Experimental study on multiple characteristics of multi-source AE signals during cavitation process of waterjet propulsion pump

In order to investigate the cavitation-induce acoustic emission (AE) characteristics of a waterjet propulsion pump, a synchronous experimental system consisting of a test pump, a high-speed camera, and acoustic emission sensors has been established. The characteristic parameters of AE signals are extracted to established the quantitative relationships between cavitation status and AE signals. These characteristic parameters mainly consist of the ringing count rate Rc , energy release rate Er , and Hurst index Hi . The cavitation status of the test pump is reflected in the characteristic parameters of the AE signals obtained by the sensors installed at different locations. The incipient blade suction side (SS) cavitation as well as the blade tip gap cavitation can be detected through the change of characteristic parameters of AE signals at impeller inlet. The start of head drop can be reflected by the variety of characteristic parameters of AE signals at impeller middle. Large-scale cavitation will lead to the breakdown of pump performance and dramatically characteristic parameters’ change of AE signals can be measured by the sensors located at impeller outlet.


Introduction
A relatively new type of ship propulsion known as the waterjet propulsion pump can be used to power both quiet underwater vessels and high-performance shallow draft ships [1,2].Cavitation widely exists in hydraulic machinery such as pumps, turbines, venturi tubes, and valves [3,4].Cavitation occurring in pumps is generally undesirable and most associated with performance deterioration, erosion, vibration, and noise emission [5][6][7].It is particularly harmful to the safe and stable operation of waterjet propulsion pumps.Hence, it is very necessary to carry out early warning and real-time monitoring of cavitation.
Cavitation generally occurs when the local pressure drops below the saturated vaporization pressure with the gas-filled or gas and vapor-filled cavities formed [8,9].Cavitation involves a significant number of two-phase vortex motions in addition to complex gas and liquid phase change processes [10].The study of the cavitation phenomenon has become one of the essential problems in the development and engineering using of waterjet propulsion pump [11].
Currently, hydraulic performance, vibration, pressure pulsation are the primary research methods for detecting the pump cavitation phenomenon.In engineering tests, the degree of cavitation is often characterised by the degree of decrease in hydraulic performances.Since the incipient cavitation in the pump does not lead to a significant change in head, this method has a serious hysteresis.Although the vibration approach has good generality, it is unreliable since the signal source is complex and requires a lot of post-processing [12] Sensor of vibration acceleration signal is shown in Figure 1(a).A destructive way of monitoring, the pressure pulsation method necessitates drilling holes in the pump casing.Pressure pulsation signal transducer is shown in Figure 1(b).AE, as a fast, non-destructive detection technology, has the advantages of quick reaction at high frequencies and good interference resistance at low frequencies, The sensor of AE signal is shown in Figure 1(c).It can be utilized to collect high-frequency signals generated by cavitation rupture-induced shell atomic lattice dislocations within hydraulic equipment [13]. To

Test object
The test object is an axial flow waterjet propulsion pump with a transparent Plexiglas shell, and synchronous experiment testing of cavitation structures and multi-source AE signals have been carried out through a synchronous experimental system consisting of a test pump, a high-speed camera, and acoustic emission transducer.The parameters of the tested axial flow pump are following: inlet diameter D1=300mm, outlet diameter D2=220mm, designed rated flow Qd=0.5m 3 /s, impeller blades Z1=7, guide vane blades Z2 =9, as shown in Figure 2.

Test bench
The experiment was carried out on the closed-cycle test bench of the waterjet propulsion pump at the Marine design & Research Institute of China.The inlet and outlet pressure signals were detected by dynamic pressure sensors with an accuracy of 0.1%.The flowrate was measured by an electromagnetic flowmeter with an accuracy of 0.3%.To measure the instantaneous torque and rotating speed, a dynamic torque sensor with an accuracy of 0.2% was installed between the motor and the pump.The locations and images of the instrumentations, and the test bench are shown in Figure 3.The inlet pressure is reduced gradually through a vacuum pump.The cavitation status of the waterjet propulsion pump was checked using a high-speed camera and a multi-source AE signal.

Cavitation high-speed camera method
First, a matte black oxidation treatment was performed to the impeller surface, resulting in a significant color-difference between the cavitation and the background, emphasizing the cavitation's movement and evolution.The PCO. dimax S high-speed camera was then positioned on a vertical line in the center of the impeller to ensure that the main moving part of the object was in the center of the image.Then, as indicated in Figure 4, modified the position and angle of the fill light so that it generated a diffuse reflection with the shell surface, avoiding the production of high light that interfered with the cavitation observation [14].In addition, selected the optimal lens focal length and shooting distance so that the camera could catch the cavitation phenomenon at its highest resolution.Finally, the aperture of the lens, and the shutter duration of the camera were all determined thoroughly based on the cavitation phenomenon's movement speed, and the length of the circumferential and radial trails [15].The specific parameters of the high-speed camera are set in Table 1.

Cavitation AE signal test method
To avoid the interference noise of the Plexiglas shell transmission and collect the more realistic cavitation AE signal, a metal waveguide rod was inserted into the impeller inlet, middle, and outlet positions of the Plexiglas shell, and the VS900-M AE sensor was pasted on the top of the waveguide rod with No. 3 vacuum grease to ensure that the surface of the AE sensor was in full contact with the top of the waveguide rod, as shown in Figure 5. AE signals from 10 sets of 3 positions were repeatedly acquired at 2-second intervals at each NPSH.The sampling frequency, sampling duration, and analysis threshold of the AE signal were set based on the features of the cavitation AE signal of the waterjet propulsion pump (its frequency range is 0.5-1Mhz [16], and the time domain signal features will be given in Section 2.1), the sampling law, and the analysis threshold of each channel increases by 4 db when there was no cavitation.Table 2 shows the specific sample parameters of the AE signal.

Conventional characteristic parameters
Conventional characteristic parameters are traditional AE signal analysis methods that have been in use since the 1950s, and they primarily include process and status characteristic parameters.Table 3 shows the many types of general characteristic parameters and appropriate waveforms.As illustrated in Figure 6, the cavitation AE signal is made of a large number of bursts with continuous and random instabilities, i.e., they cannot be separated in time and individual pulses cannot be identified.
The cavitation status waveform depicts AE behavior at a specific point in the cavitation process rather than a representation of the entire AE process that adheres to transient status criteria.As a result, using continuous status measures such as ringing count rate and energy release rate is more efficient.
where Vi is the signal voltage (mv), t1 is the starting time (s), t2 is the end time (s), △t is the continuous time (s).

New characteristic parameter
In addition to the conventional characteristic parameters based on signal waveform features, the Hurst index Hi with strong de-noising ability and prominent state differences can be proposed to extract quantitative information from non-stationary cavitation AE signals using statistical fractal dimension theory.
3.2.1.Hurst index.Hi can be calculated using R/S rescaling range analysis and linear regression by partitioning the signal sequence into m pairs of g groups of length ri (i=1, 2, 3..., g), non-overlapping subsequences.The mean value of each subsequence group is computed, as well as the dispersion, cumulative dispersion, range, standard deviation, and R/S values, defined as equation ( 2), ( 3), ( 4), ( 5), ( 6), ( 7): max( ) min( ) where r is the length of group.The R/S values of the respective series were averaged and defined as equation ( 8): This produces data pairs (logri, logti), where i = 1, 2, 3, ..., m.A linear regression of the data pairs provides a straight line slope of Hi with logri as the independent variable and logti as the dependent variable.Table 4 depicts the three possible ranges for Hi.
Table 4.The 3 scopes and representative meanings of Hi.

Scope
Representative meaning 0.5 A signal in the time domain is a signal generated by an independent free process.

0.5-1
The time-domain signal is persistently steady, and if the system has an increasing trend for a certain period, the system will continue to maintain this trend.0-0.5 The time domain signal shows unsteady characteristics, which is contrary to the above.

Result and discussion
According to the experimental results of the object of this study, the hydraulic performance, cavitation state and the characteristic parameters of AE signals had similar changing rules during the steady decline of NPSH under 0.9Qd, Qd and 1.1Qd operating conditions.Therefore, only the results of Qd condition is provided in this paper.

Change in hydraulic performance and cavitation status
Figure8 presents the cavitation performance curve of the tested pump, which is expressed by the relationship curve between the pump head and NPSH.The initial of cavitation has no effect on the waterjet pump head.As the inlet pressure decreased further, the hydraulic performance begins to drop rapidly.
Based on the high-speed photography results and pump head, five cavitation conditions labelled a to e were selected to demonstrate the progression of cavitation with decreasing NPSH.Conditions a to e marked on the cavitation performance curve represent no-cavitation, incipient cavitation of blade SS, occurrence of tip gap cavitation, the first critical cavitation condition (H drop about 1%), and the severe cavitation condition (H drop > 3%), respectively.The corresponding NPSH for each condition are shown in Table 5. Figure 9 shows high-speed photography images for four cavitation status mentioned above (apart from condition a).In order to further characterize each cavitation status, the phenomenological characterization of each cavitation status will be provided in sections 4.3, 4.4, 4.5 and 4.6.

Incipient cavitation of impeller blade SS
As shown in Figure 9(a), the cavitation incipient region is located at the inlet edge of the blade SS and near the tip position when the NPSH of the waterjet propulsion pump is 11.32m (condition b).A thin layer cavitation attached to the blade SS disappears swiftly at its trailing edge and practically abrupts in a line.There is no cloud-like shedding phenomenon on the great majority of its trailing edge span.As shown in Figure 10, the impeller inlet Rc and Er rise by 1.2 and 1.4 times, respectively, from condition a to condition b, while the impeller inlet Hi oscillates consistently between 0.5 and less than 0.5 at condition b.
The causes of the variations in the characteristic characteristics of the AE signal of the waterjet propulsion pump in this cavitation status were investigated.Because a tiny amount of intermittent sheet cavitation develops and a small amount of bubbles collapse blast wave forms along the trailing edge of the cavitation near the inlet edge of the impeller blade SS, so the blast wave transmitted through the liquid medium induces a few AE signal pulses that exceed the analysis threshold and a weak signal energy growth on the inner wall of the shell near the impeller inlet.In addition to energy expansion, the AE signal exhibits an anti-steady state aspect in the time domain due to the incoherence of the cavitation collapse.As a result, the incipient cavitation of the waterjet propulsion pump's impeller blade SS can be represented by the Rc, Er, and Hi characteristic parameters of the impeller inlet AE signal, and the variation pattern is as mentioned above.

Occurrence of tip gap cavitation
As shown in Figure 9(b), due to the high speed rotation of the impeller of the waterjet propulsion pump and the further reduction of the inlet pressure, the high speed leakage flow between the pressure surface (PS) and SS of the same blade in the impeller causes the continuous cavitation in the blade tip gap and eventually ends with a smaller volume of bubble traces and then disappears [17] when the NPSH drops to 8.81m (condition c).The collapsing bubbles have also been spotted moving closer to the inner wall of the shell as well as toward the impeller channel and the overlapping area of the blades.Compared with condition a, the small scale of leading edge cavitation develops to sheet cavitation attached on blade SS.As shown in Figure 10, the impeller inlet Rc and Er increase by 1.3 and 1.5 times, respectively, from condition b to condition c, whereas the impeller inlet Hi recovers to around 0.5 and then rises slightly towards condition c, and the Er at condition c in the middle of the impeller is twice as large as the Er near condition b.
The causes of the variations in the characteristic characteristics of the AE signal of the waterjet propulsion pump in this cavitation status were investigated.As the NPSH decreases, the scale of cavitation phenomenon near the impeller inlet increases.The coexistence of tip gap cavitation and sheet cavitation causes the flow between the leading edge of the blade SS and the inner wall of the shell becomes complex and continuous.So this cavitation status causes a further sustained increase in the amplitude, energy, and number of pulses above the analysis threshold of the impeller inlet AE signal, and the impeller inlet AE signal exhibits a weak steady-state behavior.Furthermore, the AE signal amplitude in the middle of the impeller progressively increases, although there is no visible phenomenon of exceeding the analytical threshold.Therefore, the occurrence of tip gap cavitation can be represented by the Rc, Er and Hi characteristic parameters of the impeller inlet AE signal combined with the characteristic parameter of Er in the impeller middle, and the variation pattern is as described above.

The first critical cavitation conditions (H drop about 1%)
As shown in Figure 9(c), the tip gap cavitation and blade sheet cavitation continue to grow and the shedding phenomenon of cloud-like cavitation can be seen near the trailing of sheet cavitation when the NPSH is 8.24m (condition d).Due to the cavitation phenomenon in the tip gap gradually developing into a large degree tip leakage vortex cavitation (TLVC) structure.The sheet cavitation on the blade SS appears to be largely masked by the TLVC, so it's hard to catch sight of blade sheet cavitation clearly.The cloud-like shedding phenomenon has been identified in the blade overlap area and to the depth of the flow channel between the blades, at this condition the flow may begin to exhibit strong instability characteristics and a strong pulsing status.And the bubble collapse area has developed to the middle of the impeller.As shown in Figure 10, Rc and Er at the impeller inlet steadily dropped from condition c to condition d.However, Rc and Er at the impeller's middle grow by 2.2 and 2 times, respectively, while Hi at the impeller's middle is out of the oscillation range around 0.5 and ascend fast near condition d.
The causes of the variations in the characteristic characteristics of the AE signal of the waterjet propulsion pump in this cavitation status were investigated.As the NPSH further continues to fall, a large number of bubbles form within the flow between the blade inlet side tip and the inner wall of the shell.Bubble as a type of isolation medium on the pump AE signal outward transmission has a certain degree of attenuation effect, resulting in a decreasing trend of the impeller inlet AE signal intensity.The cloud-like cavitation is shedding from the tail of the tip gap cavitation and forms a slender tail trail, with the trailing area of cavitation gradually extending to the next blade, causing the impeller passage between the blades will be slightly blocked, thus preventing the rotating blade from doing work to the flow and causing the hydraulic performance to deteriorate.At this condition the cavitation collapse area has essentially shifted to the middle of the impeller, and the bubble rupture presents a continuous stable process, resulting in the middle of the impeller AE signal's amplitude and energy steadily increased.Therefore, the first critical cavitation conditions can be represented by Rc, Er and Hi characteristic parameters of the impeller middle AE signal, and the variation pattern is as described above.
4.6.The severe cavitation condition (H drop >3%) Eventually as Figure 9(d) shown, the wedge-shaped tip gap cavitation has formed and the development of tip gap cavitation as well as cloud-like shedding cavitation will lead to a further blockage of the flow channel when the NPSH drops to 7.19m (condition e).The cloud-like shedding cavitation phenomenon at the trailing edge of the tip gap cavitation expands and extends to the mainline direction.A furious tumbling large-scale cavitation event as well as some perpendicular cloud-like cavitation erected to the SS of the blade are observed.And the overall cavitation entrainment and tumbling compared to the NPSH of 8.24m (condition d) is more intense.As shown in Figure 10, the impeller inlet Rc and Er decreased further near condition e, while the Rc, Er, and Hi of the impeller middle continue to grow, and the impeller outlet Rc and Er increase sharply by 2.4 and 1.9 times, respectively from condition d to condition e, the Hi also left the oscillation range of about 0.5 and increased sharply near condition e.
The causes of the variations in the characteristic characteristics of the AE signal of the waterjet propulsion pump in this cavitation status were investigated.At this condition cavitation has entirely formed in the pump, with a huge number of cavitation bubbles forming in the low-pressure area (impeller intake) and collapsing near the high-pressure area (impeller outlet).Cavitation consumes an increasing amount of the impeller passage while decreasing pump flow and the operating capacity of the blade to the flow medium, resulting in a significant reduction in hydraulic performance.Obviously the bubble collapsed area has extended to the impeller outlet, so the AE signal amplitude and energy at the impeller outlet rose dramatically.Therefore, the severe cavitation condition can be represented by Rc, Er and Hi characteristic parameters of the impeller outlet AE signal, and the variation pattern is as described above.

Conclusion
In this work, the multi-source AE signal multi-characteristic parameters variation pattern during the cavitation process of the waterjet propulsion pump is experimentally studied, and combined with the cavitation high-speed camera results and hydrodynamic theory.the cavitation status is accurately obtained while the multi-characteristic parameters of the multi-source AE signal of the waterjet propulsion pump are extracted, and a quantified relationship between the cavitation status and the multicharacteristic parameters of the multi-source AE signal of the waterjet propulsion pump is established, proving the feasibility of AE technology in the characterization of the cavitation status of the waterjet propulsion pump.It is demonstrated in this paper that AE technology can not only characterize incipient cavitation accurately, but also be sensitive to cavitation-induced hydraulic performance degradation, and can indicate the phenomenon of steep drop in hydraulic performance caused by large-scale cavitation in waterjet propulsion pumps, so AE technology can be applied to waterjet prolusion pumps performance prediction, real-time dynamic monitoring, and cavitation erosion prevention of working components.The following conclusions can be mainly drawn: (1) The waterjet propulsion pump's cavitation can be monitored using the AE technology.Three AE signal characteristics parameters (ringing count rate Rc, energy release rate Er, and Hurst index Hi) are presented to describe the cavitation status within the waterjet propulsion pump using the AE signal waveform and self-similarity feature.
(2) The incipient cavitation of blade SS and tip gap cavitation occurrence within the waterjet propulsion pump can be represented by the Rc, Er, Hi of the AE signal at the impeller inlet of the shell, and combined with the Er of the AE signal at the impeller middle of the shell.The first critical and severe cavitation conditions can be represented by the Rc, Er, Hi of the AE signal at the middle and outlet impeller of the shell, respectively.
(3) The migration of the bubble collapse zone, the frequency and scale of bubble rupture during the cavitation process of the waterjet propulsion pump are the main reasons for the variation of the characteristic parameters of the multi-source AE signal captured from the shell.So the different cavitation status of the waterjet propulsion pump can be precisely detected subjectively and quantitatively by the patterns of the characteristic parameters of the multi-source and multicharacteristic AE signal.

Figure 6 .
Figure 6.Cavitation AE signal within 120 microseconds: (a) AE signal within 1s,(b) AE signal within 120 μs.3.1.1.Ring count rate.The ringing count rate Rc, also known as the AE rate.Figure 7 depicts an example of a waveform.The number of individual rectangular pulses per unit time that exceed the analysis threshold is used to calculate the ringing count rate.

Figure 7 .
Figure 7.Typical ringing waveform.3.1.2.Energy release rate.Energy release rate Er is used to denote the amount of energy released from the AE waveform in a mathematical sense per unit time.It is defined as the root-mean-square value of the signal amplitude per unit time and is calculated as equation (1):

Figure 8 .Figure 9 .
Figure 8. Cavitation performance curve under Qd condition.Table 5.The corresponding NPSH for each condition.cavitation status NPSH condition a 12.43m condition b 11.32m condition c 8.81m condition d 8.24m condition e 7.19m

Table 1 .
High-speed camera parameter settings.

Table 3 .
The types and applicable waveforms of general feature parameters.