Influence Analysis on the Floating Underwater Monitoring System Applicable for Very-Low-Frequency Acoustic Measurement

Ocean vector acoustic measurement is feasible affected by the hydrodynamic interference caused by the flow fluctuations and structural vibrations, especially in the very-low-frequency monitoring. Hence, a novel horizontal floating platform including a horizontal floating cable, vertical mooring cable and floating main body is proposed and described in this paper. It has the advantages of good maneuverability along with the current and multi-stage vibration isolation. The main application of this platform is to measure the ocean ambient noise coming from the wave fluctuation and the deterministic acoustic signals such as aquatic organisms, underwater targets and sailing vehicles. The influence of the current fluctuation on the attitude angle and flow induced vibration of cables and main body are analyzed with some previous sea test data. Moreover, the comparison between the vertical type platform used before and the horizontal type platform is also discussed. It is concluded that there is obvious relevance between the attitude angle and ocean current variation. Meanwhile, the abnormal influence on the main body is caused by the vibration transmission from the fluctuation of cables. There will be the influence on the accuracy of the acoustic measurement above 100 Hz, and the inherent vibration characteristic of the main body is the primary reason.


Introduction
Ocean vector acoustic measurement system can be used to carry out the long-term and continuous observations in underwater sound field, and it has the advantages of good stealthiness, stability and maneuverability. Hence, it plays an important role in marine environmental monitoring and underwater goal-seeking. During a very-low-frequency acoustic measurement, several numerical models are established to predict the sound field characteristics and the location of underwater targets more precisely (Mavrakos and Chatjigeorgiou, 1997;Naluai et al., 2007). Nevertheless, there are some frequency overlaps between the acoustic measurement and current fluctuation caused by the flow past blunt body and vortex-induced-vibration (VIV) (Khorasanchi and Huang, 2014;Hawkes and Nehorai, 2001). It follows that the hydrodynamic characteristics of the platform, elastic suspension of sensors and fairing structures need to be researched in the very-low-frequency acoustic measurement.
Some existing underwater monitoring platforms, such as the seabed based system, towed AUV and submerged buoy, consisted of support bodies, sensor parts, electronic systems, power supply systems and anchors (Paul and Soler, 1972;Caccia et al., 2000). According to the practical re-quirements, the monitoring platform can be launched on the seafloor or some other positions underwater. There are two main types of structures, namely, bottom-mounted type and floating type, and they each have their merits and faults.
The bottom-mounted platform must be placed on the seafloor so as to reduce influences from the sea conditions and cable vibrations, meanwhile the probability of being dragged away is quite small. But it is easy to be covered by silt or marine benthos after long time operation and then the acoustic reception characteristics are affected. In addition, submarine topography differences will weaken the stability of the platform and the seismic wave propagation along the seafloor still induces some interference to the accuracy and precision of the monitoring system (Wagstaff and Aitkenhead, 2005;Shchurov, 1991). The floating platform has the advantages of adjustable depth, good maneuverability and flexible deployment, and the shape optimization can be carried out based on the AUV, UUV design experiences. But their main disadvantages are the large influences from the current and cable.
Several previous studies on the bottom-mounted platform have been finished and described by Wang et al. (2018). This paper will introduce the floating platform based on the analysis of the experimental results in the South China Sea. It is indicated from the sea test that there are obvious relevances between the acoustic signal and ocean tide especially in the frequency range less than 100 Hz. Since the dimensions and hydrodynamic characteristics of the floating body and slender cable are different in turbulent flow field, there will be increasing interference caused by the coupling vibration between the cable and body, and then the signal to noise ratio (SNR) will be greatly reduced.
Due to the military sensitivity of the underwater object positioning, very limited reports are available to the public and recent literatures mainly focus on the signal processing algorithm and recognition method. The relevant research works can be referred to Sarpkaya (2004) and Kandasamy et al. (2016). Sarpkaya (2004) summarized the VIV research progress of mostly circular cylindrical structures and reviewed the numerical models and experimental methods. Kandasamy et al. (2016) provided the vibration control methods for different marine offshore structures which include tension leg platforms (TLPs), floating production platforms and riser structures. Zhao and Guo (2016) discussed the VIV of cylinder with different cross-sections, and proved the vibration suppression effect of tail plate. Shchurov et al. (2011) described the vortex features of the acoustic intensity vector placed in the cigar-shaped measuring module. Glegg et al. (2001) described the structure and acoustic tracking performance of ambient noise sonar.
This paper firstly describes a novel floating platform structure applied for underwater acoustic monitoring. Then, the main factors which can reduce the localization accuracy and SNR are analyzed. And the theoretical relations between the sound pressure and velocity, together with the influence on the vector response are analyzed. At last, the sea test data are analyzed to determine the main influence sources through the comparison of the attitude angle variation and flow induced vibration.

Structure introduction
The horizontal floating platform consists of two main parts, namely, the cable and floating body according to the fixed-point mooring work pattern.
(1) Cable The cable has the characteristics of flexibility, large length to diameter ratio and low mass ratio. Thus, the flow field is under turbulence condition, and the Reynolds number is between 7400 and 43000 calculated from the operating flow field conditions, in which the current velocity is from 0.5 m/s to 1.5 m/s, the diameter of cable is 0.02 m, and the temperature is from 10°C to 20°C (Shi et al., 2012). Based on Karman Vortex Street theory, the frequency range of the VIV is where Sr is the Strouhal Number and Sr≈0.2 when Re< 1×10 5 , v is the current velocity (m/s), d is the cable diameter and d=20 mm. When the current velocity changes from 0.5 m/s to 1.5 m/s, the frequency range of the VIV is 2.6-13 Hz, and hence there will be the resonance vibration between the cable and flow fluctuation.
Based on the above considerations, the cable numbers should be as few as possible and the length should be short on the premise of ensuring the platform stable enough.
(2) Floating body There are several types of structures for the floating body, which are cylindrical, spherical and streamline. In order to avoid the VIV effects caused by the flow past blunt body, a simulation analysis to compare the hydrodynamic effects of the vertical and horizontal body has been carried out and shown in Fig. 1 and Fig. 2.
As shown in Fig. 1, the drag force and lift force applied on two bodies increase similarly with the current velocity. Because the vertical type body has the larger area in the incident flow direction, it suffers much larger drag force than horizontal type body. Because of the large difference between the current velocities in the upper part and lower part, the lift force applied on the vertical type body is still larger than that on the horizontal type body. Therefore, the attitude of the horizontal type body is more stable. Fig. 2 shows that the distribution area of the turbulence kinetic energy on the horizontal type body is smaller than that on the vertical type body and closer to the surface. Therefore, the flow past the body surface will dissipate quickly except the only one vortex generated behind the tail, which means that the influence caused by the flow fluctu-ation on the horizontal type body is less than that on the vertical body.
From the above analysis, the horizontal type body is a better choice due to its stability.
Consequently, the structure of the horizontal floating platform is shown as follows.
The cable consists of the vertical cable and horizontal cable, and the vertical cable is tensioned by the floating balls from the top, and connects the horizontal floating body through the horizontal cable (Shchurov et al., 2010(Shchurov et al., , 2011. Fig. 3 gives the configuration.
In order to maintain the horizontality, the net buoyance offered by the buoyance material around the whole body needs to be smaller than 10 N. In addition, the horizontality can also be adjusted through the weight change of the lead blocks inside the body, and the tail spoiler part and heading angle and pitch angle are tuned by the vertical and horizontal spoiler respectively.
The sensor fixed module is wrapped by the fairing consisting of the metallic framework and screen cloth to reduce the flow fluctuation. The technical parameters are listed in Table 1.

Acoustic measurement application
The platform is used to measure the ocean ambient noise coming from the wave fluctuation and deterministic acoustic signals such as the aquatic organisms, underwater targets and sailing vehicles. And then, the spatial characteristics, generation mechanism and position determination of the ambient noise are analyzed.
In order to measure the sound pressure and particle vibration velocity of the same point in the underwater sound field synchronously, the vector hydrophone including the sound pressure sensor and velocity sensor is adopted. The generation mechanism of the ambient noise is researched with the vector-phase analysis method, and the noise generation reason is analyzed to establish the numerical model of the very-low-frequency vector sound field.
The sound pressure in the plane wave sound field is p(r, t) = A(r)e j(ωt−kr) .
(2) The vibration velocity can be obtained by the Euler equations as: ∇ where ρ is the medium density, c is the sound propagation velocity, ω is the radial frequency, k=ω/c is the wave number, and is the gradient operator (Lei et al., 2014). The wave impendence ρc exists between p and v from Eq. (3).
Besides, the directivities of the sound pressure channel and velocity channel can be derived as: The sound pressure sensor is mainly used to measure the signals through its piezoelectric sensing element, and the velocity sensor is based on the principle of moving coil (von Winkle, 1979). The vector hydrophone is suspended on the platform by the elastic elements, which is shown in Fig. 4.
In order to improve the measurement performance, the suspension device has to meet the following requirements.
(1) The hydrophone needs to be suspended flexible enough to maintain its sensitivity; (2) The resonance frequency of the elastic suspension device should be smaller than the lower-frequency limit of the hydrophone; (3) The vibration influences and shocks from the platform or the current fluctuation should be restrained.

Analysis of the hydrodynamic noise
The frequency range of the ocean ambient noise is definitely limited from 1 Hz to 100 kHz. And the slope of the frequency spectrum above 800 Hz is -(5-8) dB/oct, while the power spectrum density level will be more than 80 dB re 1μPa/Hz. Fig. 5 gives the relation between the spectrum and the energy of the ocean ambient noise (Wenz, 1962).
As shown in Fig. 5, the turbulence noise caused by hydrodynamic force mainly exists in the frequency range of 1 Hz to 100 Hz. Under the dynamic force of flow around the platform, the vibration generated by the vector hydrophone will be transmitted to the sound receiver along the structure and hence causes the interference. Meanwhile, the flow around the hydrophone can also generate the vibration which is the pseudo noise in the response signal of the sound receiver.
According to the restriction of the sensor, the vector hydrophone is more sensitive to the hydrodynamic interference than the sound receiver. The primary reason for this phenomenon is the Karman vortex street caused by the flow around the hydrophone. The signal to noise ratios of different receivers are expressed as: where SNR V and SNR P are the signal to noise ratios of the vector and sound receiver, respectively; V 0 is the velocity; P 0 is the sound pressure; S is the surface area of the receiver;   Q 0 is the volume velocity; ω is the signal frequency; ρ is the water density; and a is the radius of the receiver (Lee, 1973).
In the case of the same size, the ratio of SNR V to SNR P is ka/3, where k is the wave number. When f=10 Hz, the value is -57 dB which corresponds to the experimental results from Keller (Keller, 1977).
The interference level will reach 30-40 dB under the extreme condition, which exceeds the interference level of the sound receiver. In order to illustrate the different responses from the vector receiver and sound receiver to the ambient noise, Fig. 6 gives the comparison of the anti-interference performance between the vector and sound receiver (Keller, 1977). The vertical coordinate shows the signal to noise ratios of different receivers. It is shown that the vector response has worse anti-interference performance in the average value and mean square error value, which is 15-45 dB less than the sound receiver in the frequency range of 8-130 Hz.
In the measurement of the vector-phase underwater signal below the frequency of 100 Hz, the interference is mainly caused by the vortex separation and wave oscillation, where the interference frequency has close relation with the flow velocity. Fig. 7 gives the comparison of flow induced noise between the vector receiver and sound receiver (Piggott, 1964). It is shown that the vector receiver has higher noise response especially in the low frequency range from 1 to 20 Hz and the magnitude increases with the flow velocity. Fig. 8 gives the spectrum of the sound pressure caused by ocean turbulence in the low frequency (Wenz, 1962). It is shown that the ambient noise magnitude increases with the flow velocity, which means that the turbulence noise caused by the vortex shedding is proportional to the flow velocity. Therefore, it is effective to improve the signal to noise ratio through reducing the flow velocity.

Relationship between the sound pressure and velocity responses in far field
As mentioned in Section 2.3, the response from the vector channel of the platform is more easily influenced by the hydrodynamic interference. Hence, in order to evaluate the anti-interference performance of the monitoring platform, the responses from the vector hydrophone and sound pressure sensor should be compared during the monitoring process. Nevertheless, the relationship between the sound pressure and velocity responses is discussed in advance.
According to Yang (2003), the wave equation of the small oscillation in the homogeneous fluid medium can be expressed as: ∇ where c is the sound propagation velocity, p is the sound pressure, t is the time, and 2 is the Laplace operator. v As for one particle in the fluid medium, the relationship of its velocity and p is deduced with Eq. (7) and Eular equation as: ∇ where is the gradient operator, and ρ is the density.
According to Hawkes and Nehorai (2001), as to the spherical wave in the infinite homogeneous medium, the sound pressure can be expressed as: where A is a constant, ω is the radian frequency, k is the wave number, k=ω/r, and r is the distance. With Eqs. (8) and (9), there is   where is the unit vector, and . Therefore, the wave impedance of the spherical wave Z 0 (kr) is expressed as: According to the far field condition and the dimension of the hydrophone, the distance between the transmitting position and receiving position should be several kilometers. In that case, the spherical wave can be treated as the plane wave in the calculation in which the sound pressure and velocity are in-phase, and Eq. (11) can be derived as: Z 0 (kr) = ρc.
(12) As to the practical monitoring in ocean environment, the measurement distance is more than 5 km in usual, and the phase offset is nearly 0°, hence Eq. (12) is applicable to evaluate the anti-interference performance of the monitoring platform.
With the evaluation criterion mentioned above, the experimental data collected from the sea test with the horizontal floating platform in the South China Sea are processed and shown from Fig. 9 to Fig. 12.
It is illuminated from Fig. 9 that the sound pressure response has almost the same variation tendency in the fre-quency domain as the velocity after the wave impendence conversion. Nevertheless, the amplitude of the velocity response is 2-5 dB larger than the sound pressure when the frequency is smaller than 25 Hz which is the main influence range of the hydrodynamic noise.
In order to extend the difference in frequency to a long period, Fig. 10 gives results in the time-frequency domain. It is indicated that there are more spectral lines in the velocity result image especially when the frequency is smaller than 30 Hz, which means that the sound pressure channel is less affected by the flow interference.
Moreover, the sound source tests have been carried out to verify the impact of the ambient noise on the single fre- Fig. 10. Comparison of the responses to ocean ambient noise from the sound pressure and velocity channels in the time-frequency domain. Fig. 11. Comparison of the responses to single-frequency signal from the sound pressure and velocity channels in the time-frequency domain, and the amplitude of signal is smaller than that of the ambient noise. Fig. 9. Comparison of the responses to ocean ambient noise from the sound pressure and velocity channels. quency signal measurement. Fig. 11 gives the results when the amplitude of the signal is smaller than that of the ambient noise. And several signal spectral lines clearly show that the frequencies are 23 Hz, 31 Hz, 56 Hz and 68 Hz. Fig. 12 gives the results when the amplitude of the signal is larger than that of the ambient noise. Only the signal spectral lines of 56 Hz and 68 Hz are clearly shown and the obvious hydrodynamic influence makes the low frequency area obscure entirely.

Influence analysis on the acoustic measurement
The hydrophone is quite sensitive to the flow induced noise and the interference level of the velocity channel can reach 30-40 dB which is higher than the output level of the sound pressure channel (Hawkes and Nehorai, 2001). While it is essential to carry out the distance measurement in the very-low-frequency acoustic research, the structural vibration characteristics of the platform need to be analyzed.

Analysis of the attitude angle
There is obvious relevance between the heading angle of the floating platform and the current direction from the attitude angle analysis of the sea test. Figs. 13-15 give the comparisons of the attitude angles and current direction. The results were obtained from the sea test in the South China Sea where the depth was 100 m and the maximum flow velocity was 0.5 m/s. The attitude angles were meas-ured by the attitude recorder and the current information was obtained by the ADCP fixed on the vertical cable. As shown in Fig. 13, there are almost the same variation tendencies in the fluctuation magnitude and rate between the heading angles of two platforms and the current direction.
Meanwhile, the relationships among the pitch angle, roll angle and current velocity are also obviously shown in Fig.  14 and Fig. 15. Hence, the fluctuations from the unsteady current will be transferred to the internal hydrophone through the platform support and then generate noise.

Analysis of the structural flow induced vibration
The vibration transmission along the cable and floating Fig. 12. Comparison of the responses to single-frequency signal from the sound pressure and velocity channels in the time-frequency domain, and the amplitude of signal is larger than that of the ambient noise. Fig. 13. Comparison between the heading angle and current direction.  body will induce the interference to the sensor inside of the platform, as well the flow past a blunt body will also generate the hydrodynamic noise. In the above-mentioned sea test, the accelerometers are fixed on the cables and floating body to measure the vibration responses of the structure under the flow impact. The reference coordinate direction is shown in Fig. 16. As shown in Fig. 16, there are three test points, and one of them is fixed on the vertical cable, another one is on the horizontal cable and the third one is on the floating body. The triaxial accelerometer is adapted to measuring the acceleration response in each point.
The comparisons of test data of the three points are shown in Fig. 17.
As shown from the time domain results in Figs. 17a-17c, the largest acceleration magnitudes in the X-axis and Zaxis are both test point-1, while in the Y-axis is test point-2. It indicates that the fluctuations normal to the axial direction of both vertical and horizontal cables have greater influence. Besides, Figs. 17d-17f give the results that the peak frequency coincidence exists in three directions between the floating body and cables, meaning that there is a great vibration influence on the stability of the platform from the cables.
In order to analyze the structural characteristics of the floating body in complex accidental loads, the vibration test is carried out in the coastal test station and shown in Fig. 18.
During the test, the floating body was launched and hung on the water by the crane. Three accelerometers were fixed on the front part, middle part and tail part, respectively. The average current velocity was about 0.6 m/s during the test. The comparisons of the test data in three direc-tions are shown in Fig. 19.
As shown in Fig. 19, several peaks exist in all three directions. The peak frequencies in the Y-axis and Z-axis are both higher than 135 Hz, while in the X-axis it is higher than 300 Hz. Hence, it is indicated that the inherent structural vibration characteristic of the floating body will induce the interference to the monitoring results in the frequency range above 100 Hz.
The flow induced vibration is the main influence source and relates to the surface roughness, shape type and structural resonance vibration of the platform body (Dewi et al., 1999;Ghasemloonia et al., 2015). Moreover, because the sectional areas of the cable and floating body are different and the current velocity is time-varying, the fluctuation frequency will change accordingly. Therefore, it is essential to research and propose the efficient vibration restraint methods.

Comparison between different types of floating platforms
The horizontal floating platform has the advantages of low-resistance and better hydrodynamic stability compared  with the vertical type body mentioned in Section 2.1, in which the configuration is shown in Fig. 20. In order to demonstrate the different performances during underwater monitoring, the experimental data from the same tests in the South China Sea are analyzed with the evaluation method proposed in Section 2.4. Both the vertical and horizontal platforms are launched to the depth of 70 m, and the sea depth is 90 m.
The analysis results are compared in Fig. 21, including the time-frequency spectrums of the sound pressure, velocity and their values.
The analysis period lasted 22 hours from July 12 12:00 to July 13 10:00. It is indicated from Fig. 21a to Fig. 21d that the vertical type has more obvious interference from the ocean ambient noise especially when the frequency is lower than 40 Hz. Although both platforms have clear response to the single frequency signal, the signal to noise ratio of the horizontal type is higher.
And then the difference values of horizontal type in Fig. 21f approach 0 dB in the frequency range higher than 20 Hz, while the vertical type has negative values in the wider frequency range from Fig. 21e. Meanwhile, the same tendency is illustrated in Fig. 22. Furthermore, the flow induced vibration has severer interference with the vertical type and both platforms are comparatively weak to withstand disturbances when the frequency is lower than 20 Hz.

Conclusions
A novel horizontal floating measurement platform for underwater acoustic monitoring with very-low frequency is proposed to improve the accuracy and stability of the underwater acoustic monitoring. Several conclusions are drawn through the analysis of the hydrodynamic noise source and the sea test.
(1) The platform consists of the cable part and floating part. Among them, the vertical cable is tensioned by the floating balls at the top, and connects the horizontal floating body through the horizontal cable. The application includes the ocean ambient noise measurement and DOA estimation of the underwater targets.
(2) The main hydrodynamic influence sources are the VIV from the cables and floating body together with the current fluctuation.
(3) There are obvious relevancies between the heading angle and ocean current direction, as well the pitch angle, roll angle and ocean current velocity.
(4) The abnormal influence on the main body is caused by the vibration transmission from the fluctuation of cables. There will be influence on the accuracy of the acoustic measurement above 100 Hz, and the inherent vibration characteristic of the main body is the primary reason.
(5) Through the comparison between the vertical type  WANG Zhen et al. China Ocean Eng., 2019, Vol. 33, No. 3, P. 373-383 381 platform used before and the horizontal type platform is discussed with the experimental data, it is indicated that the ocean ambient noise has more obvious interference with the vertical type and both platforms are unstable when the frequency is lower than 20 Hz.