863

2016,28(5):863-872
DOI: 10.1016/S1001-6058(16)60687-X

Analyzing hydro abrasive erosion in Kaplan turbine: A case study from India*

Anant Kr. RAI, Arun KUMAR

Alternate Hydro Energy Centre, IIT Roorkee, Roorkee, India, E-mail: anantkrai@gmail.com

(Received August 13, 2015, Revised October 3, 2015)

Abstract: Sediment flow through hydro turbine causes erosion of hydraulic components resulting in drop of turbine efficiency, parti-
cularly in hydropower plants of the Himalayan region. The measurement of erosion and monitoring of sediment flow in turbine are
major concerns in erosion study. Attempts have been made to study erosion mainly in Pelton and Francis turbines. In this study, a
simple and effective method has been presented to measure erosion in a Kaplan turbine of a run-of-river scheme Chilla hydropower
plant in foothills of Himalaya. Recent techniques were used to measure sediment parameters like concentration, size, shape and
mineral content. A standard erosion model is applied to estimate the erosion in Kaplan turbine blade, runner chamber and draft tube
cone. A calibration factor has been proposed to apply the erosion model for site specific conditions. It has been found that the outer
trailing edges of the turbine blade and upper runner chamber are most erosion prone zones. Sediment analysis revealed that effective
operation can reduce erosion in turbine components. The estimated erosion values from model are found to be consistent with measu-
red values. Finally, suggestions for design improvements and effective operation of erosion affected hydropower plants are given.

Key words: turbine, sediment, erosion, measurement, IEC 62364, hydropower

Introduction parameters for hydro-abrasive erosion have been ide-

The sediment flow causes reduction in the active
life of reservoirs, erosion of hydro-mechanical equip-
ment and civil engineering structures. It is very diffi-
cult to remove all the silt before passing through hydro
turbine especially in run-of-river hydropower plants.
The silt in Himalayan region contains mainly quartz
(65%-81%), an extremely hard mineral with hardness
7 on Mohs scale causing severe damage to hydraulic
machinery[1-3]. This erosion of turbine parts abets cavi-
tation, pressure pulsation, vibration, mechanical failu-
res and associated frequent shut downs. The outcome
of this is a gradual reduction of the output efficiency
of the plant[3-6]. Increased global deployment of hydro,
a renewable and sustainable energy, calls for the strong
need for hydro-erosion related studies. The hydro-
abrasive erosion due to sediment laden water is a com-
plex process and depends on many factors such as
sediment characteristics, flow characteristics and pro-
perties of substrate materials (hardness, surface mor-
phology, properties of the coating)[7-9]. Although the
* Biography: Anant Kr. RAI (1985-), Male, Ph. D. Candidate
ntified, to which extent these parameters contribute to
the hydro-abrasive erosion is not fully under-
stood[6,7,10].
One of the major difficulties in erosion study is
the measurement of erosion[11,12]. In laboratory set-ups,
the specimens are small in size and flat in shape, which
allow relatively easy measurement of erosion by wei-
ght loss, volume loss, surface roughness or deforma-
tion dimension methods[8,11]. However, erosion mea-
surements in prototype plants are difficult due to large-
sized and non-flat components. The formation of ero-
sion ripples makes the measurement of local thickness
difficult. The thickness reduction is measured using
general calipers during maintenance period of plants[11].
However, a major problem with this kind of erosion
measurement is in locating the reference point for mea-
surement. Moreover, higher accuracy of erosion data
requires erosion measurement at more points on the
profile. Recently, researchers have initiated and mea-
sured the erosion with optical 3-D scanners in proto-
type plants[4,5].
IEC 62364[13] outlines the information about the
erosion study in hydro-turbines but does not provide
the details of method of measurement for erosion thou-
gh factors such as location, minimum number of ero-
sion point measurements and number of turbine blades
864

to be inspected for Francis, Pelton and Kaplan turbine

have been provided. Few terms used in the IEC
62364[13], like the flow coefficient ( K f ) and exponent
of RS ( p) , are not prescribed for Kaplan turbine ero-
sion. Several studies have produced estimates of ero-
sion in prototype Francis[7,9,10] and Pelton turbi-
nes[3-5,14], but studies for Kaplan turbine are still insu-
fficient. Monitoring suspended sediment properties
like concentrations, size distributions, shape and mine-
ral content along with erosion of turbine is still not
common practice, which leads to inadequate informa-
tion on erosion phenomenon[13].
In the present study, a run-of-river hydropower
plant, located on Ganga river, India, is considered for
erosion study. As per literature, the studies have focu-
sed mainly on the erosion of turbine blades rather than
components of turbine i.e., runner chamber, draft tube
etc. The objective of this case study is to present a
simple approach to measure the erosion on the turbine
blades, the runner chamber and the draft tube cone of
a prototype Kaplan turbine unit. The content of this
paper is composed of five parts. The first part deals
with information of hydropower plant under study. The
second part explains the methodology adopted for
measuring erosion in different components of turbine
unit, sediment properties measurement and process of
estimation of erosion model parameters. The third part
presents the results and explains the findings in context
of literature. The fourth part explains the uncertainty
involved in the study and the fifth concludes with steps
to improve the erosion affected hydropower plants.
The current study shall be useful for hydropower plant
managers and researchers to measure erosion and de-
velop a strategy for handling erosion issues effectively.

1. Study area
The Chilla hydropower project is a run-of-river
scheme on river Ganga located upstream of the city of
Haridwar in the foothills of Himalayas and was com-
missioned in the year 1980-1981. There are 4 Kaplan
vertical shaft turbines of 36 MW capacities each with
design head of 32.5 m at 187.5 rpm. A diversion
barrage with head regulator is located at Pashulok,
5 km downstream of Rishikesh town, for diversion of
water to a 14.3 km long, 565 m3/s capacity lined power
channel. Since commissioning of the power plant, the
turbine components faced severe erosion[15]. The high Fig.1 Eroded components of Chilla hydropower plant
erosive silt from river Ganga caused extensive damage
to the underwater parts of hydro turbine units, cooler ponents of Chilla hydropower plant from maintenance
tubes, drainage pump impellers and valve seat. In ini- records during initial days. However, the erosion sce-
tial years of the power plant, the blades were found to nario improved significantly in recent years[16] due to
be extremely eroded and cracked in trailing side after the enhanced hydrological regulation in upper basins
few monsoon seasons[15]. The runner chamber also ob- owing to commissioning of 239.5 m high over river
served extensive erosion requiring heavy maintenance bed earth rock fill Tehri dam in the year 2006. The
after every few years. Figure 1 shows the eroded com- hydropower plant is owned and operated by state utili-

ty Uttarakhand Jal Vidyut Nigam Ltd., which was in-
corporated from the Uttar Pradesh State Electricity
board[17].
To minimize the entry of large size sediment in
power channel, silt excluders have been provided in
the river itself below the under sluice portion of the
diversion barrage. Further, coarser sand from the di-
verted water is removed through a silt ejector constru-
cted in the bed of the power channel 200 m down-
stream of the head regulator. The silt ejector has been
designed to remove all sediment particle size greater
than 0.5 mm[17].
2. Methodology
The erosion of Kaplan turbine blades as well as
other parts of the turbine like runner chamber and draft
tube cone were calculated from maintenance reports
of the Chilla hydropower plant for the years 2007 and
2009[18]. The details of measurement procedures and
various parameter calculations are provided in follo-
wing paragraphs.
2.1 Erosion measurement
During maintenance of turbine, an ultrasonic thi-
ckness measurement device operating on resonance
principle was used to measure the thickness of turbine
components before and after repair[18]. The ultrasonic
thickness gauge was calibrated with step block of ide-
ntical material. The surface profiles of the eroded
zones were restored with the help of templates. The
erosion depth is calculated using Eq.(1).
Tafter repair Tbefore repair = TErosion (1)
where Tafter repair is the thickness of restored profile,
Tbefore repair is the thickness of profile before repair and
TErosion is the erosion depth measured.
Fig.2 A turbine blade number 3 showing the grid lines
In order to create reference points for the measu-
865
rements, all the six turbine blades of a unit were divi-
ded into grid cells drawn with paint before repair and
then, the remaining grid lines after repair were conne-
cted to get the initial grid lines (Fig.2). Only half of the
blade was marked on the pressure side towards trai-
ling end due to erosion prevalent mainly in that por-
tion. The distance between two adjacent grid lines was
decided by the extent of erosion on blades. Figure 3
shows a simplified sketch of a blade with grid cells. A
total of 4 readings of thickness were recorded in each
25 grid cells on pressure side surface of a blade, both
before and after repair. The thickness measurements
were performed on 6 turbine blades with 100 measu-
ring points on each. The numbers of thickness measu-
rement inside a grid cell were averaged to obtain a
single value of erosion depth inside a grid cell. The
erosion depth of a blade was obtained by averaging
erosion depth of grid cells in two zonestrailing side
(A1, B1, C1, D1 and E1 cells) and tip side (A2, A3,
A4 and A5 cells).
Fig.3 Kaplan blade of Chilla turbine showing grid cells. Modi-
fied from Ref.[18]
Fig.4 Grid divisions and different parts of runner chamber and
draft tube cone
866

Table 1 Erosion model associated parameters as per IEC-62364[13]

Parameter Definition Governing equation

The depth of metal that has been removed

S / mm S = W 3.4 PLK m K f / RS p
from a component due to particle abrasion

A characteristic velocity defined for Wrunner = (u2 2 + c2 2 )0.5 , u2 = n D ,

W / ms1 each machine component and used to
quantify particle abrasion damage. c2 = 4Q /( D 2 )

The particle concentration, C , integrated PL PL (Cn K size, n K shape, n K hardness, nTs , n )

PL / kghm3
over the time, T , that is under consideration for all n

K size / mm K size is median diameter of particles, K shape is

Factor that characterizes how abrasion
K shape relates to the size, shape and hardness
of the abrasive particles respectively
K hardness
function of particle angularity, varies from 1 to
2 from round to sharp, K hardness is fraction of
particles harder than the material at the surface

Factor that characterizes the abrasion relates K m = 1 for martensitic stainless steel,
Km
to the material properties of the base material K m = 2 for carbon steel

K f is determined for each case to get agreement

K f / mms3.4 (kghm)1
Factor that characterizes how
the abrasion relates to the water
flow around each component
between calculated and observed erosion value.
Here = (0.4 p ) is used to balance unit of
Kf

Turbine reference size for calculation For Kaplan turbines, RS is

RS
curvature dependent effects of erosion the runner chamber diameter

Exponent that describes the size dependent It is used as a calibration factor to obtain
p good agreement between calculated
effects of erosion in evaluating RS
and observed erosion value

In runner chamber and draft tube cone, the verti-
cal grid lines, which originated from each guide vane
vertically downwards, were equidistant. A total of 24
vertical grid lines corresponding to 24 guide vanes, as
shown in Fig.4, were drawn. For the purpose of mea-
surement, 5 horizontal grid divisions were made in the
runner chamber corresponding to 5 major parts of ru-
nner chamber, i.e., upper cone, middle cone, lower
cone, mild steel (MS) cone and compensating cone as
shown in Fig.4. Similarly, the draft tube cone was di-
vided into 4 horizontal grid divisions, i.e., top, middle
1, middle 2 and bottom. The erosion depth in each
grid cell of runner chamber and draft tube cone was
obtained from averaging 4 measured points.
2.2 Sediment properties measurement
In order to measure sediment concentration, ma-
nual samples were collected from downstream of silt
ejector during the year 2007 and 2009 by the plant
owner. The frequency of sampling was twice daily
during monsoon period from mid-June to mid-October.
However, the sampling frequency increased to three
times daily during high sediment flow conditions to
capture the variation in sediment concentration. The
sediment after passing through the desilting arrange-
ment of silt ejector was considered to be identical with
the sediment inflow into the turbines, as no deposition
of sediment was observed in entire power channel and
forebay tank. The particle size distribution of a sedi-
ment sample was determined from sieve analysis of a
dried sample of 100.5 gm through 5 different size
sieves. The weights corresponding to each sieve class
were used to calculate the median size (d50 ) of sedi-
ment in the sample. Moreover, the sizes of particles
were also re-affirmed visually from scanning electron
microscope (SEM) photographs. Sediment mineral
composition and content was determined from powde-
red X-ray diffraction (XRD) analysis. The sediment
shape was determined from SEM photographs as re-
commended in IEC-62364[13].
2.3 IEC parameters calculation
The IEC model of erosion is applied to calculate
average loss of material from turbine blade and runner
chamber zones for 2007 and 2009. The parameters of
the model and their respective definitions are listed in
Table 1.
 867

The parameters K f and p are calibration fa- from this table that average erosion depth showed va-
[13] riation in time, i.e., 2007 and 2009. The maximum
ctors for the IEC 62364 erosion model to fit site
depth of erosion measured were 6.9 mm, 9.4 mm,
specific conditions for specific eroded component.
9.4 mm, 18.2 mm, 2.5 mm and 12.8 mm on blade
However, IEC 62364[13] does not specify these values
number 1-6 respectively in 2007. The corresponding
for components of Kaplan turbines. Moreover, for a
values of maximum depth of erosion in the year 2009
specific hydropower plant, the factors K f and p were 13.8 mm, 11.1 mm, 10.1 mm, 11.5 mm, 6.5 mm
cannot be calculated separately for a particular com- and 10 mm. Further analysis of the erosion data revea-
ponent. This is because: RS remains the same for a led that the major erosion was found on the blade
hydropower plant component, so, a change in other outlet cells, whereas minor erosion was observed on
variables such as S , PL , W and K m will always the blade radial end, i.e., blade tip. Figure 5 presents
this pattern of erosion of turbine blade surface. Other
keep ( K f / RS p ) factor intact. Hence, a new calibra- cells of the turbine blade showed no erosion. This fin-
tion factor (CF), combining K f and p is introduced ding of the present study is consistent with the findi-
ngs of Kumar and Tyagi[19].
for the model as given in Eq.(2).
Kf
CF = (2)
RS p
The CF was calculated from erosion depth in 2007 and
2009 using equation in Table 1. The obtained CF value
was used to calculate the erosion depth. The erosion
depth obtained from model was then compared with
measured erosion depth and the percentage of the di-
fference between these values of erosion was calcula-
ted using Eq.(3).

The percentage of difference in

TEstimated TMeasured Fig.5 Zones of erosion and shape of turbine blade. Adapted
TErosion = 100% (3)
TEstimated from Tyagi[15], (I: Simple bending about trunnion in Z -
direction, II: Torsion about nodal line (twist can be obse-
rved in viewing direction), III: Chord-wise bending with
where TEstimated is calculated erosion depth from IEC two nodal lines)
62364 and TMeasured is measured erosion depth. Table
Table 3 Turbine blade erosion depth for 2007 and 2009
2 shows relevant data for one turbine unit in Chilla (Source: Calculated from Chilla hydropower plant
hydropower plant. maintenance report for 2007 and 2009[18])
Table 2 Relevant data from Chilla HPP for erosion mode- Turbine Turbine blade erosion
lling blade
number Average erosion of Average erosion of
Parameter Values for one Chilla 5 cells (A1, B1, C1, 4 cells (A2, A3, A4
hydropower plant unit D1 and E1) at blade and A5) at blade
outlet/mm tip/mm
Discharge (Q ) 131 m3/s 2007 2009 2007 2009
year year year year
Diameter ( RS ) 4m
1 1.63 4.42 0.95 2.09
Rotational speed (n) 187.5 rpm 2 2.27 3.84 1.10 3.52
3 2.81 4.79 0.58 1.77
Variable (data available
Sediment concentration (C )
for 2007 and 2009 ) 4 6.41 4.64 0.00 3.40
5 0.46 2.43 0.07 1.36
6 1.88 3.87 2.84 2.75
3. Results and discussion
The erosion of 6 different turbine blades of same Average 2.58 4.00 0.92 2.48
unit encountered same erosion conditions over time. % std. dev. 72 20 103 33
However, the amount of erosion depth of these blades
Max value 18.20 13.80 12.80 11.50
varied considerably as shown in Table 2. It is apparent
868
Table 4 Runner chamber erosion depth for 2007 and 2009[18]
Runner chamber erosion/mm
Upper cone Middle cone
2007 2009 2007 2009
Average 1.89 2.56 3.73 3.88
% std. dev. 83 85 44 66
Max value 14.40 11.50 17.80 12.50
For better understanding, the shape and respective
orientation of the turbine blade apart from zones of
erosion on the blade surfaces have been shown in Fig.5.
To distinguish between these two erosion zones, i.e.,
blade outlet and blade tip, their erosion analysis were
considered separately in this study. Table 3 presents
the average values of erosion in each blade outlet, i.e.,
cells A1, B1, C1, D1 and E1, for years 2007 and 2009.
Cell A1 was the most severely affected due to erosion.
The past experience revealed that the cell A1 was even
removed due to erosion and cracking as shown in
Fig.1 at few occasions[19]. Table 3 also shows the ave-
rage erosion of blade tip, i.e., cells A2, A3, A3 and
A4.
The standard deviation of erosion depth of blade
outlet was found to be 72% and 20% in years 2007
and 2009 respectively as shown in Table 3. The high
values of standard deviation suggest the variation in
erosion depth. The variation of blade radial outlet ero-
sion in 2007 is as high as 103%, which can be explai-
ned in part by the localized cavity formation due to
other processes like cavitation. This finding corrobo-
rates the finding of IEC 62364[13], which stated the
deviation of erosion depth of the order of 40 % in the
runner blade outlet of the Francis turbine. This varia-
tion in erosion values have been also explained in IEC
62364[13] with lack of erosion data. However, as per
authors understanding, the data variation was more
attributed to the nature of metal removal phenomenon
rather than the availability of data. The removal of
surface material might have taken place due to other
activities like cavitation, corrosion or synergy effect of
these processes with erosion. The other reason might
be that there were minor geometrical differences of
the blades after so many years of repairs. Such differe-
nces might be responsible for different flow distribu-
tions and accordingly different erosion. A similar va-
riation in erosion depth values can be seen in parts of
runner chamber as shown in Table 4. Due to erosion
in a few cells in MS cone region, a very high value of
standard deviation is observed, which can be negle-
cted.
Table 4 shows the average, standard deviation in
percent, the maximum and minimum values of erosion
depth in different parts of runner chamber. It is evide-
nt from the Table 4 that the most severely affected
parts of the runner chamber were the upper cone and
Lower cone MS cone Compensating cone
2007 2009 2007 2009 2007 2009
0.66 0.35 0.04 0 3.72 0.49
169 271 480 - 14 131
12.40 8.00 1.7 0 6.10 2.50
middle cone, whereas the MS cone was least affected.
A possible explanation for no erosion in MS cone
might be that the MS cone zone has diverging shape,
as given in Fig.6, which made it less prone to erosion.
The convergent shape and proximity to turbine blades
turbulence increased the erosion of middle cone. Other
three zones, i.e., upper cone, lower cone and compen-
sating cone were cylindrical in shape. The erosion in
the upper cone was higher due to turbulence created
by the nearby turbine blades. The lower cone shows
minor erosion damage, though it is of the order of
10 mm at a few locations. The compensating cone
with material mild steel is more eroded in the year
2007 but mildly eroded in 2009. It is difficult to ex-
plain this variation, but it might be related to operation
of turbine unit in cavitation regime, which resulted in
synergy effect of erosion-cavitation to yield higher va-
lues of erosion depth.
Fig.6 Salient features of the turbines installed in the Chilla
hydropower plant. Adapted from Tyagi[15] (Runner cham-
ber: 1Upper cone, 2Middle cone, 3Lower cone,
4MS cone, 5Compensating cone. 6Runner hub,
7Draff tube cone, 8Runner cone, 9Bottom ring,
10Stay ring, 11Stay vanes, 12Sprial casing,
13Runner servomotor, 14Runner blade)
As per plant maintenance report, eroded runner
chamber with erosion marks are shown as Fig.7. The
erosion marks on this surface of runner chamber were
around 450 with vertical grid lines as shown in Fig.7.
The observed inclination of erosion marks could be
attributed to the relative velocity of water flow with
respect to runner chamber. The rotation of turbine
blade and continuous flow of water in axial direction
resulted in such relative velocity. The draft tube cone,

the portion joining the draft tube with runner chamber,
got negligible amount of erosion due to diverging
shape as shown in Fig.6. The maximum depth in a few
cells was found to be of the order of 1 mm whereas
most of the cells observed no erosion at all. This is
due to the fact that the erosion velocity in draft tube
cone is reduced.
Fig.7 Eroded runner chamber with grid lines
Fig.8 SEM photograph of sediment sample from Chilla HPP
The measured sediment concentration values th-
roughout the year revealed that a maximum concentra-
tion of 3.6 g/l and 5.3 g/l passed through turbine blades
in 2007 and 2009 respectively., From the analysis of
sediment concentration data recorded by the plant
owner, it was found that 47% of the sediment load
passed through only 20 days in 2007 in the study hy-
dropower plant. The corresponding values in 2009
were even higher, i.e., 49.2% of the sediment load in
just 12 d. It is encouraging to compare this figure with
that found by Chakrapani and Saini[20] hat the annual
sediment load of 70%-90% is transported during
14 d-24 d of the monsoon season in the Himalayan
region. Rai[16] presented a detailed study of variation
of sediment concentration in power channel with river
discharge. The mean value (d50 ) of particle size ob-
tained from sieve analysis was 0.174 mm. In the sam-
ple, 98.6% the particles measured had particle size
below 0.5 mm, the designed size exclusion limit of de-
869
silting arrangement of the study hydropower plant.
However, 95.2% of particles were below 0.355 mm in
size, which indicated proper functioning of silt ejecto-
rs. This also showed that mostly fine sand and silt
particles had passed through turbine unit of the study
hydropower plant. The XRD analysis is used to identi-
fy the main minerals and to calculate their respective
percentage presence in the sediment sample. The main
minerals found in the sample were quartz, feldspar,
biotite, muscovite, calcite, rutile, pargasite and clay
minerals, which were similar to the findings of
Virmani and Tyagi[21]. 73% of minerals were found to
be quartz which is harder than the turbine material.
SEM image revealed that most of the particles are
semi-angular in nature as shown in Fig.8.
Due to semi angular nature of sediment particles,
the sediment shape factor given in IEC-62364[13] was
chosen as 1.5. The concentration value was used with
other relevant values to calculate PL and other para-
meters of IEC-62364[13] erosion model as shown in
Table 5.
Table 5 Calculated values for Chilla HPP
Parameter Definition Values for study
hydropower unit
Characteristic velocity
Wrunner in turbine blades
40.63 m/s
Characteristic velocity
Wrunner in runner chamber
18.00 m/s
K size 0.174
K shape 1.5
Defined in Table 1
K hardness 0.73
Km 1
PL2007 PL for 2007 334.89
PL2009 PL for 2009 263.43
The obtained CF values for different components
are presented in Table 6. The estimates of erosion va-
lues as per IEC 62364[13] and respective % deviation
were also provided in the Table 6. The deviations in
erosion estimates were as high as 55 % for turbine
blade tip and 14% for runner chamber middle cone.
This variation might be caused due to many reasons
such as unconsidered processes like cavitation, varia-
tion in microstructure of eroded surface, exposure of
different erosion condition at micro-level and less fre-
quent measurement of sediment properties. The devia-
tion is expected to decrease with conducting more
number of erosion studies. Similar deviations in obtai-
ning flow coefficient are reported in IEC 62364[13].
870

Table 6 Calculated value of erosion in different parts

Erosion zones Calibration factor (CF) Erosion depth as per IEC 62364[13]/mm % deviation in estimated
and measured values of
2007 year 2009 year erosion depth
8
Turbine blade outlet 3.8810 3.83 3.01 33
Turbine blade tip 2.06108 2.04 1.60 55
7
Runner chamber upper cone 4.1410 2.57 2.02 26
Runner chamber middle cone 6.97107 4.33 3.40 14

The value of erosion depth varied considerably
for different components like turbine blade and runner
chamber for same year as well as different years. Even
for the same component of turbine, the magnitude of
erosion depth varied at different locations on the com-
ponent surface. This variation is attributed to the natu-
re of material removal. The erosion depends on many
factors such as flow parameters, sediment parameters
and surface properties. This dependence on many pa-
rameters involves inaccuracy to intrude in different
ways. Some factors like microstructure of the compo-
nent surface, operational variation during the year, ca-
vitation, corrosion etc. are not considered in this study.
Moreover, the size of the components is large and the
amount of erosion is substantially less, like erosion
depth of order of 4 mm for the runner diameter of 4 m.
The erosion phenomenon can be reduced with
different approaches suggested by researchers like
catchment area treatment, design modification of the
minimized. For location with expected erosion issues
minimizing Wrunner will enhance the erosion perfor-
mance of the blades. Another approach to reduce ero-
sion can be to stop plant operation for the days of high
sediment flow. However, for this decision a cost/bene-
fit analysis considering losses due to stoppage of plant
against cost and time of repair has to be carried out.
4. Uncertainty analysis
The uncertainty in the study is computed with
help of uncertainty analysis formulae given by Klein
and Mcclintok[23], which estimates the uncertainty in a
parameter as given below.
Y = f ( X 1 , X 2 , X 3 , X n ) (7)
Y 2
Y
2
Y
2

hydraulic components, material improvement of the Y = X1 + X2 + X3 +

X 1 X 2 X 3
surfaces and settling of sediments before entry to tur-
bine unit[8-11,19,22]. Here, a new approach for designing
the turbine to reduce erosion in Kaplan turbine blade 2

0.5

is discussed with following set of Eqs.(4)-(6). Y

+ Xn (8)
X n
u2 = n D (4)
Y
4Q The percentage of Y = 100% (9)
c2 = (5) Y
D 2
where Y is absolute uncertainty in measurement of
Wrunner = (u2 2 + c2 2 )0.5 (6) Y , which is a function of independent variables X 1 ,

where u2 is peripheral velocity and c2 is axial ve- X 2 , X 3 , X n . The values X 1 , X 2 , X 3 , X n are

locity of water in turbine. The characteristics velocity
(Wrunner ) is very important parameter in calculation of
erosion with exponent of 3.4. From Eq.(6), Wrunner de-
pends on u2 and c2 values, which are obtained from
design considerations for a site i.e., runner diameter
( D) and discharge (Q ) of Kaplan turbine. As evi-
dent from Eq.(4) and Eq.(5), u2 increases whereas
c2 decreases with increase of D . Hence, there will
be an optimum value of D for which Wrunner will be
the uncertainties in measurement of basic independent
variables X 1 , X 2 , X 3 , X n . The main parameters and
calculated uncertainty in their measurement are as
follows: average erosion = 3.37 %, sediment size =
1.92 %, sediment concentration = 2 %, sediment para-
meter PL = 2.77%.
The estimation of sediment properties like size,
shape and mineral composition is based on a single
sample. Due to non-availability of data during the ero-
sion model fitting, it is assumed that sediment with
similar properties flowed throughout the year. More

frequent measurements of these sediment properties
will increase the accuracy of model fitting.
5. Conclusions
This paper has presented a case study of erosion
in Kaplan turbine of run-of-river hydropower plant on
Himalayan river in India. The erosion of turbine com-
ponents was measured by dividing the eroded portions
of the turbine unit into smaller grid cells and by mea-
suring the erosion thickness with an ultrasonic thick-
ness gauge during maintenance. The results of this
study showed high variation in erosion depth in the
various turbine components, especially on each of the
6 blades. This study has showed that the trailing side
of the turbine blade is most prone to erosion whereas
furrows of erosion are caused in tip side. The other
major finding is that the runner chamber portion near
the turbine blade is susceptible to erosion, hence ero-
sion resistant material must be used for this region.
This method can be easily implemented at other power
stations to measure erosion thickness, which will allow
better data collection and hence better information for
effective decision making in erosion issues. This type
of erosion measurement is less costly and can be done
along with in-situ repair of the turbine.
The measurements of another important parame-
ter of erosion besides erosion thickness, i.e., sediment
properties are also presented. The results suggested
that high erosion producing sediment concentration
have passed through within a few days of the year. The
operation of plant can be adjusted during these days to
reduce erosion. The erosion model proposed by IEC
62364[13] has been applied to the study hydropower
plant. The erosion model has been found to work sati-
sfactorily, however, deviations were observed between
calculated and measure values of erosion depth due to
availability of only 2 years of data. For precise erosion
prediction, it is required to conduct more erosion stu-
dies which would enhance the knowledge of influe-
ncing parameters and their quantitative influence.
Acknowledgements
The authors would like to thank Ministry of
Human Development Resource (MHRD), India for the
financial support. Further, this study was made possi-
ble by help of Uttarakhand Jal Vidyut Nigam Ltd. and
Chilla hydropower plant. We thank Mr. Adesh Tyagi,
who provided insight and expertise that greatly assi-
sted the research.
References
[1] DRAGANITS E., GIER S. and HOFMANN C. C. et al.
Holocene versus modern catchment erosion rates at
871
300 MW Baspa II hydroelectric power plant (India, NW
Himalaya)[J]. Journal of Asian Earth Sciences, 2014,
90(4): 157-172.
[2] BAJRACHARYA T. R., ACHARYA B. and JOSHI C. B.
et al. Sand erosion of Pelton turbine nozzles and buckets:
A case study of Chilime Hydropower Plant[J]. Wear,
2008, 264(3-4): 177-184.
[3] SINGH M., BANERJEE J. and PATEL P. L. et al. Effect
of silt erosion on Francis turbine: A case study of Maneri
Bhali Stage-II, Uttarakhand, India[J]. ISH Journal of Hy-
draulic Engineering, 2013, 19(1): 1-10.
[4] WINKLER K. Hydro-abrasive erosion: Problems and so-
lutions[J]. IOP Conference Series: Earth and Environ-
mental Science, 2014, 22(5): 052022.
[5] ABGOTTSPON A., STAUBLI T. and FELIX D. et al.
Hydro-abrasive erosion of Pelton buckets and suspended
sediment monitoring[C]. Proceedings of the Hydro
Vision International. Denver, USA, 2013.
[6] ABGOTTSPON A., STERN P. and STAUBLI T. et al.
Measuring turbine abrasion and efficiency decrease: First
results of the case study at HPP fieschertal[C]. Hydro
2013. Innsbruck, Austria, 2013
[7] THAPA B. S., DAHLHAUG, O. G. and THAPA B. Sedi-
ment erosion in hydro turbines and its effect on the flow
around guide vanes of Francis turbine[J]. Renewable and
Sustainable Energy Reviews, 2015, 49: 1100-1113.
[8] RAI A. K., KUMAR A. and STAUBLI T. Developing a
test rig to measure hydro-abrasive erosion in Pelton turbi-
ne[C]. International Conference on Hydropower for
Sustainable Development. Dehradun, India, 2015.
[9] THAPA B. S., THAPA B. and DAHLHAUG O. G. Cu-
rrent research in hydraulic turbines for handling sedime-
nts[J]. Energy, 2012, 47(1): 62-69.
[10] THAPA B. S., THAPA B. and DAHLHAUG O. G. Empi-
rical modelling of sediment erosion in Francis turbines[J].
Energy, 2012, 41(1): 386-391.
[11] THAPA B. Sand erosion in hydraulic machinery[D].
Doctoral Thesis, Trondheim, Norway: Norwegian Univer-
sity of Science and Technology, 2004.
[12] DUAN C. G., KARELIN V. Y. Abrasive erosion and
corrosion of hydraulic machinery. Series on hydraulic
machinery[M]. London, UK: Imperial College Press,
2003.
[13] International Electrotechnical Commission (IEC). Hydrau-
lic machines. Guide for dealing with hydro-abrasive ero-
sion in Kaplan, Francis, and Pelton turbines[R]. Interna-
tional Electrotechnical Commission (IEC), 2013.
[14] BOES R. M. Real-time monitoring of suspended sediment
concentration and particle size distribution in the head-
water way of a high-head hydropower plant[C]. Procee-
dings of the 33rd International Association for Hydro-
Environment Engineering and Research (IAHR) Con-
gress. Vancouver, Canada, 2009, 4037-4044.
[15] TYAGI S. K. Operating experience of 436 MW Chilla
power station[C]. BHELHydro customers meet, 18th.
Chandigarh, India, 1998.
[16] RAI A. K. A study of sediment transport and its impact
on hydro-turbine in hydropower project[D]. Master
Thesis, Roorkee, India: Indian Institute of Technology,
2011.
[17] Uttarakhand Jal Vidyut Nigam Ltd. Short report of Chilla
hydropower plant (436 MW)[R]. Dehradun, India:
Uttarakhand Jal Vidyut Nigam Ltd., 2004.
[18] Chilla hydropower plant maintenance report. Job comple-
tion reportIn situ repair of underwater parts of unit-4,
Chilla hydropower plant[R]. 2007 and 2009.
872

[19] KUMAR A., TYAGI S. K. Problem of heavy damage to [22] NAIDU B. S. K. Silting problems in hydro power plants
runner chamber of Chilla power station and remedial mea- and their possible solutions[M]. London, UK: CRC Press,
sures[C]. 55th Research and Development Session of 2004.
Central Board of Irrigation and Power. Hyderabad, [23] KLINE S. J., MCCLINTOCK F. A. Describing uncertai-
India, 1991. nties in single sample experiments[J]. Mechanical Engi-
[20] CHAKRAPANI G. J., SAINI R. K. Temporal and spatial neering, 1953, 75(1): 3-8.
variations in water discharge and sediment load in the
Alaknanda and Bhagirathi Rivers in Himalaya, India[J].
Journal of Asian Earth Sciences, 2009, 35(6): 545-553.
[21] VIRMANI S. S., TYAGI S. K. Silting problem in hydro-
power stations in Uttar Pradesh: Case study of Chilla
power station[C]. Workshop on silt damages to equip-
ment in hydropower stations and remedial measures,
CBI&P. New Delhi, India, 1996.