Renewable Energy 130 (2019) 667e676

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Experimental study on the vibrational performance and its physical

origins of a prototype reversible pump turbine in the pumped hydro
energy storage power station
Yuning Zhang a, *, Xianghao Zheng a, Jinwei Li b, Xiaoze Du a
a
Key Laboratory of Condition Monitoring and Control for Power Plant Equipment (Ministry of Education), School of Energy, Power and Mechanical
Engineering, North China Electric Power University, Beijing, 102206, China
b
China Institute of Water Resources and Hydropower Research, Beijing, 100048, China

a r t i c l e i n f o a b s t r a c t

Article history: In the present paper, the vibrational performance and its physical origins of a prototype reversible pump
Received 15 February 2018 turbine in the pumped hydro energy storage power station is experimentally investigated. Specifically,
Received in revised form the vibrations of the unit in the X, Y and Z directions of the top cover, the upper and the lower brackets
2 May 2018
were all measured for three water heads (from 48% to 90% in terms of non-dimensional values) and nine
Accepted 14 June 2018
Available online 14 June 2018
load conditions (from 34% to 96% of the rated power) together with the pressure measurement at several
typical monitoring points. Based on our analysis, it was found that the vibrations of the top cover are
mainly induced by the fluid flow inside the reversible pump turbine while the vibrations of the upper
Keywords:
Reversible pump turbine
and the lower brackets are generated by the mechanical aspects of the rotor. For the top cover, three
Pumped hydro energy storage regions are proposed with their characteristics and its physical origins fully demonstrated with the aid of
Characteristic frequency several typical examples. In region I (with the low partial loads), the vibrational level of the unit is the
Vibration highest and its physical origin is the pressure fluctuation in the vaneless space with the blade passing
Water head frequency. In region II (with the medium loads), the vibrational level of the unit is medium and its
Pressure fluctuation physical origin is the swirling vortex in the draft tube. In region III (with the high partial load), the
vibrational level of the unit is the lowest and its physical origin is twofold: the pressure fluctuation in the
vaneless space with the harmonics of blade passing frequency and the mechanical aspects of the rotor.
For different water heads, transitions between aforementioned categorized regions could be observed
with the primary characteristics maintained. At last, comparing with the cases of the top cover, the vi-
brations of the upper and the lower brackets are less affected by the load and the water head variations.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction
Pumped hydro energy storage power stations (PHESPS) are
being widely built to enhance the stability of the electric grid
through providing the large-scale energy storage and the auxiliary
services [1]. The reversible pump turbine, which is a key compo-
nent in the PHESPS, can be operated in both the pumping and the
generating modes, respectively. For the pumping mode, the
reversible pump turbine mainly utilizes the surplus electricity of
the power grid to pump the water from the lower reservoir to the
upper reservoir for the purpose of the energy storage. During the
peak hour of the electricity usage, the reversible pump turbine
* Corresponding author.
E-mail address: y.zhang@ncepu.edu.cn (Y. Zhang).
works in the generating mode through releasing the water from the
upper reservoir to the lower reservoir for the power generation
during the peak period. However, the commissioned units of the
reversible pump turbine often suffers from the significant vibra-
tions, leading to the abnormal operations of the unit or even the
unplanned stop. The vibrational performance and its physical ori-
gins of a prototype reversible pump turbine are quite complex and
have not been fully investigated yet [2,3]. The primary origins of the
vibrations are twofold: one induced by the abnormal fluid flow
inside the turbine (e.g. the complex flow passing the blade and the
vortex rope in the draft tube section); another one induced by the
mechanical aspects of the rotor. Due to the seriousness of the
problem, it is badly needed to give a quantitative description of the
contributions of the above two mechanisms to the vibrations of the
unit. In the present paper, the vibrations of a prototype reversible

https://doi.org/10.1016/j.renene.2018.06.057
0960-1481/© 2018 Elsevier Ltd. All rights reserved.
668 Y. Zhang et al. / Renewable Energy 130 (2019) 667e676
Nomenclature
Roman letters
f characteristic frequency
g gravitational acceleration
H water head in the experiment
p pressure fluctuation
t time
T impeller rotation cycle
V vibration
A1 amplitude of characteristic frequency in the
frequency spectrum of the vibration signal
A2 non-dimensional amplitude of the characteristic
frequency in the frequency spectrum of the pressure
signal
f1 primary dominated frequency in the frequency
spectrum
pump turbine are experimentally investigated based on a series of
on-site measurements together with related signal processing.
In the literature, the reversible pump turbine has been inten-
sively investigated recently and a brief review of the previous in-
vestigations relating with the present work will be given below.
Generally speaking, the published papers in this field could be
categorized into the following topics e.g. on-site monitoring [4e6],
rotating stall [1,7e10], S-shaped instability [11e17], rotorestator
interaction (RSI) [18e20], hysteresis [21,22], and hump regions
[23,24]. For recent reviews, Zhang et al. [1] summarized the main
features of the complex fluid flow inside the reversible pump tur-
bine operated in generating mode with their generation mecha-
nisms. Egusquiza et al. [6] analyzed the vibration data of several
prototype reversible pump turbines and identified several para-
mount topics for the on-site monitoring (e.g. data interpretation).
Comparing with other hydroturbines (e.g. Francis turbine), the
vortex flow inside the reversible pump turbine is quite significant.
Recently, Zhang et al. [25] systematically reviewed the existing
vortex identification methods with their application examples in
hydroturbines. For the experimental measurement, Zhang et al. [4]
systematically performed an on-site measurement of a large-scale
prototype reversible pump turbine in the full load range and
identified the primary characteristics of the pressure fluctuations of
the unit in different load ranges. Rodriguez et al. [20] proposed the
usage of on-board sensors rotating with the shaft to detect the RSI
characteristics in a pump-turbine to reduce the error. With the
existence of the cavitation, a strong micro-jet will be generated
during the collapse of the cavitation bubbles [26,27], leading to the
erosive damage of the reversible pump turbine [28]. Furthermore,
the cavitation bubbles could deteriorate the inner fluid flow inside
the turbine and also interact with the pressure wave [29] through
various kinds of bubble oscillations [30,31], leading to the reduction
of the turbine performance [32].
For the fluid flow inside the reversible pump turbine, the
pressure fluctuation in the S-shaped region has been systematically
simulated by several researchers e.g. Wang et al. [11], Sun et al. [12]
with the misaligned guide vanes, Xia et al. [13] with the runaway
instability, and Zeng et al. [14] with the guide-vane closing
schemes. Multi-objective optimizations (e.g. efficiency, cavitation,
pressure fluctuation) have been also intensively conducted for the
parametric design of high-profile turbines [33e35]. Based on the
literature review, the features of the vibrations of a prototype
fn rotational frequency of the impeller
Hmax maximum water head of the reversible pump turbine
Hmin minimum water head of the reversible pump turbine
P1* load condition separating regions I and II
P2* load condition separating regions II and III
Poutput output power of the prototype reversible pump
turbine
Prated rated power of the prototype reversible pump
turbine
Vpeak
to
peak the peak-to-peak values of vibration
Greek letters
r water density
Superscript
* non-dimensional parameters
reversible pump turbine together with its physical origins have not
been fully revealed.
In the present paper, the vibrations of a prototype reversible
pump turbine in the X, Y and Z directions of the top cover, the upper
and the lower brackets were all measured for three water heads
(from 48% to 90% in terms of non-dimensional values) and nine
load conditions (from 34% to 96% of the rated power) together with
the pressure measurement at several typical monitoring points.
Effects of the load and the water head variations on the vibrational
performance are analyzed in great detail with their physical origins
identified quantitatively. The structure of the present paper is
organized as follows. In section 2, the experimental setup is
introduced with the basics of the turbine together with the
experimental instruments, operational conditions and procedures.
In section 3, the vibrational performance of the top cover and its
physical origins are investigated and identified based on the on-site
measurement data with three characteristic regions proposed. In
section 4, the vibrational performance of the upper and lower
brackets are demonstrated. In section 5, the primary findings of the
present paper are given with suggestions of future work.
2. Experimental setup
In this section, the experimental setup of the present paper will
be introduced in detail together with a brief introduction of the
basic parameters of the unit, instruments and experimental pro-
cedures. Furthermore, the symbols employed in the present paper
will be also defined (also referring to the nomenclature) and non-
dimensionalized for the reader's convenience.
2.1. Basics of the reversible pump turbine
The experiment was carried out on a prototype reversible pump
turbine located within the PHESPS. Fig. 1 shows a schematic view of
the investigated prototype reversible pump turbine. During the
experiments, only the operational conditions in generating mode
were tested. The basic information of the fluid flow inside the
reversible pump turbine and the structural components of the unit
will be briefly introduced as follows. In generating mode, the fluid
enters into the turbine through the inlet of the spiral casing, then
passing through the guide vane (for adjusting the flow rate),
vaneless space, impeller and finally enters the draft tube cone and
 Y. Zhang et al. / Renewable Energy 130 (2019) 667e676 669

Table 2
Details of the vibration and pressure monitoring points in the prototype reversible
pump turbine.

Monitoring points Direction Type

Top cover X Vibration

Y Vibration
Z Vibration
Upper bracket X Vibration
Y Vibration
Z Vibration
Lower bracket X Vibration
Y Vibration
Z Vibration
Vaneless space e Pressure
Cone section of draft tube e Pressure

Fig. 1. A schematic view of the prototype reversible pump turbine together with the
primary components. The monitoring points for the pressure measurement were
marked as the blue dots.
readers are referred to Fig. 1 for details. Fig. 2 shows the schematic
view of the impeller with nine blades of the prototype pump
turbine.
Now, the experimental instrument will be briefly introduced
with the specific parameters given. For the vibrational measure-
ment, a low-frequency sensor (with the type of DPS-0.5-8 series)
was employed during the whole data collection process. The
sensitivity of the sensor is 8 V/mm and the frequency response
range is 0.5e200 Hz. For the pressure experiment, the pressure
transmitter MPM480 (with ±0.5% accuracy in full scale) was
employed. In order to capture the related signals accurately, the
sampling frequency was set as 1 kHz. For each data collection
process, the number of data samples is over 60,000 points, which is
fair enough for the present analysis. For more details of the

elbow sections. Top cover is just above the turbine, which is an

important section for the measurement of the vibrational status of
the unit. The other important structural components include the
lower bracket and the upper bracket (with the rotor and stator of
the generator inside). The shafts of different components are con-
nected using the shaft coupling.
Table 1 shows several paramount parameters of the investigated
prototype reversible pump turbine e.g. geometrical parameters (the
impeller diameter, the number of impeller blades and the number
of active guide vanes) and the operational parameters (the rated
power and the rated water head).

2.2. Experimental instrument and procedures

In our experiments, both the vibration and pressure of the

prototype reversible pump turbine were measured in detail. For the
vibrations, the movement of the unit in the X, Y and Z directions of
the top cover, the upper bracket, and the lower bracket were all
measured for three water heads and nine load conditions (referring
to section 2.4 for details). Table 2 shows the details of the vibration
and pressure monitoring points in the prototype reversible pump
turbine. For the schematic locations of pressure monitoring points,

Table 1
Paramount parameters of the investigated prototype reversible pump turbine.

Parameter Value Unit

Impeller diameter 4.565 m

Number of impeller blades 9 e
Number of active guide vanes 20 e
Rated rotational speed 333.3 rpm
Rated power 306.12 MW
Rated water head 308.0 m
Maximum water head 329.7 m
Fig. 2. A schematic view of the impeller with nine blades of the prototype pump
Minimum water head 287.9 m
turbine.
670 Y. Zhang et al. / Renewable Energy 130 (2019) 667e676
pressure meausrement and related characteristics, readers are
referred to our previous work [4,5].
The experimental procedure contains two parts including the
load variation and the water head variation tests respectively.
Firstly, the data collection process was carried out at nine different
load conditions with a fixed water head to investigate the in-
fluences of load variations on the vibrational performance. Then
the similar measurements were conducted again for the other two
water heads (one corresponding to the rated water head) to illus-
trate the influences of water head variations on the vibrations and
pressure fluctuations. The details of the load conditions and water
head conditions conducted in the present paper are presented in
section 2.4. In the present paper, the influences of the non-
condensable gases (e.g. air) on the phenomenon [36,37] are not
considered and those dissolved gases could also affect the pressure
wave propagation inside the turbines [38].
2.3. Definitions of symbols
For reader's convenience, according to our previous in-
vestigations in this field [5], several proper non-dimensional pa-
rameters will be employed as follows. The non-dimensional time
(t*) could be defined as
t
t* ¼ : (1)
T
Here, superscript “*” refers to the non-dimensional parameter
employed in the following analysis; t is the time; T is the impeller
rotation cycle.
The characteristic frequency (f) in the obtained signal is non-
dimensionalized using the impeller rotational frequency (fn) as
follows:
f are categorized into three regions. For each region, the character-
f* ¼ : (2)
fn
The output power of the reversible pump turbine (Poutput) is
non-dimensionalized using the rated power (Prated) of the present
power station as follows:
Poutput
P* ¼ : (3)
Prated
Here, the variations of the Poutput for the tests conducted with
different water heads are quite limited and also could be safely
ignored. For the convenience, an averaged value of the measured
values obtained with three different water heads is employed in Eq.
(3).
The water head of the power station is non-dimensionalized as
follows:
H
Hmin
H* ¼ : (4)
Hmax
Hmin
Here, H is the actual water head in the present experiment con-
ducted in the given power station; Hmax and Hmin are the maximum
and the minimum water heads of the power station determined by
the authority for the purpose of the safe operations, respectively.
After the non-dimensional calculations, the rated water head of the
present power station corresponds to the non-dimensional water
head with value H* ¼ 0.48.
For the pressure signal, the amplitude of the pressure signal (p)
in the time or frequency domain is non-dimensionalized using the
rated water head as follows:
p
p* ¼ : (5)
rgH
2.4. Operational conditions for the measurement
Table 3 shows the non-dimensional values of the operating
conditions measured in the present paper in terms of load and
water head variations. For the load variations, nine load conditions
were tested covering from the low partial load conditions to the
nearly rated load conditions. The tested non-dimensional load
conditions of the reversible pump turbine (P*) include 0.34, 0.42,
0.49, 0.58, 0.65, 0.74, 0.82, 0.90, and 0.96. For the water head var-
iations, three water heads were considered (H* ¼ 0.48, 0.71, 0.90
respectively) with H* ¼ 0.48 being the rated water head. Hence,
with the increase of the H* in the present paper, the condition will
be away from the design point.
For each condition, both the vibration and pressure measure-
ments were conducted simultaneously. For the vibrational tests, a
total of nine monitoring points were employed in our experiments
covering X, Y and Z directions of the top cover, the upper bracket,
and the lower bracket of the unit. For details, readers are referred to
Table 2.
3. Vibrational performance of the top cover and its physical
origins
In this section, the vibrational performance of the top cover and
its physical origins are investigated. Firstly, effects of the load var-
iations are demonstrated based on the time-domain signal and the
peak-to-peak values of vibration. Secondly, depending on the
seriousness of vibrations and their dominant frequencies in the
frequency spectrum, the vibrational performance of the top cover
istics of the region are fully demonstrated with the aid of several
examples. Furthermore, the physical origins of vibrations in three
regions are discussed through employing the pressure measure-
ment data. For more details of the pressure fluctuations of the
reversible pump turbine, readers are referred to our previous work
[4,5]. Finally, the influences of the water head variations on the
divisions of three regions are further discussed in detail.
3.1. Influences of load variations
In this section, the influences of the load variations on the
vibrational performance of the top cover are shown and discussed
in terms of the time domain signals and the peak-to-peak values of
vibration ðVpeak
to
peak Þ: For the convenience, water head with
H* ¼ 0.90 will be employed for the discussions. For other cases,
readers are referred to section 3.3.
Figs. 3e5 show the effects of load variations on the variations of
vibrations ðVÞ in the time domain in the X, Y and Z directions of the
top cover. For three directions, the trends are quite similar. Based on
Fig. 3, one can find that during low partial load conditions (e.g.,
Table 3
Non-dimensional values of operating conditions measured in the present paper in
terms of load and water head variations.
Name Non-dimensional value
Load conditions (9) P* ¼ 0.34, 0.42, 0.49, 0.58, 0.65, 0.74, 0.82, 0.90, 0.96
Water head conditions H* ¼ 0.48, 0.71, 0.90
(3)
 Y. Zhang et al. / Renewable Energy 130 (2019) 667e676
Fig. 3. Effects of the load variations on the vibrational signal of the X direction of the
top cover in the time domain. H* ¼ 0.90. P* ¼ 0.34, 0.42, 0.49, 0.58, 0.65, 0.74, 0.82, 0.90
and 0.96, respectively.
Fig. 4. Effects of the load variations on the vibrational signal of the Y direction of the
top cover in the time domain. H* ¼ 0.90. P* ¼ 0.34, 0.42, 0.49, 0.58, 0.65, 0.74, 0.82, 0.90
and 0.96, respectively.
Fig. 5. Effects of the load variations on the vibrational signal of the Z direction of the
top cover in the time domain. H* ¼ 0.90. P* ¼ 0.34, 0.42, 0.49, 0.58, 0.65, 0.74, 0.82, 0.90
and 0.96, respectively.
671
P* ¼ 0.34 and 0.42), serious vibrations will appear. With the in-
crease of the loads, the vibrational level will decease and for
P*>0.74, no much difference could be observed among different
load conditions.
Fig. 6 shows the peak-to-peak values of vibration ðVpeak
to
peak Þ
versus load variations in the X, Y and Z directions of the top cover,
respectively. For the Z direction, with the increase of the load from
“P* ¼ 0.34”, the Vpeak
to
peak firstly decreases until “P* ¼ 0.74” and
then only varies very slightly. The trends of the X, Y and Z directions
of the top cover are quite similar. Quantitatively speaking, the vi-
brations in the X and Y directions are quite close while the vibra-
tions in the Z direction are a little bit different. For example, the
vibrations in the Z direction are a little lower than those in the X
and Y directions for the loads P* in the range of [0.34, 0.74] while are
slightly higher than those in the X and Y directions for the loads P*
in the range of [0.74, 0.96]. The difference between X, Y and Z di-
rections will be further explained in section 3.2 based on the fre-
quency spectrum.
3.2. Characteristics of the three regions
In this section, three regions will be proposed to demonstrate
the main features of the vibrations of the top cover. In this section,
water head with H* ¼ 0.90 will be still taken as an illustrating
example. As shown in Fig. 7, the primary characteristics of the
proposed three regions are demonstrated with the detailed char-
acteristics (e.g. the vibrational amplitude, dominant frequency in
the frequency spectrum and the physical origins of the frequency).
Figs. 8e11 show the examples of the three regions together with
their physical origins. In Fig. 7, the load conditions for separating
the three regions ðP1* and P2* Þ are also marked. Table 4 shows the
ranges of the load conditions and the dominant frequency ðf1* Þ for
each region with H* ¼ 0.90.
The characteristics of the three regions will be briefly intro-
duced as follows:
Region I. The load range of region I covers the low load condi-
tions (e.g. the condition P* ¼ 0.34 in our experiment). The
vibrational level of region I is the highest one among the three
regions. The physical origin of vibration is the pressure fluctu-
ation in the vaneless space (between the stationary guide vanes
Fig. 6. Effects of the load variations on the peak-to-peak values of the vibrational
signal ðVpeak
to
peak Þ. The data were collected in the X, Y and Z directions of the top
cover. H* ¼ 0.90. P* ¼ 0.34, 0.42, 0.49, 0.58, 0.65, 0.74, 0.82, 0.90 and 0.96, respectively.
672 Y. Zhang et al. / Renewable Energy 130 (2019) 667e676
Fig. 7. The main characteristics of the proposed three regions together with the
dominant frequencies and the vibrational status. In the figure, P1* refers to the load
condition separating regions I and II and P2* refers to the load condition separating
regions II and III.
Fig. 8. The time-domain plot and frequency spectrum of the vibration in region I and
its physical origins. (a) shows the time-domain plot of the vibration signal in the X
direction of the top cover. (b) shows the time-domain plot of the pressure signal in the
vaneless space. (c) shows the frequency spectrum of the vibration in the X direction of
the top cover with the dominant frequency being blade passing frequency (f* ¼ 9). (d)
shows the physical origins of the frequency f* ¼ 9 in terms of the pressure fluctuation
in the vaneless space. H* ¼ 0.90. P* ¼ 0.34.
and the rotational impeller) with the dominant frequency being
the blade passing frequency (BPF, f* ¼ 9). As shown in Fig. 8, the
dominant frequency of both the vibration and pressure signals
are f* ¼ 9 (BPF), which are generated by the flow passing the
blade of the impeller and then are transmitted to the top cover.
Other characteristic frequency includes the harmonics of BPF
(f* ¼ 18). However, the amplitudes of f* ¼ 18 are quite limited in
the frequency spectrum as shown in Fig. 8. In Fig. 8, there are
also other low-frequency components shown in the frequency
domain. However, their amplitudes are rather limited.
Fig. 9. The time-domain plot and the frequency spectrum of vibration in region II and
its physical origins. (a) shows the time-domain plot of the vibration signal in the X
direction of the top cover. (b) shows the time-domain plot of the pressure signal in the
draft tube cone section. (c) shows the frequency spectrum of vibration in the X di-
rection of the top cover with the dominant frequency being f* ¼ 0.35. (d) shows the
physical origins of the frequency f* ¼ 0.35 in terms of the pressure fluctuation in the
draft tube cone section. H* ¼ 0.90. P* ¼ 0.58.
Fig. 10. The time-domain plot and frequency spectrum of vibration in region III and its
physical origins. (a) shows the time-domain plot of the vibration signal in the Z di-
rection of the top cover. (b) shows the time-domain plot of the pressure signal in the
vaneless space. (c) shows the frequency spectrum of vibration in the Z direction of the
top cover with the dominant frequency being the harmonics of the blade passing
frequency (f* ¼ 18). (d) shows the physical origins of the frequency f* ¼ 18 in terms of
the pressure fluctuation in the vaneless space. H* ¼ 0.90. P* ¼ 0.96.
Region II. The load range of region II covers the medium load
conditions (e.g. the conditions P* ¼ 0.42, 0.49, 0.58, and 0.65 in
our experiment). The vibrational level of region II is medium
comparing with the other two regions. The physical origin of
vibration in this region becomes the swirling vortex in the draft
tube with the characteristic frequency being a typical low-
frequency component (f* ¼ 0.34e0.36). Taking case with
P* ¼ 0.58 as an example (as shown in Fig. 9), the dominant
 Y. Zhang et al. / Renewable Energy 130 (2019) 667e676 673

3.3. Influences of water head variations

In this section, the influences of water head variations are

investigated with a focus on its effects on the aforementioned
newly proposed three regions. In total, three water heads were
tested (H* ¼ 0.48, 0.71, 0.90 respectively).
Tables 5 and 6 show the characteristics of the pre-defined three
regions in terms of load ranges and primary frequencies for me-
dium water head (H* ¼ 0.71) and low water head (H* ¼ 0.48),
respectively. It should be emphasized that H* ¼ 0.48 corresponds to
the rated water head and with the increase of the water head, the
Fig. 11. The time-domain plot (a) and the frequency spectrum of the vibration signal
(b) in region III. The dominant frequency is the impeller rotational frequency (f* ¼ 1)
condition will be away from the design point. Figs. 12e15 show two
and the secondary dominant frequency is the harmonics of the blade passing fre- typical load conditions (P* ¼ 0.34 and 0.74) demonstrating the ef-
quency (f* ¼ 18) respectively. The data were collected in the X direction of the top fects of water heads on the vibrations in the X, Y and Z directions.
cover. H* ¼ 0.90. P* ¼ 0.90. Overall, the characteristics of the three regions revealed in section
3.2 are still suitable for the other two water heads in terms of
physical origins of the vibrations. Specifically, the load ranges of
three regions are affected by the water head variations, which could
Table 4
Characteristics of the defined three regions in terms of load ranges and primary
be summarized briefly as follows.
frequencies for high water head (H* ¼ 0.90). In the table, vibrational signals in the X, Firstly, with the increase of the water head, the load range of
Y and Z directions of the top cover of the reversible pump turbine are employed for region I increases while the load range of region II decreases. The
the analysis. load range of region III does not vary much with the water head
H* ¼ 0.90 Region I Region II Region III variations. Take the region I as an example. For the rated water head
(H* ¼ 0.48 as shown in Table 6), there is no region I in the X, Y and Z
P* f1* P* f1* P* f1*
directions of vibrations. For the medium water head (H* ¼ 0.71 as
X P*0.34 9 0.34＜P*0.74 0.34e0.36 0.74＜P*0.96 1 shown in Table 5), region I shows in the X and Y directions of vi-
Y P*0.34 9 0.34＜P*0.74 0.34e0.36 0.74＜P*0.96 1
brations but not in the Z direction of vibrations. For the high water
Z P*0.34 9 0.34＜P*0.65 0.34e0.36 0.65＜P*0.96 18
head (H* ¼ 0.90 as shown in Table 4), region I shows in all the X, Y
and Z directions of vibrations.
Secondly, for the low water head (H* ¼ 0.48), the lower limit of
the dominant frequency in the region II is a little bit smaller (e.g.
frequency of both the vibration and pressure signals are 0.20 in Table 6). As shown in Fig. 13, load condition “P* ¼ 0.34”
f* ¼ 0.35. Other characteristic frequencies (e.g. f* ¼ 1 or 18) could belongs to the region II for H* ¼ 0.48 and the low-frequency
be safely ignored as shown in Fig. 9. As shown in Table 4, the P2* component (f* ¼ 0.20) is induced by the swirling vortex in the
for separating regions II and III are slight different between X, Y draft tube. Comparing with the f* ¼ 0.34e0.36 shown in the other
and Z directions. Specifically speaking, P* ¼ 0.74 is within region loads in region II (e.g. 0.34＜P*0.65 as shown in Table 4), the
II in the X and Y directions while is within region III in the Z f* ¼ 0.20 is lower because the operational condition (P* ¼ 0.34) is far
direction. The above difference is caused by the transitions be- away from the design point, leading to the further decrease of the
tween the two regions with significant amplitudes of the two swirling frequency of vortex in the draft tube.
characteristic components f* ¼ 0.35 and f* ¼ 18. In section 3.3, Thirdly, with the increase of the water head, the transitions
detailed discussions on this aspect will be given for three between different regions will be demonstrated with the aid of two
different water heads.
Region III. The load range of region III covers the high (partial)
load conditions (e.g. the conditions P* ¼ 0.82, 0.90, and 0.96 in Table 5
our experiment). The vibrational level of region III is quite low Characteristics of the defined three regions in terms of load ranges and primary
frequencies for medium water head (H* ¼ 0.71). In the table, vibration signals in X, Y
comparing with the other two regions. In region III, the physical
and Z directions of the top cover of the reversible pump turbine are employed for the
origin of vibration depends on the directions. As shown in analysis.
Table 4 and Fig. 10, in the Z direction, the dominant frequency in
H* ¼ 0.71 Region I Region II Region III
the frequency spectrum is the harmonics of the blade passing
frequency (f* ¼ 18), which is generated by the pressure fluctua- P* f1* P* f1* P* f1*
tion in the vaneless space (referring to Fig. 10d). However, in the X *
P 0.34 9 *
0.34＜P 0.65 0.34e0.36 *
0.65＜P 0.96 1
X and Y directions as shown in Fig. 11, the dominant frequency in Y P*0.34 9 0.34＜P*0.65 0.34e0.36 0.65＜P*0.96 1
the frequency spectrum is the impeller rotational frequency Z e e P*0.58 0.21e0.36 0.58＜P*0.96 18
(f* ¼ 1), which is caused by the mechanical aspects of the rotor
(e.g., the unbalanced quality). This mechanism is quite different
with the aforementioned flow-induced vibration (f* ¼ 9, 0.35 or Table 6
Characteristics of the defined three regions in terms of load ranges and primary
18). In fact, in the frequency spectrum of the X and Y directions,
frequencies for low water head (H* ¼ 0.48). In the table, vibration signals in X, Y and
the amplitude of f* ¼ 18 is also of significant value, which is Z directions of the top cover of the reversible pump turbine are employed for the
much larger than the harmonics of the impeller rotational fre- analysis.
quency (f* ¼ 2) but is less than the impeller rotational frequency
H* ¼ 0.48 Region I Region II Region III
(f* ¼ 1). The characteristic frequency being f* ¼ 2 is also caused
by the mechanical aspects of the rotor (e.g., the misalignment of P* f1* P* f1* P* f1*
the shaft couplings). Hence, both the mechanical aspects of the X e e *
P 0.65 0.33e0.36 *
0.65＜P 0.96 1
rotor and internal unstable fluid flow contribute to the vibra- Y e e P*0.65 0.33e0.36 0.65＜P*0.96 1
tions in the X and Y directions. Z e e P*0.65 0.20e0.36 0.65＜P*0.96 18
674 Y. Zhang et al. / Renewable Energy 130 (2019) 667e676
Fig. 12. Effects of water head variations on the vibrational signals in the time domain.
The data were collected in the X, Y and Z directions of the top cover. P* ¼ 0.34.
H* ¼ 0.48, 0.71 and 0.90, respectively.
Fig. 13. Influences of water head variations on the frequency spectrum of vibrational
signals in the X, Y and Z directions of the top cover with low load conditions (P* ¼ 0.34).
H* ¼ 0.48, 0.71 and 0.90, respectively.
Fig. 14. Effects of water head variations on the vibrational signal in the time domain.
The data were collected in the X, Y and Z directions of the top cover. P* ¼ 0.74.
H* ¼ 0.48, 0.71 and 0.90, respectively.
Fig. 15. Influences of water head variations on the frequency spectrum of the vibration
in the X, Y and Z directions of the top cover with high partial load conditions
(P* ¼ 0.74). H* ¼ 0.48, 0.71 and 0.90, respectively.
examples (as shown in Figs. 12e15). Fig. 13 shows the influences of
water head variations on the frequency spectrum of vibrations in
the X, Y and Z directions of the top cover during P* ¼ 0.34 condition
with its corresponding time-domain plot in Fig. 12. For H* ¼ 0.48,
the dominant frequency ðf1* Þ is the low-frequency component (e.g.
0.20 in the Z direction) and P* ¼ 0.34 condition belongs to region II.
With the increase of the water head, it can be clearly spotted that
both the amplitudes of blade passing frequency (f* ¼ 9) and its
harmonics (f* ¼ 18) increase dramatically. For H* ¼ 0.90, the domi-
nant frequency ðf1* Þ is 9 and P* ¼ 0.34 condition belongs to region I
for this water heads. Figs. 14 and 15 show another example of the
transition. In the X (or Y) directions, the dominant frequency ðf1* Þ is
1 (region III), 1 (region III) and 0.35 (region II) for H* ¼ 0.48, 0.71 and
0.90 respectively. Hence, with the increase of the water head, a
transition from region III to region II is shown in the X and Y
directions.
4. Vibrational performance of the upper and lower brackets
In this section, the vibrational performance of the upper and
lower brackets are shown and discussed. Figs. 16 and 17 shows the
Fig. 16. The time-domain plots of vibration signal in the X, Y and Z directions of the
upper and lower brackets. H* ¼ 0.90. P* ¼ 0.74.
 Y. Zhang et al. / Renewable Energy 130 (2019) 667e676
Fig. 17. The frequency spectrum of the vibration in the X, Y and Z directions of the
upper and lower brackets. H* ¼ 0.90. P* ¼ 0.74.
time-domain plot and the frequency spectrum of the vibrations in
the X, Y and Z directions of the upper and lower brackets with
H* ¼ 0.90 and P* ¼ 0.74. As shown in Fig. 17, the dominant frequency
for all the directions is the impeller rotational frequency ðf1* ¼ 1Þ.
Other characteristic frequencies include the harmonics of the
impeller rotational frequency (f* ¼ 2) and the low-frequency
component. And, the blade passing frequency and its harmonics
(f* ¼ 9 and 18 respectively) could be safely ignored. Hence, the vi-
brations of the upper and lower brackets are mainly induced by the
mechanical aspects of the rotor.
Among the three directions, vibrations in the X and Y directions
are high while vibrations in the Z direction are quite low. Hence, in
the following discussions, our attention will be focused on the X
and Y directions. Fig. 18 further shows the effects of load variations
on the amplitudes of the dominant frequency f1* ¼ 1 in the X and Y
directions. Based on Fig. 18, one can find that the effects of load
variations are quite trivial noticing that the primary origin of the
vibrations is the mechanical aspects (e.g. f1* ¼ 1) rather than the
fluid flow. Similarly, the effects of water head variations on the
Fig. 18. Influences of the load variations on the amplitudes of the dominant frequency
being the impeller rotational frequency (f* ¼ 1) in the frequency spectrum in the X and
Y directions of the upper and the lower brackets. H* ¼ 0.90. P* ¼ 0.34, 0.42, 0.49, 0.58,
0.66, 0.74, 0.82, 0.90 and 0.96, respectively.
675
vibrational performance of the upper and the lower brackets are
also negligible.
5. Conclusions
In the present paper, both the vibrations and the pressure of the
prototype reversible pump turbine were measured simultaneously
in detail for different load and water head variations. Specifically,
the vibrations of the unit in the X, Y and Z directions of the top
cover, the upper bracket, and the lower bracket were all experi-
mentally tested for three water heads (from 48% to 90% in terms of
non-dimensional values) and nine load conditions (from 34% to 96%
of the rated power) together with the pressure measurement at
several typical monitoring points. The primary findings could be
summarized as follows:
(1) The vibrations of the top cover are mainly induced by the
fluid flow inside the reversible pump turbine while the vi-
brations of the upper and the lower brackets are generated
by the mechanical aspects of the rotor (e.g., the unbalanced
quality of the rotor and the misalignment of the shaft
couplings).
(2) For the top cover, both the load and water head variations
could affect the vibrations. Based on our data analysis of the
top cover in the full load range, three regions are proposed
with their characteristics fully demonstrated. For different
water heads, transitions between three categorized regions
could be also observed.
(3) The effects of load and water head variations on the vibra-
tions of the upper and lower brackets are almost negligible.
Due to the security issues of the prototype unit, the operating
conditions with extremely low partial loads or low water heads
could not be achieved in our experiments. However, the features
and findings presented here have clearly demonstrated the vibra-
tional performance of the prototype reversible pump turbine and
its physical origins within a wide range of operational conditions. In
the future, the similar methodology will be employed for the study
of the reversible pump turbine in pumping mode or other related
fluid machineries (e.g. pumps [39,40]).
Acknowledgement
This work was financially supported by the National Natural
Science Foundation of China (Project No.: 51506051).
References
[1] Y. Zhang, Y. Zhang, Y.L. Wu, A review of rotating stall in reversible pump
turbine, Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 231 (7) (2017)
1181e1204.
[2] Y. Wu, S. Li, S. Liu, H.S. Dou, Z. Qian, Vibration of Hydraulic Machinery,
Springer Netherlands, 2013.
[3] P. Do € rfler, M. Sick, A. Coutu, Flow-induced Pulsation and Vibration in Hy-
droelectric Machinery: Engineer's Guidebook for Planning, Design and Trou-
bleshooting, Springer Science & Business Media, 2012.
[4] Y. Zhang, T. Chen, J. Li, J. Yu, Experimental study of load variations on pressure
fluctuations in a prototype reversible pump turbine in generating mode,
ASME J. Fluids Eng. 139 (7) (2017) 074501.
[5] J. Li, Y. Zhang, J. Yu, Experimental investigations of a prototype reversible
pump turbine in generating mode with water head variations, Sci. China
Technol. Sci. 61 (4) (2018) 604e611.
[6] E. Egusquiza, C. Valero, D. Valentin, A. Presas, C.G. Rodriguez, Condition
monitoring of pump-turbines. New challenges, Measurement 67 (2015)
151e163.
[7] V. Hasmatuchi, M. Farhat, S. Roth, F. Botero, F. Avellan, Experimental evidence
of rotating stall in a pump-turbine at off-design conditions in generating
mode, ASME J. Fluids Eng. 133 (5) (2011) 051104.
[8] C. Widmer, T. Staubli, N. Ledergerber, Unstable characteristics and rotating
stall in turbine brake operation of pump-turbines, ASME J. Fluids Eng. 133 (4)
676 Y. Zhang et al. / Renewable Energy 130 (2019) 667e676
(2011) 041101.
[9] G. Cavazzini, A. Covi, G. Pavesi, G. Ardizzon, Analysis of the unstable behavior
of a pump-turbine in turbine mode: fluid-dynamical and spectral character-
ization of the S-Shape characteristic, ASME J. Fluids Eng. 138 (2) (2016)
021105.
[10] F. Botero, V. Hasmatuchi, S. Roth, M. Farhat, Non-intrusive detection of
rotating stall in pump-turbines, Mech. Syst. Signal Process. 48 (1e2) (2014)
162e173.
[11] Z. Wang, B. Zhu, X. Wang, D. Qin, Pressure fluctuations in the S-shaped region
of a reversible pump-turbine, Energies 10 (1) (2017) 96.
[12] H. Sun, R. Xiao, W. Liu, F. Wang, Analysis of S characteristics and pressure
pulsations in a pump-turbine with misaligned guide vanes, ASME J. Fluids
Eng. 135 (5) (2013) 051101.
[13] L. Xia, Y. Cheng, J. You, X. Zhang, J. Yang, Z. Qian, Mechanism of the S-Shaped
characteristics and the runaway instability of pump-turbines, ASME J. Fluids
Eng. 139 (3) (2017) 031101.
[14] W. Zeng, J. Yang, J. Hu, J. Yang, Guide-vane closing schemes for pump-turbines
based on transient characteristics in S-shaped region, ASME J. Fluids Eng. 138
(5) (2016) 051302.
[15] Z. Zuo, S. Liu, Flow-induced instabilities in pump-turbines in China, Engi-
neering 3 (4) (2017) 504e511.
[16] G. Olimstad, T. Nielsen, B. Børresen, Dependency on runner geometry for
reversible-pump turbine characteristics in turbine mode of operation, ASME J.
Fluids Eng. 134 (12) (2012) 121102.
[17] G. Olimstad, T. Nielsen, B. Børresen, Stability limits of reversible-pump tur-
bines in turbine mode of operation and measurements of unstable charac-
teristics, ASME J. Fluids Eng. 134 (11) (2012) 111202.
[18] Z. Li, Z. Wang, X. Wei, D. Qin, Flow similarity in the rotorestator interaction
affected region in prototype and model Francis pump-turbines in generating
mode, ASME J. Fluids Eng. 138 (6) (2016) 061201.
[19] C.G. Rodriguez, E. Egusquiza, I.F. Santos, Frequencies in the vibration induced
by the rotor stator interaction in a centrifugal pump turbine, ASME J. Fluids
Eng. 129 (11) (2007) 1428e1435.
[20] C.G. Rodriguez, B. Mateos-Prieto, E. Egusquiza, Monitoring of rotor-stator
interaction in pump-turbine using vibrations measured with on-board sen-
sors rotating with shaft, Shock Vib. (2014) 2014.
[21] D. Li, H. Wang, Y. Qin, X. Wei, D. Qin, Numerical simulation of hysteresis
characteristic in the hump region of a pump-turbine model, Renew. Energy
115 (2018) 433e447.
[22] D. Li, H. Wang, Y. Qin, L. Han, X. Wei, D. Qin, Entropy production analysis of
hysteresis characteristic of a pump-turbine model, Energy Convers. Manag.
149 (2017) 175e191.
[23] Y. Xiao, Y. Yao, Z. Wang, J. Zhang, Y. Luo, C. Zeng, W. Zhu, Hydrodynamic
mechanism analysis of the pump hump district for a pump-turbine, Eng.
Comput. 33 (3) (2016) 957e976.
[24] J. Yang, G. Pavesi, S. Yuan, G. Cavazzini, G. Ardizzon, Experimental charac-
terization of a pumpeturbine in pump mode at hump instability region, ASME
J. Fluids Eng. 137 (5) (2015) 051109.
[25] Y. Zhang, K. Liu, H. Xian, X. Du, A review of methods for vortex identification
in hydroturbines, Renew. Sustain. Energy Rev. 81 (Part 1) (2018) 1269e1285.
[26] Y. Zhang, Y. Zhang, Z. Qian, B. Ji, Y. Wu, A review of microscopic interactions
between cavitation bubbles and particles in silt-laden flow, Renew. Sustain.
Energy Rev. 56 (2016) 303e318.
[27] X.W. Luo, B. Ji, Y. Tsujimoto, A review of cavitation in hydraulic machinery,
J. Hydrodyn. Ser. B 28 (3) (2016) 335e358.
[28] T. Cencîc, M. Ho^ ^
cevar, B. Sirok, Study of erosive cavitation detection in pump
mode of pumpestorage hydropower plant prototype, ASME J. Fluids Eng. 136
(5) (2014) 051301.
[29] Y. Zhang, Z. Guo, X. Du, Wave propagation in liquids with oscillating vapor-gas
bubbles, Appl. Therm. Eng. 133 (3) (2018) 483e492.
[30] Y. Zhang, Y. Gao, Z. Guo, X. Du, Effects of mass transfer on damping mecha-
nisms of vapor bubbles oscillating in liquids, Ultrason. Sonochem. 40 (Part A)
(2018) 120e127.
[31] Y. Zhang, Y. Gao, X. Du, Stability mechanisms of oscillating vapor bubbles in
acoustic fields, Ultrason. Sonochem. 40 (Part A) (2018) 808e814.
[32] G. Lu, Z. Zuo, Y. Sun, D. Liu, Y. Tsujimoto, S. Liu, Experimental evidence of
cavitation influences on the positive slope on the pump performance curve of
a low specific speed model pump-turbine, Renew. Energy 113 (2017)
1539e1550.
[33] B. Zhu, L. Tan, X. Wang, Z. Ma, Investigation on flow characteristics of pump-
turbine runners with large blade lean, ASME J. Fluids Eng. 140 (3) (2018)
031101.
[34] L. Liu, B. Zhu, L. Bai, X. Liu, Y. Zhao, Parametric design of an ultrahigh-head
pump-turbine runner based on multiobjective optimization, Energies 10 (8)
(2017) 1169.
[35] W. Yang, R. Xiao, Multiobjective optimization design of a pumpeturbine
impeller based on an inverse design using a combination optimization
strategy, ASME J. Fluids Eng. 136 (1) (2014) 014501.
[36] Y. Zhang, Y. Zhang, S. Li, Combination and simultaneous resonances of gas
bubbles oscillating in liquids under dual-frequency acoustic excitation,
Ultrason. Sonochem. 35 (2017) 431e439.
[37] Y. Zhang, Y. Zhang, Chaotic oscillations of gas bubbles under dual-frequency
acoustic excitation, Ultrason. Sonochem. 40 (Part B) (2018) 151e157.
[38] Y. Zhang, Z. Guo, Y. Gao, X. Du, Acoustic wave propagation in bubbly flow with
gas, vapor or their mixtures, Ultrason. Sonochem. 40 (Part B) (2018) 40e45.
[39] L. Tan, B. Zhu, S. Cao, Y. Wang, B. Wang, Numerical simulation of unsteady
cavitation flow in a centrifugal pump at off-design conditions, Proc. Inst.
Mech. Eng. Part C J. Mech. Eng. Sci. 228 (11) (2014) 1994e2006.
[40] L. Tan, B. Zhu, S. Cao, Y. Wang, Cavitation flow simulation for a centrifugal
pump at a low flow rate, Chin. Sci. Bull. 58 (8) (2013) 949e952.