Thesis Jets Chi Fung Wu

Erosion Mechanisms
below a vertical plain Jet

on a Non-Cohesive Soil Bed
by
Chi-Fung Wu
to obtain the degree of Master of Science
at the Delft University of Technology,
to be defended publicly on August 7, 2018 at 2:00 PM.
Student number: 1373749

Thesis committee: Prof. dr. ir. C. van Rhee TU Delft
Dr. ir. G.H. Keetels TU Delft
Dr. ir. M. van Damme TU Delft
This thesis is confidential and cannot be made public until December 31,
2019.
An electronic version of this thesis is available at

http://repository.tudelft.nl/.
Abstract
This experimental research is focused on fail mechanisms in non-cohesive soil

beneath the jet with a Jet Erosion Test(JET). A vertical plain jet impinges
fluid to a sand bed, which causes zero shear stress beneath the jet averaged
over time. Many erosion theories would say that no pick-up happen at that
point.
Ec = √ √u
H/d ∆gD50
Large range of erosion parameter was done in this research with video
recordings to the activities beneath the jet. Tests varies with those param-
eters outflow velocities, stand off distances(SOD) and grain sizes. Compar-
isons are made to Rajaratnam, who has executed similar JET with varies
erosion parameter too. The cavity depths and hill heights correspond well
as function of the erosion parameter. Cavity widths have deviations for
values above 4.0, which might cause by different grain properties or by the
limitation of flume width. Secondary flow limits the width growth over
time.
Weak deflected jet showed only surface erosion by micro turbulence at
the soil bed till erosion parameter 1.0. One smooth bed was formed between
two hill tops. Static bed did not exceed the concerning sand during and after
start. Increasing further the erosion parameter (1.0-5.0), soil deformation
was also seen with ejection beneath the jet. Bearing capacity by Prandtl’s
theory was insufficient and failed the soil bed. Consequence, the inner cavity
exceeds the internal friction angle of sand with dynamic erosion. Erosion
parameter larger than 5.0 had a wider cavity. Grains in soil translate and
rotate in the start phase by insufficient bearing capacity. Vortices sweep
around the jet with chaotic particle transport in the inner cavity.
Theories of Van Rijn, Meyer Peter Müller and Van Rhee are not suit-
able for JET. The pick-ups rates are underpredicted in all cases, which are
based on shear stress. The impinging jet has different flow field and failure
mechanism, which can be better related to the jet momentum.
i
Acknowledgements
This report contains Chi-Fung Wu’s master thesis, which is part of the mas-
ter Civil Engineering at Delft University of Technology. The experiments
and documentation were done in collaboration with Dredging Engineering
at the faculty 3ME of the same institute.
I am grateful to the members of my committee Prof.dr.ir C. van Rhee,

dr.ir. G.H. Keetels and dr.ir. M. van Damme. Their expertise and insight
contributed significantly to this thesis.
I am thankful to Andrew for keeping me accompany at the laboratory dur-

ing the experiments. I would also like to thank mr. Yang for lots of editing
tips and Annie Wong for the English check.
Also, my gratitudes go to my parents, relatives and friends for their support

during my education in Delft.
Chi-Fung (Charles) Wu
ii
Contents
Abstract i
Acknowledgements ii
List of Figures vii
List of Tables ix
List of Symbols xi
1 Introduction 1
1.1 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Outline of thesis . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Literature review of f low processes 4

2.1 Velocity profiles . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1.1 Behaviour of free jet . . . . . . . . . . . . . . . . . . . 4
2.1.2 Processes around impinging flow . . . . . . . . . . . . 5
2.2 Flow development from turbulent jet . . . . . . . . . . . . . . 6
3 Literature review of soils 10

3.1 Properties of non-cohesive soil . . . . . . . . . . . . . . . . . . 10
3.1.1 Properties of glass beads . . . . . . . . . . . . . . . . . 11
3.1.2 Properties of bentonite . . . . . . . . . . . . . . . . . . 12
3.2 Bearing capacity . . . . . . . . . . . . . . . . . . . . . . . . . 12
4 Literature review of erosion 14

4.1 Erosion mechanics . . . . . . . . . . . . . . . . . . . . . . . . 14
4.1.1 Forces on individual particles . . . . . . . . . . . . . . 14
4.1.2 Processes around bulk erosion . . . . . . . . . . . . . . 15
iii
4.1.3 Pick-up of sand . . . . . . . . . . . . . . . . . . . . . . 16
4.1.4 Sedimentation . . . . . . . . . . . . . . . . . . . . . . 17
4.2 Theories about sediment transport . . . . . . . . . . . . . . . 17
4.2.1 Van Rijn . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.2.2 Meyer-Peter and Müller . . . . . . . . . . . . . . . . . 18
4.2.3 Van Rhee . . . . . . . . . . . . . . . . . . . . . . . . . 19
5 Set-up of the experiment 21

5.1 Experimental set-up . . . . . . . . . . . . . . . . . . . . . . . 21
5.2 Execution of experiments . . . . . . . . . . . . . . . . . . . . 23
5.3 Determination of results from experiments . . . . . . . . . . . 25
6 Hypotheses 27
7 Analysis of experiments 30
7.1 Analysis particles . . . . . . . . . . . . . . . . . . . . . . . . . 30
7.1.1 Sand . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
7.1.2 Glass beads . . . . . . . . . . . . . . . . . . . . . . . . 32
7.2 Flow characteristics of jet . . . . . . . . . . . . . . . . . . . . 32
7.2.1 Impinging jet . . . . . . . . . . . . . . . . . . . . . . . 32
7.2.2 Free jet . . . . . . . . . . . . . . . . . . . . . . . . . . 34
7.3 Experimental results compared to similar tests . . . . . . . . 36
7.4 Reproducibility of the results . . . . . . . . . . . . . . . . . . 38
7.5 Processes at bed in jet erosion test . . . . . . . . . . . . . . . 39
7.5.1 Processes on bed at low erosion parameter . . . . . . . 39
7.5.2 Processes on bed at mid-range erosion parameter . . . 41
7.5.3 Processes on bed at large erosion parameter . . . . . . 46
7.6 Influence of density . . . . . . . . . . . . . . . . . . . . . . . . 50
7.7 Influence of particle shape . . . . . . . . . . . . . . . . . . . . 54
7.8 Influence of permeability . . . . . . . . . . . . . . . . . . . . . 55
7.9 Momentum as function of pick-up rate . . . . . . . . . . . . . 57
8 Conclusion and recommendations 59

8.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
8.2 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . 60
A Grain size analysis A1
B Water system design B1
C Validation system C1
iv
D Test Results D1
v
List of Figures
2.1 Impinging jet (Gauntner et al. [1970]) . . . . . . . . . . . . . 5

2.2 Theoretical shear stress . . . . . . . . . . . . . . . . . . . . . 7
2.3 Schematic flow in the cavity . . . . . . . . . . . . . . . . . . . 8
3.1 Schematization of bearing capacity (Verruijt and Baars [2007]) 13
4.1 Forces on particle at slope . . . . . . . . . . . . . . . . . . . . 19
5.1 Side view of test setting . . . . . . . . . . . . . . . . . . . . . 22

5.2 Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.3 Model for pick-up rate . . . . . . . . . . . . . . . . . . . . . . 26
6.1 Cavity with hill . . . . . . . . . . . . . . . . . . . . . . . . . . 27

6.2 Weak deflected jet . . . . . . . . . . . . . . . . . . . . . . . . 28
6.3 Bed deformation with ejection . . . . . . . . . . . . . . . . . . 28
6.4 Bed deformation with sweep . . . . . . . . . . . . . . . . . . . 29
7.1 Particle Size Distribution . . . . . . . . . . . . . . . . . . . . 31

7.2 Jet erosion tests comparison to dimensionless depth . . . . . 36
7.3 Jet erosion tests comparison to dimensionless width . . . . . 38
7.4 Jet erosion tests comparison to dimensionless hill height . . . 39
7.5 Consistency of the pump . . . . . . . . . . . . . . . . . . . . . 40
7.6 Weak defected jet . . . . . . . . . . . . . . . . . . . . . . . . 41
7.7 Test 1 in start phase . . . . . . . . . . . . . . . . . . . . . . . 42
7.8 Cavity depth development in time . . . . . . . . . . . . . . . 44
7.9 Pick-up rate from test 1 . . . . . . . . . . . . . . . . . . . . . 44
7.11 Test 3 in start phase . . . . . . . . . . . . . . . . . . . . . . . 47
7.14 Cavity in loose and bed state in dynamic stable state . . . . . 52
vi
7.15 Pick-up rate from test 5(dense bed) . . . . . . . . . . . . . . . 53
7.16 Test 6 with glass beads . . . . . . . . . . . . . . . . . . . . . . 54
7.17 Similar test with sand . . . . . . . . . . . . . . . . . . . . . . 54
7.20 Permeability in soil bed . . . . . . . . . . . . . . . . . . . . . 56
7.18 Sand soil without bentonite . . . . . . . . . . . . . . . . . . . 56
7.19 Mixture of sand and bentonite . . . . . . . . . . . . . . . . . 56
7.21 Pick-up rate comparison with bentonite mixture and sand only 57
7.22 Pick-up as function of jet momentum . . . . . . . . . . . . . . 58
A.1 Grain distribution Dorsilit 5F . . . . . . . . . . . . . . . . . . A4

A.2 Grain distribution Dorsilit 7 . . . . . . . . . . . . . . . . . . . A5
A.3 Grain distribution Dorsilit 8 . . . . . . . . . . . . . . . . . . . A8
A.4 Particle Size Distribution: small beads . . . . . . . . . . . . . A12
A.5 Particle Size Distribution: small beads . . . . . . . . . . . . . A14
C.1 Normalised pressure distribution . . . . . . . . . . . . . . . . C6
vii
List of Tables
3.1 Classification for gravel, sand and silt . . . . . . . . . . . . . 11
7.1 Sieve result of sand . . . . . . . . . . . . . . . . . . . . . . . . 31

7.2 Sieve result of glass beads . . . . . . . . . . . . . . . . . . . . 32
7.3 Stagnation pressure at z = 15mm and z = 50mm beneath
the plain jet . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
7.4 Fluctuations around averaged pressure . . . . . . . . . . . . . 33
7.5 Stagnation pressure for impinging jet at several stand off dis-
tances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
7.6 Stagnation pressure in normal directions for free jet . . . . . 35
7.7 Stagnation pressure for free jet at several stand off distances . 35
7.8 Initial deepening in dense bed compared to loose state . . . . 51
7.9 Deepening in dense bed compared to loose state . . . . . . . 51
A.1 Sieve results with Dorsilit 5F . . . . . . . . . . . . . . . . . . A3

A.2 Particle sieve analysis with Dorsilit 5F . . . . . . . . . . . . . A3
A.3 Sieve results with Dorsilit 7 . . . . . . . . . . . . . . . . . . . A4
A.4 Particle sieve analysis with Dorsilit 7 . . . . . . . . . . . . . . A5
A.5 Sieve results with Dorsilit 8 . . . . . . . . . . . . . . . . . . . A6
A.6 Particle sieve analysis with Dorsilit 8 . . . . . . . . . . . . . . A7
A.7 Test results with small glass beads . . . . . . . . . . . . . . . A11
A.8 Test results with small glass beads . . . . . . . . . . . . . . . A11
A.9 Test results with large glass beads . . . . . . . . . . . . . . . A13
A.10 Test results with large glass beads . . . . . . . . . . . . . . . A13
A.11 Sieve result of sand . . . . . . . . . . . . . . . . . . . . . . . . A14
C.1 Stagnation head beneath plain jet z = 15mm and z = 50mm C2

C.2 Stagnation head in normal directions . . . . . . . . . . . . . . C2
C.3 Vertical velocity in impinging jet . . . . . . . . . . . . . . . . C3
C.4 Vertical velocity in free jet . . . . . . . . . . . . . . . . . . . . C3
viii
C.5 Stagnation head at plain jet x = 0mm . . . . . . . . . . . . . C4
C.6 Stagnation head at plain jet x = 5mm . . . . . . . . . . . . . C4
C.7 Stagnation head at plain jet x = 10mm . . . . . . . . . . . . C5
C.8 Stagnation head at plain jet x = 20mm . . . . . . . . . . . . C5
C.9 Stagnation pressure in normal directions for free jet . . . . . C5
D.1 Observations for accuracy . . . . . . . . . . . . . . . . . . . . D2

D.2 Test 1-25 results . . . . . . . . . . . . . . . . . . . . . . . . . D3
D.3 Test 26-50 results . . . . . . . . . . . . . . . . . . . . . . . . . D4
D.4 Test 51-75 results . . . . . . . . . . . . . . . . . . . . . . . . . D5
D.5 Test 76-100 results . . . . . . . . . . . . . . . . . . . . . . . . D6
D.6 Test 101-125 results . . . . . . . . . . . . . . . . . . . . . . . D7
D.7 Test 126-150 results . . . . . . . . . . . . . . . . . . . . . . . D8
D.8 Test 151-166 results . . . . . . . . . . . . . . . . . . . . . . . D9
ix
List of Symbols
Symbol definition unit

A Area m2
cb Concentration above bed kg/m3
Cd Drag coefficient −
Cf Friction coefficient −
Cl Lift coefficient −
Cu Uniformity coefficient −
D50 Median of grain size m
d Diameter of the nozzle m
E Pick up kg/sm3
Ec Erosion parameter −
f Friction coefficient −
Fr Froude number −
g Gravitational acceleration m/s2
H Manometric head m
h Placing height m
i Hydraulic gradient −
k Permeability m/s
L Length m
m Mass kg
n Porosity −
p Pressure N/m2
p̃ Mean pressure N/m2
p' Pressure by turbulence N/m2
Q Discharge m3 /s
R Perimeter m
Re Reynolds number −
s Impinging distance m
S Source kg/sm2
x
t Time s
T Transport −
u0 Outflow velocity at the nozzle m/s
ũ Vertical averaged water velocity over time m/s
u' Vertical water velocity by turbulence m/s
Vs Volume solids m3
Vt Volume total m3
Vv Volume voids m3
ṽ Horizontal averaged water velocity over time m/s
v' Horizontal water velocity by turbulence m/s
ve Erosion rate m/s
vg Permeability at upper layer m/s
W Width m
ws Fall velocity m/s
x Cavity width (distance top to top) m
αrec Entrainment coefficient −

β Bed slope angle ◦
Cavity depth m
∆ Specific gravity −
∆ Hill height m
θ Shields parameter −
µ Kinematic viscosity m2 /s
ν Dynamic viscosity kg/ms
ρ Density kg/m3
ρi Insitu density kg/m3
ρs Density solid sand kg/m3
ρw Density water kg/m3
τ̃ Averaged shear stress over time N/m2
τ' Shear stress by turbulence N/m2
τb Bed shear stress N/m2
τc Critical shear stress N/m2
φ Internal friction ◦
% Percentage −
◦ Degree ◦
◦C Degree temperature ◦
xi
Chapter 1
Introduction
As the population keeps growing and cities are becoming more populated,
we see an increase in the number of people who move to the coastal areas for
living, work and leisure. Investing and researching in good coastal infras-
tructures are therefore important. Cables and pipelines are vital components
in this infrastructure. One of the cost efficient methods is the application
of a moving jet for the construction method. Many studies have used with
Jet Erosion Tests (JETs) to gain a better understanding of the processes
that are going on beneath the nozzle. These studies give insights in how
the removal of the sand bed makes it possible to build infrastructures more
efficiently.
Pipelines can either be build on or in the bed to connect both sides of the
sea. Salt, vessels and organisms can harm the infrastructure in the sea dur-
ing its lifetime. To prolong the lifetime of the infrastructure, local conditions
are taken into account when building the infrastructure. A dredging vessel
can have a remote high-pressure water jet, which displaces the grains else-
where via bed or suspension load transport. Cables are put in the excavated
bed, then the cavity is covered back with soil to protect the infrastructure
against the hostile environment.
Erosion theories are mainly based on shear stresses above soil bed. A
lot is known about jet outflow and soil, but little is known about processes
around stagnation point. Shear stresses refer to the forces on the sand
particles by parallel flow in waterways. Beneath the vertical jet has no
shear stresses with zero flow velocity in ideal impinging condition, because
all kinetic energy will be converted to pressure at a certain distance from
1
the impinging jet. This would mean that no pick-up would occur in the bed
at the stagnation point. In practice sediment transport does occur in the
alluvial bed. Hence, this study focuses on the processes that erode the sand
bed at the stagnation point.
1.1 Objective
This master thesis describes the physical processes around the stagnation
point with a Jet Erosion Test (JET) in various test settings. Different water
flow velocity, height/diameter ratio, diameter particles, soil bed density and
permeability are analysed at the dredging laboratory in Delft University of
Technology at the faculty Mechanical, Maritime and Materials Engineer-
ing (3ME). This leads to the following subquestions to achieve the main
objective of this research.
What is known in the literature?
What happens at the stagnation point in jet erosion test?
How does the cavity develop over time using a stationary plain jet in
several sand bed conditions?
What are the sand bed failure mechanisms under different test set-
tings?
1.2 Outline of thesis

This report proceeds as follows.
Part one is the literature study, where various JET processes and models
are described. Findings about grain forces and sediment transport are also
included.
Part two describes the research method. The test plan is outlined for the
JET along with the various parameters. Considerations for the equipments
were dependent on the limits of the flume.
In part three, the analysis is presented. This includes a description of

the experiment preparation and how the equipment was used and validated
in the laboratory. In addition, the experiment was run and analysed. Some
2
test runs are discussed in this part too.
The final part consists of the conclusion and recommendations.
3
Chapter 2
Literature review of f low

processes
In this chapter, several flow processes are described that have emerged after
long vertical submerged impinging water jet. Starting with two extreme
cases are the free jet and impinging jet, where the first mentioned has no
blockage and the latter has a full impervious plate beneath the jet.
2.1 Velocity profiles

2.1.1 Behaviour of free jet
In the free jet condition, water flow exits unhindered from the nozzle. A
mixing layer is then formed because of velocity differences in the viscous
fluid after leaving the nozzle. Surrounded fluid is carried along with the jet,
this is called entrainment.
The vertical jet outflow (u0 ) exercises in the full width and length from
the smooth contraction nozzle. The mixing layer becomes larger in width
along streamwise direction from the turbulent jet, while the original vertical
velocity becomes smaller in propagating direction according to Rajaratnam
[1976]. This wedge-shape region is called the potential core with the vertical
velocity u0 in undeveloped regime. The length of the region is called stand
of distance (SOD). This term can made dimensionless stand of distance H/d
by original impinging distance divided to the nozzle width. Distances larger
than the potential core are called fully developed flow regime, where only
a mixing layer is present. The velocity difference induces turbulence over a
4
small distance in viscous water.
Affected area to the soil layer becomes larger further in the propagating
direction. The jet momentum spreads over a larger horizontal area. The
momentum per unit of area decreases along the propagating jet distances
for free jet.
2.1.2 Processes around impinging f low
Figure 2.1: Impinging jet (Gauntner et al. [1970])
An impinging flow deflectes from a vertical to horizontal direction as

it hits the impermeable wall. Vertical fluid propagates unhindered perpen-
dicular towards the horizontal impermeable wall. Fluid bends into parallel
directions to the wall with conservation of volume. Vertical kinematic en-
ergy converts to pressure head as it reaches the bottom as can be explained
5
through Bernouilli equation 2.1. The pressure difference with the ambient
water bends the flow in lateral directions. In addiction, the volume balance
forces the incompressible water from incoming vertical to sideways direc-
tions. A perfect impinging jet is shown without entrainment in Figure 2.1.
P u2
z+ + (2.1)
ρg 2g
The energy losses is assumed to be zero for the relatively small impinging
distances. Another assumption is that all kinematic head is converted to
pressure head at the bed with incompressible fluid. The stagnation pressure
beneath the bed can be written as:
1
P = ρgh + ρu2 (2.2)
2
Wall jet is the parallel flow above the impervious wall. Velocity-in-
depth profile becomes a square root form perpendicular to the wall with
zero velocity (rm ax) at the bed bottom. Shear stress is determined by the
depth-averaged velocity between the eye and instant bed geometry. The
roughness of the wall effects the velocity profiles in the water column. If
shear stress exceeds the resistance, particles will erode until the bed is in
equilibrium.
2.2 Flow development from turbulent jet

Water pumps vary in discharge in seconds for several reasons because of
turbulence. de Vriend et al. express velocity using two terms in analytical
terms to the instant actual value as in equation 2.3. The first term (ũ) is
the mean value over time and thesecond term (u') is the instant fluctuating
component. These combined are shown for flow velocity as following:
u = ũ + u' (2.3)
Shear stress causes by water is defined in equation 2.4 as:
τ = 0.5cf ρw u2 (2.4)
Shear stress is a function of the average shear velocity above the sand
bed and the vortex eye, which has the shape as shown in Figure 2.2 for an
impinging jet at the original bed level.
6
Figure 2.2: Theoretical shear stress
Entrainment increases the jet discharge along the viscous water flow.
Ambient water is sucked into the propagating direction. The discharge in-
creases in the mixing layer. The volume transport increases in the propa-
gating flow as: s
Hnozzle
us = f1 u0 (2.5)
Himp
s
δQ Hnozzle
= αent Rus = αent 2Lf1 u0 (2.6)
δs Himp
Alberson et al. [1950] derived the entrainment coefficient with value of
αrec = 2f1 2 = 0.091 for a plain impinging jet. f1 is an empirical constant
1
between 2.35 and 2.45 as stated by Fischer et al. [2013]
where Hnozzle is the nozzle diameter. s is the distance from the jet until
the sand bed. R is the length of the hypotenuse.
The depth averaged velocity is retrieved as follows. The discharge grows
in propagating water because of entrainment. Then, the volume splits on
two sides of the plain jet, where in assumption half of the discharge flows
to one side of the centreline. The half discharge is divided by the flow area
between the eye of the vortex and the insitu bed geometry.
7
Figure 2.3: Schematic flow in the cavity
1 Qimp
um = (2.7)
2 Bhvortex
The motion of an unit fluid can be expressed by the momentum equation
Battjes [2002].
∂u ∂u ∂u ∂u 1 ∂P ∂2u ∂2u ∂2u

+u +v +w = + ν( 2 + 2 + 2 ) (2.8)
∂t ∂x ∂y ∂z ρ ∂x ∂x ∂y ∂z
The first term in equation 2.8 is not relevant since the discharge is as-
sumed to be constant without any source of fluid. The second term cannot
be neglected, since the streamline changes direction from the longitudinal
to the lateral directions. The third term means that the stagnation pressure
is different at inconsistent impinging distances. The last term is the turbu-
lence accounting for energy losses during the flow.
Volume balance can be described as:

∆Q ∆u
+A =0 (2.9)
∆t ∆x
Turbulence is present in any jet in viscous fluid. Propagating water
exercises some instabilities after leaving the jet in the mixing layer because
of velocity differences over short distances, which change the terms in the
momentum. The stresses in non-uniform flow is influenced by the following
8
parameters: fluid density, sediment density, kinematic viscosity, grain size
and bed shear stress. These factors cause turbulence to the water stream
and result into shear stress. The interaction makes the water stream and
the bed as vortex slightly variable. Reynolds numbers are used to describe
the quality of turbulence. The equation is the ratio of inertia force divided
by the viscous force:
inertiaf orce U 0D
Re = = (2.10)
viscousshearf orce ν
A method to describe the varying turbulence flow is the Reynolds Av-
eraged Navier Stokes(RANS). The vertical outflow velocity shows some
fluctuation beneath the je. Any fluctuations (2.3) are left out during test
to make it easier to approximate the hydraulic turbulence.
9
Chapter 3
Literature review of soils
In this chapter, theories are investigated for non-cohesive soil as grain prop-
erties and pick-up. This report will mainly cover sand soil, including small
section about properties of glass beads and bentonite.
3.1 Properties of non-cohesive soil

Nature is capricious as every particle is unique due to differences in diame-
ter, shape and chemistry content. Environmental factors impact saturated
rocks high in the mountain through internal tensions caused by temperature
differences. Further downstream, rocks abrade in a propagating river to the
size of a sand grain in the ocean. Sand is a non-cohesive soil, which means
that the particles do not stick together.
Solid sand particles have void spaces in an unit of soil. Voids leads to
permeability of the sand bed so that fluid can be passed through in rela-
tive low velocity, if pressure gradient is present. A saturated sand bed and
a sudden large load can cause dilatations or compaction in incompressible
fluid. Soil properties changes temporarily with failure consequences on the
bed.
Sand is a common soil type that occurs in nature, it comes from the
wearing of larger rocks from mountains upstream that ends up int he river.
A particle with a diameter between 63µm and 2mm is classified as sand, the
sand size is between that of thesilt and gravel soil type as can be seen in Table
3.1 from Verruijt and Baars [2007]. The chemical compositions comprise
from siliciumdioxide and quartz. The interaction between particles depends
10
on the dimension, shape of the particles, density, particles size distribution,
chemical composition and porosity.
Type Subclass Diameter (mm)

Gravel Medium gravel 8.0-16.0
Gravel Fine gravel 4.0-8.0
Gravel Very fine gravel 2.0-4.0
Sand Coarse sand 0.63-2.0
Sand Medium sand 0.2-0.63
Sand Fine sand 0.063-0.2
Silt Coarse silt 0.02-0.063
Silt Medium silt 0.0063-0.02
Silt Fine silt 0.002-0.0063
Table 3.1: Classification for gravel, sand and silt
Particle sieving analysis is a useful method to examine sand soil. Prop-

erties like the median grain diameter D50 can be determined from a repre-
sentative sand sample. Additional analysis is possible to get the uniformity
coefficient, the internal friction angle and fall velocity.
A qualitative description to a unit soil volume is porosity, which is defined
as the ratio of voids volume to the total soil volume.
Vv
n= (3.1)
Vt
3.1.1 Properties of glass beads

Glass beads are made from the manufacture Sigmund Lindner GmbH in
Germany. Beads are generally used for abrasive blasting in industrial sec-
tor. The transparent beads smooths mechanically the object surface from
any dust and contamination to a clean material. The surface of object will
strengthen and smoothen after blasting the material. The beads are made
from silicium, which hardly reacts to other elements.
The beads are made in a controlled manufacture. The particles size

distribution is generally very bad graded. Besides, the roundness would
classified as spherical. Consequently, the beads result to an homogeneous
bed state.
11
3.1.2 Properties of bentonite
Bentonite reduces interaction between the voids of sand. Permeability is
reduced by the swollen bentonite with the characteristics of sand soil. Ben-
tonite is characterized as small particles with a diameter of 0.2 − 2µm. This
material is considered as impermeable soil in wet state. A mixture of sand
and bentonite takes time to collaborate for being a impervious soil bed.
Voids are barely connected to each other, which makes permeability almost
impossible.
3.2 Bearing capacity

Hypothese 3 is the soil movements by insufficient bearing capacity. The
book of Verruijt and Baars [2007] describes the vertical downward pressure
as Prandtl’s zones. The movements are schemetised in Figure 3.1 and can
be divided in three zones. The vertical hydraulic stresses are represented by
the vertical and horizontal forces in zone III, displayed as a triangle shape.
Excess stress in that wedge pushes the soil upwards in region I to above
the original bed level. This is caused by insufficient resistance of region I.
Zone II is the transition between zone I and III. Shear stress is at maxi-
mum between those zones. The bearing capacity can be determined by the
theory of Prandtl. Brinch Hansen extended the formula with the factors in-
clination, shape, cohesion and friction angle for a two-dimensional soil bed.
This theory has the following assumptions: weightless material, homoge-
neous bearing material, no volume change, uniform distributed loading and
plastic behaviour during deformation. The acting force on the soil can be
inclined. p is the vertical vector and t is the horizontal component in Figure
3.1.
Bearing capacity according to Prandtl of soil can be expressed using the

following formula.
1
p = ic sc cNc + iq sq qNq + iγ sγ γBNγ (3.2)
2
If cohesion in the bed is equal to zero, the first term equal to zero. The
second term is also equal to zero. A flat sand level has no surcharge around
the bearing area, which reduces the resistance of a burst. Only, the third
term is left in the equation. The vertical downward hydraulic momentum
12
Figure 3.1: Schematization of bearing capacity (Verruijt and Baars [2007])
acts as a strip load on the sand bed, which might lead to potential soil
failure.
Prandtl includes the following factors for the bearing capacity of soil:
1
Nc = (Nq − 1) (3.3)
tanφ
1 + sinφ πtanφ
Nq = e (3.4)
1 − sinφ
Nγ = 2(Nq − 1)tanφ (3.5)
Shape factors B ≤ L
B
Sc = 1 + 0.2 (3.6)
L
B
Sq = 1 + sinφ (3.7)
L
B
Sγ = 1 − 0.3 (3.8)
L
Inclination of forces are determined with these factors:
t
ic = 1 − (3.9)
c + ptanφ
iq = i2c (3.10)
iγ = i3c (3.11)
13
Chapter 4
Literature review of erosion
4.1 Erosion mechanics

Particles are moved over distances by a driven force as individual grain or
as bulk volume. Particles are picked up individually when there subject to
lower propagating hydraulic forces. Individual grains are exposed to the hy-
draulic forces and experience resistance on the sand bed. Larger momentum
leads to erosion in bulk or soil deformation. Other distinction is the erosion
in vertical or horizontal direction.
Several types of force are introduced to an individual grain from reader

Molenaar and Voorendt [2012] first. Then, bulk transport is described in
the second part.
4.1.1 Forces on individual particles

Drag is the horizontal force acting on a grain through viscous fluid in the
same direction. The force is a function of the drag coefficient, density, flow
velocity and the exposed area of the grain.
1
FD = CD ρu2 A (4.1)
2
Lift is the force in thevertical upward direction caused by horizontal flow.
The velocity difference and the grain shape lead to pressure gradient on a
single particle between the upper and lower surface, which could lift up the
particle.
1
FL = CL ρu2 A (4.2)
2
14
The weight of a single particle excites vertical downward force because
of gravity.
Fg = mg = ρV g (4.3)
Force gradient occurs if pressure gradient is present on particles in the top
layer. Dilatation or compaction can result in pressure differences between
the upper and lower surface of an individual particle for a short time under
those circumstances.
F1 = F0 + ∆F (4.4)
Vertical incipient motion starts when the combination of driven forces is
larger than the resistance forces on a single particle. The following equation
defines incipient motion.
FD + FL < Fg (4.5)
Soil erodes particles at the micro level, when loading forces is just larger
than the resistance. The pick-up rate is very small in general. The transport
modus of scoured particles is bed load or suspended load for short distances.
Very loose soil beds have a relatively large void space per unit of volume.
Solid particles can settle in the little voids because of compaction with ex-
ternal hydraulic force beneath the jet. If the sand soil is packed, particles
settle less likely by leak of void space. Solid grains have no space to pack
the soil bed, instead it will pick-up or push as active soil beneath the jet.
Since every grain has a different position, the characteristics for individ-
ual particle differ too in terms of:
Position compared to other particles

Weight of a single particle
Exposed area of grain in fluid
Chemical content of a particle
4.1.2 Processes around bulk erosion

This section describes the pick-up rate of bulk volume from a larger jet
momentum. A layer of sand are transported closed to the bed in a high
concentration.
Bisschop [2018] distinguishes two regime in horizontal erosion, namely

saltation and bulk erosion. Saltation entrained a few grains on the soil sur-
face into the flow. This process happens with the Shields parameter just
15
above the critical value. Bulk erosion moves the toplayer of grains as a
bed load in horizontal directions. The grains are displaced at large Shields
parameter.
Pick-up can be calculated by dividing the eroded bed volume by the

bed density. The pace in height is called the erosion velocity. The insitu soil
is a mixture of solids and void. Pick-up rate is calculated by dividing density.
E
ve = (4.6)
ρs (1 − n0 )
4.1.3 Pick-up of sand

Shields parameter is used to calculate the incipient motion of grains on a flat
bed with parallel flow. The nondimensional parameter is an empirical func-
tion of the shear force and the submerged weight of a single particle. Pick-up
takes place, when the critical Shields parameter exceeds the threshold value
according to the book by Schiereck [2003]:
τb
θ= (4.7)
(ρs − ρw )gD50
Rajaratnam [1981] adapted the Shields parameter and translated it to an

erosion parameter by adding the initial impinging distance SOD-term, then
dividing it by the square root. Rajaratnam defined the erosion parameter
for plain jet as:
U
Ec = p √ (4.8)
H/d ∆gD50
The erosion parameter is a good indicator for the final cavity dimension
of a JET. A linear correlation has been made for the erosion parameter as
function of dimensionless cavity depth, width or hill height.
The bed state can be distinguished in two geometries: weak deflected

jet and strong deflected jet. The first mentioned has a smooth cavity shape
from hill to hill. The deflected jet is attached to the bed along the entire
erosion zone. The grains are transported radially outward as bedload. The
strong deflected jet has bed deformation inside the outer region. The inner
cavity has a steep slope until the flow separation.
16
Turbulence has flows with different velocities and directions. The insta-
bility causes turbulent behaviour such as ejections and sweeps at the sand
bottom Nezu et al. [1993]. Ejection is characterised by a local vertical and
horizontal velocity lower than the depth-averaged flow velocity. Sweep has
a larger local horizontal velocity than the depth-averaged flow velocity.
4.1.4 Sedimentation
Sedimentation settles the particles in suspended load on the sand bed. In-
crease in bed height can be seen as a reverse of erosion. Causation is gen-
erally the leak of a driven force to the particle, which settles by gravity.
High sand concentration hinders temporarily settlement of suspended grains,
which increases the sedimentation time. Fluid can not escape through the
voids with the decreasing space.
This thesis mainly focuses on erosion. Any sedimentation during the

tests were not included in the pick-up and transport rate. A simple sedi-
ment equation is stated as:
S = cb ws (4.9)
Settling velocity is defined as:
r
4g∆D50
ws = (4.10)
3Cd
4.2 Theories about sediment transport

Well-known erosion theories base particle transport on the horizontal stress
load, see for example the theories by Van Rijn, Meyer-Peter Müller and Van
Rhee. Empirical experiments are mainly based on shear stress and particle
properties at the bed. Particle transport with an increasing or decreasing
bed height could be regarded as pick-up in vertical direction.
4.2.1 Van Rijn

Many experiments were done with pick-ups of single particles in the range
between 0.13 and 1.5mm, according to Rijn [1984]. Those tests were with
shear flow velocity up to 1m/s. Particles start when hydraulic force ex-
ceeds the resistance on a grain. The empirical transport and dimensionless
diameter is defined as:
17
τb − τb,cr
T = (4.11)
τb,cr
The transport function could also be written as:
θ − θcr
T = (4.12)
θcr
Dimensionless particle diameter is defined as a function of:
∆g 1/3
D∗ = D50 ( ) (4.13)
v2
Combining equations 4.11 and 4.13 result in the empirical pick-up rate:
p
E = 0.00033ρs ∆gD50 D∗0.3 T 1.5 (4.14)
This function is described as the mass of eroding soil per unit of area and
time Rhee [2015]. This equation is a good approximation for the pick-up
rate caused by the hydraulic shear flow above the sand bed.
The erosion velocity scoured in depth, is the distance per unit of time.
The pick-up is divided by the insitu bed density to get the erosion velocity
in bed, which is defined as:
E−S
Ve = (4.15)
ρs (1 − n)
4.2.2 Meyer-Peter and Müller

Experiments with sand bed-load transport have been conducted with parti-
cles larger than 0.4mm. This theory defines the incipient motion differently
from the previous model. The mobility parameter is as follows de Vriend
et al.:
Φ = 8(θ − θc )1.5 (4.16)
With dimensionless mobility number, bed load transport can be deter-
mined with:
E
Φ= q (4.17)
( ρsρ−ρ
w
w
)gD 3
50
The pick-up function is then:

r
1.5 ρs − ρw 3
E = 8(θ − θc ) ( )gD50 (4.18)
ρw
18
The main difference is the dimensionless transport parameter compared
to the theory by Van Rijn. The first part of the term is to the power of 1.5,
but without dividing to θcr . Critical theta is smaller than the value one,
which lowers the pick-up rate on the one hand, but multiplies with factor 8
on the other hand.
4.2.3 Van Rhee

This theory deals better with the high velocity shear flow as a consequence
of pore volume increase in a short time (Rhee [2015]). The theory extends
the theory of Van Rijn with a correction for high flow. A sudden shear
movement in the sand bed would trigger temporal dilatation. Pore pressure
decreases suddenly in the soil layer, which results in hydraulic gradient on
the bed. Incompressible fluid cannot fill up the increased pore volume due
to low permeability. Water pressure hinders erosion in the bed for a short
time.
Another relevant mechanism is the stability of a particle on a slope. The
actual maximum slope angle β could be larger than the internal friction
angle φ, if an external force is present against active sliding soil. Shear
stress caused by water leads to drag force Fs and hydraulic gradient Fi in
soil, which takes into account the high water shear flow. The submerged
weight of sediment is defined as Fg . The forces are shown on a single particle
at a slope:
Figure 4.1: Forces on particle at slope
The force of a single particle is critical, when the internal friction force
is fully used. Incipient motion of a single particle starts, when the angle
19
between the normal force and resulting force exceed φ angle. Critical force
balance is defined as:
Fs + Fg sinβ = (Fg cosβ + Fi )tanφ (4.19)

Rhee and Bezuijen [1992] approaches the stability with two different
methods: Single particle and Continuum.
Single particle
Fg = 1/6πD50 3 (ρs − ρw )g (4.20)
Fi = 1/8πD50 3 ρw gi (4.21)
Continuum
Fg = (1 − n0 )(ρs − ρw )g (4.22)
Fi = ρi gi (4.23)
Incipient motion starts on a flat surface, if actual Shields parameter

exceeds the adapted critical parameter:
θ > θcr,rhee (4.24)
Pick-up rate E is almost the same as Van Rijn with an adapted Shield
parameter, but including permeability. Erosion rate can be defined as:
E−S
ve = (4.25)
ρs (1 − n0 − cb )
sin(φ − β) ve ni − n0 A
θcr,rhee = θcr ( + ) (4.26)
sinφ k 1 − ni ∆
p
E = φ g∆D50 (4.27)
φ = 0.00033D∗0.3 T 1.5 (4.28)
θ − θcr,Rhee 1.5 p
E = 0.00033D∗0.3 ( ) ∆gD50 (4.29)
θcr,Rhee
Van Rhee includes permeability in the adapted Shields parameter, which
is important at high flow velocity. The pick-up rate decreases due to per-
meability term in equation 4.26. A gradient in top soil layer is defined as:
vg ni − n0
i= (4.30)
k 1 − ni
20
Chapter 5
Set-up of the experiment
The design and the set-up of the jet erosion test are described in this chapter.
The first section shows the required equipments of the experiments. The
execution is described in the second section for several test series. The main
experiments took place in the acrylic flume at the Dredging laboratory of
faculty Mechanical, Maritime and Materials Engineering in Delft University
of Technology.
5.1 Experimental set-up

All experiments were executed in the front compartment in the fully width
of the flume. An extra divider with a height of 0.7m was built to reduce
lateral flow around the jet. The water level was in all compartments at
0.9m. Figure 5.1 shows the schematic side view of the experiment including
the water system. The flume is located in a controlled environment with
a temperature around 20◦ C, the water temperature was also around 20◦ C.
The directions are defined as xyz-directions. X is longitudinal, y is lateral
and z is normal direction.
These following items were used to execute the experiment:

Pump
Splitter G1 41
Water discharge meter (Flowtec Variomag)
Valves G1 14
Tube G1 14
Jet
Hose
21
Impinging table
Pressure gauges; Rosemount 1151 Smart
Sand
– Dorsilit 5F
– Dorsilit 7
– Dorsilit 8
Glass beads
– Small
– large
Camera
– GoPro Silver Hero 3+
– Fastcam APX RS Photron
Figure 5.1: Side view of test setting
The flume has the dimensions 1.5x0.75x1.0mm (LxWxH) with a 0.7m

divider in the front compartment. The walls and the divider are made from
transparent acrylic of 10mm thickness. A frame holds the stainless tube and
the jet in place. The nozzle has a two-quarter funnel shape with a 2.5mm
22
Figure 5.2: Directions
outflow width. The length of the jet is a compromise between minimise sed-
iment placements on the drag head and the hydraulic interaction on both
side of the jet. Valves can be adjusted beforehand for the right outflow ve-
locity, a constant momentum from the nozzle. Jet height could be adjusted
to meet the test requirements via two bolts in the rigid overhanging frame.
Technical drawings for jet and the frame can be seen in B along with more
details about the set-up. Soil is filled to 0.2m height in all test cases and
spread evenly over a homogeneous bed.
Two pressure gauges (Rosemount) measured in the range between 0.0

to 7.0kP a beneath the jet for the flow characteristics. The computer can
show instantly the vertical local pressure beneath the jet. Two hoses are
connected with the gauges for validating the absent of lateral flow at 0.2m
height. The first gauge is located right beneath the nozzle at the same
level as the bed at the half length of the plain jet (50mm) from the front
wall. The second gauge is placed beneath the jet, 15mm behind the front
wall. Before executing the experiment, gauge hoses are filled with fluid, to
make sure that no air is trapped in the measuring system for reliable results.
5.2 Execution of experiments

This research examines what happens for different variables inside the ero-
sion parameter. The variables are outflow velocity, stand off distance (ab-
breviation SOD), grain size and density of soil. Each run has to be checked
first on the following points: valves, frame, water jet, the sand bed and the
camera. After going through the checklist, the test could be executed until
the cavity has reached the dynamic equilibrium.
23
The first series of tests were performed with a fully impermeable imping-
ing jet and free jet, which differ in flow characteristics. Two gauges were
mounted to measure the pressure beneath the jet. Before the tests were
run, outflow velocity, SOD and equipments were checked. Extra attention
was paid whether the gauges were placed correctly beneath the pain jet at
15mm and 50mm behind the front wall.
The second set of experiments was conducted with various velocities on

permeable loose sand bed. The flat homogeneous sand bed was carefully
prepared for each run. The sand mass was measured before each run to
know the density of the bed. The soil bed was made homogeneous and
flat before every run. One test setting were executed three times for more
reliable comparisons. The valve on the jet was essential to control for the
outflow velocity.
The third series of tests was executed with different SODs. The imping-
ing height could be adjusted via the two bolts on the overhanging rigid frame.
The fourth tests sequence were tests with three different particle diam-
eters, namely D50 0.63, 0.93 and 1.41mm. The sand volumes were then
further analysed for more detailed information. This analysis can be found
in appendix A. The front compartment was cleaned every time after chang-
ing for a different sand bed.
The fifth set was tested on dens soil bed. The loose soil state was tested
first before a dens test was executed. A long vibrator was used to pack the
soil bed. Extra grains were added in the specified soil volume, which leads
to a denser state. After all the preparations, it was assumed that there was
a homogeneous dens bed of the soil.
The sixth series were executed with perfect spherical glass beads under
several conditions. Beads were sieved in StevinLab II at the faculty Civil
Engineering in order to measure the diameter and distributions. More infor-
mation about this can be found in Appendix A. A black curtain was used
in the background to brighten it up and increase contrast from the beads.
The last series was executed with bentonite. Bentonite was added with
a 6.0% volume percentage to the non-cohesive sand soil, which is equal to
3.6% mass percent of sand weight. Permeability decreases significantly with
24
this bed mixture, according to Foortse [2016]. The well mixed sand and
bentonite was measured in dry state and put in the front compartment.
Water was carefully added, so that the bentonite would not move through
the voids. This set-up had to rest at least 24 hours to reduce the permeabil-
ity in sand.
5.3 Determination of results from experiments

Important aspect is the pick-up rate as function of shear stress. Calculation
for the shear stress is described in this section. The mean discharge from
the jet is measured with the electromagnetic flow meter. The discharge will
increase in the distance between the nozzle and the sand bed by entrain-
ment. (see eq. 2.5) The discharge at the bed is assumed to split half to one
side, the other half to another side. The mean shear velocity is determined
as followed, the half discharge is divided by the flow area between the eye
of the vortex and the insitu bed geometry (see eq. 2.7). A schematic flow
pattern can be seen in Figure 2.3. The bed shear stress is, then, calculated
with equation 2.4 with a constant cf = 0.004.
s
Hnozzle
us = f1 u0 (5.1)
Himp
Qs = αent 2sus (5.2)
1 Qs
um = (5.3)
2 BHvor
τ = 0.5cf ρw u2m (5.4)

This research deals only with pick-up without any sedimentation. The
scoured cavity is in the form as two triangle. The pick-up rate is the width
times the depth for a consecutive time step. This method is done till re-
circulated suspended grains start settling in the cavity. Sedimentation can,
therefore, set to the value zero. Then, the pick-up rate is calculated by
rewriting equation 4.6, which will be analysed in the next chapter.
E = ve ρs (1 − n) (5.5)
25
Figure 5.3: Model for pick-up rate
26
Chapter 6
Hypotheses
I expect particle movements beneath the jet by because of one of the follow-
ing reasons:
1. Small hill inside the cavity will be formed, pick-up occurs only in
presence of shear stress. The shear stress entrained the particles up to
form a cavity around the centerline as in Figure 6.1. Zero shear stress
averaged over time would not pick-up any particles.
Figure 6.1: Cavity with hill
2. Pick-up is caused by shear stress on the bed surface. A smooth cavity

is formed in the soil as in Figure 6.2. Turbulence part of velocity
erodes the hill beneath the jet.
27
Figure 6.2: Weak deflected jet
3. Vertical flow pushes the soil bed and dragged the particles as ejection
away. Sediments going downwards would be observed beneath the jet
as Prandtl. Surrounded soil can not resist the bed deformation. Fluid
follows the cavity geometry. The eroded particle are dragged along
the bed. Dynamic bed will be as strong deflected jet.
Figure 6.3: Bed deformation with ejection
4. The particles are pushed away as a ejection in combination with sweeps.

Grains are deformed in the bed at the beginning of the experiment.
Surrounding soil has insufficient weight to resist any bed deformation.
The grain movement will translate and rotate away from the center-
line. The final bed state takes as strong deflected jet.
28
Figure 6.4: Bed deformation with sweep
29
Chapter 7
Analysis of experiments
This chapter is divided in three parts to answer the research objectives.

The first part presents the particle sieving analysis. The grain properties
are derived from Dorsilit. The second part presents the comparison with
plain impinging jet from the study of Rajaratnam [1981]. The third part
presents the observations from the processes for low and large erosion pa-
rameter values beneath the JETs. Videos and results were analysed in the
start, asymptotic and end phase. Pick-up rates are considered for some test
cases. The fourth part present the experiments with glass beads and ben-
tonite. Pick-up processes are compared with Dorsilit sand soil.
7.1 Analysis particles

7.1.1 Sand
Dorsilit 5F, 7 and 8 were used in this experiment, which was distributed
in 2015 by Sibelco in Dordrecht. These sand packages were tested in dry
conditions with a sieve machine at the faculty of Civil Engineering at Uni-
versity of Delft. The sieve experiment results of the three used grains are
shown in the Particle Size Diagram shown in Figure 7.1. The details of the
sieve tests are attached in Appendix A along with the manufacture sheets.
Table 7.1 shows the results of the three sand packages with results of the
D50 , density, porosity, internal friction and uniformity coefficient.
Each sand package was tested three times before the start of the exper-
iment to ensure results as described in Appendix A. The content exists of
non-cohesive sand from more than 98% of siliciumdioxide. These three sand
30
packages were of a grey-ish color mixed with some white particles. The den-
sity in the dry state condition was for these three sand types 2630kg/m3 .
The D50 , was obtained by linear interpolating.
Figure 7.1: Particle Size Distribution
Type sand Diameter Density Porosity Internal Uniformity

[mm] [kg/m3 ] [-] friction angle [◦ ] coefficient [−]
Dorsilit 5F 1.41 1971.3 0.38 33 1.42
Dorsilit 7 0.93 1903.5 0.35 33 1.34
Dorsilit 8 0.63 1938.2 0.43 31 1.52
Table 7.1: Sieve result of sand
D50 has a little variation within the 10% margin from the prescript. The
deviation is so small that the D50 in the Table A.11 will be used for further
calculations in this study. The internal friction angle for sands was between
the 31◦ and 33◦ , therefore particles can be classified as sub roundness.
Each sand package have an uniformity coefficient of 1.34 and 1.52. Sand
is usually sorted at the manufacturer, since grains occurring in nature are
mostly badly graded. Uniformity coefficients are valued near to one, which
means that there is a higher permeability of sand soil.
31
7.1.2 Glass beads
The beads were used for this research as an artificial sand bed. The shape
of each bead was assumed to be perfectly spherical and smoothly shaped.
The manufacturer can produce beads with small error due to the reliable
machines in controlled area.
Type Glass beads Diameter [mm] Internal friction [◦ ]

Small 1.13 14
Large 1.45 13
Table 7.2: Sieve result of glass beads
Small and large glass beads have respectively a D50 of 1.13 and 1.45mm
as shown in appendix A. The research equipment was insufficient to provide
some accurate results. A large part of the beads got stuck in the two sieves,
which analysing the diameter size difficult. The diameter was, therefore,
assumed to be D50 = 1.13mm and D50 = 1.45mm as derived from the pro-
duction sheets. The internal friction angels are respectively 14◦ and 13◦ .
The solid density was 2500kg/m3 in dry state.
7.2 Flow characteristics of jet

Normal flows were tested with the stationary plain jet in impinging and free
state before the main experiments were conducted with sand soil. The two
gauges measurements would not vary too much, if a 3D effect is not present
during the tests.
Two pressure gauges measured the flow field beneath the jet. Each run
had a duration of one minute for impinging and free jet. During that time,
the averaged and turbulent part has been recorded for the stagnation pres-
sure.
7.2.1 Impinging jet

The detailed results of the stagnation pressure are shown in Table C.2 for
impinging jet with its main and fluctuations values. Sensors were placed at
32
15 and 50mm length from the front panel at the impermeable flat surface
for measuring various velocities and H/d distances. The pressure gauges
were mounted in a flat smooth plate at the bed level, which were connected
to a computer for measuring instant local stagnation pressure. Results for
SOD = 10 are shown in table 7.3 along with the averaged and fluctuation
values for several outflow velocities.
Velocity Front Middle Front Middle Difference

[m/s] gauge gauge fluctuation fluctuation [%]
[kP a] [kP a] [kP a] [kP a]
1.2 0.62 0.64 0.10 0.09 3.2
1.6 1.17 1.21 0.17 0.22 3.4
2.0 2.12 2.16 0.27 0.34 1.9
2.4 2.58 2.65 0.29 0.43 2.7
Table 7.3: Stagnation pressure at z = 15mm and z = 50mm beneath the

plain jet
The stagnation pressures from the middle gauge had larger values than
the front gauge for all tests. The values had a maximum difference of 3.4%.
A net momentum towards the front panel can be concluded, possibly be-
cause of small seepage between the impinging table and the front panel that
caused pressure differences. Also, wall friction could affect the hydraulic
propagation near the wall. But, the small normal flow can be neglected as
the deviation percentage is small. The normal flow is neglected in all further
tests since the difference between the middle and front gauges is small.
Velocity Front Middle Front Middle Difference Difference

[m/s] gauge gauge fluctuation fluctuation [%] [%]
1.2 0.62 0.64 0.10 0.09 16 14
1.6 1.17 1.21 0.17 0.22 15 18
2.0 2.12 2.16 0.27 0.34 13 16
2.4 2.58 2.65 0.29 0.43 11 16
Table 7.4: Fluctuations around averaged pressure
33
The fluctuation term of velocity (u') decreased at a larger outflow veloc-
ity in a fully developed regime, yet the middle gauge shows more turbulence
in absolute value. The turbulence is at least 11% of the average stagnation
pressure. u˜0 and u' is u0 as stated in Chapter 2. Averaged plus peeks could
trigger incipient motion of a single particle, if resistance is smaller than the
hydraulic force.
U0 = 2.0m/s Stagnation Fluctuation Difference

at H/d= pressure pressure [%]
[−] [kP a] [kP a]
10 2.16 0.34 16
20 1.22 0.31 25
30 0.86 0.12 14
Table 7.5: Stagnation pressure for impinging jet at several stand off distances
Three tests were done at SOD 10, 20 and 30 for one certain momentum
outflow. Stagnation pressure correlates inversely with SOD, which is in line
with the erosion parameter (equation 4.8). A larger SOD leads to smaller
averaged stagnation pressure and fluctuations.
7.2.2 Free jet

The experiments were executed for the flow characteristics with free jet.
Setting was almost similar as the impinging water jet, except that the im-
pervious surface was removed. Water was free to propagates towards the
bottom of the flume. Free jet can be seen as a 100% permeable bed with-
out any blockage, although the small frontal area of gauge can have had a
negative effect to the measured stagnation pressure.
Free jet has a vertical pressure difference beneath the plain jet. Normal
flow is likely with larger averaged stagnation pressure in the middle gauge.
This means a net force is present towards the front panel as in the impinging
jet condition.
Fluctuation u' was also present in free jet in all outflow velocities dur-
ing experiments. The instant stagnation pressure varied around 15% of the
averaged value. The turbulence term (u') contributes the peeks for a short
34
Velocity Front Middle Front Middle Percentage Percentage
[m/s] gauge gauge fluctuation fluctuation Front-middle fluctuation
[kP a] [kP a] [kP a] [kP a] [%] [%]
1.2 0.76 0.77 0.10 0.10 13 13
1.6 1.16 1.13 0.18 0.19 16 17
2.0 2.10 2.14 0.24 0.26 11 12
2.4 2.51 2.56 0.43 0.48 17 19
Table 7.6: Stagnation pressure in normal directions for free jet
time, which might cause incipient motion for a single particle.
U0 = 2.0m/s Stagnation Fluctuation Difference

at H/d= pressure pressure [%]
[−] [mm] [mm]
Free Jet
10 2.14 0.26 12
20 1.15 0.19 17
30 0.72 0.11 15
Table 7.7: Stagnation pressure for free jet at several stand off distances
The average pressure has an inverse correlation with SOD. The imping-
ing distance-term is in line with the erosion parameter from the equation
4.8. The turbulence part makes up between 12 and 15% of the stagnation
pressure.
Conclusion
No significant difference is observed in the turbulence term between the
impinging and free jet. Free jet has a stagnation pressure in various outflow
velocities and SODs.
All stagnation pressures in the free flow were slightly lower than in the
impinging jet.
The average pressure at the stagnation point differed slightly from the
middle and front gauge. The impinging jet indicated a 3.4% pressure
35
difference in normal directions, while the free jet had a deviation of
2.0% in normal directions.
Fluctuation was observed at the averaged stagnation pressure at all

outflow velocity around 15% for impinging and free jet.
Stagnation pressure was inversely related to SOD. The correlation is

in line with the erosion parameter.
7.3 Experimental results compared to similar tests

The bed dimensions of these experiments are compared to other similar
tests. The goal for the comparison is to validate the plain jet and the sand
bed geometry inside the flume. Detailed data are shown in Appendix D in
the dynamic stable state. The results are processed in the Figures 7.2 to 7.4.
In the graphs, two lines can be observed. The red line represents the
regression line of this research and the black line represents the results from
a comparable experiments conducted by Rajaratnam [1981].
Cavity depth
Figure 7.2: Jet erosion tests comparison to dimensionless depth
36
The data points seem to follow the trendline from Rajaratnam. Yet,
some variations can be observed between 4.0 to 6.5 of the erosion parame-
ter. The regression line has a transition point around the value 5.0, where
the tangent of the cavity depth changes as well as the cavity width. The data
points display some discrepancy compared to the trendline of Rajaratnam,
which could be the result of the differences in grain properties, like round-
ness or chemistry content. The cavity width clearly differs at higher values
of the erosion parameter. The report did not state clearly how Rajaratnam
measured the cavity width, but the black trendline does seem to follow the
experimental points from this study. Deviation in cavity width might causes
by the limited dimension of the flume. It should also be noted that many
observations from this experiment were in a wide range of erosion parameter.
The red regression line consists of the experimental observations with a

transition point at erosion parameter 5.0. Dimensionless depth as a function
of erosion parameter 1.0 until 5.0 is empirically defined as:

= 0.29 ∗ Ec + 0.211 (7.1)
H
Dimensionless depth as function of erosion parameter from 5.0 is defined as:

= 0.89 ∗ Ec − 3.08 (7.2)
H
Cavity width
The dimensionless width shows a similar trend as the function of erosion
parameter, but the results display lower values along the erosion parameter.
Results from Rajaratnam show almost similar results up to the value of 4.0.
From that value on, the normalised width shows smaller values with higher
values of the erosion parameter. One possibility could be the limited flume
width, which affects the hydraulic flow pattern in the dynamic stable state.
Secondary flow was not seen, but did affect the cavity growth.
The dimensionless cavity width is empirically correlated to the erosion

parameter: Dimensionless width as a function of erosion parameter 0.4 until
5.0 is defined as:

= 0.55 ∗ Ec + 0.28 (7.3)
x
Dimensionless width as function of erosion parameter from 5.0 is defined as:

= 1.31 ∗ Ec − 3.93 (7.4)
x
37
Figure 7.3: Jet erosion tests comparison to dimensionless width
Cavity hill height

The dimensionless hill height shows comparable results with other experi-
ments conducted with plain jet. A small overestimation could be observed
from the experiments until thee erosion parameter reaches 3.5, hereafter
lower values than that of Rajaratnam can be observed. A clear transition is
not observed around erosion parameter 5.0 from this research experiments.
Overall, the absolute length of hill heights were relatively low.
The dimensionless hill height has a correlation with erosion parameter:

The dimensionless hill height as a function of erosion parameter 0.5 until
5.0 is defined as:
∆
= 0.1 ∗ Ec + 0.15 (7.5)
H
The dimensionless hill height as a function of erosion parameter from 5.0 is
defined as:
∆
= 0.28 ∗ Ec − 0.83 (7.6)
H
7.4 Reproducibility of the results

Test D50 = 0.63mm, u0 = 2m/s and SOD = 10 has been tested three times
to check for reproducibility of the basic experiment. The test was conducted
38
Figure 7.4: Jet erosion tests comparison to dimensionless hill height
to ensure the reliability and consistency of the submerged pump. As can be

seen in equation 2.3, the turbulence term affects the reliability of the pump.
The results from Table D.1 is shown in Figure 7.5.
The three lines in the graph show the observation points, which do not
vary much from each other. The consistency of the pump is therefore ver-
ified with these three tests. Further experiments are therefore assumed to
be consistent with representable outcomes for the specific test setting and
the erosion velocity. The maximum deviation was at t = 0.5s from the first
and second tests.
7.5 Processes at bed in jet erosion test

This section described the observed processes at bed for a wide range of
erosion parameter. Low erosion parameter values are discussed first. Then,
higher values are considered.
7.5.1 Processes on bed at low erosion parameter

This test was conducted with an erosion parameter up to 1.0. Main charac-
teristic is the low activity beneath the jet for the weak deflected jet.
39
Figure 7.5: Consistency of the pump
At the start phase, the grains started rolling in lateral direction beneath
the jet. The low momentum outflow let particles transport away from the
centerline with a low take-off angle. Pick-up was likely triggered by micro
hydraulic turbulence beneath the jet, because the particles were displaced
over small distances. The cavity grows gradually to the final bed geometry.
Recirculation of suspended grains started quickly, where entrainment is im-
portant to bring the particles back in the dynamic erosion. The suspended
load is settled in the dynamic cavity after some time. Bed load transport
was seen, if the bed eroded. Suspended load started along with bed load
with an erosion parameter close to the value one for a while.
Particle movements stopped slowly to zero after some time. The sus-
pended load is settled in the static cavity. Some times particles vibrated on
the eroded cavity bed, this happened when hydraulic force is equal to resis-
tance of the submerged particle. Cavity geometry shows a weak deflected
jet with slopes smaller than the internal friction angle of particles.
At the end phase, the particles are all settled on the weak deflected
cavity. The bed remains in the same geometry position after the pump has
been switched off, so the the pick-up rate and particle transport were valued
to zero. The stagnation point remained the deepest point of cavity in the
test. Hills are formed with slopes equal to or smaller than the internal angle
of the sand.
The pick-up rate was hard to measure, since singularity was observed
around the stagnation point with individual particle movement. Pick-up
40
Before test t=1.0s
Figure 7.6: Weak defected jet
rates were therefore too small to be determined.
7.5.2 Processes on bed at mid-range erosion parameter

In this test erosion parameter from 1.0 to 5.0 during the test. Main charac-
teristic is the strong deflected jet with the pick-up in the inner cavity.
In the start phase, strong deflected jet showed soil deformation and sur-
face erosion. Grains at stagnation point started pick-up with low take-off
angle, similar to the test with low values of the erosion parameter. Mi-
cro turbulence is only a minor part of the pick-up in the start phase. In
addition, large momentum from jet caused bed deformation at the bed as
ejections. Grains were dragged along the curved cavity by the deflected
hydraulic flow. An inner cavity appears with dynamic erosion in the outer
cavity. Particles recirculate the suspended grains quickly after start, which
effects the dynamics of turbulence in impinging jet. This is in line with the
theory of Prandtl. Bearing capacity of soil was likely to be insufficient for
the vertical force from the impinging plain jet. Calculation checks are also
discussed int his section.
41
Initial bed condition Bed condition at t=0.1s
Bed condition at t=0.2s Bed condition at t=0.3s
Figure 7.7: Test 1 in start phase
In dynamic stable state, mid-range erosion parameter shows dynamic

erosion above the inner cavity inside the outer cavity. A swaying stagnation
point was observed at certain distances from the jet depending on the test
settings. This was not seen in the start phase. The dominated hydraulic
pressure stowed the inner cavity geometry to a more internal friction angle.
The impinging jet on impervious bed tests showed the fluctuated stagnation
pressure, which declares the dynamic stable cavity depth. At the same time,
minor and dominant vortices switched frequently from side of the jet. The
vortices seemed to want to break out from the jet. These sweeps could not
propagate further in the stream, so the eye of the vortices distanced a certain
distance from the jet. The dynamic inner cavity shape enclosed the vortices
42
by the steep slope in the inner cavity, which exceeded the internal friction
angle of the soil. The wall jet eroded in the inner cavity. The scoured flow
separated at the transition of the inner and outer cavity. Active sliding sand
bed added discontinuously bulk erosion to the jet system, which suspended
the grains by the wall flow in the inner cavity. Small grains of the gradation
were deposited at the outer slope, where the flow velocity is lower in the
ambient water.
When the pump was switched off, dynamic erosion stopped instantly due
to the lack of hydraulic force. Suspended grains were deposited on the bed
by gravity. Steep inner cavity collapsed instantly because of absent of hy-
draulic force. The grains at the outer gravity rolled to the lower part of the
cavity until the sand bed had reached its equilibrium. The dynamic stag-
nation point was covered with the suspended particles, which shortened the
cavity depth compared to the asymptotic state. The deepest cavity point
ends around the centreline but at a higher location. In contradiction, the
settlement shortened the cavity hill height with a few millimeters from the
original bed. The final bed had slopes with the internal friction angle.
Test case
Two test cases were checked for the bearing capacity, erosion rate and the
pick-up rate in the mid-range of the erosion parameter.
The first test case has an erosion parameter with a value of 3.6. The jet
has a SOD = 30 and the outflow velocity is 2.0m/s.
The bearing capacity was measured from impinging jet 0.77kP a. This
value exceeds the bearing capacity of 0.26kP a from Prandtl’s equation,
which is in line with the observed soil failure. The assumption was made
that 15% of vertical force would be in horizontal directions. Vertical move-
ment is likely because of the insufficient bearing capacity of the soil. A zone
III wedge could not be observed.
The graph shows the cavity depth development over time for both strong
deflected and dispersing jet, which can also be seen as erosion velocity.
43
Figure 7.8: Cavity depth development in time
This test case is plotted as the red line with the diamond shapes.
The experimental pick-up rate was checked with the theories of Van Rijn,
Meyer-Peter Müller and Van Rhee as shown in Figure 7.9.
Figure 7.9: Pick-up rate from test 1
The pick-up rate from this experiment was with a factor ten larger com-
pared to Van Rijn, Meyer-Peter Müller and Van Rhee. The erosion param-
44
eter was 3.6, but the pick-up rates were slightly larger. Pick-up rates were
approximately 12000g/s/m with a small increase as a function of the shear
stress. The propagating energy is reduced by viscous flow before hitting the
sand bed. The momentum spread in the hydraulic mixing layer over a larger
area beneath the jet.
The second test case had an outflow velocity of 1.2m/s with SOD = 10.
The erosion parameter is 3.8 was comparable value of test case 1. The mea-
sured pressure is 0.64kP a from table 7.3. The theoretical bearing capacity
is 0.26 calculated with equation 3.2. Vertical movement happened because
of insufficient bearing capacity of the soil which was also observed this test.
The erosion velocity was in test 2 smaller than in test 1. Smaller jet
momentum resulted in lower pick-up rate as in the yellow plot in Figure 7.8.
The pick-up rate from this experiment was with a factor ten larger com-
pared to Van Rijn, Meyer-Peter Müller and van Rijn. Shear stress is likely
not to be related to the pick-up rate. The erosion parameter was 3.8 with
pick-up rate around 11000g/s/m regardless of the shear stress. The lower
jet momentum has a lower pick-up rate than in test 1.
45
7.5.3 Processes on bed at large erosion parameter
Erosion parameter is treated from 5.0 till 10.6 for all the relevant processes
during test. Main characteristic is the strong deflected jet with the pick-up
in the inner cavity.
Dispersing jet showed bed deformation and surface erosion at the start-
ing phase. High jet momentum hits the soil bed. Sand particles translated
and rotated quickly in the bed in this regime after the start as showed in
Figure 7.11. Ejections were observed because of insufficient bearing capacity
according to Prandtl’s theory. Turbulence caused that the sweeps got into
deeper soil, which widens the cavity at the same time in the starting phase.
The stagnation point was at t = 0.1s to the right and at t = 0.4s to the
left. The plastic zone and the passive zone were seen on the videos just like
in the theory of Prandtl. Soil wedge (zone III, see 3.1) beneath stagnation
point could not be clearly observed. It could be possible that the turbu-
lence on the bed surface diminished the force balance in the triangle. The
inclination factor increased as described in Chapter 3.10. Recirculation of
suspended grains started just milliseconds after the start. A shield of grains
was formed above the dynamic erosion, which indicated insufficient erosion
capacity for a short period of time. Cavity grows more in depth than under
strong deflected jet.
46
Initial bed condition Bed condition at t=0.1s
Bed condition at t=0.2s Bed condition at t=0.3s
Figure 7.11: Test 3 in start phase
Dispersing jet had a dynamic stable cavity after a while. Many sim-
ilarities with strong deflected jet could be observed in this stage. A big
difference is that the cavity depth-width ratio is larger in the dispersing jet
condition. Dynamic erosion had a more chaotic flow pattern. The shield of
particles above inner cavity had been disappeared before reaching this stage.
Small gradation in particles were seen during these tests. Minor and dom-
inant vortices switched continuously from one side of the center line. The
minor vortex had a greater supply of grains from dynamic erosion, while the
dominant vortex lost suspended particles. This activity switched from side
when the minor vortex had a larger concentration of grains. Along the steep
slopes eroded with the active soil failure as a sliding sand volume. Smaller
particles could displace relatively over larger distance from center line by
experiencing smaller drag forces, while larger grains deaccelerate faster by
47
large frontal area during the transport.
Dynamic erosion stopped when the pump was switched off. Suspended
grains deposited on the bed by gravity. Steep inner cavity collapsed along to
lower part of the bed. The dynamic stagnation point was covered with the
suspended particles, which shortened the cavity depth compared to asymp-
totic state. Settling was hindered, if high concentrations of grains fell down.
The deepest cavity point ends at a higher location while the cavity hill height
shortened with a few millimeters. The final bed had slopes with the internal
friction angle of the sand soil.
Test cases
Test case 3 has a erosion parameter of 6.3 with outflow velocity of 2.0m/s
in SOD = 10.
Maximum bearing capacity is 0.26kP a as calculated from equation 3.2.
The measured stagnation pressure was 0.77kP a averaged over time. This
value exceeds the maximum, which is in line with the bed deformation from
observations.
The eroded depth was the largest of all analysed experiments. This
means also the largest erosion velocity in the first second after start. Jet
momentum is equal to that of test 1, but with a lower impinging distance.
This is displayed as a red plot with triangle shapes in Figure 7.8.
48
No considered theories can explain the results of the pick-up rates from
this experiment. The theories underestimate the pick-up by a factor 100.
The experimental rates are approximately 12000g/s/m. Shear stress is
larger at the starting phase compared to test case 1 with larger SOD and
lower pick-up rate. Shear stress is apparently not a function of pick-up rate.
The momentum outflow did not vary, so a likely correlation is made with
the jet momentum and the pick-up rate.
Test case 4 has a bed with larger particles of D50 as in test 1. Grain size
is increased from 0.63mm to 0.93mm. The jet momentum is equal to test 1
and 3, but with an erosion parameter of 4.2.
Soil moved as Prandtl’s theory explains because of insufficient bearing
capacity. The measured stagnation pressure 0.77kP a exceeded the maxi-
mum bearing capacity of 0.25kP a. A temporary shield grain was seen on
top the cavity, which appeared by the small sediment discharge capacity.
The erosion velocity is similar as test 1 and 3 in the first 0.5s. The same
jet momentum resulted in the same order of erosion velocity. This displayed
as the blue plot in Figure 7.8.
49
The experimental pick-up rates had larger values than what the theories
prescribe. This considered test setting had the same jet momentum as test
case 1 and 3. The erosion parameter is 5.2 with pick-up rates that were
approximately 12000g/s/m in the starting phase. In this test case, rates
were slightly lower because of the larger grain size and larger weight than
smaller particles. But still, jet momentum is a good indication of pick-up
rate regardless of particle sizes.
7.6 Inf luence of density

Qualitative research was done for the loose and dense sand bed state from
the plain jet. The interesting part was that density was not directly related
to the erosion parameter, while permeability did correlate with density and
erosion parameter. Also, a certain jet energy had to pick-up more particles
per unit of eroded bed volume. A larger pick-up rate meant more turbu-
lence along the propagating water stream, which resulted in a smaller cavity
dimension. Erosion parameter should be smaller in dense bed state.
50
Test Depth dense Depth dense Depth dense Depth loose Depth loose Depth loose
[−] t = 1s t = 2s t=∞ t = 1s t = 2s t=∞
[mm] [mm] [mm] [mm] [mm] [mm]
D63V040HD10 8 10 19 8 11 19
D63V120HD10 30 42 48 31 37 49
D63V120HD30 22 36 45 26 39 45
D63V200HD10 54 65 80 59 72 82
D63V200HD30 54 62 85 59 69 88
D93V120HD10 29 33 43 30 34 43
D141V120HD10 33 35 38 27 36 38
D141V120HD20 30 33 34 28 32 34
D141V120HD30 16 24 32 20 22 32
D141V200HD10 51 56 57 53 xx 57
D141V200HD20 48 52 54 49 53 54
D141V200HD30 41 43 46 43 45 48
Table 7.8: Initial deepening in dense bed compared to loose state
Overall, the penetration depth was smaller in the dense bed state than
in the loose sand state. More recirculating particles reduced propagating
momentum by turbulence toward the stagnation point. Less energy was
available to infiltrate the sand bed for a deeper cavity. Cavity depth de-
velopment takes a longer time to reach the asymptotic state as can be seen
from Figure 7.8.
Test Width dense Width dense Width dense Width loose Width loose Width loose
t = 1s t = 2s t = inf ty t = 1s t = 2s t=∞
[−] [mm] [mm] [mm] [mm] [mm] [mm]
D63V040HD10 54 58 69 48 52 67
D63V120HD10 116 122 140 102 120 136
D63V120HD30 112 152 182 114 142 176
D63V200HD10 150 196 366 122 174 358
D63V200HD30 196 xx 326 204 xx 318
D93V120HD10 108 136 138 106 126 132
D141V120HD10 61 67 96 51 56 92
D141V120HD20 66 71 80 67 72 77
D141V120HD30 72 81 82 65 76 78
D141V200HD10 78 80 96 70 xx 92
D141V200HD20 88 94 114 79 91 112
D141V200HD30 95 108 120 88 95 105
Table 7.9: Deepening in dense bed compared to loose state
The cavity width was larger in the first seconds in the dense sand con-
ditions compared to when there was less packed sand soil. Possibly, the
dense sand bed acted as an impermeable shield above vortices, which con-
verted more longitudinal momentum to lateral directions by the hydraulic
51
flow. Dense bed reduces the permeability, which is in line with the Shields
parameter. The pick-up rate should be lower according to Van Rhee, which
was not in line with the cavity width observations. The final cavity width
has increased in the dense bed compared to loose bed state. The eroded soil
volume was enlarged from packed bed to loose density at the bed level.
Test 3 in loose bed Test 6 in dens bed
Figure 7.14: Cavity in loose and bed state in dynamic stable state
The erosion parameter has no parameter for density, but bed geometry
is in all cases different with dens soil. The cavity depth is just a few mil-
limetres smaller, in contrary the cavity width is a few millimetres larger.
Test case 5
Test 5 had a higher density with similar jet setting like in test case 3. The
density went from 1938.2kg/m3 in loose state to 1994.5kg/m3 in the satu-
rated dense bed state. Porosity decreased from 0.43 to 0.38.
The erosion velocity is similar to that in test 1, 3 and 4 in the first 0.5s.
The same jet momentum resulted in the same order of erosion velocity. This
is displayed as in the purple plot in Figure 7.8. The original dense bed state
had less cavity depth in dynamic stable state.
52
Figure 7.15: Pick-up rate from test 5(dense bed)
Actual pick-up rates were larger than the three considered erosion mod-
els. Shear stress is also not determinative in dens bed. The jet momentum
is similar to that in test case 3. The pick-up rates were in the same order
magnitude. The same amount of particles had to be displaced to reach a
certain jet momentum, which leads to a smaller cavity dimension in the first
second after the start. The cavity had slightly less depth in the dynamic bed
state, but the width and hill heights were larger because of the increasing
soil volume. Initial dense bed increased in volume to lose soil after deposi-
tion.
Conclusion
A denser bed resulted to slightly smaller cavity depth, but a larger
width. One unit of eroded dense soil had lower density after deposition,
which widens the cavity.
An eroded dense bed would turn into loose state after depositing on
the original bed level. The loose deposition from a dense bed was
larger than the original loose bed.
53
7.7 Inf luence of particle shape
Glass beads had effects on the dimensions of cavity. The comparison was
made with large glass beads and Dorsilit 5F, that had a similar D50 .
The expectation would be that it would have the same eroded depth as
sand soil. The cavity width and hill height are likely to be different, since
the internal friction angle is smaller. Width would be larger and hill height
smaller.
Test case 6
Figure 7.16: Test 6 with glass Figure 7.17: Similar test with
beads sand
Beads beneath the jet started with surface erosion and bed deformation.
Micro turbulence causes surface erosion in the start phase. Bed deformation
was observed in line with the theory of Prandtl.
The bed geometry had some significant differences in dimensions. The
cavity depth was deeper than the sand bed. This might be related to a lower
solid density or the roundness of glass beads. The volume from the inner
cavity had spread above the original bed level. Particles were deposited at
the hills by impinging jet flow, while grains eroded to the lower part until
the hill slopes take the internal friction angle of 13◦ . The hill height was
clearly lower compared with sand soil. The dimension of the hills was much
wider due to a lower internal friction angle than sand grains.
Transport could hardly be observed on camera. No clear figures could

be made for the processes of glass beads. Pick-up rate for glass beads was
analysed for that reason.
Two theoretical processes are important to the pick-up rate of glass
beads compared to sand soil. Firstly, the shape of the glass beads reduces
54
the resistance between particles because of a smaller contact area, which
increases the pick-up rate around the stagnation point. Secondly, the skin
friction of the beads decreases the hydraulic drag coefficient, which reduces
the pick-up easier.
7.8 Influence of permeability

Dry sand volume was mixed with 6.0% bentonite. Conversion to dry mass
has been applied, which meant a 3.6% mass percentage of dry sand. This
percentage did block the permeability, but it will not behave as silt. Sand
and bentonite were mixed well manually in dry condition. A small portion is
spread evenly in the dry flume to a height of 0.20m. The flume is slowly filled
with water to avoid any stirring and turbidity. The actual execution of the
test happened after resting 24hours, so that the bentonite could bulge with
water between the sand particles. The mixture reduces the permeability to
90% compared to the sand bed.
The expectation was that the bentonite mixture will erode on the surface
beneath the jet. Dilation is important, since voids are locked up. Bed defor-
mation will immediately be blocked by the underpressure in void space. The
pick-up rate was expected to be smaller. Van Rhee relates the permeability
as inverse to the Shield parameter (eq. 4.26). The pick-up is correlated
inversely with Shield, so a smaller permeability should lead to a reduction
in the pick-up rate. Surface erosion would be expected by micro turbulence,
but critical shear stress is higher because of stickiness between the particles.
Sand bed might erode as chips if cohesiveness is high with a large hydraulic
force.
55
Clear water from bentonite
Bentonite flow in bed
mixture
Figure 7.20: Permeability in soil bed
Figure 7.18: Sand soil without Figure 7.19: Mixture of sand and
bentonite bentonite
The results were not as expected, the cavity geometry in bentonite was
larger than the non-cohesive sand soil after twenty minutes. The perme-
ability and stickiness should increase the resistance between the particles
and lead to smaller cavity. The contradiction was observed a larger cavity
depth, width and hill height as can be seen in Figure 7.18 and 7.19. The
pick-up rate was larger, which contradicts with the lower permeability and
the lower pick-up rate. The soil bed compacted as loose sand into dens bed
with appearance of small ejection. Possibly the valve was not set correctly.
A larger momentum leads unfortunately to a larger cavity geometry.
Ground water flow was observed, which was assumed to be zero, which
is visible with concentrated bentonite as a pressure wave after start as in
Figure 7.20. The wave moved with around 0.1m/s through sand bed for
seconds. After some time, the concentration was diluted with clear water.
56
Test case 7
Figure 7.21: Pick-up rate comparison with bentonite mixture and sand only
The pick-up rate was surprisingly higher with bentonite soil than without
at the starting phase. The dimensions of the cavity were larger for depth,
width and hill height in the dynamic stable state. It could be possible that
the mixture test setting was configured wrongly with a higher fluid outflow.
A larger cavity geometry indicated a higher jet momentum from the imping-
ing jet during the experiment. The consequence was a larger pick-up rate
because of larger outflow energy at the starting phase.
7.9 Momentum as function of pick-up rate

One of the hypothesis is the pick-up rate as function of jet momentum. This
is shown in Figure 7.22.
57
Figure 7.22: Pick-up as function of jet momentum
Similar jet momentum had almost the same magnitude of pick-up rate
in the start phase. This proved the strong correlation between the jet mo-
mentum and the pick-up rate. The jet momentum determined the pick-up
rate. Parameters like grain sizes and SODs had not a significant effect to the
pick-up rate. Larger jet momentum resulted in larger deviation in pick-up
rate. The pick-up flux decreased with the increasing jet momentum. Hin-
dered erosion reduced the growth in the erosion velocity.
Dens bed (test 5) showed a slightly larger pick-up rate as for the loose
bed. Same energy outflow had moved more amount of solid particles. The
specific energy was lower for the solids, despite that the pick-up was larger
for the dense bed.
58
Chapter 8
Conclusion and
recommendations
8.1 Conclusion
The following conclusions can be made from the experiments in the labora-
tory:
The first grain movement was a surface erosion beneath the jet in
all jet erosion tests caused by turbulences inbetween the two vortices.
The instant shear stress eroded soil bed into a weak deflected jet. The
turbulence varied the shear stress in time, so the stagnation point was
swaying beneath the jet. A small hill was not formed inside the cavity
as in my hypothesis as in Figure 6.2 for all experiments.
Erosion parameter from 1.0 showed strong deflected jet with pick-
up inside the inner cavity. Ejection and bed deformation was the
main causation of the cavity pick-up. The turbulence caused sweeps
for sediment motion beneath the jet, which resulted in instant shear
stress on the soil bed. Bed and suspended load followed above the
dynamic bed geometry as injection. The momentum on bed from the
jet exceeded the bearing capacity from the theory of Prandtl, which
translated and rotated the particles in the bed in the start phase. The
occurrence was soil failure.
Pick-up rate was correlated to the jet momentum. The experimental
tests showed a fixed jet momentum result in a pick-up rate with weakly
dependency with SOD and grain size. Jet momentum caused bed
deformation, if maximum bearing capacity was exceeded. Particle
translation mainly determined the pick-up rate.
59
Erosion theories are underpredicted the pick-up rate for the jet ero-
sion tests. Soil failure is because of the loading exceeding the bearing
capacity.
Dense soil bed should be corrected for the permeability. Pick-up rate
was smaller than in the original dens bed by higher resistance between
particles compared to the loose bed.
Similar tests to that of Rajaratnam were reproduced with experimental
JET in non-cohesive soil. Cavity depths and hill heights were similar,
but the cavity widths were significant lower for the larger values of the
erosion parameter.
8.2 Recommendation
Normal flow was measured with gauges in the validation phase. This
means a small normal net force is present beneath the plain impinging
jet, while the focus was 2-dimensional. A jet without front casing
could avoid the normal water flow.
The dimensionless width as function of the erosion parameter differs
from similar tests. Hence, tests with a wider flume is recommended
for preventing secondary flow, which might have affected the cavity
width.
The test with bentonite mixture has a larger cavity dimensions than a
normal sand bed. This could not be clarified. The jet momentum was
likely larger because of wrong test settings. Accurate outflow settings
are recommended to get a good result.
60
Bibliography
M.L. Alberson, W.D. Baines, D. Citrini, S. Corrsin, Y.B. Dai, H.R. Henry,
J.S. Holdhusen, R.A. Jensen, H. Rouse, and A. Streiff. Diffusion of sub-
merged jets-discussion. TRANSACTIONS OF THE AMERICAN SOCI-
ETY OF CIVIL ENGINEERS, 115:665–697, 1950.
J.A. Battjes. Vloeistofmechanica. Collegedictaat CT2100, 2002.
F. Bisschop. Erosion of sand at high flow velocities. 2018.
H.J. de Vriend, H. Havinga, B.C. van Prooijen, P.J. Visser, and Z.B. Wang.
River engineering. Collegedictaat CIE4345.
H.B. Fischer, J.E. List, C.R. Koh, J. Imberger, and N.H. Brooks. Mixing in
inland and coastal waters. Elsevier, 2013.
B. Foortse. Retardation of breach growth under high flow velocities. 2016.
J.W. Gauntner, P. Hrycak, and J.N.B. Livingood. Survey of literature on

flow characteristics of a single turbulent jet impinging on a flat plate.
1970.
A. Koched, M. Pavageau, and F. Aloui. Vortex structure in the wall region

of an impinging plane jet. Journal of Applied Fluid Mechanics, 4(2), 2011.
W.F. Molenaar and M.Z. Voorendt. Manual hydraulic structures 1. Col-

legedictaat CIE3330, 2012.
I. Nezu, A. Tominaga, and H. Nakagawa. Field measurements of secondary

currents in straight rivers. Journal of Hydraulic Engineering, 119(5):598–
614, 1993.
N. Rajaratnam. Turbulent jets, volume 5. Elsevier, 1976.
61
N. Rajaratnam. Erosion by plane turbulent jets. Journal of hydraulic Re-
search, 19(4):339–358, 1981.
C. van Rhee. Dredging processes 2. Collegedictaat COE4727, 2015.
C. van Rhee and A. Bezuijen. Influence of seepage on stability of sandy

slope. Journal of geotechnical engineering, 118(8):1236–1240, 1992.
L.C. van Rijn. Sediment pick-up functions. Journal of Hydraulic engineering,

110(10):1494–1502, 1984.
G.J. Schiereck. Introduction to bed, bank and shore protection. CRC Press,
2003.
A. Verruijt and S. van Baars. Soil mechanics. VSSD Delft, the Netherlands,
2007.
62
Appendix A
Grain size analysis
This appendix chapter is aimed to support the main body of this report with
a particle sieve analysis conducted at the faculty Civil Engineering in Delft.
The first part of this chapter describes the properties of the respective sand,
which was ordered at the company Sibelco Europe located at Papendrecht
in the Netherlands. The second part is about the characteristics of glass
beads that were also sieved in the laboratory. This material was bought
from the firm Sigmund Lindner GmbH located in Germany.
Method
The diameter of the sieves are chosen based on that mesh sizes are around
the same D50 as described on the product sheet. Each sieve and dustpan
are weighted before the sieving tests start. The sieves with the largest mesh
piles up at the top of the sieve column. Smaller grain openings are placed
lower in the sieve column. A dustpan is placed to obviate the very small
particles between the sieve column and apparatus. The sand bag is mixed
well before it is put into the sieve machine. The machine shakes for ten
minutes, so grains can fall through the sieves. The mass of each sieve and
dustpan are weighted with the filtered particles inside. The result can be
proceed for the particle sieve analysis.
Each soil package is tested three times to ensure reliable results. The
measured weights are noted to proceed the cumulative mass and passing
cumulative mass percentage. A graph shows the grain diameter on the
horizontal log-axis and the cumulative relative mass on the vertical axis.
The analysis is also repeated for the other two sand volumes and two other
different glass beads.
A1
The weight of each batch is then compared to see, whether the error
is within the 1% before and after each run, to ensure reliable sieve results.
The outcome is the average weight of each particle package from the three
tests, which uses for a proper semi-log graph for each soil volume. Another
method is linear interpolating of the two nearest data points for the D50 in
the semi-log graph.
Procedure:
Read D50 sand from production sheet
Choose several sieves around D50
Measure mass of each empty sieve
Take a volume sand
Measure the mass of the volume sand
Place the sieves correctly on the sieving apparatus
Start the machine and set the timer to 10 minutes
Wait
For each sieve layer measure the mass
Calculate the genuinely particle mass of each sieve
Show results in table and graph
Sieving results: Dorsilit 5F

The sand Dorsilit 5F comes from crystal silica sands. The material is packed
in a bag of around 25kg each. Sand particles are sub rounded with gray-ish
color and compose about 99.1% of silicon dioxide (SiO2 ). The solid density
is 2630kg/m3 . The production sheet (page A9) shows that the majority of
the grains should be between 1.0 and 1.8mm. Therefore, sieves with these
meshes were chosen: 0.85, 1.0, 1.168, 1.4, 1.6, 1.7, and 2.0mm.
The D50 is 1.43mm, which was determined by linear interpolating the

points between 1.4 and 1.6mm. The determined value fitted well with the
production sheet from the manufacturer. The internal friction was about
33◦ with loose saturated density of 1971.2kg/m3 for this sand batch. The
uniformity coefficient is determined by linear interpolation, Cu = D D10 =
60
1.477
1.039 = 1.42. Dorsilit 5F is classified as poorly graded soil with almost sim-
ilar grain size.
A2
Sieve Sieve diameter Test 1 [g] Test 2 [g] Test 3 [g] Average
[mm] mass[g]
1 2.0 0.3 0.2 0.3 0.3
2 1.7 70.4 59.6 74.2 68.1
3 1.6 50.1 44.9 60.3 51.8
4 1.4 163.2 152.7 186.4 167.4
5 1.168 148.3 159.4 196.7 168.1
6 1.0 52.4 59.6 72.0 61.3
7 0.85 8.5 11.0 12.1 10.5
8 dustpan 25.3 44.5 23.8 31.2
total 518.5 531.9 625.8 558.7
Table A.1: Sieve results with Dorsilit 5F
Sieve Sieve diameter Average Cumulative Passing cumulative

[mm] mass [g] mass [g] mass percentage [%]
1 2.0 0.3 0.3 99.9
2 1.7 68.1 68.4 87.8
3 1.6 51.8 120.2 78.5
4 1.4 167.4 287.6 48.5
5 1.168 168.1 455.7 18.4
6 1.0 61.3 517.0 7.5
7 0.85 10.5 527.5 5.6
8 dustpan 31.2 558.7 0
Table A.2: Particle sieve analysis with Dorsilit 5F
The fall velocity is around the range of 1.95 to 3.99 seconds, which means
the particles have a diameter of 4.3 and 1.04mm. The earlier found D50 cor-
responds within these values.
Sieving results: Dorsilit 7

Dorsilit 7 is crystal silica sand, which is packed in shrink foil of around 25kg
each. Sand particles are sub rounded with gray and white particle color and
composes around 99.1% of silicon dioxide. The production sheet on page A9
shows that the grain sizes should be between 0.6 and 1.2mm. Sieves were
A3
Grain distribution Dorsilit 5F
100
cumulative percentage 80
60
40
20
0
10−1 100 101
graindiameter
Figure A.1: Grain distribution Dorsilit 5F
used for particle size distribution with meshes 0.5, 0.6, 0.71, 0.85, 1.0, 1.168
and 1.4mm.
[mm] mass [g]
1 1.4 0.6 0.6 0.7 0.6
2 1.168 11.0 9.1 12.2 10.7
3 1.0 183.7 174.6 190.4 182.9
4 0.85 180.4 174.7 183.7 179.6
5 0.71 125.7 115.8 130.3 123.9
6 0.6 26.4 25.3 27.5 26.4
7 0.5 9.5 8.5 8.5 8.3
8 dustpan 18.7 14.5 12.5 15.2
total 556.0 523.1 565.8 547.6
Table A.3: Sieve results with Dorsilit 7
According to the manufacturer the sub round grain diameter should be

between 0.6 and 1.2mm. Linear interpolation is applied to two surrounding
data points. Resulting in D50 of 0.93mm, which is a reasonable value and it
A4
1 1.4 0.6 0.6 99.9
2 1.168 10.7 11.3 97.9
3 1.0 182.9 194.2 64.5
4 0.85 179.6 373.8 31.7
5 0.71 123.9 497.7 9.1
6 0.6 26.4 524.1 4.3
7 0.5 8.3 532.4 2.8
8 dustpan 15.2 547.6 0
Table A.4: Particle sieve analysis with Dorsilit 7
Grain distribution Dorsilit 7

100
cumulative percentage
80
60
40
20
0
10−1 100 101
graindiameter
Figure A.2: Grain distribution Dorsilit 7
will be used in further investigations. The internal friction is 33◦ . The loose
saturated density is 1903.5kg/m3 in dry solid state. Uniformity coefficient
is Cu = D 0.979
D10 = 0.730 = 1.34, which means the regarding sand package is
60
graded as poorly graded soil.
The fall velocity is approximately between the 2.26 and 5.14 seconds,
which means the particles have a diameter of 3.2 and 0.62mm. The earlier
A5
found D50 corresponds with these values.
Sieving results: Dorsilit 8

Dorsilit 8 contains crystal silica sand, which is packed in bags of around
the 25kg each. Properties of sand particles are sub rounded with gray and
white particle color and composes around 98,9% of silicon dioxide, according
to the manufacturer’s description. The production sheet (page A10) shows
that the grain sizes has to be D50 around 0.57mm. Sieves were chosen with
meshes 0.125, 0.25, 0.297, 0.425, 0.5, 0.6, 0.85, and 1.0mm.
[mm] mass [g]
1 1.0 0.2 0.1 0.0 0.1
2 0.85 1.3 1.3 0.7 1.1
3 0.6 354.9 307.6 287.2 316.6
4 0.5 94.4 104.2 93.3 97.3
5 0.425 102.0 114.6 102.0 106.2
6 0.297 16.9 17.1 14.2 16.1
7 0.25 2.7 2.7 2.5 2.6
8 0.125 2.4 2.0 2.1 2.2
9 dustpan 0.0 0.2 0.3 0.2
total 574.8 549.8 502.3 542.4
Table A.5: Sieve results with Dorsilit 8
The D50 is 0.63mm after interpolating between the 0.6 and 0.85mm
sieves. The sieved D50 is slightly larger than the prescript, the deviation is
within the 10%. This error could have several reasons, it could e because
of inaccurate test equipment or the manufacturer could have used differ-
ent parts of their bulk volume. Sorting of grains diameter are unavoidable
during processes or transportation. The deviation is not remarkably large
and is assumed to have only little influence to the main results. The inter-
nal friction is 31◦ with loose saturated density of 1938.2kg/m3 . Uniformity
coefficient is Cu = D 0.680
D10 = 0.448 = 1.52. Dorsilit 5F is classified as poorly
60
graded soil.
The fall velocity is around between the 2.49 and 7.50 seconds, which
A6
1 1.0 0.1 0.1 99.9
2 0.85 1.1 1.2 99.8
3 0.6 316.6 317.8 41.4
4 0.5 97.3 415.1 23.5
5 0.425 106.2 521.3 3.9
6 0.297 16.1 537.4 0.9
7 0.25 2.6 540.0 0.4
8 0.125 2.2 542.2 0.04
9 dustpan 0.2 542.4 0
Table A.6: Particle sieve analysis with Dorsilit 8
means the particles have a diameter of 2.6 and 0.29mm. The earlier found
D50 corresponds to these values.
A7
Grain distribution Dorsilit 8
100
cumulative percentage
80
60
40
20
0
10−1 100 101
graindiameter
Figure A.3: Grain distribution Dorsilit 8
A8
Dorsilit
Kristal kwartszand
Omschrijving Dorsilit kristal kwartszand is een gewassen zand die zich door een hoog SiO2
gehalte en afwezigheid van humusachtige stoffen en andere
verontreinigingen onderscheid. Na het drogen wordt het nauwkeurig
uitgezeefd.
Toepassing Dorsilit kristal kwartszand is onder andere geschikt als vulstof en

instrooimateriaal voor kunststofgebonden troffel- en gietvloeren.
Eigenschappen Vorm : rondvormig

Kleur : grijs / wit / gebroken wit / beige
Hardheid : 7 Mohs
3
Soortelijk gewicht : 2,63 kg/dm
3
Stortgewicht : 1,60 kg/dm
Chemische Analyse SiO2 : 99,1 %

(Slechts een indicatie) Al2O3 : 0,42 %
Fe2O3 : 0,024 %
TiO2 : 0,020 %
CaO : 0,02 %
MgO : 0,02 %
K2O : 0,13 %
Na2O : 0,01 %
Gloeiverlies : 0,20 %
Korrelmaten Nr. 2F 3,0 – 5,0 mm Nr. 7 0,6 – 1,2 mm

Nr. 3 2,0 – 3,5 mm Nr. 8 0,3 – 0,8 mm
Nr. 5G 1,6 – 2,5 mm Nr. 9 0,1 – 0,5 mm
Nr. 5F 1,0 – 1,8 mm
Andere korrelmaten op aanvraag
Verpakking  In plastic zakken van 25 kg op eenmalige pallets van 1000 kg met

krimpfolie.
 Geweven polipropyleen big bags
 Losgestort in bulkauto
Bovenvermelde informatie is gebaseerd op gemiddelde waarden. De typische eigenschappen en chemische analyses zijn
bedoeld als voorbeelden en kunnen niet beschouwd worden als vervanging voor eigen testen en onderzoek in alle
omstandigheden waarbij eigenschappen en chemische samenstellingen kritische factoren zijn.
TDS. 2014-Dorsilit 1/1
Worldwide Sibelco Europe

Tel: +31 (0)78 6546770 PO Box 184
[email protected] NL-3350 AD Papendrecht
The Netherlands
A9
Dorsilit 8
Kristal kwartszand
Omschrijving Dorsilit kristal kwartszand is een gewassen zand die zich door een hoog SiO 2
gehalte en afwezigheid van humusachtige stoffen en andere verontreinigingen
onderscheid. Na het drogen wordt het nauwkeurig uitgezeefd.
Toepassing Dorsilit kristal kwartszand is onder andere geschikt als vulstof en instrooimateriaal
voor kunststofgebonden troffel- en gietvloeren.
Eigenschappen Vorm : rondvormig

Kleur : grijs / wit / gebroken wit / beige
Hardheid : 7 Mohs
3
Soortelijk gewicht : 2,63 kg/dm
3
Stortgewicht : 1,60 kg/dm
Chemische SiO2 : 98,9 %

Analyse Al2O3 : 0,60 %
(slechts een indicatie) Fe2O3 : 0,01 %
Ti2O : 0,04 %
CaO : 0,02 %
MgO : 0,01 %
K2O : 0,10 %
Na2O : 0,01 %
Gloeiverlies : 0,20 %
Korrelverdeling > 1,0 mm Slechts sporen

> 0,8 mm 01%
0,63 - 0,8 mm 31%
0,315 - 0,63 mm 67%
< 0,315 mm 01%
D50 0,57 mm
Andere korrelmaten op aanvraag
Verpakking  Zakken van 25 kg op eenmalige pallets van 1000 kg met krimpfolie.

 Bigbags van 1500 kg op eenmalige pallets
 Losgestort
Bovenvermelde informatie is gebaseerd op gemiddelde waarden. De typische eigenschappen en chemische analyses zijn
bedoeld als voorbeelden en kunnen niet beschouwd worden als vervanging voor eigen testen en onderzoek in alle
omstandigheden waarbij eigenschappen en chemische samenstellingen kritische factoren zijn.
TDS. 2014-Dorsilit 8 1/1
Worldwide Sibelco Europe

Tel: +31 (0)78 6546770 PO Box 184
[email protected] NL-3350 AD Papendrecht
The Netherlands
A10
Sieving results: Small glass beads
The certified glass beads were produced from Sigmund Lindner GmbH in
Germany. The roundness of particles are spherical, which results to a small
internal friction angle of 14◦ . The beads are transparent and non-cohesive.
The grain size should be 90% of the mass between 0.8 and 1.0mm grain
diameter 1 . Therefore, sieves were chosen with openings 0.6, 0.85, 1.0, 1.168
and 1.4mm for this test run.
[mm] mass [g]
1 1.4 0.0 0.0 0.0 0.0
2 1.168 0.1 0.1 0.1 0.1
3 1.0 57.2 72.0 54.5 61.2
4 0.85 105.9 87.4 98.1 97.1
5 0.6 3.9 3.6 7.8 5.1
6 dustpan 0.0 0.0 0.0 0.0
total 185.6 153.9 148.6 162.8
Table A.7: Test results with small glass beads

1 1.4 0.0 0.0 100.0
2 1.168 0.1 0.1 99.9
3 1.0 61.2 61.3 62.5
4 0.85 97.1 158.4 3.1
5 0.6 5.1 163.5 0
6 dustpan 0.0 163.5 0
Table A.8: Test results with small glass beads
The density in dry state was 2500kg/m3 . The internal friction angle wa
14◦ .
The D50 was 1.13mm by interpolating the sieve meshes between 1.0
and 1.168mm. The grain grading was very homogeneous. Major beads fell
1
http://www.sigmund-lindner.com/
A11
into two sieves, which can be considered as unreliable. This method was
insufficient to get an accurate grain size distribution. In addition, the par-
ticles size was not aligned with the manufacture sheet. The roundness was
spherical due to the artificial production process, which characterised the
homogeneous package of this material. As a result, a relative steep grain
size distribution appeared with only three valid points in semi-loggraph.
Additional analysis for beads was not possible with this sieving method.
Figure A.4: Particle Size Distribution: small beads
For more accurate beads diameter method of laser diffraction analysis

or imaging particle analysis.
Sieving results: Large glass beads

The certified glass beads were produced from Sigmund Lindner GmbH in
Germany. The roundness of particles are spherical, which results in a small
internal friction angle of 13◦ . The beads are transparent and non-cohesive.
The grain size should be 90% of the mass between 1.25 and 1.65mm grain
diameter 2 . Sieves with meshes 1.0, 1.168, 1.4, 1.65 and 1.8mm were used
for this test runs.
The density in dry state is 2500kg/m3 . The internal friction angle is 13◦ .
2
http://www.sigmund-lindner.com/
A12
[mm] mass [g]
1 1.8 0.0 0.0 0.0 0.0
2 1.65 4.3 3.2 4.9 4.1
3 1.4 104.7 92.4 101.9 99.7
4 1.168 76.3 58.0 41.4 58.7
5 1.0 0.3 0.3 0.3 0.3
6 dustpan 0.0 0.1 0.1 0.0
total 167.1 163.1 160.5 163.5
Table A.9: Test results with large glass beads

1 1.8 0.0 0.0 100.0
2 1.65 4.1 4.1 97.5
3 1.4 99.7 103.8 36.2
4 1.168 58.7 162.5 0.2
5 1.0 0.3 162.8 0.0
6 dustpan 0.0 162.8 0.0
Table A.10: Test results with large glass beads
The grain size was between 1.25 and 1.65mm according to the pre-
script. The D50 was 1.45mm by interpolating sieve openings between 1.4
and 1.65mm. Major beads fell into two sieves, which causes to unreliable
results. The particles sieve analysis was insufficient to get an accurate grain
size distribution graph, although the D50 was within the given limits. The
roundness is spherical due to the artificial production process. Also, a high
percentage of the beads fell only in two sieves, which proved for the homo-
geneous package of this material. A narrow gradation appeared with the
data points in the semi-loggraph, which made it an inaccurate outcome for
a grain size distribution analysis.
A13
Figure A.5: Particle Size Distribution: small beads
Conclusion
The following D50 are shown in Table A.11 for sand and glass beads. Cal-
culation were made with the values from the particle sieve analysis or the
production sheets.
Type sand Diameter [mm] Internal friction angle Uniformity coefficient

Dorsilit 5F 1.41 33 1.42
Dorsilit 7 0.93 33 1.34
Dorsilit 8 0.63 32 1.52
Large glass beads 1.45 13 xx
Small glass beads 1.13 14 xx
Table A.11: Sieve result of sand
The sand packages were classified as sub round with internal friction an-
gle of around 32◦ . Gradation was very poorly graded, since the uniformity
coefficient was close to one. Glass beads were even worse graded with an
unknown uniformity coefficient. The beads fell only in two sieves by the ho-
mogeneous particle size. The internal friction angles were respectively 14◦
and 13◦ for small and large glass beads.
A14
Appendix B
Water system design
This appendix described the design process with a starting point and con-
siderations to get the final jet system. An acrylic flume is assigned to the
experiments with a vertical submerged hydraulic plain jet. Many aspects
were considered to get the final experiment set-up. The experiments were
executed at the Dredging laboratory at faculty 3ME at Delft. The plain jet
was produced by Demon in Delft.
Starting point
The main objective was to see the 2D effect beneath the hydraulic plain jet.
The starting point was the plain jet with a width of 2.5mm nozzle diame-
ter, which is a reasonable dimension for the test flume. Also, one goal was
to try to duplicate the experiment from Rajaratnam [1981] with the same
jet width. The form of jet was free to accommodate the research preferences.
Considerations for jet

The 3D effects should be excluded in the test system(normal-direction).
The normal length must be at least twenty times the width of the nozzle
Koched et al. [2011]. The nozzle width was 2.5mm. The length should be
at least 20 ∗ 2.5 = 50mm. The length was doubled to create some space for
preparations in the soil before each run. The nozzle was, therefore, designed
with an inner length of 100mm.
B1
The thickness of the hull was determined at 2mm for all the walls of the
draghead. The jet and the tube were made from aluminium. The jet barely
bend as calculated with the forget-me-not formulas. The assumption was
made with the outflow of 2.5m/s, which was converted to the pressure via
equation 2.1. The bending in the draghead was calculated to 8.7 ∗ 10−4 m,
which was neglectable and did not caused any problem during the tests.
2
5 q ∗ l4 5 ( 1000∗2.5 ) ∗ 0.14
w= = 2
= 8.7 ∗ 10−4 m (B.1)
384 EI 384 70 ∗ 109 ∗ 0.1∗0.0023
12
Other consideration was to clamp the jet to the frame.
1 ∗ F ∗ l3
w= (B.2)
3 ∗ EI
2
1 ∗ ( 1000∗2.5 ) ∗ 1.33
w= 2
3∗70∗((0.104∗0.0543 )−(0.104−2∗0.002)(0.104−2∗0.002)3
= 0.4 ∗ 10−4 m (B.3)
12
The inner dimension of the nozzle width is 50mm. The vertical velocity
in the draghead must be much larger than the vertical outflow from the
nozzle. u0 >> w0 with in mind the volume balance Qd = Qn . b0 ∗ l ∗ u0 =
b1 ∗ l ∗ w0 The broader draghead created an hydraulic drop at the opening
with acceleration in the funnel, which was curved as two quarter circle with
a radius of 25mm. This shape provided the most uniform outflow with less
turbulence than other nozzle shapes.
The draghead had a rectangular shape with a length of 300mm in lon-
gitudinal direction. Suspended particles could accumulation on top of the
head during test. A longer head should avoid any deposition of grains on the
jet, but more length separates hydraulic interaction on both sides of the jet.
Compromise was made on top of the draghead one meter long round alu-
minium pipe, where the fluid can flow around and the jet arm had sufficient
length to clamp for several SOD.
The final dimensions of plain jet and frame could be found on page B5.
Considerations for f lume

The overall outer dimensions of the flume is 1.5*0.75*1.0m (lxbxh), which
was manufactured from transparent acrylic. The walls, bottom, weirs and
ridges have a wall thickness of 10mm. The flume had initially two com-
partments with a 0.9m height weir in the middle of the normal length. The
B2
water level was set to 0.9m during the experiments. Any waves would damp
out by the overflow of the high weir.
The water jet had to fit the concerning cell. The hydraulic jet was made
from a 2mm thick plate by cutting, bending and welding. This results in an
outer dimension of jet 100 + 2 ∗ 2 = 104mm. The added weir had 110mm
space behind the inner front wall. The larger length was chosen for avoiding
any jams between the jet and the added acrylic walls.
The flume is divided in three compartments with respectively 0.7 and
0.9m height transparent weirs between the front, middle and back cells.
The 0.7mm height weir is a bit lower, but still the holds the majority of the
suspended particles inside the front compartment in the high values of the
erosion parameter. The higher 0.9mm weir holds smaller particles from the
back compartment.
The limitation to this research is the width of the flume with 0.75mm.
An experiments should avoid any particle depositions against the side wall.
This could affect the outcomes of the dimensions of the eroded cavity. There-
fore, assumptions was made that the soil bed remain at original bed level
for about one-third of the flume width.
Other considerations
The experiments are executed in the front compartment with the hanging
plain jet on a stiff frame. The frame has a clamp system for the jet, so
the stand off distances can be adjusted to the required impinging distance
between the nozzle and the sand bed.
As mentioned, the cavity width has a maximum of 0.5m to avoid any
deposition against the side walls. The sand hole has a ratio of 0.25 in
relation to depth-width (Rajaratnam [1981]). The cavity depth would be
at the maximum of 0.125m. The sand bed is prepared to 0.2m level for all
experiments.
The discharge meter was of the brand Flowtec Variomag from type Dis-
comaq DMT 6530, which is placed between the Y-splitter and the jet. This
meter is an electromagnetic flow meter. Two pipes are mounted before and
after the discharge meter for an uniform flow for an accurate measurements.
The lengths of both tubes are ten times the diameter of the hose.1 The me-
ter has a range from 0 to 4.0m3 /h with an accuracy of 0.03m3 /h
1
https://www.erniegraves.com/egc-news/minimum-straight-pipeline-lengths-for-10-
popular-flowmeters
B3
A constant submerged pump is used for several experiments with a range
of discharges. An Y-splitter was placed to direct a certain discharge to the
jet, the rest recirculated back in the overflow compartment. The valves need
to set up, if other outflow velocity is required. The submerged pump needs
at least a discharge of 0.000625m3 /s. The design outflow was calculated for
erosion parameter five with the largest grain size.
u0
Ec = p √ (B.4)
H/d ∗ g∆D50
Rewriting results in:

p p √ √
u0 = Ec ∗ H/d ∗ g∆D50 = 5 ∗ 10 ∗ 9.81 ∗ 1.65 ∗ 0.00141 = 2.36m/s
(B.5)
The value has been rounded up to uo = 2.5m/s for a conservative approach.
The minimum discharge is calculated to Q = u ∗ A = 2.5 ∗ 0.0025 ∗ 0.1 =
6.25 ∗ 10−4 m3 /s. Head losses require at least 1.25m head with discharges
around 6.4 ∗ 10−4 m3 /s. An Ebara water pump fulfils those test require-
ments. The production sheet shows a maximum discharge of 240l/min =
0.250m3 /min = 4.2 ∗ 103 m3 /s at H = 2m.
The pump is not adjustable in revolutions. The discharges can be ad-
justed to the experiment specifications by the two valves. One valve has been
placed above the jet to lower the discharge, the second valve is mounted at
the recirculation hose to increase the jet momentum for taking out the air
each morning.
Two cameras were used in the experiments, namely an actioncam and a
high speed camera. The actioncam was a GoPro Hero3+silver. The camera
can reach a maximum of 50 frames per second with quality of 1028*960i.
Videos of ten minutes were made with this device. The purpose of this cam-
era was to capture the dimensions of cavities at the dynamic stable state
at 0.5m in front of the flume. The second camera was a Fastcam APX
Photron(grey color). The camera has an 8Gb RAM memory with frame
quality of 1024*1024 and 1500 frames per second. The actual width is
140mm, so each millimeter is about 7.3pixels (1024/140 = 7.31pixel/mm).
The film rate per second is set to 1500 for all tests, so that the smallest
particles in high velocity outflow do not blur in a frame. This camera is
mainly used in purpose for the start phase, since the record timespan is just
6.67 seconds.
A table for the pressure gauges were used to test the potential flow in
the normal direction at an impinging level. The static head should not differ
B4
D5 0[mm] Pixel per grain[-]
0.63 4.6
0.93 6.8
1.41 10.3
in case of a plain jet. The pressure gauge is placed at 15mm and 50mm
behind the front wall.
Test
The experiments are as followed prepared in the flume. Fresh water is filled
in the three compartments to a water level of 0.9m. All test attributes are
coupled and set-up for the experiments. This setting rests for 24hours to
adapt the environmental temperature. The pump is started for pushing out
air in the water system every morning. After some time, preparations are
made for a flat soil bed.
Before the main experiment, the jet has to be tested for any outflow
in the normal directions. The set-up is made to prove the plain jet with
the pressure gauges in plastic flat table. The pressure opening is at the
soil bed height of 0.2m at 15mm or 50mm from the front panel. Another
test is a pressure gauge in free flow, which is secured to a fixed standing pole.
After each test run, the bed has to be prepared for the next test. The
soil was flattened to an homogeneous bed state with a peddle. The jet stand
off distance or discharge are adjusted to the test requirements. The cameras
started before the pump is turned on.
If dens bed is tested, a long vibrator is used to dense the soil bed by
adding extra weighted grains to the specified soil volume. The vibrator can
only be turned on/off to one amplitude and frequency. An homogeneous
bed is assumed in the soil in length, width and depth after preparations.
The aim is whether the erosion parameter is related to density of soil.
Bentonite is mixed in dry state of 3.6% of a sand mass. Mixture is put
in the front compartment without water. Flume is slowly filled with water
to 900mm height and let it rest for 24 hours.
B5
B6
B7
B8
Water jet
The aims of frame and water jet were to get quantitative measuring results
and its distribution in several conditions.
So test set-up can vary in:
Outflow velocity
Stand of distance
Grain size
Density
These parameters results in:
Cavity depth
Cavity width
Cavity hill height
Cavity development in time
B9
Appendix C
Validation system
This research deals with a plane jet. An important subject is that the jet
behaves only in 2-dimensional (longitudinal and lateral) directions. The nor-
mal effect has to be avoided to get representable results. Any 3-dimensional
are tested with the impermeable impinging and free jet.
Impinging jet
Two pressure gauges has been placed at 15mm and 50mm beneath the noz-
zle in normal direction in impinging and free jet-setting with the plane jet.
Vertical pressure is measured thirty seconds for getting averaged and band-
width value at stationary phase. This test was done with D50 = 0.63mm
in SOD=10. These values are converted to stagnation pressure in Table C.2.
The vertical middle gauge values are slightly higher for several fluid
outflow in impinging jet than the front gauge. Normal effect can be ne-
glect for this impinging jet. Percentage differences are around 3% of the
measured values between front and middle gauges. Two exception of 13%
and 9% were at the outflow velocity of 0.4m/s and 0.8m/s, but in absolute
values are neglectable difference. Very small leakage between the imping-
ing and the front wall lowers the pressure. This means a small net force is
present from center in normal directions. Also the wall friction could have
a minor effect to the results of the front gauges. Assumption could be made
that bearly any water slipped through the joint between the surface and the
front wall.
C1
Velocity Front Middle Front Middle Difference
[m/s] gauge gauge stagnation head stagnation head [%]
[kP a] [kP a] [mm] [mm]
0.3 0.06 0.06 6 6 0
0.4 0.08 0.09 8 9 13
0.8 0.31 0.34 32 35 9.7
1.2 0.62 0.64 63 65 3.2
1.6 1.17 1.21 119 123 3.4
2.0 2.12 2.16 216 220 1.9
2.4 2.58 2.65 263 270 2.7
Table C.1: Stagnation head beneath plain jet z = 15mm and z = 50mm
Small fluctuation are expected from a turbulent impinging jet. Measured

values are presented in the Table below:

[m/s] gauge gauge fluctuation fluctuation [%] [%]
0.3 0.06 0.06 0.02 0.02 33 33
0.4 0.08 0.09 0.03 0.02 38 22
0.8 0.31 0.34 0.07 0.08 23 24
1.2 0.62 0.64 0.10 0.09 16 14
1.6 1.17 1.21 0.17 0.22 15 18
2.0 2.12 2.16 0.27 0.34 13 16
2.4 2.58 2.65 0.29 0.43 11 16
Table C.2: Stagnation head in normal directions
Turbulence is around 20% of the mean vertical stagnation pressure. The

presents of fluctuation shows a small instability of impinging jet during the
test. Vortices occurs on both side from the centerline with a swaying stag-
nation point. The turbulence indicates some flow processes in the lateral
directions.
Tables C.3 and C.4 represent the results of pressure head of impinging
and free jet at different SOD.
C2
H/d Outflow velocity [m/s] Average static head[m] Fluctuation[m]
10 0.4 0.09 0.02
10 1.2 0.64 0.09
10 2.0 2.16 0.34
20 0.4 0.03 0.02
20 1.2 0.40 0.09
20 2.0 1.22 0.31
30 0.4 0.03 0.02
30 1.2 0.25 0.07
30 2.0 0.77 0.12
Table C.3: Vertical velocity in impinging jet
H/d Outflow velocity Average static head Fluctuation

10 0.4 0.10 0.03
10 1.2 0.77 0.1
10 2.0 2.14 0.26
20 0.4 0.04 0.01
20 1.2 0.38 0.07
20 2.0 1.15 0.19
30 0.4 0.07 0.02
30 1.2 0.31 0.07
30 2.0 0.75 0.11
Table C.4: Vertical velocity in free jet
The flow characteristics of plain impinging jet are considered here in

the lateral directions at SOD = 10. The Table C.5, C.6, C.7 and C.8 are
the measured vertical pressure for test at SOD = 10. Figure C.1 gives the
overview of those tables. A normalised Gaussian distributed is the pressure
distribution beneath the jet.
Free jet
Two pressure gauges are mounted at a firm screw-threads at the bottom of
the flume. These are placed at the same location as the impinging jet from
the previous tests. The gauges have measured the pressure for 30 seconds
at stationary phase as stated in Table below. The tests were executed at
C3
Velocity Theoretical Theoretical Stagnation Stagnation Normalised
[m/s] pressure head pressure head stagnation head
[kP a] [mm] [kP a] [mm] [−]
0.3 0.046 4.6 0.02 2.0 0.444
0.4 0.086 8.2 0.03 3.1 0.375
0.8 0.326 32.6 0.18 18.3 0.563
1.2 0.734 73.4 0.40 40.8 0.555
1.6 1.305 130.5 0.68 69.3 0.531
2.0 2.248 224.8 1.22 124.4 0.553
2.4 2.936 293.6 1.38 140.7 0.479
Table C.5: Stagnation head at plain jet x = 0mm

[kP a] [mm] [kP a] [mm] [−]
0.3 0.046 4.6 0.01 1.0 0.222
0.4 0.086 8.2 0.03 3.1 0.375
0.8 0.326 32.6 0.13 13.3 0.406
1.2 0.734 73.4 0.28 28.5 0.389
1.6 1.305 130.5 0.49 49.9 0.383
2.0 2.248 224.8 0.91 92.8 0.413
2.4 2.936 293.6 0.99 100.9 0.344
SOD = 10 for several outflow velocities.
The deviations are all smaller than 5% of the vertical averaged pressure.
Between the front and the middle gauges are slightly different, which means
the occurrence of normal flow beneath the free jet. The instant measure-
ments show that the water stream varies at all time for all jet momentum
with 15% of the averaged pressure, but smaller velocities has relatively large
deviation.
C4
[kP a] [mm] [kP a] [mm] [−]
0.3 0.046 4.6 0.01 1.0 0.222
0.4 0.086 8.2 0.02 2.0 0.250
0.8 0.326 32.6 0.06 6.1 0.188
1.2 0.734 73.4 0.12 12.2 0.167
1.6 1.305 130.5 0.21 21.4 0.164
2.0 2.248 224.8 0.36 36.7 0.163
2.4 2.936 293.6 0.40 40.8 0.139

[kP a] [mm] [kP a] [mm] [−]
0.3 0.046 4.6 0.00 0.0 0.000
0.4 0.086 8.2 0.00 0.0 0.000
0.8 0.326 32.6 0.04 4.1 0.125
1.2 0.734 73.4 0.05 5.1 0.069
1.6 1.305 130.5 0.07 7.1 0.055
2.0 2.248 224.8 0.09 9.2 0.041
2.4 2.936 293.6 0.10 10.2 0.035

[m/s] gauge gauge fluctuation fluctuation Front-middle fluctuation
[kP a] [kP a] [kP a] [kP a] [%] [%]
0.3 0.08 0.02 0.08 0.02 25 25
0.4 0.10 0.03 0.10 0.03 30 30
0.8 0.33 0.04 0.34 0.05 12 15
1.2 0.76 0.10 0.77 0.10 13 13
1.6 1.16 0.18 1.13 0.19 16 17
2.0 2.10 0.24 2.14 0.26 11 12
2.4 2.51 0.43 2.56 0.48 17 19
Table C.9: Stagnation pressure in normal directions for free jet
C5
0,7
0,6
0,5
2.4m/s
0,4
P/P0[-]
2.0m/s
0,3
1.6m/s
0,2 1.2m/s
0,1 0.8m/s
0
0 2 4 6 8 10
x/d[-]
Figure C.1: Normalised pressure distribution
C6
Appendix D
Test Results
All the results of the experiment for this research can be found here. The ex-
periment has been done carefully to produce accurate outcome. The tables
shows an extensive data, which has been produced during the test period.
Only the good representative results has been added to table for a good
solid conclusion. Test failures, like wrong setting of outflow velocity, is fil-
tered out after thoughtful thinking. Please note that some test has a higher
density of bed or has a mixture with bentonite.
Experimental results
First specific test D50 = 0.63mm, u0 = 2m/s and SOD = 10 are tested
thrice for an accuracy test.
Dynamic bed state occurs at many test, which varies the maximum depth
continuous in time. The deepest cavity depth has been visually measured
after 20 minutes.
D1
Time [s] Test 1 [mm] Test 2[mm] Test 3[mm]
0 0 0 0
0.05 5 4 5
0.10 12 10 10
0.15 17 18 18
0.20 26 27 24
0.25 31 34 30
0.30 35 41 37
0.35 42 45 39
0.40 44 46 42
0.45 41 51 43
0.50 43 52 45
0.55 46 54 47
0.60 53 55 50
0.65 54 57 48
0.70 55 57 51
0.75 55 58 53
0.80 58 58 53
0.85 59 59 54
0.90 59 60 56
0.95 58 61 58
1.0 59 62 58
Table D.1: Observations for accuracy
D2
Test H/d Grain Specific Nozzle Erosion Hill height Cavity depth Cavity width Reynolds
[−] size[mm] gravity [−] velocity [m/s] parameter [−] [mm] [mm] [mm] number [−]
1 30 0.63 1.6 2.00 3.62 31 78 159 5000
2 10 0.63 1.6 1.20 3.75 30 49 136 3000
3 10 0.63 1.6 2.00 6.26 43 82 179 5000
4 10 0.93 1.6 2.00 5.15 27 78 129 5000
5 10 0.63 1.6 2.00 6.26 37 81 167 5000
6 10 1.45 1.6 2.00 4.12 10 65 118 dens
7 10 1.41 1.6 2.00 4.20 19 61 95 5000
8 10 1.41 1.6 2.00 4.20 17 57 92 5000
9 30 1.41 1.6 0.95 1.16 25 34 87 2350
10 30 1.41 1.6 0.55 0.69 11 12 49 1400
11 30 1.41 1.6 0.35 0.43 0 0 0 875
12 30 1.41 1.6 0.70 0.89 15 18 64 1800
13 30 1.41 1.6 0.75 0.92 21 13 63 1875
D3
14 30 1.41 1.6 2.10 2.61 30 55 116 5300
15 30 1.41 1.6 0.45 0.54 6 9 39 1100
16 30 1.41 1.6 0.40 0.52 5 4 34 1050
17 30 1.41 1.6 4.90 6.04 48 143 230 12250
18 30 1.41 1.6 2.40 3.01 28 63 129 6000
19 30 1.41 1.6 5.05 6.22 57 153 241 12625
20 30 1.41 1.6 5.10 6.26 58 155 249 12700
21 30 1.41 1.6 3.60 4.41 39 98 178 8950
22 30 1.41 1.6 3.00 3.72 33 88 156 7550
23 30 1.41 1.6 2.40 2.93 28 64 130 5950
24 30 1.41 1.6 2.00 2.4 27 48 105 5000
25 30 1.41 1.6 0.75 0.90 17 19 63 1825
Table D.2: Test 1-25 results

26 30 1.41 1.6 1.20 1.5 19 32 78 3000
27 30 1.41 1.6 0.25 0.31 0 0 0 625
28 30 1.41 1.6 0.40 0.5 3 1 24 1000
29 30 1.41 1.6 0.80 0.99 16 17 68 2000
30 20 1.41 1.6 2.00 3.0 33 54 112 5000
31 20 1.41 1.2 3.20 4.83 36 104 153 8000
32 20 1.41 1.6 3.70 5.58 36 114 171 9250
33 20 1.41 1.6 4.50 6.79 53 158 231 11250
34 20 1.41 1.6 2.60 3.92 30 73 124 6500
35 20 1.41 1.6 1.20 1.8 25 34 77 3000
36 20 1.41 1.6 0.60 0.91 11 10 43 1500
37 20 1.41 1.6 0.40 0.6 3 4 23 1000
38 20 1.41 1.6 0.30 0.42 0 0 0 700
39 10 1.41 1.6 2.00 4.20 19 57 92 5000
D4
40 10 1.41 1.6 1.20 2.5 9 37 53 3000
41 10 1.41 1.6 0.85 1.66 3 26 42 2125
42 10 1.41 1.6 0.30 0.58 1 2 20 750
43 10 1.41 1.6 1.90 3.66 16 63 82 4700
44 10 1.41 1.6 2.70 5.22 22 85 106 6700
45 10 1.41 1.6 3.30 6.45 30 99 143 8250
46 10 1.41 1.6 4.00 7.83 39 124 184 10050
47 10 1.41 1.6 4.65 9.06 48 153 222 11625
48 10 1.41 1.6 4.80 9.35 56 167 246 12000
49 30 0.93 1.6 2.0 3.0 42 74 176 5000
50 30 0.93 1.6 1.85 2.75 38 61 149 4625

51 30 0.93 1.6 1.2 9.8 24 38 93 3000
52 30 0.93 1.6 1.00 1.46 23 33 96 2450
53 30 0.93 1.6 0.70 1.46 24 29 84 1750
54 30 0.93 1.6 0.40 0.6 6 12 52 1000
55 30 0.93 1.6 0.30 0.37 4 15 55 625
56 30 0.93 1.2 0.15 0.22 3 13 54 375
57 30 0.93 1.6 0.10 0.12 0 0 0 200
58 30 0.93 1.6 2.80 4.16 46 92 189 7000
59 30 0.93 1.6 3.25 4.82 47 107 203 8100
60 30 0.93 1.6 3.70 5.50 49 116 218 9250
61 30 0.93 1.6 4.30 6.39 57 142 247 10750
62 30 0.93 1.6 3.90 5.80 51 126 224 9750
63 20 0.93 1.6 3.90 7.22 46 129 212 9750
64 20 0.93 1.6 4.20 7.77 54 145 236 10500
D5
65 20 0.93 1.6 3.40 6.29 39 108 184 8500
66 20 0.93 1.6 2.0 3.6 30 68 133 5000
67 20 0.93 1.6 1.70 3.15 25 54 112 4250
68 20 0.93 1.6 1.2 2.2 18 39 86 3000
69 20 0.93 1.6 0.85 1.57 15 27 64 2125
70 20 0.93 1.6 0.40 0.7 10 14 44 1000
71 20 0.93 1.6 0.25 0.46 2 2 25 625
72 20 0.93 1.6 0.10 0.19 0 0 0 250
73 10 0.93 1.6 2.0 5.15 27 78 129 5000
74 10 0.93 1.6 1.60 4.19 21 51 87 4000
75 10 0.93 1.6 1.2 3.1 12 38 66 3000

76 10 0.93 1.6 0.80 2.09 8 25 45 2000
77 10 0.93 1.6 0.50 1.31 4 18 34 1250
78 10 0.93 1.6 0.40 1.0 3 10 24 750
79 10 0.93 1.6 0.10 0.26 2 2 14 250
80 10 0.93 1.6 0.05 0.13 0 0 0 125
81 10 0.93 1.6 2.90 7.59 30 94 148 7250
82 10 0.93 1.6 3.25 8.51 34 108 172 8125
83 10 0.93 1.6 3.70 9.68 42 123 199 9250
84 10 0.93 1.6 4.10 10.73 48 143 238 10250
85 30 0.63 1.6 2.0 3.62 31 78 159 5000
86 30 0.63 1.6 1.60 2.94 34 64 154 4000
87 30 0.63 1.6 1.2 2.2 19 45 88 3000
88 30 0.63 1.6 0.80 1.47 25 39 119 2000
89 30 0.63 1.6 0.4 0.7 17 16 37 1000
D6
90 30 0.63 1.6 0.30 0.55 2 13 55 750
91 30 0.63 1.6 0.20 0.37 0 0 0 500
92 30 0.63 1.6 0.25 0.46 0 0 0 625
93 20 0.63 1.6 2.0 4.4 39 79 162 5000
94 20 0.63 1.6 1.60 3.60 34 56 133 4000
95 20 0.63 1.6 1.2 2.7 30 40 114 3000
96 20 0.63 1.6 0.80 1.80 18 32 84 2000
97 20 0.63 1.6 0.50 1.12 18 28 80 1250
98 20 0.63 1.6 0.40 0.9 8 16 61 1000
99 20 0.63 1.6 0.15 0.34 0 0 0 375
100 20 0.63 1.6 2.40 5.40 43 85 182 6000

101 20 0.63 1.6 2.80 6.80 49 106 212 7000
102 10 0.63 1.6 2.0 6.26 43 82 179 5000
103 10 0.63 1.6 1.60 5.09 20 63 118 4000
104 10 0.63 1.6 1.2 3.75 30 49 136 3000
105 10 0.63 1.6 0.80 2.54 13 29 66 2000
106 10 0.63 1.6 0.40 1.3 17 19 67 1250
107 10 0.63 1.6 0.30 0.95 10 6 28 750
108 10 0.63 1.6 0.15 0.48 0 0 0 375
109 10 0.63 1.6 2.40 7.63 39 96 175 6000
110 10 0.63 1.6 2.80 8.90 40 110 197 7000
111 10 0.63 1.6 2.00 6.26 37 85 159 5250
112 20 0.63 1.6 2.00 4.72 41 76 163 5250
113 30 0.63 1.6 2.00 3.86 50 74 193 5250
114 10 0.63 1.6 2.00 6.26 39 81 167 5250
D7
115 10 0.63 1.6 0.30 0.95 11 4 25 750
116 10 0.63 1.6 2.40 7.63 41 96 183 6000
117 30 0.63 1.6 2.40 4.41 55 82 223 6000
118 30 0.63 1.6 2.00 3.86 50 68 183 5250
119 30 0.63 1.6 0.30 0.55 13 2 39 750
120 20 0.63 1.6 2.00 4.72 41 74 162 5250
121 30 1.13 1.6 2.00 2.88 23 53 167 5250
122 20 1.13 1.6 1.60 2.19 18 38 132 4000
123 20 1.13 1.6 1.20 1.64 16 26 110 3000
124 20 1.13 1.6 0.80 1.10 13 15 89 2000
125 20 1.13 1.6 0.25 0.34 1 1 52 625

126 20 1.13 1.6 2.00 3.52 19 62 146 5250
127 20 1.13 1.6 1.60 2.68 14 44 108 4000
128 20 1.13 1.6 1.20 2.01 11 28 88 3000
129 20 1.13 1.6 0.80 1.34 7 19 64 2000
130 20 1.13 1.6 0.30 0.50 3 3 27 750
131 10 1.13 1.6 2.00 4.98 24 58 134 5250
132 10 1.13 1.6 2.00 4.98 19 63 123 5250
133 10 1.13 1.6 1.60 3.79 14 43 86 4000
134 10 1.13 1.6 1.20 2.85 9 31 65 3000
135 10 1.13 1.6 0.80 1.90 6 16 38 2000
136 10 1.13 1.6 0.25 0.59 3 4 18 625
137 10 1.13 1.6 0.40 0.95 5 7 23 1000
138 10 1.13 1.6 2.40 5.69 21 76 148 6000
139 10 1.13 1.6 2.70 6.40 26 94 179 6750
D8
140 30 1.45 1.6 2.00 2.42 17 51 139 5000
141 30 1.45 1.6 2.00 2.54 17 52 138 5250
142 30 1.45 1.6 1.60 1.93 16 41 106 4000
143 30 1.45 1.6 1.20 1.45 11 26 84 3000
144 30 1.45 1.6 0.80 0.97 9 20 75 2000
145 20 1.45 1.6 0.40 0.48 2 4 31 1000
146 30 1.45 1.6 0.25 0.30 0 0 0 625
147 30 1.45 1.6 2.40 2.90 22 62 149 6000
148 30 1.45 1.6 2.80 3.38 26 78 165 7000
149 30 1.45 1.6 2.00 3.11 12 56 127 5250
150 30 1.45 1.6 1.60 2.37 10 43 93 4000

151 20 0.63 1.6 1.20 1.77 6 31 72 3000
152 20 0.63 1.6 0.80 1.18 7 23 59 2000
153 20 0.63 1.6 0.40 0.59 3 9 32 1000
154 20 0.63 1.6 0.25 0.37 0 0 0 625
155 20 0.63 1.6 2.40 3.55 16 68 126 6000
156 20 0.63 1.6 2.80 4.14 17 80 147 7000
157 10 0.63 1.6 2.00 4.12 10 65 118 5250
158 10 0.63 1.6 1.60 3.35 5 49 73 4000
159 10 0.63 1.6 1.20 2.51 3 32 50 3000
160 10 0.63 1.6 0.80 1.67 2 22 38 2000
D9
161 10 0.63 1.6 0.40 0.84 1 9 22 1000
162 10 0.63 1.6 0.25 0.52 1 4 12 625
163 10 0.63 1.6 2.40 5.02 15 72 122 6000
164 10 0.63 1.6 2.80 5.86 17 88 148 7000
165 10 0.63 1.6 3.20 6.69 21 97 164 8000
166 10 0.63 1.6 2.0 6.26 44 86 170 bentonite

Thesis Jets Chi Fung Wu

Uploaded by

Copyright:

Available Formats

Thesis Jets Chi Fung Wu

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thesis Jets Chi Fung Wu

Uploaded by

Copyright:

Available Formats

Erosion Mechanisms

below a vertical plain Jet

Student number: 1373749

An electronic version of this thesis is available at

This experimental research is focused on fail mechanisms in non-cohesive soil

I am grateful to the members of my committee Prof.dr.ir C. van Rhee,

I am thankful to Andrew for keeping me accompany at the laboratory dur-

Also, my gratitudes go to my parents, relatives and friends for their support

List of Figures vii

2 Literature review of f low processes 4

3 Literature review of soils 10

4 Literature review of erosion 14

5 Set-up of the experiment 21

8 Conclusion and recommendations 59

A Grain size analysis A1

B Water system design B1

2.1 Impinging jet (Gauntner et al. [1970]) . . . . . . . . . . . . . 5

3.1 Schematization of bearing capacity (Verruijt and Baars [2007]) 13

4.1 Forces on particle at slope . . . . . . . . . . . . . . . . . . . . 19

5.1 Side view of test setting . . . . . . . . . . . . . . . . . . . . . 22

6.1 Cavity with hill . . . . . . . . . . . . . . . . . . . . . . . . . . 27

7.1 Particle Size Distribution . . . . . . . . . . . . . . . . . . . . 31

A.1 Grain distribution Dorsilit 5F . . . . . . . . . . . . . . . . . . A4

C.1 Normalised pressure distribution . . . . . . . . . . . . . . . . C6

3.1 Classification for gravel, sand and silt . . . . . . . . . . . . . 11

7.1 Sieve result of sand . . . . . . . . . . . . . . . . . . . . . . . . 31

A.1 Sieve results with Dorsilit 5F . . . . . . . . . . . . . . . . . . A3

C.1 Stagnation head beneath plain jet z = 15mm and z = 50mm C2

D.1 Observations for accuracy . . . . . . . . . . . . . . . . . . . . D2

Symbol definition unit

αrec Entrainment coefficient −

 What is known in the literature?

 What happens at the stagnation point in jet erosion test?

1.2 Outline of thesis

In part three, the analysis is presented. This includes a description of

The final part consists of the conclusion and recommendations.

Literature review of f low

2.1 Velocity profiles

2.1.2 Processes around impinging f low

Figure 2.1: Impinging jet (Gauntner et al. [1970])

An impinging flow deflectes from a vertical to horizontal direction as

2.2 Flow development from turbulent jet

Shear stress causes by water is defined in equation 2.4 as:

∂u ∂u ∂u ∂u 1 ∂P ∂2u ∂2u ∂2u

Volume balance can be described as:

Literature review of soils

3.1 Properties of non-cohesive soil

Type Subclass Diameter (mm)

Table 3.1: Classification for gravel, sand and silt

Particle sieving analysis is a useful method to examine sand soil. Prop-

3.1.1 Properties of glass beads

The beads are made in a controlled manufacture. The particles size

3.2 Bearing capacity

Bearing capacity according to Prandtl of soil can be expressed using the

Literature review of erosion

4.1 Erosion mechanics

Several types of force are introduced to an individual grain from reader

4.1.1 Forces on individual particles

 Position compared to other particles

What is known in the literature?

What happens at the stagnation point in jet erosion test?

Position compared to other particles

Fluctuation was observed at the averaged stagnation pressure at all

Stagnation pressure was inversely related to SOD. The correlation is