facilities...........................................................................................................24
Instrum entation ................................................................................................28
Velocity measurement..............................................................................28
Water level measurement.............................................................. ...........29
Temperature measurement.............................................................................29
Acoustic transducers ................................................................................30
Video measurements ..................................................................... ...........32
Measurement setup ..................................................................................35

Point gauging ...................................................................................................36
M odels....................................................................................................................36
Sedim ent ................................................................................................................38

4 EXPERIMENTAL DATA REDUCTION AND ANALYSIS...................................41

Experim mental Pprocedure.............................................................................................41
Data Reduction.......................................................................................................43
Processed Data Results .......................................................................................... 44
Discussion of Results...................................................................................................56

5 COMPARISON OF LOCAL SCOUR PREDICTION EQUATIONS ........................ 58

Scour Prediction Equations..........................................................................................58
Com pilation of Local Scour Data ................................... .........................................65
Com parative Analysis ..................................................................................................67
Field Data Example................................................................................................85

6 SCALING MODEL SCOUR DEPTHS TO PROTOTYPE CONDITIONS ...............89

7 SUM M ARY AND CON CLU SION S .................................................................96

Sum m ary ................................................................................................................96
Conclusions..................................................................................................................98
Recom m endations for Future W ork............................... .......................... ..........99

APPENDICES

A EFFECTS OF THE W EIR ........................................................................................100

B FLOW UNIFORM ITY ..............................................................................................103

C EXPERIM EN TAL DATA ........................................................................................105

D COMPARISON OF METHODS CALCULATING CRITICAL DEPTH
AVERAGE VELOCITY .................................. ................. ................................109

REFEREN CES .......................................................................................................... 14

BIOGRAPHICAL SKETCH ........................................ ................ ........................... 118

LIST OF TABLES

Table page

2.1 Shape factor for pier nose, taken from HEC-18 (1995).........................................22

2.2 Other factors for pier shapes ............................................................................... 22

2.3 Correction factor K2 for angle of attack 0 of flow, where L is the length of the pier
and b the width, taken from HEC-18 (1995) ................................... ........... 23

3.1 Table with values of (-log (D50/b)) before sand availability..................................37

3.2 Table with values of (-log (D5o/b)) for the piles and sediment used in the tests......38

4.1 Parameters for the experiments conducted .........................................................41

6.1 Cases for model and prototype..............................................................................93

A.1 Prototype and model dimensions ......................................................................100

C .1 Experim ental data............................................................................................. 105

LIST OF FIGURES

Figure page

1.1 Effects of local scour on a bridge pier
(taken from University of Louisville web site) .....................................................

2.1 Schematics of the vortices around a cylinder..................................... .............. 8

2.2 General relationship between equilibrium scour depth and
water depth when other parameters are held constant .......................................12

2.3 General relationship between equilibrium scour depth and velocity
when other parameters are held constant......... ............................. ........... 14

2.4 Dependence of scour depth on mean velocity for two types of sediment ..............16

2.5 Dependence of equilibrium scour depth on ratio of sediment diameter
to structure diameter for 1 > U/Uc > 0.9 and yo/b > 2.5 .....................................18

2.6 Influence of a on scour depth, taken from Melville and Sutherland (1988)............19

2.7 Graphic relationship between a and KI, taken from Chiew (1984) .......................20

2.8 Alignment factor K2 for rectangular piers, taken from Laursen (1958)...................21

3.1 Aerial photograph of the S.O. Conte Laboratory........................................... ..25

3.2 Schematic figure of the setup for the experiments..............................................26

3.3 A test pile with the instrumentation......................................................................31

3.4 Detailed view of the arrays for the small structures...........................................31

3.5 Detailed view of the arrays for the large structures ....................................... ..32

3.6 Picture of the cameras for the 0.114 m (4.5 in) pile..............................................33

3.7 Picture of the cameras for the 0.305 m (12 in) pile............................... ........... 34

3.8 Picture of the cameras for the 0.92 m (36 in) pile..............................................34

3.9 Diagram of the measurement system ....................................................................35

3.10 Cross-Section of the 0.92m (36 in) pier............................... .................................37

3.11 Gradation Curve for Sand No. 1 (Dso = 0.22 mm)...........................................39

3.12 Gradation Curve for Sand No. 2 (D50 = 0.80 mm)..............................................40

4.1 Elevation contour plot of the equilibrium scour hole for Experiment No.1
(b = 0.114 m, D50 = 0.22 mm, yo = 0.186 m and U = 0.290 m/s) ........................45

4.2 Elevation contour plot of the equilibrium scour hole for Experiment No.2
(b = 0.305 m, Ds0 = 0.22 mm, yo = 0.190 m and U = 0.305 m/s) ......................45

4.3 Elevation contour plot of the equilibrium scour hole for Experiment No.3
(b = 0.920 m, D50 = 0.22 mm, yo = 2.268 m and U = 0.325 m/s) ......................46

4.4 Elevation contour plot of the equilibrium scour hole for Experiment No.4
(b = 0.920 m, D50 = 0.80 mm, yo = 2.402 m and U = 0.454 m/s) ........................46

4.5 Elevation contour plot of the equilibrium scour hole for Experiment No.5
(b = 0.920 m, D50 = 0.80 mm, yo = 0.866 m and U = 0.335 m/s) ......................47

4.6 Elevation contour plot of the equilibrium scour hole for Experiment No.6
(b = 0.305 m, D50 = 0.80 mm, yo = 1.305 m and U = 0.381 m/s) ......................47

4.7 Elevation contour plot of the equilibrium scour hole for Experiment No.7
(b = 0.114 m, D50 = 0.80 mm, yo = 1.280 m and U = 0.388 m/s) ......................48

4.8 Equilibrium profiles of scour hole for Experiment No.1
(b = 0.114 m, D50 = 0.22 mm, yo = 0.186 m and U = 0.290 m/s)
a) Inline profile, and b) Normal profile.............................................................49

4.9 Equilibrium profiles of scour hole for Experiment No.2
(b = 0.305 m, Dso = 0.22 mm, yo = 0.190 m and U = 0.305 m/s)
a) Inline profile, and b) Normal profile.............................................................50

4.10 Equilibrium profiles of scour hole for Experiment No.3
(b = 0.920 m, D50 = 0.22 mm, yo = 2.268 m and U = 0.325 m/s)
a) Inline profile, and b) Normal profile..............................................................51

4.11 Equilibrium profiles of scour hole for Experiment No.4
(b = 0.920 m, D50 = 0.80 mm, Yo = 2.402 m and U = 0.454 m/s)
a) Inline profile, and b) Normal profile ............................................................52

4.12 Equilibrium profiles of scour hole for Experiment No.5
(b = 0.920 m, D50 = 0.80 mm, yo = 0.866 m and U = 0.335 m/s)
a) Inline profile, and b) Normal profile................................................ ............53

4.13 Equilibrium profiles of scour hole for Experiment No.6
(b = 0.305 m, D50 = 0.80 mm, yo = 1.305 m and U = 0.381 m/s)
a) Inline profile, and b) Normal profile................................................ ............54

4.14 Equilibrium profiles of scour hole for Experiment No.7
(b = 0.114 m, D50 = 0.80 mm, yo = 1.280 m and U = 0.388 m/s)
a) Inline profile, and b) Normal profile................................................ ............55

5.1 Local scour depth dependence on velocity, as proposed by Sheppard (1997).........65

5.2 Comparison of equations for clearwater data ...........................................................69

5.3 Comparison of equations for live-bed data....................................................70

5.4 Comparison of measured and calculated scour for clearwater data for
(A) Garde et al. (1993), and (B) Ahmad (1953) equations................................71

5.5 Comparison of measured and calculated scour for clearwater data for
(A) Breusers et al. (1977), and (B) Chitale (1962) equations ............................72

5.6 Comparison of measured and calculated scour for clearwater data for
(A) Froehlich (1988), and (B) Gao et al. (1992) equations................................73

5.7 Comparison of measured and calculated scour for clearwater data for
(A) Hancu (1971), and (B) HEC-18 (1995) equations.......................................74

5.8 Comparison of measured and calculated scour for clearwater data for
(A) Inglis-Blench (1962), and (B) Jain and Fischer (1979) equations.................75

5.9 Comparison of measured and calculated scour for clearwater data for
(A) Laursen and Toch (1956), and (B) Melville (1997) equations....................76

5.10 Comparison of measured and calculated scour for clearwater data for
(A) Shen et al. (1969), and (B) Sheppard et al. (1995) equations......................77

5.11 Comparison of measured and calculated scour for live-bed data for
(A) Ahmad (1953), and (B) Breusers et al. (1977) equations............................78

5.12 Comparison of measured and calculated scour for live-bed data for
(A) Chitale (1962), and (B) Froehlich (1988) equations............................ ..79

5.13 Comparison of measured and calculated scour for live-bed data for
(A) Gao et al. (1992), and (B) Hancu (1971) equations............................. ..80

5.14 Comparison of measured and calculated scour for live-bed data for
(A) HEC-18 (1995), and (B) Inglis-Blench (1962) equations ...........................81

5.15 Comparison of measured and calculated scour for live-bed data for
(A) Jain and Fischer (1979), and (B) Laursen and Toch (1956) equations..........82

5.16 Comparison of measured and calculated scour for live-bed data for
(A) Melville (1997), and (B) Shen et al. (1969) equations................................83

5.17 Comparison of measured and calculated scour for live-bed data
for Sheppard et al. (1995) equation........................... ........................................84

5.18 Comparison of equations for Experiment No. 3
(b = 0.920 m, Dso = 0.22 mm, yo = 2.268 m and U = 0.325 m/s) ......................86

5.19 Comparison of equations for example of field conditions............................. ..87

6.1 Fitting curves for the scale factor using equation (6.1).................................. ..91

6.2 Regions for the model and prototype cases........................................................92

6.3 Data fitting curves for the scale factor .................................................................94

A.1 Positions of velocity profile collection ................................................................ 101

A.2 Flow profiles of the model at certain distances from the entrance of the flume ....101

B.1 Contour plots of velocity for the 1.22 m (4 ft) water depth .................................104

B.2 Contour plots of velocity for the 2.44 m (8 ft) water depth ...................................104

C.1 Scour history for Experim ent 1 ................................... ........................................105

C.2 Scour history for Experim ent 2 ................................... ........................................106

C.3 Scour history for Experim ent 3 ............................................................ ................106

C.4 Scour history for Experim ent 4 ................................... ........................................107

C.5 Scour history for Experim ent 5 ................................... ........................................107

C.6 Scour history for Experim ent 6 ............................................................................108

C.7 Scour history for Experiment 7 ............................................................................108

D.1 Relationship between critical shear velocity and median grain diameter .............109

D.2 Comparison of the methods for calculating Ue in respect to
(A) Grain Median Diameter, (B) Temperature, and (C) Relative Roughness ...111

KEY TO SYMBOLS

b pile width (diameter)

bm model pile width

bp prototype pile width

B channel width

c exponent related to bed load

Cwr rectangular weir coefficient

D16 sediment size for which 16 percent of bed material is finer

Dso median sediment diameter

D5o m model median sediment diameter

D84 sediment size for which 84 percent of bed material is finer

D50 p prototype median sediment diameter

dse equilibrium scour depth

dse(c) equilibrium scour depth for a sediment with a given a

dse m model equilibrium scour depth

dse p prototype equilibrium scour depth

Fr Froude number

Frc critical Froude number

g acceleration of gravity

H head over the weir

K pier shape factor

K1 factor for shape of pier nose

K2 factor for angle of attack of flow

Kd sediment-size factor

KI flow intensity factor

Ks pier-shape factor

KyD flow depth-pier width factor

K, pier-alignment factor

K, factor for the gradation of sediment

L length of pier

Lm model scale length

Lp prototype scale length

Pw weir height

q discharge per unit width

Q discharge

R ratio of model to prototype Dso/b

Rep pier Reynolds number

Rew wall Reynolds number

s relative sediment density

S scale factor

SG geometric scale factor

U mean depth average velocity

Ue critical depth average velocity

U' velocity for the initialization of sediment movement around the

structure

u* bed shear velocity

u*e critical bed shear velocity

Vc velocity for the initialization of sediment movement around the

structure

yo depth of flow

Ca, Ca2 coefficients of scale factor, SF, related to b/Dso of prototype

0 angle of attack of flow

It dynamic viscosity of water

v kinematic viscosity of water

p density of water

Ps density of sediment

C gradation of sediment

T bed shear stress

A factor for pier shape

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

LOCAL SEDIMENT SCOUR AT LARGE CIRCULAR PILES

By

Athanasios Pritsivelis

August 1999

Chairman: Dr. D. Max Sheppard
Major Department: Coastal and Oceanographic Engineering

Local sediment scour experiments were performed with three circular pile diameter

(0.114 m, 0.305 m, and 0.920 m) and two sediment grain sizes (D50 = 0.22 mm and 0.80

mm) under clearwater scour conditions. The results were analyzed and combined with

data from other researchers to form two data sets, one for clearwater and one for live-bed

conditions. These data sets were used to evaluate and compare 14 different scour

prediction equations. The 0.92 m diameter pile was then considered to be a prototype pile

and the other, smaller piles models of the prototype. The model and prototype data were

then used to obtain the proper model to prototype scale factor for computing the

equilibrium scour depth for the prototype from the model results. The scale factors are

represented in plots of scale factor versus D50 (model)/ D50 (prototype) for various values of

sediment to pile diameter ratio for the prototype.

CHAPTER 1
INTRODUCTION

When designing hydraulic structures, one must consider many aspects, which are

important for the functionality and mostly for their safety and the safety of the people

depending on them.

Some of the aspects considered are (Pilarczyk, 1995):

1. Function of the structure; erosion is not always the problem as long as the

structure can fulfill its function,

2. Physical environment; the structure should offer the required degree of

protection against hydraulic loading, with an acceptable risk and, when possible,

meet the requirements resulting from landscape, recreational and ecological

viewpoints,

3. Construction method; the construction costs should be minimized to an

acceptable level and legal constrictions must be adhered to,

4. Operation and maintenance; it must be possible to manage and maintain the

hydraulic structure.

From all these aspects, the cost of construction and maintenance is generally the

controlling factor. In the case of bridge design, local scour around the piles of the bridge

is one of the factors that contribute to the total construction and maintenance cost and is

important in the estimation of stability of the structure. Underprediction can result in

costly bridge failure and possibly in the loss of lives, while overprediction can result in

millions of dollars wasted on a single bridge.

Scour in general is caused by changes in the characteristics of flow, which lead to

changes in the sediment transport in the area of interest. There is an equilibrium scour

depth associated with a given set of environmental conditions (water depth, sediment

density and size, flow velocity, etc.). When the environmental conditions change (e.g. a

change in flow velocity), the scour depth progresses toward the new equilibrium depth. If

the new conditions are maintained for a sufficiently long period of time the new

equilibrium scour depth is reached. Scour caused by a structure (e.g. a pile) can be

divided into two categories: general (or contraction) and local. As a first approximation,

the scour caused by each process can be added to obtain the total scour. Of the two types

of scour, local scour is the least understood.

General scour happens when the cross sectional area of the flow at a particular

location is reduced due in most cases to the presence of a new structure. This results in an

increase in flow velocity and bed shear stress and increases the potential for erosion of the

area. This type of scour is well understood and there are a number of methods for

predicting these scour depths.

Local or structure induced scour is caused by the structure itself. A picture of local

scour around a structure (bridge pier) is shown on Figure 1.1. The main causes of local

scour are

1) an increase in mean flow velocities in the vicinity of the structure,

2) the creation of secondary flows in the form of vortices and

3) the increased turbulence in the local flow field.

Two kinds of vortices are observed: Wake vortices, downstream of the points of flow

separation on the structure and horizontal vorticies at the bed and free surface, due to

pressure variations along the face of the structure. Those phenomena, although relatively

easy to observe, are difficult to quantify mathematically. Some researchers (Shen et al.,

1969) have attempted to describe this complex flow field mathematically but with little

success. A number of numerical solutions have also been attempted but with limited

success.

Figure 1.1 Effects of local scour on a bridge pier
(taken from a Bridge Scour Site, 1998)

Historically, scientists and engineers have used scale models in order to predict

complex prototype phenomena. This requires a knowledge and understanding of the

modeling laws for the particular situation (i.e. a knowledge of the pertinent dimensionless

groups for the processes involved), in order to extrapolate the measured values for the

model to prototype conditions. Ideally, the values of the salient dimensionless groups are

made the same in the model as they are in the prototype. This, however, is not always

possible and when it is not the extrapolation process becomes more difficult. For local

structure induced scour, the processes are characterized by three dimensionless groups,

Y U and Ds- where b is the diameter of the pile, yo is the water depth, U in the
b Uc b

mean depth average velocity, Uc is the critical depth average velocity and Ds0 is the

median sediment diameter. It is normally easy to configure the model so as to make the

first two groups the same for the model and prototype but due to lower limits on the

sediment diameter (before the sediment becomes cohesive) the third group usually cannot

be properly scaled. In many locations, such as Florida, the sediment is sand with a

prototype Ds0 between 0.1 and 0.4 mm and the structures can be large (on the order of 10

to 20 m in width). For a geometric scale of 40, the model sediment Ds0 would have to be

as small as 0.0025 mm, which is in the cohesive range, where the scour properties are

significantly different (and less understood). The problem is then, how to predict the

prototype scour depth from the scour depth measured in the model study. The most

common procedure used to date has been to simply use the prototype-to-model geometric

5

scale (i.e. dse = dsem ). Note that this ignores the differences in D0 for the model
me L b

and prototype and the impact of this difference on the scour depth.

Most scour prediction relationships, such as the Colorado State University (CSU)

relation, (currently used in the FHWA Hydraulic Engineering Circular No. 18) and the

University of Florida (UF) equation (Sheppard et al., 1995) are empirical. Many of these

equations are relatively accurate in predicting local scour for laboratory scale structures

but are very conservative in their estimates of scour depths for prototype structures. There

is a significant amount of scour data in the literature for small structures but very little for

prototype size structures in a controlled environment.

One of the objectives of the work reported in this thesis is to provide local scour data

for large circular piles. Laboratory experiments were conducted for three pile diameters,

two different sediment sizes and a range of water depths. The results of those experiments

along with other data from various researchers were then used to compare 14 different

scour prediction equations. This data set was also used to obtain the scale factor for

estimating prototype scour depths from measured model scour depths.

To better understand the problem of local scour, Chapter 2 reviews the current state

of knowledge of local scour. The primary mechanisms are reviewed first followed by the

formulation of the most important dimensionless groups for local scour.

Chapter 3 gives a detailed description of the experiments performed and the

procedures used. Seven experiments were conducted with three pile diameters and two

sediment sizes. The experiments were run for approximately 24 hours beyond the point at

which the scour depth ceased to increase.

Chapter 4 presents the experimental results and the data processing procedures. In

line and normal scour hole profiles are presented along with contour plots of the scour

holes.

Chapter 5 presents a comparison of a number of scour prediction equations using

clearwater, live-bed and a typical example of field data. The equations are compared on

their ability to predict those data and the results are shown with line and scatter plots.

In Chapter 6 the problem of scaling model scour depths to prototype conditions is

examined. A scaling factor that depends on D50 and D50 is presented.
D50-p b ) prototype

Finally, a summary of the work conducted for this thesis, along with some

conclusions and recommendations for future work is given in Chapter 7.

CHAPTER 2
BACKGROUND AND LITERATURE REVIEW

Bridge scour has received a lot more attention this past decade and a number of

advances have been made. Many researchers have conducted laboratory experiments in

order to improve the accuracy of equilibrium scour depth prediction. Most of these tests

were done with simple structures, like a single pile, in a horizontal sand bed subjected to

a steady flow. These experiments succeeded in identifying the physics of the local scour

processes. Recently, there has been an increase in the research on local scour at more

complex piers.

2.1 Description of the Flow Field Around a Pier

This section discusses the flow field near a cylindrical pile in a steady flow, as seen

from various researchers. The flow field in the immediate vicinity of a structure is quite

complex, even for simple structures such as circular piles. The dominant feature of the

flow is the large-scale eddy structure, or system of vortices that develop about the pier.

That system is the basic mechanism of local scour, and has been recognized by many

investigators (e.g. Shen et al., 1966, Melville, 1975).

It has been found that depending on the type of pier and flow conditions, the eddy

structure can be composed of one or more of the basic systems, which comprise the

horseshoe vortex and the wake vortex systems. It has also been identified that the

acceleration of the flow around the pier is a factor in the early stages of local scour.

8

The horseshoe vortex system is developed at the base of the structure. The term

"horseshoe" is derived from the shape that the system takes as viewed from above. It

wraps around the structure and tails downstream, as shown in Figure 2.1. Shen et al.

(1966) describe the horseshoe vortex system very accurately. The main cause of this

system is the stagnation pressure difference developed at the nose of the structure. If the

difference is large enough, there is a downward flow, which rolls up ahead of the toe of

the structure to form the horseshoe vortex system. The system in its simplest form is

composed of two vortices, a large one, adjacent to the structure, with a small counter

vortex. A more complicate system consists of multiple small counter vortices, which are

unsteady and periodically shed and are swept downstream. Clearly, the geometry of the

structure is important in determining the strength of the system, although it is not steady

for all kinds of structures. Blunt nosed structures create the most energetic vortex system.

Wake Reaion

'.TOP VIEW..
T P Vo .W *- -. '

SJ.OE VIWM

Figure 2.1 Schematics of the vortices around a cylinder

Melville (1975) measured mean flow directions, mean flow magnitude, turbulent

flow fluctuations, and computed turbulent power spectra around a circular pier for

flatbed, intermediate and equilibrium scour holes. He found that a strong vertical

downward flow developed ahead of the cylinder, as the scour hole enlarged. The size and

circulation of the horseshoe vortex increased rapidly and the velocity near the bottom of

the hole decreased as the scour hole was enlarged. As the scour hole develops, the

intensity of the vortex decreases and reaches a constant value at the equilibrium stage.

Although the horseshoe vortex is considered the most important scouring

mechanism for steady flows, the wake vortex system is also important. Wake vortices are

created by flow separation on the structure. Large scour holes may also develop

downstream from piers, when the horseshoe vortex system does not form, as

demonstrated by the experiments of Shen et al. (1966). With their vertical component of

flow, wake vortices act somewhat like a tornado in removing the bed material, which is

then carried downstream by the mean flow.

2.2 Scour Formulation for Steady Flow

In order to formulate quantitative relationships to predict local scour, there has to be

a distinction between clearwater and live-bed scour. Clearwater scour is the local scour

that occurs when the flow velocity is below the value needed to initiate sediment motion

on the flat bed upstream from the structure. Live-bed scour occurs when the flow velocity

exceeds the value needed to initiate sediment motion upstream of the structure. The

sediment is assumed to be cohesionless and the term local scour is assumed to be the

depth of the deepest point around the structure.

The equilibrium scour depth depends on the fluid and sediment properties, the flow

parameters, and characteristics of the structure, e.g.

dse = f [, p, g, D50, u, p,, yo, U, b, h(pier)], (2.1)

where

dse is the equilibrium scour depth,

p and ps are the density of water and sediment respectively,

g is the dynamic viscosity of water (depends on temperature),

g is the acceleration of gravity,

Ds5 is the median diameter of the sediment,

a is the gradation of sediment,

yo is the depth of flow upstream of the structure,

U is the depth average velocity,

b is the pier diameter/width normal to the flow,

h(pier) is a function that depends on the shape and the alignment of the pier in the

flow.

The most important dimensionless groups for local scour can be obtained from the

quantities given in Equation 2.1. These eleven quantities can be expressed in terms of

three fundamental dimensions: force, length and time. According to the Backingham I

theorem, 11 3 = 8 independent dimensionless groups exist for this situation. An example

of these eight groups is shown in Equation 2.2.

d U p, Ub U D50, h(pier) (2.2)
b bt *4 p v U b

Of these eight groups, the following six are considered the most important as shown in

Equation (2.3)

dse o b,U o-, h(pier)J, (2.3)
b b 'Ub b

where Uc is the critical depth averaged velocity (the velocity required to initiate sediment

motion on a flat bed).

2.3 Effect of the Various Parameters

The aspects of scour processes described by the various dimensionless groups are

understood as a result of numerous laboratory experiments. These groups are described in

the following paragraphs.

2.3.1 Effect of Aspect Ratio yv/b

The effect of flow depth on scour depth is significant. Observations (Ettema 1980,

Chiew 1984) have shown that, for shallow flows, the scour depth increases with

increasing depth of flow up to a point beyond which no effect is observed. The above is

generally accepted by many researchers e.g. Laursen and Toch (1956), Neill (1973),

Cunha (1970). As Ettema (1980) states, the effect of the flow is due to the impact it has

on the formulation of the surface and bottom vortices located at the upstream side of the

pier. The pressure differences along the nose of the pier cause a surface roller at the top

and the horseshoe vortex at the bottom. The two rollers rotate to opposite directions. In

principle, as long as they do not interfere, the local scour depth is independent of the

depth of the flow. With decreasing depth, the surface roller interferes with the downflow,

making it weaker. Other reasons are the influence of the sediment bar behind the structure

for small values of yo/b and the fact that the portion of the flow that is diverted into the

hole, diminishes for low values of yob (Ettema, 1980).

Another researcher, Bonasoundas (1973) concluded that the effects of the flow depth

were insignificant for yo/b > 1 to 3. However, his experiments were only run for two

hours and thus equilibrium depths could not have been achieved. Basak (1975), Hancu

(1971), White (1973), Chabert and Engeldinger (1956) and Jain and Fischer (1979)

observed the same general behavior for live-bed conditions.

Laboratory data (Ettema 1980, Chiew 1984, etc.) indicate that, if the other two

parameters (U/Uc and Dso/b) are held constant, the equilibrium scour depth increases

rapidly with yo/b until a value of 2.5 to 3 and then remains constant. This is shown

schematically in Figure 2.2.

2.5-

2-

1.5-

1-

0.5-

0---------------
0 1 2 3 4
yo/b

Figure 2.2 General relationship between equilibrium scour depth and water depth
when other parameters are held constant

2.3.2 Effect of Velocity Ratio U/U_

The critical depth average velocity, Uc, is important since it is the velocity at which

sediment movement is initiated on a flat bed upstream of the structure. A better

understanding of critical velocity can be achieved, by examining the forces acting on a

sediment particle subjected to a steady flow (Shields, 1936). Shields states that a critical

bed shear stress exists, above which there is movement of the sediment at the bottom.

Comparing the drag force on a grain with its submerged weight, Shields developed an

empirical formula for the critical shear stress. From the shear stress, the shear velocity,

u = can be computed. For a fully developed velocity profile, the Prandtl-Von

Karman formula can be used to compute the depth averaged velocity as a function of the

shear velocity and bed roughness (Sleath, 1984). From the above, the critical depth

averaged velocity, Uc, can be calculated. For flow velocities higher than this value, the

local scour is defined as "live-bed scour." "Clearwater scour" is defined as the local scour

that occurs for velocities less than the critical value.

Several different methods have been used by researchers to estimate the value of

critical depth averaged velocity and their results can differ significantly. Note that in

comparing data from different researchers, it is important that the same method be used

for all the data. A brief description of the more common methods for computing Uc is

given in Appendix D along with a comparison of their predicted values.

The depth of equilibrium local scour is closely related to the mean depth average

velocity. It is evident from published literature that under clearwater conditions, the local

scour increases almost linearly with the approach velocity. Breusers et al. (1977)

developed the relationship between scour depth and the ratio of U/Uc by dimensional

analysis, but applied no detailed analysis into the nature of the dependency. Hanna (1978)

found out that scour is initiated around u*/u*, = 0.5, where u* is the bed shear velocity and

u*, is the critical bed shear velocity and reaches a maximum value for u*/u*c equal to

unity. Other data, that show that the initiation of scour depth starts at around 0.4-0.5Uc,

for circular piles, come from Chabert and Engeldinger (1956) and Ettema (1980). Most of

the laboratory data seem to follow Figure 2.3.

Clearwater Live-bed
Conditions Conditions
<--- --->

-0.4-0.5

1
U/UC

Figure 2.3 General relationship between equilibrium scour depth and velocity
when other parameters are held constant

As for live bed conditions (U/Ue > 1.0); the results are even more difficult to obtain.

The difficulty resides in the fact that the experiments are difficult to perform and the

results hard to interpret. Early researchers related the relative scour depth dse/b to the

Froude number. Most of the conclusions drawn state that for a given flow depth, the

scour depth increases with increasing velocity. Later researchers e.g. Chabert and

Engeldinger (1956), Laursen (1962) etc. found that as the velocity increased beyond the

critical velocity, Uc, the scour depth decreases by about 10%. Further increases in

velocity increase the scour depth until a second peak is reached. The decrease of scour

depth after the initiation of live-bed conditions was believed to be caused by infill of the

scour hole due to movement of the bed. Hancu (1971) conducted a series of live-bed

scour experiments and concluded that the local scour depth is independent of the flow

velocity for live-bed conditions. This conclusion is similar to that of Breusers et al.

(1977).

Ettema (1980) stated though that there is a dependence of both clearwater and live-

bed equilibrium local scour depth on velocity. He concluded that the maximum value for

live-bed local scour can be lower or higher than the one for the clearwater depending on

the type of sediment (Figure 2.5). He thought that for ripple-forming sediments (D50 < 0.6

mm), the deepest scour hole is for live-bed conditions, while for non ripple-forming

sediments, the deepest scour depth occurs at transition from clearwater to live-bed

conditions. His explanation for the above distinction is that "shallower maximum depths

of local scour generally occur for ripple-forming than for non ripple-forming bed

materials, because the formation of ripples on the approach bed alters the roughness of

the bed, creates a low level of sediment transport into the scour hole, affects the boundary

layer separation at the pier and consequently changes the strength of the horseshoe

vortex". Sheppard (1997) concluded that the condition that defines which peak

(clearwater or live-bed) is higher depends on the ratio of median diameter of the sediment

and pier diameter, D50/b, and not on whether the sediment was ripple-forming.

Non ripple-forming sediment

\

Ripple-forming sediment

-0.4- 0.5
1

U/UC

Figure 2.4 Dependence of scour depth on mean velocity for two types of sediment

2.3.3 Effect of Sediment to Pier Size D5o/b

The effect of this parameter was not recognized until recently. Laursen and Toch

(1956) stated "exactly the same depth of scour should result in the model, no matter what

velocity or sediment is used, as long as there is general bed load movement and the

Froude number is everywhere less than unity". However they added "because secondary

effects of velocity and sediment size which could not be detected in the limited range of

the laboratory data may become important at large scale, the validity of this conclusion

can only be tested by model-prototype conformity studies". The results of the model study

by Chitale (1962), Ahmad (1962) support the conclusions drawn by Laursen and Toch

(1956) concerning the effect of sediment on scour depth. Krishnamurthy (1970) also

stated that the effect of sediment size is negligible for high Froude number and large pile

sizes.

The data from Chabert and Engeldinger (1956) show a small effect of sediment size

on scour depth. Nicollet and Ramette (1971) extended the experiments of Chabert and

Engeldinger (1956) and showed that the sediment size has a considerable effect. Raudkivi

and Ettema (1977) decided to hold U/Ue and yo/b constant and change D50/b. Even though

their data showed effects of sediment size, they attributed the changes to ripple formation

and it was not given further thought. Baker (1986) correlated scour depth with b/D50

using the data from Raudkivi and Ettema (1977) and found that there is a correlation

between the two values. Ettema's (1980) clearwater data showed that the influence of

sediment size is significant, if the ratio of b/D50 is less than 20-25 and that for higher

values the scour depth is independent of sediment size. A similar study to determine the

scour depth dependence on D50/b was conducted by Sheppard and Ontowirjo (1994).

Using their data and that due to Ettema (1980) and Chiew (1984), they found that the

effect of sediment size is as shown in Figure 2.5. The value of dse/b increases and then

decreases with increasing values of D50/b. Sheppard (1997) assumed that one possible

explanation for why the laboratory data correlate so well with the parameter D50/b is that

it is actually the ratio of two different Reynolds numbers, one based on the sediment grain

diameter (and the associated near bed shear velocity) and one on the structure diameter

and the depth average velocity. Both Reynolds numbers are important in characterizing

the flow and sediment transport in the vicinity of the structure.

3-

2.5 +
+ +
+t

S(E + ttema (
2 + '

0 +A

0.5 -
0 + Ettema (1980)
0 UF/USGS
0-
-4 -3 -2 -1
log(Ds0/b)

Figure 2.5 Dependence of equilibrium scour depth on ratio of
sediment diameter to structure diameter for 1 > U/Uc > 0.9 and yo/b > 2.5

2.3.4 Effect of Sediment Gradation a

For a given structure and flow, the equilibrium scour depth is very dependent on the

sediment gradation. Experiments carried out by Nicollet and Ramette (1971) and a more

extensive study by Ettema (1976) showed that the equilibrium scour depth decreases as

the standard deviation of the particle size distribution increases for clearwater scour

conditions. For live-bed tests conducted by Baker (1986) and plotted by Melville and

Sutherland (1988), there was a reduction as well but not as much as for clearwater and for

values of U/Uc > 4, the effect was almost independent of the gradation of the sediment

(Figure 2.6). It is believed that the main effect of a is in the formation of an armor layer

around the upstream perimeter of the pier reducing the scour hole depth.

2.0

dse/h b

0 1.O

U/Uc

Figure 2.6 Influence of a on scour depth, taken from Melville and Sutherland (1988)

20

Ettema (1980) replotted his previous data as K, versus a, where KI is the coefficient

in the equation

= K dse (2.9)
b b

where a- = 8
D16

de ( is the equilibrium clearwater scour depth for a sediment with a given a,
b

dse is the equilibrium scour depth for uniform sediment (a < 1.6).
b

The graph is reproduced in Figure 2.7. The data is grouped by sediment type (ripple-

forming, D50 < 0.6 mm or non ripple-forming, D50 > 0.6 mm). KI varies from 1.0 for

uniform sediments to less than 0.25 for sediments with a large gradation.

3
0-g

Figure 2.7 Graphic relationship between o and IK, taken from Chiew (1984)

2.3.5 Effect of Pier Properties (Shape and Alignment)

The shape of the pier and its orientation to the flow can have a significant effect on

the equilibrium scour depth. The shape and orientation effect can be accounted for with

coefficients multiplied times the equation for a circular cross-section pier.

db
dse KK2( dse (2.10)
b b circular pier

Values for K1 for common geometric pier cross-sections are given in Tables 2.1 and

2.2. The effect of flow skew angle (pier orientation) can be seen in Table 2.3 and Figure

2.8, taken from HEC-18 (1995) and Laursen (1958) respectively.

7 I 1 I II 1 I I i I a s

5 r10

0 15 30 45 60 75 90
ANGLE OF ATTACK 0 (Degrees)

Figure 2.8 Alignment factor K2 for rectangular piers, taken from Laursen (1956)

Table 2.1 Other factors for pier shapes
Laursen Chabert
And And
Tison Toch Engeldinger Neill Venkatadri
Shape in plan L/a (1940) (1956) (1956) (1973) (1965)

Lenticular

Parabolic nose

Triangular nose, 600

Triangular nose, 900

Elliptic

Ogival

Joukowski

Rectangular

Circular

0.67
0.41

0.86

0.76

1.40

1.00

0.56

0.75

1.25

Table 2.2 Shape factor for pier nose, taken from HEC-18 (1995)
Shape of pier nose KI
Square Nose 1.1
Round Nose 1.0
Circular Cylinder 1.0
Group of cylinders 1.0
Sharp Nose 0.9

1.00

0.97
0.76

0.91
0.83

1.11

1.11

0.73

0.92

0.86

1.11

1.00

1.00

23

Table 2.3 Correction factor K2 for angle of attack 0 of flow,
where L is the length of the pier and b the width, taken from HEC-18 (1995)
Angle L/b = 4 L/b = 8 L/b = 12
0 1.0 1.0 1.0
15 1.5 2.0 2.5
30 2.0 2.75 3.5
45 2.3 3.3 4.3
90 2.5 3.9 5.0

CHAPTER 3
EXPERIMENTAL APPROACH

This chapter includes a detailed description of the experiments performed as part of

this work, as well as a description of the equipment and instrumentation used.

3.1 Facilities and Instrumentation

The equipment used in the experiments is divided in four categories. The first is the

facilities, which includes the flume and the equipment used to place and remove the

sediment. The second is the instrumentation used to measure the water properties, flow

conditions and scour depth. The third category is a description of model piers and the last

category is a description of the sediment used in the experiments.

3.1.1 Facilities

All experiments were conducted in a flume located at the U.S. Geological Survey

Biological Research Division, S.O. Conte Anadromous Fish Research Laboratory

(referred to here as USGS-BRD Laboratory) in Turners' Falls, Massachusetts. The

primary purpose of this laboratory is to study the behavior and biology of anadromous

fish and to conduct research on fish passages. This is the first time the flume has been

used for sediment scour research. An aerial view of the laboratory is shown on Figure 3.1.

The flume area of the laboratory has three parallel open channels. The main channel,

located in the middle, has a width of 6.1 m (20 ft). The two side channels have widths of

3.05 m (10 ft). All three channels have a length of 38.6 m (126.5 ft) and a depth of 6.4 m

(21 ft). A not-to-scale, schematic drawing of the flume area in the Engineering Building is

shown in Figure 3.2. Only the 6.1 m wide main channel was used for the work reported

here.

Figure 3.1 Aerial photograph of the S.O. Conte Laboratory

The laboratory is located between a hydropower stationcanal and the Connecticut

River. The flow in the flume is generated by the head difference between the canal and

the Connecticut River. There is an intake pipe connecting the flume and the canal. The

flow passes through grates to filter the water entering the flume. The flow is controlled by

four sluice gates with dimensions of 1.22 m x 1.22 m (4 ft x 4 ft). Two of the gates are

located at the north wall and one on either side at the north end. The gates are controlled

with electric motors that raise or lower them individually. Discharge as large as 350 cfs

Flow Intake from Reservoir

126 I

NOT TO SCALE
All dimensions in feet

Flow Discharge
To Connecticut
River

Clearwater Scour Test Setup

NOT TO SCALE
All dimensions in feet

TestSediment 6

Section A-A

1- 126 1
I I 1

Filter Material

Base Sediment

Test Sediment

Base Sediment

Section B-B

Figure 3.2 Schematic figure of the setup for the experiments

i 4

can be achieved with the two main sluice gates with an additional 50 cfs from the other

two gates (depending on the canal and river elevations).

For the purpose of the work reported in this thesis, a weir was placed at the

downstream end of the flume to control the water level and volumetric discharge. Model

tests were performed to determine how far upstream of the weir, the velocity profiles

were affected (see Appendix A). It was determined that the weir has minimal effect on the

velocity profiles in the test section.

The discharge, Q, over a rectangular weir occupying the entire width of the flow B

can be computed using the equation

Q= Cwr 2 BH3, (3.1)
3

where Cwr is the rectangular weir coefficient. From dimensional analysis arguments, it is

expected that Cw is a function of Reynolds number (viscous effects), Weber number

(surface tension effects) and the ratio of water head over the weir to the weir height,

H/Pw. In most practical situations, the Reynolds and Weber numbers effects are

negligible, and the following expression can be used (Rouse 1946, Blevins 1984)

Cw, = 0.611+0.075-. (3.2)
Pw

More precise values of C,, can be found in the literature (Henderson, 1966).

At the upstream entrance of the flume, a 5.5 m (18 ft) high flow straightener was

installed to maintain uniformity of the flow over the width of the flume. The flow

straightener consists of vertical wood slats with a 25% opening. The setup was successful

in that it produced near uniform flow across the flume for the range of water depths and

velocities used in the experiments (see Appendix B for details).

In order to reduce the volume of the relatively expensive, uniform diameter sand,

gravel with a D50 of approximately 0.48 cm (3/8 in) was used as a filler away from the

test area (Figure 3.2). A filter cloth with a mean diameter opening of 0.1 mm was placed

over the drains in the floor of the flume and between the sand and the gravel. An

additional advantage of the gravel was the reduction of the drain time for the flume. A

total of 360 m3 of gravel and 205 m3 of test sediment were used. In order to prevent

sediment transport near the entrance of the flume, where the water was jetting from the

openings in the flow straightener, gravel was placed on the bed for the first 8.5 m (28 ft).

This hastened the development of a fully developed velocity profile and prevented the

formation of sand dunes that would have occurred due to the increased velocities near the

flow straightener.

3.1.2 Instrumentation

The instrumentation for the data collection is composed of the instruments, for the

measurement of the water temperature and elevation, flow velocity and scour depth. The

instruments are described below.

3.1.2.1 Velocity measurement

Two commercially available electromagnetic current meters were used to measure

the velocity during the tests [Marsh-McBimey Models 523 (0.5 in sensor) and 511 (1.5 in

sensor)]. The water velocity was measured at the same two horizontal locations upstream

of the test structure for all the experiments. The meters were located a distance of 1.52 m

(5 ft) from the sides of the flume and approximately 5 meters upstream from the center of

the test structures. The vertical position of the velocity sensors were such that they were

at the point of depth averaged velocity for a fully developed velocity profile, which is

approximately yo3 from the bed. The time over which the velocity was averaged was

increased until the measurement was steady. This value was found to be one minute.

Velocities at the same elevation of the sensors were also measured using an impeller type

current meter (Ott-meter). This instrument was also used during the experiments to check

the electromagnetic meters. The duration of the measurement at each location was one

minute. The accuracy of the electromagnetic and impeller meter measurements was

estimated to be 1 cm/s and 0.5 cm/s respectively.

3.1.2.2 Water level measurement

A water pressure sensor was used to measure the water level during the tests. The

water level was measured at the same location for all the experiments downstream of the

test structure and approximately seven meters upstream of the weir. The time over which

the water level was averaged was one minute. The accuracy of the water level

measurement was estimated to be 0.5 cm.

3.1.2.3 Temperature measurement

A temperature sensor was used to measure the water temperature during the tests.

The temperature was measured at the same location for all the experiments, which was

close to the test structures. The meter was located a distance of approximately three

meters downstream from the test area and close to the wall of the flume. The time over

which the temperature was averaged was one minute. The accuracy of the temperature

measurement was estimated to be 0.050C.

3.1.2.4 Acoustic transducers

Two different transducer arrays were used for the temporal measurement of the

scour, one for the smaller structures and another one for the large structure. Both arrays

consisted of three elements, each of which contained four crystals. The transducers are

called Multiple Transducer Arrays or simply MTAs. They were positioned at the front of

the structure and at angles of 830 from the front. Their height from the sandy bottom

varied according to the water depth but they were always underwater and as close as

possible to the surface so as not to interfere with the scour process. For most of the

experiments, the transducers were located 10 cm (4 in) from the water surface. The

MTAs were custom built for this application by SeatekTM. For a more detailed description

of the system, see Jette and Hanes (1997).

The arrays for the small pile diameters consisted of four 2.25 MHz transducers with

4 cm separation between the elements. The transducers were 2.5 cm in diameter. The

footprint at a range of 0.9 m (3 ft) was 5 cm (which means a 1.5 degrees spread angle of

the acoustic beam). The arrays for the large pile diameters consisted of four 2.25 MHz

transducers with 8 cm separation between elements. The transducers were 4.0 cm in

diameter. The footprint at a range of 2.7 m (9 ft) was 8.7 cm. The arrays were made of

anodized aluminum.

The MTAs were mounted on each structure with an aluminum ring. The rings were

fastened at a certain height and each of the three arrays was positioned on the ring at

predetermined positions. Figure 3.3 shows the mounting of the transducers and cameras

and Figures 3.4 and 3.5 show the details of the transducers.

-I-

Camera
Traverse Mechanism--

Video Cameras

Test Sand

Acoustic Transdui

- Test Pile

I I

Figure 3.3 A test pile with the instrumentation

MTA for small pile diameters (0.5 to 1 ft. diameter)

side view (coss-section)

mounting ring

2.25 MHz transducers

front view (cross-section)

('27

0,375 l
(0,95 m)

j lSEATEK
u---eed i64In(e3 a n ----D- r raw by: Chris Jette' 3-24-97

Figure 3.4 Detailed view of the arrays for the small structures

FLOW

Plan View

1

MTA for large pile diameters (greater than 1 ft.) 1/2 scale

side view (cross-section) front view (cross-section)

mounting nng

0 5 In

top view

i..---0-- n(o.5, [ SEATEK
rawn by: Chris Jette' 3-25-97

Figure 3.5 Detailed view of the arrays for the large structures

3.1.2.5 Video measurements

The video measurement consisted of the video equipment, the mechanisms to

control them and the stepper motor for the movement of the cameras inside the piers.

Although it was used as a backup measurement, it proved to be very reliable. Two

cameras were used to monitor the rate of scour by moving vertically inside the piers. The

mechanism for support and moving of the cameras was the same for all the piers but the

carriage for the cameras changed according to the diameter of the pier (Figures 3.6 3.8).

The controller for the traverse mechanism was designed to allow traverse speeds from

10m/h to 1 mm/h. The mechanical part consisted of a threaded rod, which supported the

cameras, and was connected to a stepper motor, the speed of which could be set by the

controller. The rod had the appropriate length so the cameras could follow the sand-water

interface during the scour process.

The output from the video cameras was sent to the VCR. A second controller was

programmed to turn on the lights and record one minute of video at specified intervals

throughout the experiments. After the end of the recording time, the video and lights were

turned off and the system put in a stand-by mode until the next recording time. The

controller also switched between cameras during the recording session.

Figure 3.6 Picture of the cameras for the 0.114 m (4.5 in) pile

Figure 3.7 Picture of the cameras for the 0.305 m (12 in) pile

Figure 3.8 Picture of the cameras for the 0.92 m (36 in) pile

3.1.2.6 Measurement setup

Two personal computers with 486 processors were used for data acquisition. One

was used for velocity, water elevation and temperature and one for the acoustic transducer

measurements of the scour depth. The first computer was programmed to take one-minute

samples of velocity, water elevation and temperature every 30 minutes. The data were

written to a file on the hard drive. The second personal computer was connected to the

acoustic transducers through a control box (SeaTek Control Box). The purpose of this

box was to gather the signals from all 12 crystals and convert them to distances from the

transducer to the bed. The communication between the personal computer and the

acoustic control box was done with the software, CrosstalkTM, a serial/parallel/modem

communication package. The data was viewed on a computer screen and stored in files.

Data were sampled for ten seconds every ten minutes. A diagram of the measurement

system is shown in Figure 3.9.

PC Mechanical Monitor
#1 #2 Traverse
SSystem -
SeaTek VCR
Control Box
-Inside Camera/VCR
Cameras Control Box
Acoustic TieD
Water Digital Transducers Tme/D

Figure 3.9 Diagram of the measurement system

3.1.2.7 Point gauging

After every experiment, the scour hole was surveyed with a point gauge system. The

system was composed of a square array of steel beams that spanned the width of the

flume with a side length of 6.1 m (20 ft). The point gauge system was located on an

aluminum carriage that moved along tracks on the steel beams. The point gauge was

connected to a string potentiometer and voltage source so that the vertical position was

indicated by an electrical potential in mV. The horizontal and vertical accuracies of the

measurements were estimated to be 0.5 cm and 0.5 mm respectively.

3.1.3 Models

The models used in these experiments were piles with cylindrical cross-sections. The

pile diameters were 0.114 m (4.5 in), 0.305 m (12 in) and 0.92 m (36 in) with lengths that

exceeded the water depth. They were attached to the floor of the flume and given lateral

support near the top by a bridge that spanned the width of the flume above the water.

The 0.92 m (3 ft) diameter pile was 5.5 m (18 ft) high. It had two Plexiglas windows

fitted for scour visualization at angles of 450 from the front of the pile (Fig. 3.10). To

prevent the pile from distorting in shape, steel channels were attached to the walls inside

of the pier. Measuring tapes were glued to the inside faces of the piles at locations where

they could be seen by the inside video cameras. The 4.5 in and 12 in piles were

constructed of Plexiglas. Both were 3.35 m (11 ft) in length.

0.92 m (36 in)
diameter pipe

Aluminum
carrier

Waterproof case
with cameras

Figure 3.10 Cross-section of the 0.92m (36 in) pier

The rationale behind the pile diameter used in the experiments is as follows. The

0.92 m diameter was the largest standard diameter that could be used in the 6.1 m wide

flume without creating significant contraction scour. Even this diameter exceeded the

"rule of thumb" limit of 10% of the flume width. The second factor used in sizing the

piles was the need to maintain the same Dso/b ratio using different values of D50 and b. At

the time that the piles were being designed, the uniform diameter sand sizes that were

available were 0.20 mm and 0.65 mm. This would have produced Dso/b ratios as shown

in Table 3.1.

Table 3.1 Table with values of (-log (Dso/b)) before sand availability
Dso(mm)
b (m) 0.20 0.65
0.920 3.66 3.15
0.305 3.18 2.67
0.114 2.76 2.24

By the time the sand was purchased, the available uniform sand sizes changed to

0.22 mm and 0.80 mm. This resulted in Dso/b ratios as shown in Table 3.2. Even though

these ratios were not as close as planned, they were sufficient to test the dependence of

equilibrium local scour depth on the parameter Dso/b.

Table 3.2 Table with values of (-log (Dso0b)) for the piles and sediment
used in the tests
Dso (mm)
b (m) 0.22 0.80
0.920 3.62 3.06
0.305 3.14 2.58
0.114 2.71 2.15

3.1.4 Sediment

Two sand sizes, 0.22 mm and 0.8 mm, were used in these experiments. Experiments

with all three piles were performed with each of the two sediments. Near uniform

sediment diameters were used, since (under clearwater conditions) the greatest scour

depths occur in uniform diameter sediments. The sand was purchased from a vendor in

Rhode Island that had a system for producing near uniform diameter sand.

The grain size distributions for the two sand sizes are shown in Figures 3.11 and

3.12. Their properties are summarized as follows:

Sand No. 1

Ds0 = 0.22 mm

-= = 1.51
1A6

Mineral content = Quartz

Mass density = 2650 kg/m3

Sand No. 2

D5o = 0.80 mm

S 8= _= 1.29
6 D

Mineral content = Quartz

Mass density = 2650 kg/m3

100 -

80-

60-

40-

0.01

D4 = 0.32 mm

D50 = 0.22nim

20-[ D16=0.14nunm

S I I I 1 1I I
0.1
diameter (mm)

1

Figure 3.11 Gradation Curve for Sand No. 1 (Dso = 0.22 mm)

40

100-

D4= 1.07 mm

80-

60-

| D5o= 0.80 mm

40-

20 D6=0.64mm

0-1
0.1 1 10
diameter (mm)

Figure 3.12 Gradation Curve for Sand No. 2 (D50 = 0.80 mm)

CHAPTER 4
EXPERIMENTAL DATA REDUCTION AND ANALYSIS

This chapter contains the results from the experiments performed at the USGS-BRD

Laboratory in Turners' Falls, Massachusetts, as part of the work for this thesis. The

techniques and procedures used in performing these experiments and in reducing and

analyzing the data are also included.

4.1 Experimental Procedure

The sediment, flow and structure parameters used in the experiments are

summarized at Table 4.1 below.

Table 4.1 Parameters for the experiments conducted
Experiment No D50 (mm) b (m) yo (m) U (m/s) Uc (m/s) U/Uc
1 0.22 0.114 1.22 0.300 0.333 0.90
2 0.22 0.305 1.22 0.300 0.333 0.90
3 0.22 0.92 2.29 0.320 0.355 0.90
4 0.80 0.92 2.29 0.447 0.497 0.90
5 0.80 0.92 0.90 0.403 0.448 0.90
6 0.80 0.305 1.22 0.414 0.460 0.90
7 0.80 0.114 1.22 0.414 0.460 0.90

To assume consistent results, a series of procedural steps were developed for the

data collection. The procedures are divided into a) the pre-test, b) during test, and c) post-

test steps.

Before scour tests could be performed in the USGS-BRD flume, the bed had to be

prepared. First the gravel was put in place. Dividers in the flume made this task easier.

The next step was to place filter material on top of the gravel to separate it from the test

sediment. The following step was to install the structure. Next the test sediment was

placed on top of the gravel and in the test section. The sediment was compacted every 20

to 30 cm with a diesel driven mechanical compactor and hand tampers in the area near the

structure. The bed was then leveled. An observation platform that spanned the width of

the flume at the test section was mounted in the flume. The platform could be moved

vertically to accommodate the range of water depths used in the experiments. The next

step was to install and test the data collection equipment. The flume was then filled with

water up to the level of the weir. Care was taken so as not to disturb the bed or initiate

scour at the test structure. The water was then allowed to stand at this level, until the air

trapped in the sediment was released. This ranged from 4 to 10 hours.

The experiment was started by opening the gates until the desired water level and

flow velocity were reached. The water level and velocity were checked throughout the

experiment and minor adjustments were made when necessary. The speed of the video

camera traversing mechanism was also adjusted during the experiment to account for the

changing rate of scour.

When no scour depth changes were observed for approximately 24 hours, the

experiment was terminated and the flume drained. The procedure took two to three hours.

Pictures of the scour hole were taken from six positions around the pile. The scour hole

was then point gauged and the structure was removed. At the end of the tests with the first

sediment, the test sand was loaded into 0.76 m3 (1.0 yd3) polypropylene bags and

removed from the flume. The second sand was then placed in the flume using the same

procedure as for the first sand.

4.2 Data Reduction

The quantities measured during these experiments were water temperature, water

elevation, flow velocity and scour depth. The scour depth was measured using two

different instruments during the test (acoustic and video) and a third method was used at

the end of each test after the water was drained from the flume (point gauge).

The data for temperature and water elevation were written to a file on the hard drive

of a personal computer (PC) used for data collection. At the end of each experiment, each

of the three data sets were averaged over the time span of the test.

The data from the two velocity meters were also written to the hard drive on the

same PC. The two data sets were averaged individually over the time duration of the tests.

The average from the individual gauges were then averaged to obtain the value associated

with that experiment. An impeller velocity meter (Ott meter) was used to check the

electromagnetic meters.

The video data were stored on one or more videocassettes for each experiment. After

each experiment, the video tapes were viewed and the scour depth versus time was

recorded.

The acoustic transducer data were written to a file on the hard drive on a second PC.

The data collected had to be filtered in order to remove those signals that did not

correspond to scour hole depth. These were caused by acoustic reflection from suspended

sediment and other particles in the flow and zero readings when the return signal was not

detected. The filtering was accomplished with the help of a computer program called

SMS, which allowed the data to be displayed and edited graphically. The acoustic

transducer ran continuously throughout the experiment and ten seconds of data was

sampled and recorded every ten minutes.

After the test was completed and the flume was drained, the entire scour hole was

surveyed with a point gauge. The (x, y) position of the tip was recorded along with the

voltage output from a string potentiometer, which was proportional to the vertical

positions of the gauge. The initial position of the bed was established by the initial tape

reading from the video cameras inside the pile. Then the voltage readings were converted

into distances from the undisturbed bed to the elevation of the scour hole at each

horizontal (x, y) position. The origin of the coordinate system was located at the center of

the pile and the initial, undisturbed bed. The survey extended to the flat portion of the bed

beyond the scour hole.

4.3 Processed Data Results

The processed data are presented in summary tables and graphs. A table in Appendix

C gives the summary of all the experiments conducted along with graphs showing the

scour depth versus time for both the video and acoustic transducers (when both are

available).

Contour plots of the equilibrium scour holes for all seven experiments are presented

on Figures 4.1 to 4.7. Equilibrium scour profiles inline and normal to the flow for all

seven experiments are given in Figures 4.8 to 4.14.

Z (in)

-1

-20 -10 0 10 20 30
X (in)
Figure 4.1 Elevation contour plot of the equilibrium scour hole for Experiment No.1
(b = 0.114 m, D50 = 0.22 mm, yo = 0.186 m and U = 0.290 m/s)

z (in) 0O o
330-
3 0 .. .a.

-3

-2
-6

-10
-1 0

-30 -20 -10 0 10 20 30 40 50 60 70
X (in)

Figure 4.2 Elevation contour plot of the equilibrium scour hole for Experiment No.2
(b = 0.305 m, D50 = 0.22 mm, yo = 0.190 m and U = 0.305 m/s)

Z (in)
5
3
1
0
-1
-3
-5
-7
-9
-11
-13
-14
-15

8N' , ~- -I

:I,4

S)0
x (in)

Figure 4.3 Elevation contour plot of the equilibrium scour hole for Experiment No.3
(b = 0.920 m, Dso = 0.22 mm, Yo = 2.268 m and U = 0.325 m/s)

60L

-18 2

-22
-26

-26 -20 40 40
-30
-34
-37 8

X (in)

Figure 4.4 Elevation contour plot of the equilibrium scour hole for Experiment No.4
(b = 0.920 m, D50 = 0.80 mm, Yo = 2.402 m and U = 0.454 m/s)

Z (in)

6

2
-2

--6

X(in)

Figure 4.5 Elevation contour plot of the equilibrium scour hole for Experiment No.5
(b = 0.920 m, D50 = 0.80 mm, yo = 0.866 m and U = 0.335 m/s)

z (in)

Q6
A

X(in)

Figure 4.6 Elevation contour plot of the equilibrium scour hole for Experiment No.6
(b = 0.305 m, D50 = 0.80 mm, yo = 1.305 m and U = 0.381 m/s)

48

Z(in) 3
2
0
-1 I0
-2 z .

-5 -2oL
-6
-30-

-7
1-8

X (in)

Figure 4.7 Elevation contour plot of the equilibrium scour hole for Experiment No.7
(b = 0.114 m, Dso = 0.80 mm, Yo = 1.280 m and U = 0.388 m/s)

10-

8-

6-

4-

-4 4 I _

I I
-30 -20

' I
-10

a) Inline profile

0 -

-30 -20 -10 0
Y (in)

10 20 30

b) Normal profile

Figure 4.8 Equilibrium profiles of scour hole for Experiment No. 1
(b = 0.114 m, Dso = 0.22 mm, yo = 0.186 m and U = 0.290 m/s)
a) Inline profile, and b) Normal profile

Flow

--------------

0-

-2 -

-4 -

10
X (in)

30
30

40
40

FloIv
0

I

Flow

-40 -30 -20 -10 0 10 20 30 40 50 60 70
X (in)

a) Inline profile

Flo\~
0

0 -4- -

-60 -40 -20 0
Y (in)

20 40 60

b) Normal profile

Figure 4.9 Equilibrium profiles of scour hole for Experiment No.2
(b = 0.305 m, D50 = 0.22 mm, yo = 0.190 m and U = 0.305 m/s)
a) Inline profile, and b) Normal profile

-70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60
X (in)

a) Inline profile

-80 -60 -40 -20 0
Y (in)

20 40 60 80

b) Normal profile

Figure 4.10 Equilibrium profiles of scour hole for Experiment No.3
(b = 0.920 m, D50 = 0.22 mm, yo = 2.268 m and U = 0.325 m/s)
a) Inline profile, and b) Normal profile

10-

0-

-10-

-20 -

-30 -

-40

a) Inline profile

-80 -60 -40 -20 0
Y (in)

20 40 60 80

b) Normal profile

Figure 4.11 Equilibrium profiles of scour hole for Experiment No.4
(b = 0.920 m, D5o = 0.80 mm, Yo = 2.402 m and U = 0.454 m/s)
a) Inline profile, and b) Normal profile

Flo\w

0
X (in)

- I
-80 -60

SI
-40

20
20

40

60

80
80

'

I *

\

\,
- - - - -

",z

' I

-80 -60 -40 -20 0 20 40 60 80
X (in)

a) Inline profile

-80 -60 -40 -20 0
Y (in)

20 40 60 80

b) Normal profile

Figure 4.12 Equilibrium profiles of scour hole for Experiment No.5
(b = 0.920 m, D50 = 0.80 mm, yo = 0.866 m and U = 0.335 m/s)
a) Inline profile, and b) Normal profile

-30 -20 -10 0 10 20 30 40
X(in)

50 60

a) Inline profile

-50 -40 -30 -20 -10 0
Y(in)

10 20 30 40 50

b) Normal profile

Figure 4.13 Equilibrium profiles of scour hole for Experiment No.6
(b = 0.305 m, D50 = 0.80 mm, yo = 1.305 m and U = 0.381 m/s)
a) Inline profile, and b) Normal profile

-20 0 20 40
X (in)

a) Inline profile

Flo\%
S

-50 -40 -30 -20 -10

0
Y (in)

10 20 30 40 50

b) Normal profile

Figure 4.14 Equilibrium profiles of scour hole for Experiment No.7
(b = 0.114 m, D50 = 0.80 mm, yo = 1.280 m and U = 0.388 m/s)
a) Inline profile, and b) Normal profile

- - - - -

4.4 Discussion of Results

The most unique aspects of the local scour experiments conducted during the work

of this thesis have to do with the size of the structure tested and the duration of the tests.

The 0.92 m (36 in) diameter pile is one of the largest, tested under controlled steady flow

conditions. Any questions regarding the equilibrium depths were eliminated by

continuing the tests until no measurable change in the scour hole was observed.

Several interesting observations were made regarding the scour hole at the large

diameter piles. The location of the deepest point in the hole was different from that for

the smaller piles. For the 0.114 m (4.5 in) and 0.305 m (12 in) diameter piles, the

equilibrium scour hole was more uniform in depth from the side to the front (upstream

edge) with the deepest point being directly in front of the pile. The deepest point in the

scour hole at the 0.92 m pile was between 450 and 600 from the front. There are several

possible explanations for this. One might be the fact that the slope of the pressure

distribution along the circumference of the pile is smaller for large diameter piles than for

small diameter piles. Another possible reason is that the effect of the horseshoe vortex

seems to be diminished for the large piles. The scour holes from the accelerated flow do

not extend to the front of the large structure, thus the horseshoe vortex is deprived of the

energy, caused by flow separation at the upstream edge of the scour hole.

Note that there was a slight asymmetry of the scour hole (see for example Figure

4.11). This is believed to be due to the slightly higher velocity on the west side of the

flume. Flow profiles for the 1.22 m (4 ft) and 2.44 m (8 ft) water depths that show this

asymmetry are given in Appendix B.

Ripples in the test section for the first two experiments can be also observed (Figures

4.1 and 4.2). The reasons for their formation are various. The sand, used in these

experiments, was very fine (0.22 mm), and ripples are common for sediments with D50 -

0.6 mm. Another reason is that those experiments were conducted at conditions very

close to live-bed conditions, so there may have been brief periods of time, when the flow

exceeded the critical value, Uc.

The other reason is the presence of a big sand dune, which did not affect the

formation of the scour holes, but helped initiate the ripples. This sand dune was formed

initially near the flow straightener due to velocities greater than Uc. The flow straightener

had an opening of 25% of the total cross section, so the velocities were up to four times

greater than the velocity downstream. Those conditions formed a sand dune that could

travel downstream but its speed was very small and it did not affect any of the results.

This problem was later corrected by substituting gravel for sand just downstream of the

flow straightener.

CHAPTER 5
COMPARISON OF LOCAL SCOUR PREDICTION EQUATIONS

5.1 Scour Prediction Equations

Over the past 40 years, numerous studies have been conducted and many equations

have been developed to predict local equilibrium scour depth. Most of these equations

were based on laboratory data. In some cases, limited field data was also used. Due to the

complexity of the processes involved in local scour and the large number of parameters

needed to describe these processes, the predictive equations differ dramatically. In some

cases, the data on which these empirical equations are based were not published. Fourteen

of the more commonly referred to equations are evaluated and compared in this chapter,

using two data sets, one for clearwater and one for live-bed scour conditions. The

clearwater data set includes the data obtained as part of the work for this thesis.

Perhaps the most commonly used scour prediction equation in the United States is

the so-called Colorado State University (CSU) equation. It is the recommended equation

in the U.S. Federal Highway Administration (FHWA) Hydraulic Engineering Circular

No. 18 (HEC-18) (1995). This equation is

( b 0.65
dse =2.0yK, K2 .Fr0.43, (5.1)

where

dse is the equilibrium scour depth,

Yo is the flow depth upstream of the structure,

K1 is a correction factor for pier nose shape,

K2 is a correction factor for angle of attack flow,

b is the pier width and

U
Fr = Froude number -
4g-o

Values from K1 and K2 can be obtained from HEC-18 and are discussed in Chapter 2.

This equation was developed by fitting laboratory data and is recommended for both live-

bed and clearwater conditions. In HEC-18, limiting values for the scour depth normalized

by the pier diameter, dse/b, are recommended as follows: dse/b = 2.4 for Fr < 0.8 and 3.0

for Fr > 0.8.

Melville (1997) developed a scour equation based on various laboratory experiments

of the form:

dse =I KdKyDKaKs, (5.2)

where

->1
1 u,
K, = flow intensity factor = for
U U
-<1
U, U,

b
-->25
1.0 D5
Kd = sediment-size factor = for

0.57 log 2.24 -- < 25
I D50 D50

b
< 0.7
2.4b Yo

KyD = flow depth-pier width factor = 2 y b for 0.7 < <5
yK
YO

4.5yo b
>5
Yo

K = pier-alignment factor, and

K, = pier-shape factor.

Values for K, and Ks can be obtained from tables given in Melville and Sutherland

(1988).

Hancu (1971) proposed the following scour prediction equation

dse = 2.42b 2- (5.3)
I Uc gb

U2
for 0.05 < _< 0.6 where Ue is calculated by the equation
gb

0.2
Uc = 1.2 (s-1)gD5o Yo .
D50

Equation (5.3) does not apply for values of less than 0.5.
UC

Laursen and Toch (1956) developed design scour depth curves that were later

described by Neill (1964) in equation form as

dse =1.35b07y03. (5.4)

Shen et al. (1966) used laboratory data and limited field data to develop the

following clearwater scour prediction equation,

dse= 0.000223 ReO619, (5.5)

Ub
where Rep = -. For live-bed conditions, Equation (5.5) was found to be too
v

conservative, so Shen (1971) recommended using the Larras (1963) equation,

dse =1.05b075. (5.6)

Breusers et al. (1977) developed the following equation (which is similar to

Equation (5.3) (Hancu, 1971))

dse = b f KK2 tanh ), (5.7)

where

U,
f =0 for -<0.5,
Uc

U U
f=2 -1 for 0.5<- <1.0,
U, Uc

U
Uc

and KI and K2 are the same as in Equation (5.1).

Jain and Fischer (1979) developed a set of equations based on their laboratory

experiments. For Fr Frc > 0.2

dse =2b(Fr Fr ).25 o (5.8)
(b

where Frc = critical Froude number = c
-gyo

For Fr-Frc <0

ds = 1.85bFr 025' (5.9)

For 0 < Fr Frc < 0.2, the larger of the two scour depths computed from Equations (5.8)

and (5.9) are to be used.

Garde et al. (1993) developed a clearwater scour equation based on various

laboratory experiments

,7 0.75(, 0.16 2 -0.4
ds =0.66 o u ,- (5.10)
Do Ds0) Ds0) (s-1)gDso

where Vc is the velocity for the initialization of sediment movement at the structure,

which is about 0.4-0.5 Uc, and s is the relative sediment density. The equation given to

calculate Vc is

V2 1, b -0. 11( Yo 0.16
(s-1)gD50 D50 D50)

Another scour prediction equation was developed by Chitale (1962), using very

limited laboratory data,

d = -0.51+ 6.65Fr 5.49Fr2. (5.11)
Yo

Based on data from previous investigations for local scour at spur dikes, Ahmad

(1953) concluded that scour does not depend on grain size for the range of his data (0.1 to

0.7 mm). He stated that this might not be accurate for the entire range of the sediment

sizes. Ahmad (1962) reanalyzed the work of Laursen (1962) with a special emphasis on

scour in sand beds in West Pakistan and developed the following equation:

dse = Kq2/3 Y, (5.12)

where K is assumed from model studies conducted by Ahmad (1962) to vary from 1.7 to

2.0 for piers and abutments. A value of 1.85 was used in the calculations made in this

thesis. Also q is the discharge per unit width of the channel and all units are in English.

Inglis (1949) performed experiments on model bridge piers and developed an

equation from his data. Blench (1962) reduced Inglis' (1949) formula to
- ,0.25
dse =1.80.25q 0.5 o -Yo, (5.13)

where all units are in English.

Gao et al. (1992) presented empirical equations for clearwater and live bed local

scour that were used by highway and railway engineers in China for more than 20 years.

The equations for clearwater and live-bed conditions are

dse = 0.78Kb0.6 015D07 U U (5.14)
U U

and dse = 0.65K b06 yo15D.07 U U (5.15)

respectively. Ks is a coefficient for pier shape and has the values of 1.0 for cylinders, 0.8

for round-nosed piers and 0.66 for sharp-nosed piers. The critical depth average velocity,

Uc, the velocity for initialization of scour at the pier, U' and the exponent c in Equation

(5.15) are calculated using

Uc= 0.14 [17.6(s-)D +6.05E 7 10 0.5

(D o0.0o53
', = 0.645-50 U ,,and
( b )

C= ( )9.35+2.23logD50
C= U\

Froehlich (1988) compiled field measurements of local scour at bridge piers from

reports of several investigations and developed the equation

dse = 0.320b Fr0.2 0.6 (5.16)
x.46 b o.08

where ) is a pier shape factor and has the value of 1.3 for square-nosed piers, 1.0 for

round-nosed piers, and 0.7 for sharp-nosed piers.

Sheppard et al. (1995) at the University of Florida developed a clearwater scour

prediction equation for cylindrical piers based on laboratory data,

b jb U, b
(5.17)
=4.81tanh [ 1-2.87 +1.43L ] 2log D exp -0.18[-log-( 2]2.9
b) t(_) I( b )

Based on his limited live-bed dataand data from other researchers, Sheppard (1997)

concluded that a second peak in the normalized scour depth (dse/b) versus normalized

velocity (U/Uc) plot occurred in the live-bed scour range and had a value of

approximately 2.1 (for yo/b > 2.5). The value of U/Uc, where the live-bed peak occurs, is

believed to coinside with the conditions, where the bed flattens. That condition has been

shown to be dependent on the Froude number and the velocity normalized by the

sediment fall velocity (Snamenskaya, 1969). For design purposes Sheppard (1997)

recommended connecting the clearwater peak to the live-bed peak with a straight line as

shown in Figure 5.1. For velocities greater than the value for the live-bed peak, the

equilibrium scour depth is assumed to be the value at the live-bed peak. For values of

yo/b < 2.5, the height of the live-bed peak is computed using the equation

e = 2.ltanh bo
db

independent of Do/b

2.1 ---- ----- -- -- -

depends on Dso/b

0.4- 0.5 1.0 (U/U),bp
U/U

Figure 5.1 Local scour depth dependence on velocity,
as proposed by Sheppard (1997)

5.2 Compilation of Local Scour Data

The clearwater and live-bed data sets used in this analysis were compiled from a

number of sources. These include Ettema (1980), Chiew (1984), Sheppard et al. (1995),

Chabert and Engeldinger (1956), Jones (1997), Melville and Chiew (1999), Graf (1995),

Shen et al. (1966), and the data obtained as part of the work for this thesis.

The data were divided into two categories, clearwater and live bed. This was done

for several reasons. The first reason is that the local scour processes are different for these

two categories. In the clearwater region, there is sediment movement only around the

structure, while in live-bed conditions there is movement over the entire bed. That may

result even at the reduction of the scour hole for velocities just over the critical value, Uc,

and is based on observations by Ettema (1980), Chiew (1984) and others. Another reason

is that many of the scour prediction equations were based on curve fits to just clearwater

data (Garde 1992, Sheppard et al. 1995), while others used both clearwater and live-bed

data.

The criteria used in selecting the data used in this analysis were as follows. The

author had to present sufficient information about the parameters of the experiment, so

that the equations used in this analysis could be evaluated. In addition the duration of the

experiments had to be sufficiently long that equilibrium (or near equilibrium) scour

depths were achieved. The time required to reach equilibrium is not well understood, but

it is known to increase with the size of the structure and (for clearwater scour) the

velocity. For the purpose of this analysis a minimum of 17 hours was chosen. Only scour

data for structures with a circular cross section were used, because the coefficients for

other shapes of structures vary for the different equations. Another criterion was that the

test sediment had to be relatively uniform in grain size. The data chosen had a gradation a

less than 1.6. Using these criteria, 215 clearwater and 244 live-bed data points were

compiled.

The scour depths were computed for the conditions of the test using Equations

(5.1) (5.17). The value for critical depth average velocity, Uc, was calculated using

Shields' diagram, except where stated otherwise, like for Hancu (1971), Melville (1984),

HEC-18 (1995) and Gao et al. (1992). For the data where temperature was not reported, a

value of 150C was assumed. Little or no data on bedforms was available for most of the

data. In the computation of Uc, a relative roughness r (ks/D5o) was assumed to be five. For

Breusers et al. (1977) and Hancu (1971), scour depths were assumed to be zero for U/Uc

< 0.5 and for U/Uc < 0.45 for Sheppard et al. (1995). Also for Hancu (1971), some data

points could not be considered due to limitations on the equation by the author.

5.3 Comparative Analysis

Predictions of the clearwater data will be discussed first. The Jain and Fischer

(1979), Laursen and Toch (1956), Gao et al. (1992) and Melville (1997) equations do a

good job of predicting clearwater scour depths for smaller diameter piles but overpredict

as the diameter increases. The Chitale (1962) and Ahmad (1953) equations, which depend

on a limited number of parameters, overpredict most of the data, as expected. The Hancu

(1971) equation underpredicts (as much as 95%) for the majority of the data points where

the equation can be applied. The Breusers et al. (1977), Froehlich (1988), Garde et al.

(1993), Inglis-Blench (1962) and Shen (1966) equations underpredict most of the data.

The best predictions were by the HEC-18 (1995) and Sheppard et al. (1995) equations

with an average overprediction of 31% and 16% respectively.

When the same equations (with the exception of Garde et al. (1992)) are used to

predict the 244 live-bed scour depths, some differences are observed. The Jain and

Fischer (1979), Gao et al. (1992) and Melville (1997) equations still overpredict as the

pier diameter increases. The Laursen and Toch (1956) equation gives better predictions

(53% as opposed to 104% for clearwater scour). The Chitale (1962) equation still

overpredicts (up to 690%) and the Ahmad (1953) equation underpredicts most of the

live-bed data. The Hancu (1971) equation, which cannot be applied to all the data, still

underpredicts by as much as 92%. The Froehlich (1988) and Inglis-Blench (1962)

equations underpredict for the live-bed data, as they did for clearwater conditions. The

Breusers et al. (1977) and Shen (1966) equations overpredict for most of the data. Again

the best predictive equations are the HEC-18 (1995) and Sheppard et al. (1995) equations

with mean overpredictions of 26% and 12% respectively. Both equations appear to be

more accurate for live-bed conditions.

The mean of the absolute values of the predictions and the range of the predictions

for all of the equations are shown in line plots of the ratio of measured to calculated scour

depths are given in Figures 5.2 and 5.3 for the clearwater, and live-bed data sets

respectively. The mean of the absolute value of the predictions for every equation is

represented with an asterisk along the range of the prediction line. The line of absolute

agreement is the dotted line passing from 1.0. The narrower the range and the closer the

mean of the absolute value is to the line of absolute agreement, the better is the equation.

Scatter plots of calculated versus measured scour depths for all the equations are given in

Figures 5.4 5.17.

C

Froelich (1988)-
Gao et al. (1992)-
Inglis-Blench (1962)-
Ahmad (1953)-
Melville (1997)-
Garde et al. (1993)-
Chitale (1962)-
HEC-18 (1995)-
Jain & Fisher (1979)-
Breusers et al. (1977)-
Hancu (1971)-
Laursen & Toch (1956)-
Shen et al. (1966)-
Sheppard et al. (1995)-

comparison of 14 equations
using 215 clearwater data
I

I I I I I I I
-4 -2 0 2 4 6 8
Ratio of calculated to measured scour
- Prediction Range
Mean of Absolute Value of Scour Predictions
- Absolute Agreement Line

Figure 5.2 Comparison of equations for clearwater data

I
10

I L1

w

I w

w

h\
I
iw

~---

I

I,
I

70

Comparison of 13 equations
using 244 live-bed data

Froelich (1988)
Gao et al. (1992)-
Inglis-Blench (1962)-
Ahmad (1953)-
Melville (1997)-
Chitale (1962)-
HEC-18 (1995)-
Jain & Fisher (1979)-
Breusers et al. (1977)-
Hancu (1971)-
Laursen & Toch (1956)-
Shen et al. (1966)-
Sheppard et al. (1995)-

Ke
-~t---

I,

-1~-
1,

~jc

-I. I

-6 -4 -2 0 2 4 6 8
Ratio of calculated to measured scour

Prediction Range
Mean of Absolute Value of Scour Predictions
- - Absolute Agreement Line

Figure 5.3 Comparison of equations for live-bed data

I
10

-

(A) Garde et al. (1993)

Ettema (1980)
Chiew (1984)
UF/USGS
Chabert & Engeldinger (1
Jones (1997)
Melville & Chiew (1999)

0 0.3 0.6 0.9
measured scour (m)

1.2 1.5

(B) Ahmad (1953)

++
++
++
++

Ettema (1980)
Chiew (1984)
UF/USGS
Chabert & Engeldinger (1956)
Jones (1997)
Melville & Chiew (1999)

0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4
measured scour (m)

Figure 5.4 Comparison of measured and calculated scour for clearwater data for
(A) Garde et al. (1993), and (B) Ahmad (1953) equations

1.5-

1.2-

- I
0 0.3

' I I
0.6 0.9
measured scour (m)

(B) Chitale (1962)

++
+ +

Ettema (1980)
Chiew (1984)
UF/USGS
Chabert & Engeldinger (1956)
Jones (1997)
Melville & Chiew (1999)

0 0.4 0.8 1.2 1.6 2 2.4
measured scour (m)

Figure 5.5 Comparison of measured and calculated scour for clearwater data for
(A) Breusers et al. (1977), and (B) Chitale (1962) equations

0.9-

0.6-

0.3 -

0-

1.2

I
1.5

(A) Breusers et al. (1977)
S + Ettema(1980)
0 Chiew (1984)
* UF/USGS
A Chabert & Engeldinger (1956)
E Jones (1997)
r Melville & Chiew (1999)

c**

2 + +
2L + +

(A) Froehlich (1988)

0 0.3 0.6 0.9
measured scour (m)

1.2 1.5

(B) Gao et al. (1992)

Ettema (1980)
Chiew (1984)
UF/USGS
Chabert & Engeldinger (1!
Jones (1997)
Melville & Chiew (1999)

0 0.3 0.6 0.9
measured scour (m)

1.2 1.5

Figure 5.6 Comparison of measured and calculated scour for clearwater data for
(A) Froehlich (1988), and (B) Gao et al. (1992) equations

(A) Hancu (1971)
Ettema (1980)
Chiew (1984)
UF/USGS
Chabert & Engeldinger (1956)
Jones (1997)
Melville & Chiew (1999)

0 0.3 0.6 0.9 1.2 1.5
measured scour (m)

(B)HEC-18(1995)

Ettema (1980)
Chiew (1984)
UF/USGS
Chabert & Engeldinger (1956)
Jones (1997)
Melville & Chiew (1999)

0 0.2 0.4 0.6
measured scour (m)

Figure 5.7 Comparison of measured and calculated scour for clearwater data for
(A) Hancu (1971), and (B) HEC-18 (1995) equations

(A) Inglis-Blench (1962)

Ettema (1980)
Chiew (1984)
UF/USGS
Chabert & Engeldinger (1956)
Jones (1997)
Melville & Chiew (1999)

0 0.3 0.6 0.
measured scour (m)

(B) Jain and Fischer (1979)

Ettema (1980)
Chiew (1984)
UF/USGS
Chabert & Engeldinger (1956)
Jones (1997)
Melville & Chiew (1999)

0 0.3 0.6 0.9 1.2 1.5 1.8
measured scour (m)

Figure 5.8 Comparison of measured and calculated scour for clearwater data for
(A) Inglis-Blench (1962), and (B) Jain and Fischer (1979) equations

(A) Laursen and Toch (1956)

2-

1.6 -

* *

I I I I I

+ Ettema (1980)
O Chiew(1984)
* UF/USGS
A Chabert & Engeldinger (1956)
Jones (1997)
r Melville & Chiew (1999)
I I I I I I

I I I I
0 0.4 0.8 1.2
measured scour (m)

(B) Melville (1997)

4 + Ettema (1980)
0 Chiew (1984)
S* UF/USGS
A Chabert & Engeldinger (1956)
ED Jones (1997)
F Melville & Chiew (1999)

0 0.3 0.6 0.9
0 0.3 0.6 0.9

1.2 1.5 1.8 2.1
1.2 1.5 1.8 2.1

measured scour (m)

Figure 5.9 Comparison of measured and calculated scour for clearwater data for
(A) Laursen and Toch (1956), and (B) Melville (1997) equations

*

*

1.2-

0.8-

0.4-

A-

2.1 -

1.8-

1.5-

1.2-

0.9-

0.6-

0.3-

A-

I

\f

t

v

(A) Shen et al. (1966)
+ Ettema (1980)
0 Chiew (1984)
* UF/USGS
A Chabert & Engeldinger (1956)
e Jones (1997)
] Melville & Chiew(1999)

*: *
++ *
a++
+- made ++ *
4gj

I 0.3
0 0.3

0.6 0.9
measured scour (m)

1.2

(B) Sheppard et al. (1995)

Ettema (1980)
Chiew (1984)
UF/USGS
Chabert & Engeldinger (1'
Jones (1997)
Melville & Chiew (1999)

1.5

0 0.2 0.4 0.6 0.8 1
measured scour (m)

Figure 5.10 Comparison of measured and calculated scour for clearwater data for
(A) Shen et al. (1966), and (B) Sheppard et al. (1995) equations

1.5 -

1.2-

0.9-

0.6-

0.3 -

0-

1 --

0.8-

F .........

V

(A) Ahmad (1953)

0 0.4 0.8 1.2
measured scour (m)

(B) Breusers et al. (1977)

A AL
+ +
+ +A,. *

+++ + +

AL

S + Shen et al. (1966)
Chiew (1984)
A Chabert & Engeldinger (1956)
Graf(1995)
/ Sheppard (1995)
'

0.1
0.1

0.2
0.2

0.4
0.4

measured scour (m)

Figure 5.11 Comparison of measured and calculated scour for live-bed data for
(A) Ahmad (1953), and (B) Breusers et al. (1977) equations

0.4-

0.3-

0.2-

0.1 -

0-

' I

(A) Chitale (1962)

0

A A A
A A A

*A A A

0 0.2 0.4 0.6 0.8
measured scour (m)

(B) Froehlich (1988)

+ Shen et al. (1966)
* Chiew(1984)
A Chabert & Engeldinger (1956)
* Graf(1995)
S Sheppard (1995)

9

0 0.2
measured scour (m)

Figure 5.12 Comparison of measured and calculated scour for live-bed data for
(A) Chitale (1962), and (B) Froehlich (1988) equations

80

(A) Gao et al. (1992)

0.8

0.6

E0
:J

0.4
C)

0.2

0

0 0.2 0.4 0.6 0.8
measured scour (m)

(B) Hancu (1971)
+ Shen et al. (1966)
Chiew (1984)
A Chabert & Engeldinger (1956)
Graf(1995)
9 Sheppard (1995)

AA A A AA

A A A
S A A *
uA A A A 4 AA
44 A& A A
AA* A A A

' I

0.1 0.2
measured scour (m)

0.3
0.3

Figure 5.13 Comparison of measured and calculated scour for live-bed data for
(A) Gao et al. (1992), and (B) Hancu (1971) equations

9 9

+
4 A

0.3 -

0.2-

0.1-

S I

81

0.4 (A) HEC-18 (1995)

0.3
+ A

o

0.2- +
4-+
o +A

0.1 4- + Shenetal.(1966)
4 0 Chiew (1984)
A Chabert & Engeldinger (1956)
Graf(1995)
e Sheppard (1995)
0 ---- I I I I
0 0.1 0.2 0.3 0.4
measured scour (m)

(B) Inglis-Blench (1962)
+ Shen et al. (1966)
Chiew (1984)
A Chabert & Engeldinger (1956)
Graf(1995)
Q Sheppard (1995)

0.2

U./ A AA
CA. / +
o +A* *
:: / +++ + +
0.1- A*
0

0 0.1 0.2 0.3
measured scour (m)

Figure 5.14 Comparison of measured and calculated scour for live-bed data for
(A) HEC-18 (1995), and (B) Inglis-Blench (1962) equations

(A) Jain and Fischer (1979)

0.1

0-

0.4

0.3

0
0
-o 0.2

0.1

0

0 0.1 0.2 0.3 0.4 0.5
measured scour (m)

(B) Laursen and Toch (1956)

A A&
+ +
+ -- Ai/ *
++t+ +,
+
At
AL .,/

0 0.1 0.2 0.3 0.4
measured scour (m)

Figure 5.15 Comparison of measured and calculated scour for live-bed data for
(A) Jain and Fischer (1979), and (B) Laursen and Toch (1956) equations

0.6

E 0.4

0
U
o

S0.2
o .

(A) Melville (1997)

A

+ A
A
-t? Q *
A A- A
A4t+ ++
AA

0 0.1 0.2 0.3 0.4 0.5
measured scour (m)

0.3

S0.2

0

0.1

0

(B) Shen et al. (1966)

+ +++ im+it k Q

AinA *

0 0.1 0.2 0.3
measured scour (m)

Figure 5.16 Comparison of measured and calculated scour for live-bed data for
(A) Melville (1997), and (B) Shen et al. (1966) equations

84

Sheppard et al. (1995)

0.4

0.3

0
S0.2

0.1

0

0 0.1 0.2 0.3 0.4
measured scour (m)

Figure 5.17 Comparison of measured and calculated scour
for live-bed data for Sheppard et al. (1995) equation

*
Ar

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
70	html_echo_mainwriter.add_text_to_page	Finished reading and writing the file