Hydraulics of Sediment Transport: March 2012
Hydraulics of Sediment Transport: March 2012
net/publication/221928160
CITATIONS READS
21 4,571
1 author:
Yousef Hassanzadeh
University of Tabriz
278 PUBLICATIONS 613 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Yousef Hassanzadeh on 31 October 2015.
1. Introduction
This chapter deals with the theoretical and experimental considerations of hydraulics of
sediment transport, involved in identifying the hydraulics formulas for fluid flow and
sediment computation in open channels, and analyzing the flow and sediment
characteristics of the water motion. In general, the field of sediment transport is very
complex, and the engineers in this field should refer to more comprehensive works to better
understand the computational basis.
The hydraulics of flow in a river and its sediment transport characteristics are the two basic
phenomena that determine it’s geometric and plan form shape. There are many variables
that affect the hydraulics of flow and the nature of sediment transport in a natural stream.
As indicated Yang et al. (1996), the Yellow River in China is notorious for enormous amount
of sediment it carries. The total average annual sediment discharge to the sea in China is
about 1.94×109 t. of which 59% comes from the Yellow River. A concentration of 911kg/m3
was measured on September 7, 1977, at the Saumenxia Station near the entrance of the lower
Yellow River.
The condition of incipient sediment motion is important in a large variety of problems
associated with sediment transport. For more than two centuries workers in this field have
attempted to formulate the conditions of incipient motion.
Many research programs have been devoted to the study of the sediment transport in
channels. Extension can be found by Vanoni (1984), Yallin (1963, 1972) and Yang (1972,
1973). Yallin (1963, 1972) developed a bed load equation incorporating reasoning similar
to Einstein (1942, 1950), but with a number of refinements and additions. Yang (1972,
1973) approached the total transport from the energy expenditure point of view and
related the transport rate to stream power. Shen and Hung (1972) derived a regression
equation based on laboratory data for the sand-sized particles. Using the same concept,
Ackers and White (1973) defined sediment transport functions in terms of three
dimensionless groups namely, size, mobility and transport rate of sediments. His
functions are based on flume data carried out with flow depths up to 0.4 meters. One of
the most extensive field and laboratory studies of sediment transport is that by Van Rijn
(1984). He has presented a method which enables the computation of the bed load
transport as the product of the saltation height, the particle velocity and the bed load
concentration.
24 Hydrodynamics – Theory and Model
More recently, Hassanzadeh (2007) based on the dimensional analysis and the Buckingham
П-theorem in reasoning and discussion of bed load phenomenon has presented a
dimensionless semi-empirical equation on the bed load. Comparisons have been made
between this formula with common ones on sediment hydraulics after their unified
descriptions. It is showed that, the latest formula agrees well with the measured data and
could be regarded as optimal, compared with other common formula.
The theoretical equation for the distribution of suspended sediment in turbulent flow has
been given by H. Rouse. Further useful information on the modification of the theory can be
found in Einstein and Chien (1955), Vanoni (1984), Hassanzadeh (1985, 1979), and many
others (Graf 1971, 1998 and Raudkivi 1976). A special study on the hyperconcentrated fluid
mud in rivers is also reported by Mei et al. (1994).
Utilizing the data obtained from the Vanyar gauging station on Adji-chai River in Tabriz,
Iran, and by means of regression analysis on the relationship between the suspended load
and water discharge, Hassanzadeh (2007) has given two empirical equations with high
regression coefficient to calculate the river’s suspended load for the wet and dry seasons.
This chapter presents a comprehensive analysis of the hydraulics of sediment transport. The
effort is focused on those aspects of the study that will produce the best overall results
within given constraint of time. This chapter briefly reviews hydraulics formulas for fluid
flow in open channels and several fundamental sediment computations and contains the
following subjects:
1. Sediment properties
2. Threshold of Particle Transport
3. Channel Roughness and Resistance to Flow
4. The Sediment Load
2. Sediment properties
The dynamic problems of liquid - solid interaction are greatly influenced by the sediment
properties. The description of the latter, however, is exceedingly complex and one is forced
to make many simplifying assumptions. The first of which is the subdivision into cohesive
and non- cohesive sediments.
In cohesive sediments the resistance to erosion depends on the strength of the cohesive
bond between the particles which may far outweigh the influence of the physical
characteristics of individual particles. The problem of erosion resistance of cohesive soils
is a very complex one and at present our understanding of the physics of it is very
incomplete.
The non- cohesive soils generally consist of larger discrete particles than cohesive soils and
the movement of these particles depends on the physical properties of the individual
particles, such as size, shape and density.
It has a direct effect on the mobility of the particle and can range from great boulders, which
are rolled only by mountain torrents, to fine clays, which once stirred uptake days to settle.
The size of particles can be determined in a number of ways. The nominal diameter refers to
the diameter of a sphere of same volume as the particle, usually measured by the displaced
volume of a submerged particle. The sieve diameter is the minimum length of the square
sieve opening through which a particle will fall. The fall diameter is the diameter of an
equivalent sphere of specific gravity δ= 2.65 having the same terminal settling velocity in
water at 24˚C.
Shape. Apart from size, shape affects the transport of sediment but there is no direct
quantitative way to measure shape and its effects. McNown (1951) suggested a shape factor
S.F. = c / ( ab) , where c is the shortest of the three perpendicular axes (a, b, c,) of the
particle. The shape factor is always less than unity, and values of 0.7 are typical for naturally
worn particles.
Density. Density of the particles is important and must be known. A large variation in
density affect sediment transport by segregation, e.g. the armoring effect of the heavy
minerals on dune crests. The mass density of a solid particle, ρs, describes the solid mass per
unit volume. The particle specific weight, S , corresponds to the solid weight per unit
volume of solid. The specific weight of a solid, S , also equals the product of the mass
density of a solid particle, S , times the gravitational accelerating, thus:
S S . g (1-1)
Submerged specific weight of a particle, S' . Owing to Archimedes' principle, the specific
weight of a solid particle, S , submerged in a fluid of specific weight, γ, equals the
difference between the two specific weights, thus,
S' S ( S ) g (1-2)
S S
2.65 (1-3)
The specific gravity is a dimensionless ratio of specific weights, and thus its value remains
independent of the system of units.
In addition it depends on the extent of the fluid in which it falls, on the number of particles
falling and on the level of turbulence intensity. Turbulent conditions occur when settling
takes place in flowing fluid and can also occur when a cluster of particles is settling. For
grain diameter d greater than 2mm, the fall velocity w can be approximated by the
following equation:
w 3. 32 d(mm) (1-4)
Falling under the influence of gravity the particle will reach a constant velocity named the
terminal velocity, when the drag equals the terminal velocity, i. e. difference of the solid and
fluid velocities, Vs- V= w; we obtain the following equation:
d2 w2 d3
CD g( S ) (1-5)
4 2 6
or
4 1
w2 gd( S ) (1-6)
3 CD
Thus, the problem reduces to finding the value of the drag coefficient, CD, for the particle in
question. For spherical particles of diameter d in a viscous fluid of dynamic viscosity μ, the
drag coefficient is fairly well defined. In laminar flow region, for Re < 0.5 and approximately
for up to 1.0, where Re wd / , we have the stokes' solution of
24
FD 3 dw , and C D (1-7)
Re
The Stokes' solution may be considered if either the viscosity of fluid is very large (heavy
oil) or the particle is very small (dust particle). Stokes, in solving the general differential
equation of Navier- Stokes, neglected the inertia terms completely. Oseen (1927) seems to
have been the first who successfully included, at least partly, the inertia terms in his solution
of the Navier- Stokes equation. The Oseen's solution in approximate form is
24 3 (1-8)
CD (1 Re)
Re 16
Goldstein (1929) provides a more complete solution for Oseen approximation and gives the
drag coefficient in the form of
24 3 19 71 (1-9)
CD (1 Re Re 2 Re 3 ...)
Re 16 1280 20480
For Re 2.
The value of drag coefficient, CD, depends strongly on the level of free stream turbulence,
apart from turbulence caused by the particle itself. It also depends on whether or not the
surface of the sphere is hydraulically smooth or rough.
Schiller and Naumann (1933) suggest a formula that gives good results for Re <800, or
Hydraulics of Sediment Transport 27
24
CD (1 0.150 Re 0.687 ) (1-10)
Re
Schiller and Naumann (1933) also multiplied Eq. (1-6) by ( d / )2 and obtained
4 S d 3
C D Re 2 g (1-11)
3 2
Olson (1961) suggests that the drag coefficient can be well represented by the following
equation, for Re <100,
1
24 3
CD (1 Re) 2 (1-12)
Re 16
susp
1 k eC (1-13)
Where
μsusp= Viscosity of the suspension
μ = Viscosity of the liquid medium
ke= Einstein's viscosity constant
C = Volumetric concentration of the solid phase.
ke = 2.5 for C < 2-3%.
m m (1-14)
m
Ft (1-15)
tan
Fn
Where Ft and Fn are the forces parallel and normal to the angle of repose . In this study Ft and
Fn are resultants of the hydrodynamic drag FD, the lift force FL and the submerged weight.
Equation (1-15) for the condition of incipient motion under the action of these three forces
becomes
W sin FD (1-16)
tan
W cos FL
Where angle is the inclination of the bed from the horizontal at which incipient sediment
motion takes place. Fig. 1.1 shows the situation of these three named forces. The drag FD and
lift FL forces may be expressed as
ub2 (1-17)
FD C DK 1d 2
2
ub2
FL C L K 2 d 2 (1-18)
2
Where ub = fluid velocity at the bottom of the channel
CD, CL= drag and lift coefficients, respectively
d= particle diameter
K1, K2 = particle shape factors
ρ = liquid density
Hydraulics of Sediment Transport 29
W K 3 ( S ) gd 3 (1-19)
With K3 being another shape factor and ρs being the solid particle density.
Introducing Eqs. (1-17), (1-18), and (1-19) into Eq. (1-16) yields
Where (ub)c is the critical bottom velocity at which incipient sediment motion takes place. The
quantity of the right- hand side in Eq. (1-20) is referred to as the sediment coefficient A ,
The sediment coefficient A depends on the particles properties, the dynamics of the
flow, the channel slope, and the angle of repose. The angle of repose is the slope angle
formed with the horizontal by granular material at the critical condition of incipient
sliding. The angle of repose which is given by Lane (1953), in Figure 1.2 depends on
particle size.
For more than two centuries the hydraulicians have attempted to formulate the conditions
of incipient motion. One of the earliest relations is due to Brahms (1753). Brahms gave the
critical velocity Vc of water as
Vc kW 1/6 (1-22)
Fig. 1.2. Angle of repose for uniform non cohesive sediment (Lane, 1953)
Fig. 1.3. Erosion and deposition criteria for uniform particles (Hjulström, 1935)
Hydraulics of Sediment Transport 31
uc2
2.50( d h )0.20 (1-23)
( S 1) gd
uc2
3.61(tan cos sin ) (1-24)
( S 1) gd
The latest is by Yang (1973, 1996) who used the conventional drag and lift concepts
combined with the logarithmic velocity distribution and arrived at
Vc 2.5 u*d
0.66 ; 0 70 (1-25)
w log u* d 0.06
Where the numerical constants are from empirical curve fitting. The above equation is
valide for the hydraulically smooth and transition zones and for the hydraulically rough
region the relationship is:
Vc
2.05 , R e* 70 (1-26)
w
It should be noted that, Eq.1.25 yields Vc w when u* d 1.48 .This would means that
particles just a little finer than 100 μm will behave as a fixed boundary because w 0 for
this grain size. It is likely that the formula will give acceptable results for the shear velocity
Reynolds number Re* > 1.5 or 2.
In the turbulent range, for Re*> 70, Yang's expression states that particles on a bed will begin
moving when the average velocity is twice the particle settling velocity.
Where Rh= hydraulic radius, and S = bed slope. In a wide open channel, the hydraulic
radius Rh is equal to the depth of flow h; hence
0 hS ghS (1-28)
The relationship between the friction velocity u* and shearing stress 0 is given by
32 Hydrodynamics – Theory and Model
u* 0 / Rh gS (1-29)
The first research in the mechanics of sediment transport, using the foregoing concept, was
reported by Shields (1936). Shields determined that the critical condition could be related to
two dimensionless parameters: the dimensionless shear stress or the Shields parameter F*,
ud
and the boundary Reynolds number or the shear Reynolds number Re * * . The Shields
parameter F* reflects the ratio of the force producing sediment motion to the force resistance
motion and may be computed by:
0
F* (1-30)
( S )d
Use of the Shields diagram requires that the critical value of the Shields parameter F*c be
determined. To facilitate computations when grain size is known, the dimensionless
diameter d* may be computed by Jullien (1995) equation:
1
( 1)g 3
d* d 2 (1-31)
Where d= sediment diameter, δ= specific gravity of sediment, and = kinematic viscosity.
The Shields diagram (Fig. 1.4) is a widely used method to determine the condition of
incipient motion based on bed shear stress. Points lying above the curve representing the
critical condition correspond to sediment motion, and points below the curve correspond to
no motion. Three somehow distinct zones can be noticed in Fig. 1.4. It should be mentioned
that, many hydraulic engineering problems deal with flow in the turbulent range and
sediment having d >1 mm, the value of Shields parameter may be considered F*c = 0.047 for
critical condition in the range of boundary Reynolds number Re* > 40 (Yalin and Karahan,
1979). For this condition Eq. (1-28) is rearranged as
c F* ( S )d 0.047( S )d (1-32)
c
0.06 (1-33)
( S )d
Zeller (1963) finds this constant to be too high, and he obtained a value of 0.047.
The simplest method for estimating the critical condition for the movement of cohesionless
sediment is using the linear relationship between critical bed shear stress τc and grain size d
given by Julien (1995).
These equations are approximately valid for d50 > 0.3 mm, and can be used as a quick check
against other methods and to help determine the hydraulic roughness regime.
Considerable field data are used by Lane (1953) to establish the critical tractive force
diagram on function of grain size, as shown in Fig. 1.5. This diagram concern of allowable
tractive force in channels as a function of grain size, for the range of fine to coarse
noncohesive materials. As can be seen in Fig. 1.5, the critical shear stress τc for clear water is
considerably lower than for water with a low or high content of sediments. The Lane
diagram summarizes much of the important research done and, hence, should prove very
helpful for the hydraulician engaged in stable channel design.
Fig. 1.5. Allowable tractive force in channels as a function of grain size (Lane, 1955).
It should be noted that the initiation of significant motion within a bed of mixed sediment
can be affected by factors such as hiding of smaller grain by the larger ones and armoring.
Graf (1971) indicated that for materials which are not uniformly sized or contain cohesive
materials, the critical bed shear stress for incipient motion should be higher than predicted
in the Shields diagram. Also, the shear stress is not distributed evenly accros a cross section.
For a straight prismatic trapezoidal channel, Lane (1953) determined the shear stress
distribution shown in Fig. 1.6 and concluded that in trapezoidal channels maximum shear
stress for the bottom and sides is approximately equal to 0.97 hS and 0.75 hS ,
respectively.
Hydraulics of Sediment Transport 35
Fig. 1.6. Shear stress distribution in a trapezoidal channel section (Lane, 1955).
36 Hydrodynamics – Theory and Model
The boundary shear stress distribution for a curved trapezoidal channel was experimentally
measured by Ippen and Drinker (1962) who found that the maximum shear stresses τmax,
occur at the outer toe of the bank immediately downstream of the curve. Shear stress in the
curved reach will be 2 to 3 times greater than the shear in a straight channel. The ratio of
maximum local boundary shear stress, τmax, in a curved reach to the average boundary shear
stress in a approach straight channel τ0, is given by
max r
2.65( c )0.5 (1-36)
0 B
Where rc = centerline radius of the bend, and B = water surface top width at the upstream
end of the curved reach.
WS2 sin 2 a 2 S2
where a = effective area of the particle, τs = unit tracrive force or shear stress on the side of
the channel, Ws = submerged weight of the particle, and θ = angle of the side slope. When
this force is large enough, the particle will move. On the other hand, the resistance to motion
of the particle is equal to the normal force Ws cosθ multiplied by the coefficient of friction, or
tan , where is the angle of repose. Hence, by the principle of friction motion in
mechanics, we have
Solving for the unit tractive force τc that causes impending motion on a sloping surface,
WS tan 2 (1-38)
S cos tan 1
a tan 2
Similarly, when motion of a particle on the level surface is impending owing to the tractive
force aτL, the following is obtained from Eq. (1-37) with θ = 0:
WS tan a L (1-39)
Solving for the unit tractive force τL that causes impending motion on a level surface
WS (1-40)
L tan
a
Hydraulics of Sediment Transport 37
The ratio of τs to τL is called the tractive or shear stress force ratio; which is an important
ratio for design purposes. From Eqs. (1-38) and (1-40) the ratio is
S tan 2
K cos 1 (1-41)
L tan 2
Simplifying
sin 2
K 1 (1-42)
sin 2
It can be seen that this ratio is a function only the inclination of the sloping side θ and of the
angle of repose of the material . The Eq. (1-41), is the form given by Lane (1953), has been
suggested for the use in channel designs. Both of these expressions give the ratio of shear stress
required to start motion on the slope to that required on the level surface of the same material.
By knowing the critical shear stress on the bottom, with the aid of Shields diagram the
critical wall shear stress can be calculated, provided information on the angle of repose is
available.
It is evident that, for a bank to be stable, the angle of the bank θ, must be less than of the
angle of repose , or in other words, for stability reasons, > θ.
V f (K h ) (1-43)
Where τV = unit stream power ( W / m 2 ), and f(Kh) is the erodibility index. Erosion will
occur when τV > f(Kh), but not when τV < f(Kh). The rate of energy dissipation per unit of
bed area is determined by:
V hSV (1-44)
Where = unit weight of water, h = depth, V = velocity and S = energy slope. For loose
granular material, 0.1 < d50 < 100 mm, the relationship between unit stream power at the
critical erosion condition τVc and the erodibility index Kh may be given by
Erosion occurs if unit stream power exceeds the value of τVc. For cohesionless granular
sediment, the erodibility index value may be related to grain diameter by:
38 Hydrodynamics – Theory and Model
Where = angle of repose (Annandale, 1994).Fig. 1.7 shows the eroding and noneroding
conditions based on the erodibility index method given by Annandale (1996).
Fig. 1.7. Eroding and noneroding conditions based on the erodibility index method
(Annandale, 1996).
slope S the hydraulic radius Rh, and a coefficient which expresses the boundary roughness.
The Chézy equation is usually written in the form:
V C RhS (1-47)
1 23
V Rh S (1-48)
n
In which n= Value in the Manning equation shows the roughness or flow-resistance
characteristics of the boundary surface with the dimensional equation of [ n] L1 3T .
By multiplying both sides of the equation by the wetted cross-sectional area A, Manning's
equation can be solved for discharge in SI units:
1
Q ARh2 3 S (1-49)
n
This equation is widely used in open channel water flow computations.
8g
V . Rh S C RhS (1-50)
or
8
V u* (1-51)
With
C 8g , and 8g C 2 (1-52)
Unlike the friction coefficient in the Chézy and Manning equations, the Darcy-Weisbach
friction factor λ is dimensionless, and can be read from Moody diagram.
40 Hydrodynamics – Theory and Model
16
d50
n (1-53)
21.1
Where d50= median sediment particle diameter in meter (Morris and Fan, 1998).
Muler (1943) proposes to calculate the roughness coefficient K which is a result of the
friction of the top layer of the grains,
26
K 16
( m1 3 / sec) (1-54)
d90
Where d90 represents the size of the sediment in the bed for which 90 percent of the material
is finer. This is certainly a reasonable diameter, since the top layer, being made up of the
largest grains and is armoring the bed.
Specific procedures that can be used to determine Manning's n values for channels and
flood plains, described by Arcement and Schneider (1989), are based on Cowan's (1956)
method, research on channel roughness by Aldridge and Garrett (1973) and flood plain
studies.
Cowan's procedure determines roughness values using a base n value which is then
modified to incorporate additional factors which influence flow resistance.
0.05 to 0.9 m. The variation in Manning's n value with respect of channel mean velocity V,
and submerged ratio ( h hv ) , for grass channels is given by Torres (1997) in Fig 1.8. In which
h and hv represents water depth and the vegetation height, respectively.
Fig. 1.8. Variation in Manning n value for grass channels (Torres, 1997)
42 Hydrodynamics – Theory and Model
4.3 Bedforms
When the sediment materials enter motion, the random patterns of erosion and
sedimentation generate very small perturbations of the bed surface elevation. In many
instances, these perturbations grow until various surface configurations known bed forms
cover the entire bed surface. Resistance to flow, which depends largely on bed form
configuration, directly affects water surface elevation in alluvial channel. As flow velocity
increases, an initially flat sand bed develops first ripples and then dunes. With additional
velocity, as illustrated in Fig.1.9, the stream subsequently transitions into a plain bed form,
and finally forms antidunes with standing waves which may or may not crest and break.
Transition from the lower flow regime, where bed forms dominant roughness, to the upper
regime produces a dramatic drop in roughness and will produce a discontinuous discharge
rating curve.
Various bed form configurations and geometry define the boundary roughness and
resistance to flow in alluvial channels. The primary variables that affect bed form
configuration and geometry are the slope of the energy grade line, flow depth, bed particle
size, and particle fall velocity. Flat bed, or plain bed, refers to a bed surface without bed
forms. Ripple shapes is small bed form, vary from nearly triangular to almost sinusoidal.
Dunes are larger than ripples and are out of phase will the water surface waves.
Bedforms are classified into lower and upper flow regimes based upon their shape,
resistance to flow, and mode of sediment transport (Simons and Richardson, 1963, 1966). A
transition zone exists between the two flow regimes, where bedforms range from washed-
out dunes to plain bed or standing waves. The relationship between bedform and stream
power developed by Simons and Richardson (1966) shown in Fig. 1.10 can be used to
determine the flow regime. Because of the non-uniform conditions in natural channels,
different flow regimes and bed forms can coexist in different areas of the same channel.
Hydraulics of Sediment Transport 43
The Simons and Richardson (1966) proposed predictor bed form encompassing both lower
and upper regimes when plotting the stream power (τ0V) as a function of particle diameter
d. Their bedform predictor based on extensive laboratory experiments is quite reliable for
shallow stream but deviates from observed bed forms in deep streams.
Fig. 1.10. Predictional bed form type based on stream power (Simons and Richardson, 1966)
44 Hydrodynamics – Theory and Model
Example 1. 1. Determine critical grain size for noncohesive sediment in a wide channel with
the following characteristics:
T= 20˚C ( 1 10 6 m 2 s )
h= 0.8 m ; V= 1.25 m/s ; n=0.03
Solution. Use Manning's equation and assume Rh=h, since the channel is wide, and
rearrange to compute: V S 1 2 h 2 3 n
Empirical Relationships:
Use Eq. 1.34 by letting τ0=τc, compute d50:
1 1
u* d ( gRhS ) 2 d (9.81 0.8 0.0019) 2 0.02
Re * 2442
1 10 6
The flow regime is strongly turbulent; i.e., Re * 70.
Yang's Method: In the turbulent range rearrange the Yang's Eq. 1.24 to obtain:
Vc
w
2.05
w 2 0.61 2
d ( ) ( ) 0.034m 34mm
3.32 3.32
It should be noted that, for particle in this grain size, viscosity effects are negligible.
Using Shields diagram and letting τ0=τc, in the turbulent flow range, from Fig. 1.4, we have
F*C= 0.047. Rearrange Eq. 1.32 to solve for diameter:
Hydraulics of Sediment Transport 45
c 14.9
d 0.019m 19mm
F*C ( s ) 0.047(25966 9810)
Determine the critical Kh value from stream power by rearranging Eq. 1.45:
Kh 6.3 10 4 1 3
d50 ( )1 3 ( ) 0.034m 34mm
20 tan 20 0.8
Example 1.2. Determine critical grain size for noncohesive sediment in a wide channel at
20˚C ( 10 6 m2 / s ) with the following characteristics:
S 2650 kg f / m 3 ; 1000 kg / m 3
c 7.85
d 0.01m 10mm
F* c ( S ) 0.047(25996 9810)
A similar value is obtained from Fig 1.5, and using Eq. (1.32). For the erodibility index
method, compute stream power:
Kh 1.46 10 4 1 3
d50 ( )1 3 ( ) 0.021m 21mm
20 tan 20 0.8
46 Hydrodynamics – Theory and Model
The above approaches all predict a reduction in the critical grain size as flow depth
increases, with average velocity remaining unchanged. However, mean velocity approaches
will predict a grain size identical to that in Example 1.1 because the mean velocity remain
unchanged between the two examples.
Extensive discussions on this subject has been given by Vanoni (1984), Yallin (1963, 1972) and
Yang (1972, 1973). Yallin (1963, 1972) developed a bed-load equation incorporating reasoning
similar to Einstein but with a number of refinements and additions. Yang (1972, 1973)
approached the total transport from the energy expenditure point of view and related the
transport rate to stream power. Shen and Hung(1972) derived a regression equation based on
laboratory data for the sand-sized particles. Using the same concept, Ackers and White(1973)
defined his sediment transport functions in terms of three dimensionless groups namely, size,
mobility and transport rate of sediments. His functions are based on flume data carried out
with uniform or near uniform sediments with flow depths up to 0.4 meters. One of the most
extensive field and laboratory studies of sediment transport is that by Van Rijn(1984). He has
presented a method which enables the computation of the bed-load transport as the product of
the saltation height, the particle velocity and the bed-load concentration.
qs 0
( , , , S ) 0. (1-56)
gd 3 sd s
qs 0 U *2
1 ; 2
gd 3
( s )d ( s 1) gd
;
3 4 S
s ( s ) s
48 Hydrodynamics – Theory and Model
where s and represent density of solids and fluid respectively, and u* 0 is shear
velocity. Using the Buckingham П-theorem procedure, the bed load equation may be
expressed as:
qs 0
( ) , (1-57)
agd 3 s )d
0 U *2
f . (1-58)
( s )d s
( 1) gd
Based on the properties of the Buckingham П-theorem and neglecting the mild slope S,
qs agd 3 may be obtained as follows:
1 qs 1 qs
. .
3 gd 3 s agd 3
The variation of dimensionless sediment discharge in unit width, qs agd 3 with respect of
f 0 ( s )d has been presented in Figure 1.11, and compared with field measured data
obtained from Vanyar guaging station on Adji - chai river (with a = 1.65 , d = 2.5 – 10 mm,
S = 1.1 *10-3 and width of B=29.9-39.35m).
The Hassanzadeh (2007) bedload equation which agrees closely with the measured data has
been expressed as follows:
qs
24 f 2.5 . (1-59)
3
agd
Comparison have been made between the last proposed equation (1.59) with common ones
on sediment hydraulics after their unified descriptions.
For comaprison reason the common formulae on sediment hydraulics after their unified
description are given in Table 1.1 and Fig.1.11.
As Figure 1.11 and Table 1.1 show, the Hassanzadeh (2007) equation (1.59) agrees well with
the measured data and could be considered as an optimum one compared with the formulae
given by the others (Graf, 1971, Julien, 1995 and Larras, 1972).
Example 1.3. Determine the rate of bed load transport in a rectangular cross section river
with the following hydraulic characteristics:
Average flow depth h= 5.87 m
River bed slope S= 6.5 × 10-4
Hydraulics of Sediment Transport 49
Width B= 46.52 m
Chézy coefficient C= 56.
Median size of bed materials d50= 0.012 m =12 mm
S 2650 kg / m 3 , 1000kg/m 3 .
Solution: Use the Hassanzadeh-type equations. Assume Rh=h, since the river is wide.
Determine the values of: f, a, agd 3 and C* as bellows respectively:
0 hS (1000)(5.87)(6.5 10 4 )
f 0.193
( S ) ( S )d (2650 1000)(0.012)
S 2650 1000
a 1.65
1000
C 56 1000
C* . . 6.75
g S 9.81 2650
Using the Hassanzadeh (2007) type of common dimensionless formulas on the hydraulics of
sediment transport, the rate of bed load has been calculated and given in Table 1.1.
qS
Author Formula QS ( kg / s) s BqS
agd 3
Meyer-Peter qS
8( f 0.047)1.5 0.446 291.05
(1934) agd 3
qS
10C ( f 0.076) f 1.5
3
Shields agd
0.67 436.68
(1936) C
C .
g S
Einstein-Brown qS
23.6 f 3 0.17 110.64
(1942) agd 3
Kalinske qS
10 f 2.5 0.164 106.72
(1947) agd 3
Bonnefille qS
5.5 f 1.5 (4.26 f 0.5 1)1.25 0.393 256.08
(1963) 3
agd
Hassanzadeh qS
24 f 2.5 0.393 256.12
(2007) 3
agd
Table 1.1. Dimensionless existing formulas to estimate bed load (Hassanzadeh, 2007)
Hydraulics of Sediment Transport 51
C hy a Z
( ) , (1-60)
Ca y ha
w w
Z or Z (1-61)
ku ku *
Where C and Ca= concentration of sediment having fall velocity w at vertical distances y
and a above the bed and h=total depth, k=0.4 is von Karman's constant. β is of the order of
unity for fine sediment and appears to decrease with increasing particle size. If β=1 and
K=0.4 the ratio w u 1 corresponds to Z=0.4. Eq.1.60 represents the state of equilibrium
between the upward rate of sediment motion due to turbulent diffusion and the downward
volumetric rate of sediment transfer per unit area due to gravity. This equation can be used
to determine the concentration C at any height y above the bottom relative to the known
concentration Ca at height a above the bed. The value of Z will decrease as fall velocity w
decreases; producing a vertical distribution will exist for each size class. Vertical
concentration distributions for several values of Z are illustrated in Fig 1.12, assigning a
concentration distribution Ca=1 at height a=0.05 h. For given shear stress, Z is proportional
to w, which means that fine-grained material has small values of Z and the particles are
fairly uniformly distributed throughout the depth, whereas coarse grains will be near the
bed.
Equation (1.60) has been found that to give a better fit to the observed distribution β must be
taken to 0.6 (Hassanzadeh, 1985 and 1979).
To formulate river sediments, data from Vanyar guaging station on Adji-chai river, near
Tabriz city, were collected for 8 years. All the suspended sediment data were collected by
the surface water departement of Tabriz water organization. The collected data consist of
the instantaneous water discharges and the corresponding sediment load.
Through the use of the instantaneous values of suspended sediment load, Qs , and water
discharge, Q, regression equations for dry and wet seasons were developed. Fig.1.13
represents the average monthly water discharge Qm (m3/s), and suspended load, Qsm
(tons/day) with respect of time in months from October 1.
The inspection of this figure shows that reasonable correlations do exist between the average
monthly water discharge Qm and suspended load Qsm for a period of 8 years, from october
52 Hydrodynamics – Theory and Model
1974 through september 1982 water years. It is obvious from this figure that at some time
during the year a good correlation between the average monthly water discharge Qm and
the sediment load Qsm exists . However, for 6 months, the sediment load was quite low even
though the water discharge remained fairly high. Thus an exact correlation between water
discharge and sediment load did not exist throughout the whole water year. It should be
stated that, in Fig.1.13 the values of monthly and annual water discharge Qm (m3/s) and the
corresponding suspended sediment load Qsm (tons/day) are related respectively to the left
and right axis.
Fig. 1.13. Relationship between Qsm and Qm with time in months for a period of 1974-1982
(Hasaanzadeh, 2007)
Fig.1.14 shows the relationship between the percentage of the cumulated annual suspended
load moving past the Vanyar guaging station in a given number of days for the four water
years. An examination of Fig.1.14 will reveal that the bulk of the sediment load has been
moved during storm events. Since the number of storm events in a water year is small and
the duration of the storm events are generally short, the bulk of the suspended load passes
through station during the relatively small number of days in a water year.
Fig.1.15 shows the variation of daily suspended load Qs (tons/day) with the corresponding
water discharge Q (m3/s) for the dry and wet seasons of the 1974-1982 water years, for the
Vanyar guaging station on the Adji-chai river. Throught the use of these instantaneous
values of suspended load, Qs , and water discharge, Q, regression equations for the wet and
dry seasons were developed.
The resulted equations are as follows:
For wet seasons (Q>15m3/s):
Fig. 1.14. Annual suspended load carried in a given number of days (Hasaanzadeh, 2007)
Fig. 1.15. Variation of suspended load Qs with respect to water discharge Q (Hasaanzadeh,
2007)
Hydraulics of Sediment Transport 55
6. Nomenclatuers
a reference level; radius of particle; particle effective area; one of the triaxial
dimensions of a particle; immersed sediment specific gravity
A wetted cross-section area
A′ sediment coefficient
b one of the triaxial dimensions of a particle
B water surface top width
c one of the triaxial dimensions of a particle
C concentration of sediment at vertical distance y, Chézy coefficient
Ca concentration of sediment at vertical distance a
CD drag coefficient
CL lift coefficient
C* Chézy coefficient en dimensionless form
d particle diameter
d* dimensionless diameter
F force
FD drag force
FL lift force
Fn normal force to the angle of repose
Ft parallel force to the angle of repose
Fr Froude number
F* Shields parameter
f hydrodynamic-immersed gravity force ratio
g gravitational acceleration
h water depth
hv vegetation height
k constant, von Karmman’s constant
K1, K2, K3 particle shape factor
K Strickler coefficient, tractive- force ratio
Kh erodibility index
n Manning coefficient
P length of wetted perimeter, pressure
Q water discharge
qs rate of bedload in volume per unit time and unit width
Qs rate of bedload in weight per unit time, daily suspended load
Qm average monthly water discharge
Qsm average monthly suspended load
r radius
Re Reynolds number
Re* boundary Reynolds number
Rec critical Reynolds number
Rh hydraulic radius
S bed slope
S.F. shape factor
Sf energy gradient
t time
56 Hydrodynamics – Theory and Model
T temperature
tan coefficient of friction
u velocity in x-direction
u* shear velocity
ub liquid velocity at the bottom of the channel
V mean velocity over depth
Vc liquid critical velocity
Vs velocity of solids
w fall or settling velocity
W grain weight
x coordinate direction
y coordinate direction
z coordinate direction
Z exponent in the suspension distribution
angle of the inclination of the bed form from the horizontal
constant
liquid specific weight
solid particle specific weight
solid particle submerged specific weight
specific gravity
angle of the side slpe
friction factor
liquid dynamic viscosity
dynamic viscosity of the suspension or mixture
liquid kinematic viscosity
kinematic viscosity of the suspension or mixture
dimensionless group
liquid mass density
solid particle mass density
unit shear stress or tractive force
critical unit shear stress
unit shear stress at a solid boundary
unit shear stress on the level surface
unit shear stress on the side of the channel
unit steam power
angle of repose of the materials
7. Subscripts
none liquid phase
b at bed or at bottom
c critical condition, center line
L level surface
Hydraulics of Sediment Transport 57
8. References
Ackers, P., and White, W.R. 1973.”Sediment Transport: New Approach and Analysis”,
Journal of the Hydraulics Division, ASCE, No. HY11.
Annandale, G.W, 1996. ”Prediction of Sediment Distribution in a Dry Reservoir: A
Stochastic Modeling Approach”, PP.I.85-I.92, proc. 6th Federal Interagency
sedimentation Conf. Las Vegas.
Annandale, G.W., 1995. ”Erodibility”, Journal of Hydraulic Research, IAHR, 33(5): 471-
494.
Arcement, G.J.J, and Schneider, V.R., 1989. ”Guide for Selecting Mannings Roughness
Coefficients for Natural Channels and Flood Plains”, USGS Water Supply Paper
2339, Washington, D.C.
Barnes, H.H, 1467. ”Roughness Characteristics of Natural Channels”, USGS water- supply
paper 1849, Washington, D.C.
Carstens, M.R., 1966. ”A Theory of Heterogeneous Flow of Solid in Pipes”, Proc. Am. Soc.
Civil Engrs., Vol. 95, no.Hy1.
Chow, V. T., 1959. ”Open Channel Hydraulics”, McGraw-Hill Book Company.
Cowan, W.L., 1956. ”Estimating Hydraulic Roughness Coefficient”, Agricultural
Engineering, 37 (7): 473-475.
Einstein, H.A., and Chien, N. 1955. “Effects of Heavy Sediment Concentration near the Bed
on Velocity and Sediment Distribution”, M.R.D. sediment series No. 8, University
of California, Institute of Engineering Research and United States Army
Engineering Division, Missouri River, Corps of Engineers, Omaha, Neb.
Fortier.S, and Scobey, F.C, 1426. ”Permissible canal velocities”, Transactions, A.S.C.E, Vol.
89,PP.940-956.
Goldstein, S., 1929. ”The Steady Flow of Viscous Fluid Past a Fixed Spherical Obstacle at 3-
Small Reynolds Numbers”, Proc. Roy. Soc., London, Vol. 123A.
Graf, W. H. and Altinakar, M. S., 1998. “Fluvial Hydraulics” John Wiley, U.K .
Graf, W.H., 1971. “Hydraulics of Sediment Transport” Mc Graw-Hill Book Company.
Hassanzadeh, Y., 1979. “Distribution des vitesses et des concentrations dans un écoulement
diphasique liquide/solide à surface libre”, La Houille Blanche, No.1, Paris.
Hassanzadeh, Y., 1985. “Hydrodynamics of Two-phase Flows” Fifteenth Congress on Large
Dams, ICOLD, Lausanne, Suisse
58 Hydrodynamics – Theory and Model
Hassanzadeh, Y., 2007. “Evaluation of Sediment Load in a Natural River” Journal of Water
International, Vol.32, No.1, Pg.145-154.
Ippen, A.T, and Drinker, P, A, 1962. ”Boundary Shear Stress in Curved Trapezoidal
Channels”, j. Hydraulics Div. ASCE, 88 (Hy5): 143-180.
Julien, P.Y., 1995. ”Erosion and Sedimentation”, Combridge university press, combridge,
U.K.
Lane, E.W, 1935. ”Progress Report on Studies on the Design of Stable Channels of the
Bureau of Reclamation”, proc. A.S.C.E.79. sep.no.280.
Lane, E.W., 1955. ”Design of Stable Channels”, trans. ASCE, 120: 1234-1279.
Larras, J., 1972. “Hydraulique et Granulats” Eyrolles, France
Leliavsky,S., 1955. ”An Introduction to Fluvial Hydraulics”, constable.
McNown, J. S., 1951. ”Particles in slow motion”, La Houille Blanche, 6, no. 5. Leipzig.
Mei, C.C., and Liu, K., and Ng. C., 1994. “Two Models for Roll waves in a Mud Layer”,
Proceeding of the ICHD, 94, Wuxi, China..
Morris, G. L. and Fan J., 1998. ”Reservoir Sedimentation Handbook”, McGraw-Hill, New
York.
Muller, R., 1943. ”Theoretische Grundlagen der Fluss- und Wildbachverbauungen”, Mitteil.
VAWE, Eidgen. Tochn. Hochschule, Zurich, no.4.
Neil,C.R., 1467. ”Mean velocity criterion for scour of coarse uniform bed material”, Int.
Assoc. Hydr. Res., 12th Congress, Fort Collins 3,46-54.
Olson, R., 1961. ”Essential of Engeneering fluid Mechanics”, chap. 11, International
Textbook, Scranton.Pa.
Raudkivi, A. J., 1976. ”Loose Boundary Hydraulics”, pergamon press Ltd., Oxford.
Schiller, L. and Naumann, A., 1933. ”Ueber die grund legenden Berechnugen beider
schwerkraftaufbereituv”, Z. d. V. D. I. 77
Shen, H.W. and Hung, C.S., 1971. ”An Engineering Approach to Total Bed Material Load by
Regression Analysis”, Proc. Sedimentation Symposium, Berkeley.
Van Rijn,L.C., 1984.”Sediment Transport, Part I: Bed Load Transport”, Journal of Hydraulic
Engineering, Vol. 110, No.10,ASCE.
Vanoni, V.A., 1984.“Fifty Years of Sedimentation” Journal of Hydraulic Engineering, Vol.
110, No. 8 ASCE.
Vanoni,V.A.,1946.”Transportation of Suspended Sediment by Water”, Trans.
A.S.C.E.,Vol.III.
Yalin,M.S., 1963.”An Expression for Bed Load Transportation”, ASCE 89, HY3.
Yalin,M.S., 1972. ”Mechanics of Sediment Transport”,Pergamon Press.
Yang,C.T., 1972.”Unit Stream Power and Sediment Transport”, Proc.ASCE,98, HY10, 1805-
26.
Yang,C.T., 1973.”Incipient Motion and Sediment Transport”, Proc.ASCE99,HY10,1679-
1704.
Yang,C.T, 1996. ”Sediment Transport: Theory and Practice”, MC Graw- Hill, New york.