The Application of Laser Velocity Meter in Detecting Incipient Cavitation
and Measurement its Intensity, Inside Axial Flow Pumps
Though the cavitation as a damaging phenomenon of hydraulic
devices, has been drawing the interest of many researchers, almost very
few investigations have been done on the cavitation measurement inside
axial flow pumps. The present study is one of the leading ones which consider
this phenomenon inside this widely used type of pumps. Oscillations of
the structure of the pump were used to measure the cavitation. An average
energy method for identification cavitation occurrence and measurement
its intensity has been developed. This is called Logarithmic Cavitation
Intensity (LCI). A statistical analysis was undertaken in a both time
and frequency domains and the LCI was proved as a proper criterion for
defining the cavitation intensity. Though being very robust, the introduced
method is very simple and does not require time consuming calculations.
This causes LCI method to be feasible by simple hardware with low sampling
frequency, resulting in reducing the computational time as well as hardware
complexity and cost.
Cavitation is rapid formation and collapse of vapor bubbles, some times occurs
within pumps or other hydraulic devices. This malfunction can be extremely destructive
to the pump. Cavitation can cause pitting of impeller, impeller vanes and pump
casing. At the other hand, the pump efficiency will decrease significantly during
cavitation (Gulich, 2008). Therefore, cavitation in pumps,
as an unacceptable phenomenon, should by all means be avoided. To do this, one
has to know the inception of cavitation and its intensity within pump, especially
for pumps working in industrial environments.
Cavitation within water pumps has been the subject of numerous studies.
According to the available literatures, there are two different approaches
to detect the onset of cavitation in a liquid:
Numerical modeling: It is often used to predict the onset of cavitation
of a single bubble and rarely within a pump. There are some well-known models,
which can be used in describing the phenomena and behavior of cavitation cores
(Tsudaa and Takagib, 2008; Zhang et
al., 2008b). However, there is no exact algorithm to calculate the noise
due to cavitation at the different operating conditions of a pump.
Engineering methods: The exact value of cavitation noise can be
obtained by one of the available engineering methods. The engineering
methods which have been utilized until now are:
||Determination of the Net Positive Suction Head (NPSH) at constant pump
speed and flow rate: According to ISO standard, drop in the total delivery
head is a good criterion to identify fully developed cavitation inside a
pump (Europump, 1999)
||Visualization of flow: Using a transparent model casing and often
a stroboscopic, visualization of cavitation is possible. This method is
suitable for a single bubble and for high-powered pumps (Lee
et al., 2008; Zhang et al., 2008a)
||Paint erosion testing: This method is based on painting the impeller
blades and shrouds and observation of the cavitation erosion by evidence
of the removal of the paint. Its application is good in combination with
the NPSH test, which shows if the cavitation occurred within the pump or
not (Gulich, 2008)
||Measurement of the static pressure within the flow: With this method,
the onset of the cavitation cores is determined indirectly by comparison
of the measured static pressure and the vapor pressure at a given temperature
of the flow (Lee et al., 2008)
||Measurement of the vibration of structure: The implosion of vapor-filled
zones (bubbles) generates pressure pulsations which excite structure vibrations;
by mounting a transducer on the pump body, the cavitation and its intensity
can be monitored. The measured signal may be contaminated and corrupted
by background noise, such as that of aerodynamical, mechanical and electromagnetic
origins, which attenuate or amplify the measured signal (Futakawa
et al., 2008; Farkas and Pandula, 2006)
||Measurement of the sound pressure: Cavitation produces broadband
high-frequent noise. This noise is emitted when the cavities collapse violently
and when the high pressure peaks are generated. Using sensitive microphones
and sound level meters, incipient cavitation can be identified (Karpiouk
et al., 2008; Alfayez and Dyson, 2005;
Hodnett and Zeqiri, 2008)
||Analysis of pump driver input variables: Measurement of electrical
characteristics of the pump electromotor (current and voltage for
example) some times gives variable information to monitor cavitation
When detecting cavitation from an operating machine, environmental disturbances
make the indirect measurements difficult. Indirect measurement methods
will especially be useful if the measured data from non-cavitation circumstances
Since, the cavitation in pump was taken into consideration, most cavitation
detection studies have been concentrated in cavitation inside centrifugal pumps
(Alfayez and Dyson, 2005; Al-Hashmi
et al., 2004; Hattori and Kishimoto, 2008).
Though the axial flow pumps are extremely used in application and industry,
much less studies have been reported about cavitation in these types of pumps.
It seems that the present study is one of the leading experiments, done in the
measurement of cavitation in an axial flow pump.
This study which can be classified into the engineering methods category,
offers a simple and yet novel technique to detect cavitation and presents
an index to show its intensity. A suitable laser velocity probe was used
to capture the oscillations of the tested pump structure. The data obtained
from this sensor were analyzed by a developed Energy Method, both in the
time and frequency domains. The output of this modeling was assigned to
an index, named by LCI. More illustrations showed that the LCI is capable
to identify the occurrence and quantify the intensity of cavitation within
pumps. The results also demonstrate that the proposed method is able to
detect cavitation by a simple hardware with low sampling frequency. This
character leads to reduction the computation time as well as hardware
complexity and cost.
Cavitation may appear to be a problem in fluid power systems. It affects
fluid power systems and components in various ways, which are usually
undesirable. The severity of the effects of cavitation varies as a function
of a machines power. Figure 1 shows an impeller that
has been severely damaged by cavitation.
Cavitation is usually divided into two classes of behavior: inertial (or transient)
cavitation and non-inertial cavitation. Inertial cavitation is the process when
a void or bubble in a liquid rapidly collapses and produces a shock wave (Brujan
et al., 2008). Such cavitation often occurs in pumps, propellers,
impellers and in the vascular tissues of plants. Non-inertial cavitation is
the process where a bubble in a fluid is forced to oscillate in size or shape
due to some form of energy input, such as an acoustic field (Suslick
and Flannigan, 2008).
Within a pump, the flow area at the eye of the pump impeller is usually smaller
than either the flow area of the pump suction piping or the flow area through
the impeller vanes. At the eye of a pump, the decrease in flow area results
in an increase in flow velocity accompanied by a decrease in its pressure. The
greater the pump flow rate, the greater the pressure drop. When the local pressure
falls below the saturation pressure of the fluid, the liquid may flash to vapor.
Any vapor bubbles formed by the pressure drop at the eye of the impeller are
swept along the impeller vanes by the flow of the fluid. When the bubbles enter
a region where local pressure is greater than the saturation pressure farther
out the impeller vane, the vapor bubbles abruptly collapse. This process of
the formation and subsequent collapse of vapor bubbles in a pump is called cavitation
(France and Michel, 2005).
||A cavitation damaged impeller
Cavitation in a pump has a significant effect on pump performance. It degrades
the performance of the pump which leads to fluctuation in the flow rate and
discharge pressure. Cavitation can also be destructive to the pump internal
components. If a pump cavitates, collapse of vapor bubbles will cause a physical
shock to the leading edge of its impeller vane. This shock will create small
pits on this area of the impeller vane. Each individual pit is microscopic in
size, but the cumulative effect of millions of these pits, over a long period
of time can eventually destroy the pump impeller (Zheng et
al., 2008; Xu et al., 2005; Guogang
et al., 2008; Hattori and Kishimoto, 2008).
Cavitation can also cause excessive pump vibration, which could damage the pump
bearings, wearing rings and seals. To avoid the cavitation, the pressure of
the fluid at all points within the pump must remain above the fluid saturation
pressure. The existence of cavitation is often very difficult to detect because
cavitation occurs typically at locations where the access for measuring instruments
THE DOPPLER EFFECT (Maulik, 2005)
The doppler effect, named after Christian Doppler, is the change in frequency
and wavelength of a wave which is perceived by an observer moving relative
to the source of the wave. A stationary light source emits a continuous
light wave with the frequency f and the wavelength λ. A wave train
with the length λ passes a stationary observer in the time T = 1/f.
If in contrast the observer moves away from the light source at the speed
v, then the wave train needs a slightly longer time T, to pass the observer.
The total distance the wave travel in the time T includes the distance
λ of the observed wave train and also the distance v, T traveled
by the moved observer in the time T. For the moving observer, the wave
vibration has the cycle duration T and because f = 1/T and λ =
c/f, this then results in:
and thus the frequency f to:
f = f (cv)/c = f (1v/c)
Therefore, if the observer moves away from the light source (v>0),
then the light frequency will be shifted to smaller values (red shift)
and if he moves towards the light source (v<0), then an increased frequency
will be measured (blue shift). Figure 2 schematically
shows this effect. The above analysis is an approximation for small velocities
in comparison to the speed of light which is fulfilled very well for practically
all technically relevant velocities and is the base of laser velocity
||Application of Doppler effect in laser velocity probes
EXPERIMENTAL SET-UP AND MEASUREMENTS
An experimental axial flow pump was chosen for the experiments. The pump
was a single volute design with a 5-vane closed-face impeller. The impeller
and shaft were supported by two rolling element bearings and were coupled
to a 5.6 kW motor with a relatively flexible coupling. The motor was driven
directly from a 480 V, 60 Hz source. This experimental pump has been designed
and well adapted to analyze the cavitation phenomenon. Its vanes angle
was changeable by which cavitation with different intensity levels could
be created inside. At the other hand, it had a transparent impeller casing
through which and utilizing a stroboscopic light, the cavitation intensity
within the pump case could be identified. This pump also was well equipped
with suitable gages to determine NPSH and its drop in cavitation condition
according to ISO standard. During the tests, the pumping loop was supplied
with water from a 5000 L tank. The loop could support up to 50 L sec-1
flow rates. A stature of the utilized experimental pump and its equipments
is shown in Fig. 3.
In all test cases a Laser Doppler Velocity Probe was used to measure velocity
of the outside body of the water pump. A frequency of 50 kHz was chosen as the
sampling frequency of the used A/D hardware. If a much higher sampling frequency
was used then there would be more data to be handled, more memory would be required
and the computation time would be longer. In practice, in order to be sure of
having no aliasing, an analogue anti-aliasing low-pass filter with a cut-off
frequency less than half of the sampling frequency was used before sampling.
At the other hand, the window effect is a common problem in spectra analysis.
It has been found (Gao et al., 1993) that the
Hamming window is the best choice for the spectra analysis.
|| The utilized experimental pump and pumping loop
||Cavitation free case. The vanes angle was set to 20°
First stage of cavitation. Increasing the vane angle
by 2° caused the first cavitation bubbles create around the impeller
Second stage of cavitation. The vanes angle was increased
by 2°. The cavitation bubbles extended and a mild cavitation noise
was hearing. The pump efficiency relatively began to drop
The data acquisition processes executed in 5 different intensity levels
of cavitation; cavitation free stage and four high and higher cavitation
intensities. At each level, the pump was run at a constant speed of 1500
rpm for a sufficiently long time to bring the flow state to an equilibrium.
Then data collection step initiated and sufficient data were saved. The
vane angle then slowly was altered and the cavitation with other intensity
level occurred in the pump. Figure 4-8 show the different
stages of experiments.
||Third stage of cavitation. The vanes angle was set at
around 26°. The cavitation extended more and more and its noise
Final stage of experiments which was related to vanes
angle of 29°. The cavitation bubbles not only enclosed the vanes,
but extended to whole impeller case also. A severe cavitation noise
similar to knocking was hearing and the pump efficiency fell dramatically
ANALYSIS AND RESULTS
The approach of this study to quantify cavitation intensity is shown
in Fig. 9. Having shown that the body vibrations are
usually made up of several frequencies, an appropriate description of
cavitation intensity is the average energy. The advantage of using the
average energy method is its equivalence in the frequency and in the time
domains. The equation of the average energy in terms of the spectra data
and in the time domain:
||The study approach to quantify cavitation intensity
where, x (n) is the filtered data and X (k) is the spectra data of the
original signal, i.e.,
In Eq. 5, N is the No. of samples used in both cases
and w (n) is the Hamming window function. The cavitation intensity was
finally defined as the Logarithmic Cavitation Intensity (LCI):
LCI = ln (average energy)
where, average energy is either from spectra data (Eq.
3) or from filtered data (Eq. 4).
Figure 10 shows the LCI calculated in the time domain,
versus cavitation intensity. It is clear from Fig. 10
that occurrence and rising the cavitation intensity leads to increase
in LCI. Thus it is possible to introduce LCI as a proper criterion in
order to reveal cavitation characteristics. Figure 11
show the cross correlation between the LCI calculated by the average energy
in the time domain and the average energy in the frequency domain (using
a Hamming window). Figure 11 clearly indicates that
using either a filter or a Discrete/Fast Fourier transform technique can
achieve the same result in the cavitation detection system. Since, the
calculation of LCI in the time domain (using the filtered data) is much
simpler to implement than using the spectra data, LCI in the time domain
utilized to quantification of the cavitation intensity.
|| LCI in time domain, versus cavitation intensity
||Cross correlation between the LCI in the time domain
and in the frequency domain
The body oscillation of a pumping set depends on its speed and load and
on the instability in the pump. It can also appear due to cavitation.
Cavitation as a source of instability causes vibration, noise, pitting
and material erosion and deterioration of pump performance.
In this study, cavitation detection and measurement its intensity in
an axial flow pump has been systematically studied from the signal processing
as well as the instrumentation point of view. A telemetry system comprising
a laser velocity probe was successfully installed and used to monitor
cavitation erosion induced oscillation. The pump structure oscillation
has been analyzed by other researchers too, but the main contribution
of this study is introduction of the LCI as an index of cavitation occurrence
and its intensity inside the pump. Utilization either a filter or a Discrete/fast
fourier transform technique can achieve the same result in the cavitation
measurement system. Since the calculation of LCI in the time domain (using
the filtered data) is much simpler to implement than using the spectra
data, LCI in the time domain utilized to quantification of the cavitation
Though the 3% drop in pump pressure head has been turned to be a standard method
to detect cavitation initiation, monitoring the LCI seems to be a better indicator.
In order to use the static pressure to monitor for cavitation, two additional
measurements (fluid velocity and fluid temperature) are required. This is because
the actual NPSHA is not solely dependent on the static pressure, but is also
a function of the average velocity and the vapor pressure of the fluid. In turn,
the vapor pressure of the fluid is a function of its temperature. LCI, on the
other hand, requires only one measurement. The other advantage of using the
LCI index reveals of the fact that whereas monitoring the NPSHA is an indirect
indicator of cavitation existence or not in the pump, the LCI is a direct indication
of cavitation, not only for its occurrence identification, but also for its
While the present study is one of the leading in cavitation measurement
inside an axial flow pump, its results and the proposed method is completely
adjustable to other pumps and hydraulic devices. The results also demonstrate
that the proposed method is capable of detecting cavitation by simple
hardware with low sampling frequency, leads to reduction the computation
time as well as hardware complexity and cost.
1: Alfayez, L. and G. Dyson, 2005. The application of acoustic emission for detecting incipient cavitation and the best efficiency point of a 60 kW centrifugal pump: Case study. J. NDT. E. Int., 38: 354-358.
2: Al-Hashmi, S., F. Gu, Y. Li, A.D. Ball, T. Fen and K. Lui, 2004. Cavitation detection of a centrifugal pump using instantaneous angular speed. Proceedings of the 7th Biennial Conference on Engineering Systems Design and Analysis, Manchester, July 19-22, 2004, American Society of Mechanical Engineers, pp: 185-190.
3: Brujan, E.A.,T. Ikeda and Y. Matsumoto, 2008. On the pressure of cavitation bubbles. Exp. Ther. Fluid Sci., 32: 1188-1191.
Direct Link |
4: Europump, 1999. NPSH for Rotodynamic Pumps: A Reference Guide. 1 Edn., Elsevier Science Ltd., UK., pp: 33-37.
5: Farkas, I. and Z. Pandula, 2006. Problems with conventional vibration measurements by detection of cavitation operating of centrifugal pumps. Proceedings of the 7th International Conference on Vibration Measurements by Laser Techniques, Ancona, June 19-22, 2006, Society of Photo-Optical Instrumentation Engineers, pp: 63450Z1-63450Z7.
6: France, J.P. and J.M. Michel, 2005. Introduction-The Main Features of Cavitating Flows. In:., Fluid Mechanics and its Applications-Fundamentals of Cavitation, Moreau, R. (Ed.). Kluwer Academic, New York., pp: 5-12.
7: Futakawa, M., H. Kogawa, S. Hasegawa, Y. Ikeda and B. Riemer et al., 2008. Cavitation damage prediction for spallation target vessels by assessment of acoustic vibration. J. Nucl. Mater., 377: 182-188.
CrossRef | Direct Link |
8: Gao, X., R. Stone, C. Hudson and I. Bradbury, 1993. The detection and quantification of knock in spark ignition engines. Proceedings of the International Fuels and Lubricants Meeting and Exposition, October 1-15, 1993, SAE International, Philadelphia, pp: 932759-.
9: Gulich, J.F., 2008. Suction Capability and Cavitation. In: Centrifugal Pumps. 3rd Edn., Springer- Verlag, Berlin, UK., pp: 298-316.
10: Guogang, W., G. Ma, D. Sun, H. Yu and H. Meng, 2008. Numerical study on fatigue damage properties of cavitation erosion for rigid metal materials. J. Univ. Sci. Tech., 15: 261-266.
11: Hattori, S. and M. Kishimoto, 2008. Prediction of cavitation erosion on stainless steel components in centrifugal pumps. J. Wear, 265: 1870-1874.
12: Hodnett, M. and B. Zeqiri, 2008. Toward a reference ultrasonic cavitation vessel: Part 2: Investigating the spatial variation and acoustic pressure threshold of inertial cavitation in a 25 khz ultrasound field. IEEE Trans. Ultrason. Ferroelectrics Freq. Contr., 55: 1809-1823.
Direct Link |
13: Karpiouk, A.B., S.R. Aglyamov, F. Bourgeois, A. Ben-Yakar and S.Y. Emelianov, 2008. Quantitative ultrasound method to detect and monitor laser-induced cavitation bubbles. J. Biomed. Opt., 135: 030501-030501.
14: Lee, H., E. Akagawa, E. Tatsumi and Y. Taenaka, 2008. Characteristics of cavitation intensity in a mechanical heart valve using a pulsatile device: Synchronized analysis between visual images and pressure signals. J. Artif. Organs., 11: 60-66.
Direct Link |
15: Maulik, D., 2005. Doppler Systems: Physical principles of Doppler Ultrasonography: Doppler Ultrasound in Obstetrics and Gynecology. 2nd Edn., Springer, New York, pp: 9-16.
16: Suslick, K.S. and D.J. Flannigan, 2008. Inside a collapsing bubble: Sonoluminescence and the conditions during cavitation. Int. Rev. Phys. Chem., 59: 659-683.
17: Tsudaa, S. and S. Takagib, 2008. A study on the growth of cavitation bubble nuclei using large-scale molecular dynamics simulations. Fluid. Dynam. Res., 40: 606-615.
18: Xu, R., X. Chen, Z. Shen, J. Lu and X. Ni, 2005. Experimental and theoretical study of cavitation- induced mechanical effect on a solid boundary. J. Hydrodyn., 6: 724-729.
Direct Link |
19: Zhang, M., G. Wang and X. Li, 2008. Design and application of cavitation flow image program. Proceedings of the International Symposium on Photoelectronic Detection and Imaging 2007: Image Processing, Beijing, China, September 9-12, 2008, The International Society for Optical Engineering, pp: 66230H-66230H6.
20: Zhang, X.B., L.M. Qiu and Y. Matsumoto, 2008. Computational fluid dynamic study on cavitation in liquid nitrogen. Cryogenics, 48: 432-438.
21: Zheng, Y.G., S.Z. Luo and W. Ke, 2008. Cavitation erosion-corrosion behaviour of CrMnB stainless overlay and 0Cr13Ni5Mo stainless steel in 0.5 M NaCl and 0.5 M HCL solutions. Tribol. Int., 41: 1181-1189.
CrossRef | Direct Link |