Wear 378-379 (2017) 126135

Contents lists available at ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

Case Study

Erosion wear on Francis turbine components due to sediment ow

Adnan Aslam Noon, Man-Hoe Kim n
School of Mechanical Engineering, Kyungpook National University, Daegu 41566, South Korea

art ic l e i nf o a b s t r a c t

Article history: Sediment particles owing through the turbine components erode the surface in interaction. Erosive
Received 20 July 2016 wear of hydro turbine components generally depends on different parameters such as concentration, size
Received in revised form and shape of the sediments particle, velocity of ow, properties of the base material of the turbine
13 February 2017
components and operating hours of the turbine. Tarbela Dam Hydel Project (TDHP) located in the Hi-
Accepted 14 February 2017
Available online 17 February 2017
malayan range in Pakistan is facing the same problem. The sediments particle have caused damage to the
plant equipment, mainly to the turbine components; stay vanes, guide vanes, runner and draft tube. As a
Keywords: result, these components are disassembled and refurbished almost every year. Analysis have been per-
Sediment particles formed on one of the Francis turbine units to predict the effect of sediment particles concentration, size
TDHP
and shape on erosion rate. Gradual removal of the base material has changed the proles of various
Francis turbine
components of the turbine and also has weaken its structure. One of the major concerns of these effects
Runner
Erosive wear is the continuous loss of turbine hydraulic efciency. The governing equations of uid ow are solved
numerically on an unstructured grid using FEM based software ANSYS CFX. Finnie erosion model is used
to compute average erosion rates. Simulation results are compared with the actual site data. The CFD
analysis showed good agreement with the results of experimental work done previously using similar
kind of geometries and operating conditions.
& 2017 Elsevier B.V. All rights reserved.

1. Introduction Numerous researches have been conducted in the eld of se-

Pakistan has a big potential of hydropower due to the existence of
large quantity of water resources originating from the snow covered
mountains and glaciers of the Himalayan range and the areas exposed
to regular monsoon rainfall. Indus River is one of them that ows
through the whole country. It carries large amount of sediments
containing high percentage of hard abrasive minerals like sand, silt
and clay. This river has great importance in terms of water storage,
irrigation and power generation for Pakistan.
Sediments owing through the dam have resulted in the reduction
of water storage capacity of the reservoir as well as damaged the plant
equipment. Due to this, abrasion and erosion wear occurs at the hy-
draulic turbine components like spiral casing, stay vanes, guide vanes,
runner, draft tube etc. These type of wear mechanisms reduce the life
of the turbine components, as a result difculties in operation arises
and increase in maintenance requirements becomes necessary and
eventually it leads to nancial losses [1]. Teran et al. [2,3] introduced a
new ow distribution methodology for the operation of turbines to
minimize wear. They showed that optimized geometries of the runner
reduced the rate of wear at the point of maximum efciency by 39.5%
but at the cost of a reduction in efciency from 95% to 90%.
n
Corresponding author.
E-mail address: manhoe.kim@knu.ac.kr (M.-H. Kim).
diment erosion in hydro turbines. Some studies have developed
mathematical model to estimate the effect of sediment erosion in
different hydraulic components of the power plant. Thapa et al. [4]
have improved the relationship developed to estimate the sedi-
ment erosion in Francis turbine. A simple erosion model was
suggested to express suitable design, operation and maintenance
plan for Francis runner at a specic site conditions. In another
study [5], it was found that the runner outlet diameter, peripheral
velocity at inlet and blade angle distribution have the major effect
on sediment erosion of Francis runner. It was also detected that
optimization of hydraulic design of blade prole alone can reduce
sediment erosion more than 30%.
Rajkarnikar et al. [6] have emphasized the requirement of new
design methodology to reduce the impacts of sediment erosion in
hydraulic turbines. Padhy et al. [7] presented the experimental
data collected for different parameters, correlations have been
developed for wear rate of Pelton turbine buckets as a function of
critical parameters, i.e. size and concentration of silt particles and
jet velocity.
In the present work, numerical analysis based erosion wear
predictions and as a consequence, efciency losses are presented
for Francis turbine components, especially the runner. After doing
the scaling and similitude analysis, Francis turbine model having a
rated capacity of 35 kW has been used. The details of important
parameters of the turbine are given in Table 1. The simulations

http://dx.doi.org/10.1016/j.wear.2017.02.040
0043-1648/& 2017 Elsevier B.V. All rights reserved.
 A. Aslam Noon, M.-H. Kim / Wear 378-379 (2017) 126135 127

Table 1 coefcient of virtual mass force, CL is the lift coefcient, and Ksf is
Parameters of the prototype and model Francis turbine. the inter-phase drag coefcient.
The one-way and fully coupled options are used depending
Parameter Prototype Model
upon the value of [11], which is dened as the ratio of the
Head [m] 97 3.1 particulate mass per unit volume ow to the uid mass per unit
Discharge [m3/s] 979 6.31 volume ow, Eq. (4) i.e.;
Power [MW] 175 0.035
Runner Diameter [m] 4.77 0.94 rp p
Runner Speed [RPM] 136 121.6 =
r f f (4)
Number of blades 11 11
Number of guide vane 22 22
Number of stay vane 10 10
where, 0.2 is the threshold value. One-way coupling is valid
for sediment particles volume concentration upto 14.86%.

2.2. Erosion modeling

have been performed in operating range from under load (40%
rated load) to over load (140% rated load) with steady state con-
dition and viscous ow turbulence Shear Stress Transport model
(SST) using ANSYS CFX code. The simulation results are compared
and validated with the experimental data.
2. Theoretical aspects
Random motion of sediments particle against a solid surface is
a source of damage for different turbine components. The prop-
erties of the sediments, uid (carrying sediments) and surface
material governs the amount of sediments erosion damage [8].
Stainless Steel is used in the manufacturing of guide vanes and
runner, while carbon steel is used in the manufacturing of stay
vanes. In general, mean velocity of uid, mass, concentration, size,
shape, hardness of the sediments particle and impact angle at
which the particles collide with the surface etc. affects the sedi-
ments erosion of turbo machinery working in water [9].
2.1. Mathematical formulation
In the Eulerian approach, velocity, volumetric concentration
and density of the solid and the carrier phase are calculated
through the volume-averaged, steady continuity Eq. (1) and mo-
mentum Eqs. (2) and (3) for each phase respectively [10].
The wear of a wall due to the erosive effect of particle impacts
is a complex function of particle impact, uid and wall properties.
There is a choice of two erosion models in ANSYS CFX, those of
Finnie and Tabakoff. The choice of one model over another is lar-
gely simulation-dependent. In general, the Tabakoff model pro-
vides more scope for customization with its larger number of input
parameters. But it requires a larger number of coefcients, which
can take more computational time and which could be a source of
generating errors in the numerical scheme. In the current study,
the simplied erosion model of Finnie has been used [12]. This
model is available in CFX for use in combination with the La-
grangian particle tracking and Eulerian-Eulerian multiphase
approaches.
However, erosion is found to vary with impact angle and ve-
locity according to the relationship:
E = kV pn f ( ) (5)
where k is the constant depends on uid properties, like density,
viscosity, slurry temperature etc.
Vp is the impact velocity, n is the velocity exponent, which is
dependent on material of the eroded surface and f() is a function
of impact angle which is discussed below:
1
f ( ) = cos2( ), if > 18.4
3 (6)

2.1.1. Continuity equation f ( ) = sin( 2 ) 3sin2( ), if < 18.4 (7)

2.3. Turbulence modeling

( )
. tt vt = 0 (1)
SST turbulence model is adopted for the ow separation pre-
where t is either solid phase s, or uid phase f. diction, which is the two equations (k- and k-) model. Eq. (7) is
used in the SST model for multiphase ow [13,14]:
2.1.2. Momentum equations
Separate momentum equations are used for liquid and solid m a1k m
Tm =
phases. max( a1m, S mF2) (8)
Momentum equation for uid phase:
where, is the density, k is the turbulent kinetic energy, is the
. ( v v ) = P + .
f f f f f f (
+ f f g + Ksf vs vf ) turbulent frequency, F2 is the blending function, T is the turbulent
viscosity and S is the shear stress. m is used for mixture of sedi-
+ C vmf f ( vs . vs vf . vf ) + CLsf ( vf vs )x( xvf ) (2) ments and water, a1 is a coefcient which determines the fraction
Momentum equation for solid phase: of uid or solid present.

( )
. sfs vfsvs = s P + . s + ss g + Ksf vf vs ( )
( ) (
+ C vmsf vf . vf vs . vs + CLsf vs vf x xvf )( ) (3) 3. Tarbela Dam Hydel Project (TDHP)

where f, f, vf, .f and s s, vs, .s are the volumetric con- TDHP consists of Francis turbines working under gross head of
centrations, densities, velocities, viscous terms of uid and solids 97 m and an average volume ow rate of 979 m3/s. The plant has a
respectively. P is the static pressure gradient, Cvm is the maximum output capacity of 3478 MW, with a total of 14 units.
128 A. Aslam Noon, M.-H. Kim / Wear 378-379 (2017) 126135

Table 2
Sediment particles characteristics.

Sediment particles 1 2 3 4 Average% Particulate diameter, dp (m) Density (kg/m3) Mohs' Hardness Color

Sand 31.09 33.17 31.77 29.37 31.35 95 1535 56 Dark grey

Silt 53.34 54.13 55.33 52.13 53.73 38 1330 5 Light grey
Clay 15.58 12.70 12.70 12.70 14.92 2.4 1170 6 Light brownish

Analysis have been performed on unit number 11, which produces 100
175 MW of power. Numerous strategies from sediments capturing 90
technique at the intake section to the surface hardening of runner
blades have been tried at TDHP to address the sediments erosion 80
problem. However, satisfactory results have not obtained [15]. Every 70

Percent Finer
year maintenance work is carried out during the shutdown period for
60
the eroded turbine blades and other components to keep the power
plant running normally. Sediment erosion studies are carried out after 50
every 3 months through the year to address the seasonal effects. 40
Particulate samples were collected from the outlet section of
30
unit number 11 (Tunnel 2). It has been found that sand, silt and
clay are present in the dam reservoir bed in variable proportions 20
[16]. Table 2 shows the sediments owing through turbine runner 10
are combinations of different particle sizes. The composition,
0
density, color and size of the sediment particles are analyzed un- 1 10 100 1000
der Scanning Electron Microscope (SEM) and Particle Analyzer in
Sediment particles diameter, dp [microns]
Materials characterization and analysis lab at GIK Institute, Paki-
stan. It has been found that 85% of the sediments particle in the Fig. 1. Particle size distribution of sediment sample for TDHP.
Indus River are solid particles like sand and silt, whose hardness
value is almost 6 on the Moh's scale.

3.1. Sediment particles concentration

A substantial increase of sediment particles in the reservoir has

been observed during the recent years. Sediments concentration at
the intake of turbine unit number 11 was quantied. A calibration
study was undertaken at three different times, i.e.in January, May
and September; to ascertain the actual concentration (by weight).
A known amount of sediments sample is weighed. Then the se-
diments are heated up on a heating plate to evaporate the water
present in the suspension. The solid particles are weighed after-
wards to get mass based concentrations of the sediment particles.
It is found that sediments concentration is much higher during
Fig. 2. The axes of an irregular shape particle.
the monsoon period from July-Oct, than in winter from Nov-Mar
season as shown in Table 3. mutually perpendicular axes, b the third axis, c the shortest
of three mutually perpendicular axes as shown in Fig. 2.
3.2. Sediment particles size and shape Nearly 30 samples examined under SEM for Albertson shape
factor, the results are presented in Table 4.
Sieve curve is plotted in Fig. 1 which shows that 90% of the With a shape factor value of 0.65, it is concluded that sediment
particles entering the turbine are less than 100 m in diameter particles are sub-rounded in shape.
and mean diameter (d50) of the particles entering runner is 40 m.
The sediment particles at TDHP are studied and analyzed. Al-
bertson shape factor is used to calculate the values of shape factor, 3.3. Drop in runner efciency
A [17] as shown in Eq. (8):
Thermodynamic efciency measurements carried out on turbine
c
A = unit number 11 by utilizing equation number (10). Fig. 3 shows the
( ab) (9)
Table 4
where A is Albertson shape factor, a the longest of three Examination using SEM for Albertson shape factor.

Table 3 Sample space A Mean A

Sediment particles concentration as a function of water head.
3 0.47 0.65
Head Months Concentration [PPM] 7 0.75
4 0.51
Low NovMar 500010,000 5 0.89
Medium AprJune 10,00030,000 5 0.34
High JulyOct 30,00060,000 6 0.78
 A. Aslam Noon, M.-H. Kim / Wear 378-379 (2017) 126135 129

90 module, CFX mesh for the computational domain as shown in

Fig. 5. Unstructured grid is used which consists of tetrahedron,
85
pyramid and prism elements. This type of mesh is anticipated to
Hydraulic Efficiency [%]

80 give more exibility for automatic generation of grid around in-

tricate geometries. Boundary layer near the blade surface is cap-
75
tured through the layer mesh. The ination option has been used
70 with the rst layer thickness of 0.00030 m as measured by Eq. (11)
[18]. Fifteen layers have been used with a growth rate of 10%.
65
y = L. y+ .74.ReL13/14 (11)
60 10th August
5th November Sensitivity of the solution has been tested for each numerical
55
mesh as shown in Table 5. To decrease the inuence of mesh size
50 on the numerical results, a grid independency test has been con-
1 2 3 4 5 6 7 8 9 10 ducted for three different mesh sizes. Minimum value of error is
Volume Flow rate, Q [m3/s] found for the 26.45 million elements grid, so this value is used in
all the simulation cases.
Fig. 3. Turbine efciency measurements at TDHP.

4.3. Parameters investigated

results of the efciency at TDHP within the interval of 12 weeks of
monsoon period to assess the erosion damage. Within the operational
Under the present work, three operating conditions: (i) over
period of one monsoon season, the average loss of efciency recorded
load ( 36), (ii) rated load ( 20) and (iii) under load (
at Best Efciency Point (BEP) is about 4%.
6), covering the whole working range of the turbine operation.
P The rated load condition is the best efciency point of the turbine.
=
gQhT (10) The important parameters on which erosion rate depends are
concentration, size and shape of sediment particles. The values of
Here, P T, is actual power produced, is the water density, g is the
these important parameters used are obtained from the plant site
gravity, Q is the volume ow rate and hT is the available turbine head.
and are given in Table 6.

4.4. Simulation procedure

4. Numerical analysis method
One-way coupling is used as the sediment particle concentra-
tion is less than the threshold value.
4.1. Francis turbine model
Four main components; spiral casing, stay vanes and guide
vanes, runner and draft tube for the Francis turbine are analyzed
under the present work which are connected through the domain
interfaces. Design Modeler module of ANSYS has been used for 3-D
model of the Francis turbine as shown in Fig. 4.
4.2. Mesh generation
The mesh is generated for each part by using the meshing
The simulations were carried out for all the four components of
turbine as shown in Fig. 6. Finite Element Method (FEM) is used
which gives strong and efcient prediction for the rotating ow in
Francis turbine. Linearized mass and momentum differential
equations are solved simultaneously through high resolution
scheme in space [19].
Mass ow rate is taken at inlet boundary for spiral casing and
pressure at outlet boundary for the draft tube respectively for the
ow analysis [20]. The particle number rate is determined from
the mass ow rate assigned to the representative particles divided
by the mass of an actual particle, which equals to 6310000 parti-
cles. However, non-slip ow conditions are used for walls.
Spiral casing, stay vanes, guide vanes and draft tube are taken
as stationary components while runner is rotating with angular
speed of 121.6 rpm. All the components are connected by domain
interface.

4.5. Hydraulic efciency

The hydraulic efciency is dened by the Eq. (12) [21] and it is

computed in ANSYS CFX by

h =
( T )
( P P ) g
ti to (12)

where, Pti and Pto are the pressure at the inlet of casing and the
outlet of draft tube respectively, is the density of uid and g is
the gravitational acceleration, is the angular speed and T is the
torque produced by runner respectively.
The normalized hydraulic efciency loss (loss_ero) of the tur-
bine was evaluated through the data obtained for high erosion and
low erosion conditions in two different season's i.e.; in January and
September by using the following equation;

Fig. 4. Computational model of Francis turbine.

loss_ero = (
low erosion )
high erosion /low erosion (13)
130 A. Aslam Noon, M.-H. Kim / Wear 378-379 (2017) 126135

Fig. 5. Model mesh (a) complete domain (b) enlarged view of the runner (c) ination layers shown near the runner wall.

Table 5
Mesh Independence test details.

Mesh % difference from experi- % difference from experi-

Size in mental data found in literature mental data obtained from
Millions for Erosion rate density site for efciency loss

1) 21.0 10.7 15.4

2) 26.5 3.3 9.8
3) 29.4 7.9 13.4

Table 6
Important parameters.

Sr. No. Parameters Value

Fig. 6. Boundary conditions at the inlet, exit and wall.
1 Sediments particle size, dp [microns] 10, 20, 30, 40, 50, 60, 70, 80
2 Sediments particles shape factor, [] 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 5.1. CFD model validation
3 Sediments concentration, Cw [%] 0.5, 1, 2, 3, 4, 5, 6

The present study is validated through the experimental work

published in the literature, details are presented in Table 7. Data
used by Rajkarnikar et al. [5] is compared with the simulations
data; comparable values for head, efciency and volume ow rate
5. Results and discussion are used for validation. The maximum error is found to be 5.48%.

The CFD analysis has been carried out to study the effect of
sediment particles concentration, size and shape on erosion rate
and consequently, the efciency loss for the Francis turbine. Main
turbine components like stay vanes, guide vanes and runner are
damaged as a result of erosion wear. Simulation results are com-
pared qualitatively through the actual site data for Francis turbine
single components, i.e. stay vane, guide vane and blade of the
runner vane.
5.2. Effect of erosion on different components
5.2.1. Stay vane
Fig. 7(a) shows the sediment ow velocity around the stay
vane. The erosion loss on the stay vane is compared with the CFD
analysis as shown in Fig. 7(b) and (c). Erosion damage is found to
be higher close to the upper and lower portions of the leading
 A. Aslam Noon, M.-H. Kim / Wear 378-379 (2017) 126135 131

Table 7 pressure side of the runner, consequently higher erosion is found

Validation of numerical results through experimental data published in the Fig. 9(c), especially at the middle of the blade. This is due to the
literature.
blade proletail vortex ow, which leads to higher erosion rate
Parameter investigated Data used in Data used in density in the blade outlet. Minor erosion spots were also ob-
experiment simulations served at the leading edge, lower part of the outlet area and on the
suction side of the blades. As runner is the most critical compo-
Head, h [m] 201.5 97
nent for erosion damage, a detailed analysis has been carried out
Efciency, [dimensionless] 96 94
Volume ow rate, Q [m3/s] 2.35 6.31 for runner blades. Average values of the erosion rate density are
Experimental value of Erosion rate 92.16 plotted against sediments particle concentration, size and shape in
density, E [mg/hr m2] the following sections.
Numerical value of Erosion rate 98.67
density, E [mg/hr m2]
% Error 6.59 5.3. Effect of sediment particles concentration

The variations in erosion rate density with concentration in

edge of the stay vane than at the middle.
5.2.2. Guide vane
The erosion damage on the guide vane at actual site is com-
pared with the CFD analysis as shown in Fig. 8. Higher velocity is
observed Fig. 8(b) at the leading edge of the guide vane, as a result
some erosion spots on the blade pressure side are found Fig. 8(c),
but the suction side has minimum erosion. Only minor erosion
damage affects are observed at the lower cover compared with the
hub.
PPM at three different sediment particles size with constant semi-
round shape are shown in Fig. 10. It is apparent that the erosion
rate density increases almost linearly on increasing sediment
concentration for all the three silt sizes. However, it increases with
higher rate with larger values of silt concentration.
Fig. 11 shows the variation in hydraulic efciency loss as a
function of sediments concentration. The efciency loss increases
with concentration at three different silt sizes with constant semi-
round particles. The rate of increase is higher for larger values of
silt concentration.

5.2.3. Runner vane 5.4. Effect of sediments particle size

The erosion wear pattern on runner blade is observed during
shutdown period at TDHP and is similar to the pattern which has The change in erosion rate density with sediments particle size
been predicted from the simulations as shown in Fig. 9(c). in microns at three different sediments concentration with con-
Higher velocity is seen Fig. 9(b) at the outlet area of the stant semi-round shape is presented in Fig. 12. It is evident that

Fig. 7. Stay vanes with covers, (a) Velocity Prole (b) at actual site (c) CFD analysis.
132 A. Aslam Noon, M.-H. Kim / Wear 378-379 (2017) 126135
Fig. 8. Guide vanes with covers, (a) Velocity prole (b) at actual site, (c) Erosion rate density prole.
within the range of sediment size i.e., 10 80 m investigated, the
erosion rate density increases with the increase in sediment size.
However, the rate of increase of erosion rate density is not same
for different values of sediments concentration. The increase in
erosion damage is higher at larger values of silt concentration. The
reason is the strong turbulence ow, which gains erosive energy,
causing damage to the blade surface.
Fig. 13 shows the variation in hydraulic efciency loss as a
function of sediments particle size. The efciency loss increases
with silt size at three different sediments concentration with
constant semi-round particles. Efciency loss is found to be more
for higher values of silt size.
5.5. Effect of sediments particle shape
The change in erosion rate density with sediments particle
shape at three different silt sizes with constant silt concentration
is shown in Fig. 14. A particle shape factor equals to one signies
that the shape of the particle is close to the spherical one or vice
versa. Erosion damage decreases as sediment particle shape factor
is increased. It is obvious that maximum erosion occurs in the case
of nonspherical shape of the particles.
Fig. 15 shows the variation in hydraulic efciency loss as a
function of sediments particle shape. The efciency loss decreases
with increase in particle shape factor at three different silt sizes
with constant sediment particles concentration. It decreases
sharply for higher values of silt size.
5.6. Effect of wear on power for different concentrations, particle
sizes and shapes
Erosion rate density is plotted against power for three different
values of sediment particles concentration, size and shape.
Fig. 16 shows the variation in wear against power for three
different sediment particles concentration. As wear rate increases
power decreases for all the sediment concentrations.
For 10,000 PPM concentration, the wear rate is low and power
generated has maximum value of 32.2 kW. For other two values of
concentration the wear rate is higher and power produced has
lower values.
Fig. 17 shows the variation in wear against power for three
different sediments particle size. For 20 m particle size, the wear
rate is low and power generated has maximum value of 32.4 kW.
For other two values of sediment size the wear rate is higher and
power produced has lower values.
Fig. 18 shows the variation in wear against power for three
different sediments particle shape. For particle shape factor,
0.47 and 0.78, the wear rate is low and power generated has
minimum values of 32 and 32.1 kW. For 0.65, the wear rate is
lower and power produced has higher value 32.4 kW.
6. Conclusions
Erosion wear at the Francis turbine components, particularly at
the runner is a complex phenomenon and needs to be examined in
 A. Aslam Noon, M.-H. Kim / Wear 378-379 (2017) 126135 133

Fig. 9. Runner at TDHP, (a) Velocity prole, (b) at actual site, (c) Erosion rate density prole.

500 5
Average erosion rate density, E [mg/hr.m2]

Normalized efficiency loss, loss_ero [%]

450 dp= 20 microns 4.5 dp = 20 microns

400 dp= 40 microns 4 dp= 40 microns

350 3.5 dp = 80 microns

dp = 80 microns
300 3

250 2.5

200 2

150 1.5

100 1

50 0.5

0 0
0 10000 20000 30000 40000 50000 60000 0 10000 20000 30000 40000 50000 60000 70000

Sediment particles concentration, Cw [PPM] Sediment particles concentration, Cw [PPM]

Fig. 10. Erosion rate density against different values of sediments particle Fig. 11. Normalized efciency loss against different values of sediments particle
concentration. concentration.

order to predict the erosion and efciency losses. Under the pre- period from July-Oct, than in winter season.
sent research, the effect of sediments particle concentration, size  SEM and particulate analysis showed that mean diameter (d50) of the
and shape for erosion and efciency loss has been investigated particles entering runner is 40microns and are semi-round in shape.
through simulations.  During the monsoon period, the average efciency loss is re-
On the basis of numerical investigations, the following con- corded as 4% at the Best Efciency Point (BEP).
clusions are drawn:  Effect of erosion damage is observed at different components of
the Francis turbine.
 Sediment concentration is much higher during the monsoon
134 A. Aslam Noon, M.-H. Kim / Wear 378-379 (2017) 126135

400 4.5
Conc. = 10000 PPM
Average erosion rate density, E [mg/hr.m2]

Normalized efficiency loss, loss_eff [%]

4 dp = 20 Microns
350
Conc. = 30000 PPM
dp = 40 Microns
3.5
300 Conc. = 50000 PPM dp = 80 Microns
3
250
2.5
200
2

150 1.5

100 1

0.5
50
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0 10 20 30 40 50 60 70 80 90
Sediments particle shape factor,
Sediment particles size, dp [microns]
Fig. 15. Normalized efciency loss against different values of sediments particle
Fig. 12. Erosion rate density against different values of sediments size. shape.

4
400
Conc. = 10000 PPM
Average erosion rate density, E [mg/hr.m2]
Normalized efficiency loss, loss_eff, [%]

3.5 Conc. = 10000 PPM

350 Conc. = 30000 PPM
Conc. = 30000 PPM
3 300 Conc. = 50000 PPM
Conc. = 50000 PPM
2.5 250

2 200

150
1.5
100
1
50
0.5
0
0 30 30.5 31 31.5 32 32.5
0 10 20 30 40 50 60 70 80 90 Power, P [kW]
Sediment particles size, dp [microns] Fig. 16. Erosion rate density against power for different values of sediments particle
concentration.
Fig. 13. Normalized efciency loss against different values of sediments particle
size.

400
450
Average erosion rate density, E [mg/hr.m2]

dp = 20 microns dp = 20 microns
Average erosion rate density, E [mg/hr.m2]

dp = 40 microns
400 350 dp = 40 Microns
dp = 80 microns
dp = 80 Microns
350 300

300 250

250
200

200
150
150
100
100 30 30.5 31 31.5 32 32.5 33
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Power, P [kW]
Sediment particles shape factor,
Fig. 17. Erosion rate density against power for different values of sediments particle
Fig. 14. Erosion rate density against different values of sediments particle shape. size.

 It has been found that erosion and efciency loss both increases
with increase in sediment particles concentration and size.  Numerical results are compared and validated through pub-
 It has also been found that erosion and efciency loss both lished experimental and site data. The results are found to be in
decreases with increase in particle shape factor. good agreement.
 A. Aslam Noon, M.-H. Kim / Wear 378-379 (2017) 126135 135

450 handling sediments, Energy 47 (2012) 6269.

Average erosion rate density, E [mg/hr.m2]

Sigh = 0.47 [5] B.S. Thapa, B. Thapa, O.G. Dahlhaug, Empirical modelling of sediment erosion
in Francis turbines, Energy 41 (2012) 386391.
400 Sigh = 0.65 [6] B. Rajkarnikar, H.P. Neopane, B.S. Thapa, Development of rotating disc appa-
ratus for test of sediment-induced erosion in Francis runner blades, Wear 306
Sigh = 0.78
350 (2013) 119125.
[7] M.K. Padhy, R.P. Saini, Effect of size and concentration of silt particles on
erosion of Pelton turbine buckets, Energy 34 (2009) 14771483.
300
[8] M.S. Bingley, D.J.O. Flynn, Examination and comparison of various erosive
wear models, Wear 258 (2005) 511525.
250 [9] M. Abid, A.A. Noon, H.A. Wajid, Erosion study of tunnel 1 of Tarbela Dam, IJST
Trans. Mech. Eng. 38 (2014) 253261.
[10] A.A. Noon, M.H. Kim, Erosion wear on centrifugal pump casing due to slurry
200 ow, Wear (2016) 103111.
[11] G. Brown, Use of CFD to predict and reduce erosion in industrial slurry piping
150 system, in: Proceedings of the Fifth International Conference on CFD in the
Process Industries. CSIRO, Melbourne, Austrailia. December, 2006, pp. 1315.
[12] I. Finnie, Some observations on the erosion of ductile metals, Wear 19 (1972)
100 3644.
30 30.5 31 31.5 32 32.5 [13] C. ANSYS, Solver Modeling Guide. Release16.0. 20062016.
Power, P [kW] [14] P.P. Gohil, R.P. Saini, Effect of temperature, suction head and ow velocity on
cavitation in a Francis turbine of small hydro power plant, J. Energy 93 (2015)
Fig. 18. Erosion rate density against power for different values of sediment particles 613624.
shape. [15] M. Abid, A.A. Noon, Turbulent ow simulations through Tarbela Dam Tunnel 2,
J. Water Resour. Prot. (JWARP) 2 (2010) 507515.
[16] M. Abid, A.A. Noon, Simulation of turbulent ow through Tarbela Dam Tunnel
References
3, IIUM Eng. J. 11 (2) (2010) 201224.
[17] Slurry Systems Handbook, B.E. Abul Naga. Mazdak International Inc. Abbots-
[1] V. Levy, The solid particle erosion behavior of steel as a function of micro- ford, USA, 2002.
structure, Wear 68 (1981) 269287. [18] S.M. Salim, S.C. Cheah, Wall y strategy for dealing with wall-bounded tur-
[2] L.A. Teran, R.D. Aponte, J. Munoz-Cubillos, C.V. Roa, J.J. Coronado, J.A. Ladino, F. bulent ows, in: Proceedings of the International Multi Conference of En-
J. Larrahondo, S.A. Rodrguez, Analysis of economic impact from erosive wear gineers and Computer Scientists, 2009.
by hard particles in a run-of-the-river hydroelectric plant, Energy 113 (2016) [19] R.J.K. Wood, D.W. Wheeler, Design and performance of a high velocity air
11881201. sand jet Impingement erosion facility, Wear 220 (1998) 95112.
[3] L.A. Teran a, C.V. Roa, J. Muoz-Cubillos, R.D. Aponte, J. Valdes, F. Larrahondo, S. [20] S. Jain, R. P. Saini, A. Kumar. CFD approach for prediction of efciency of Francis
A. Rodrguez, J.J. Coronado, Failure analysis of a run-of-the-river hydroelectric turbine, in: Procedings of IGHEM, AHEC IIT Roorkee, 2010.
power plant, Eng. Fail. Anal. 68 (2016) 87100. [21] Z. Carija, Z. Mirsa, S. Fucak, Validation of Francis turbine CFD simulation,
[4] B.S. Thapa, B. Thapa, O.G. Dahlhaug, Current research in hydraulic turbines for Strojarstvo 50 (2008) 514.