Influence Analysis and Vibration Restraint Solutions Research on the Underwater Acoustic Monitoring System

In order to reduce the hydrodynamic and structural influences on the detection accuracy especially in the very-low-frequency range, some vibration restraint methods are raised, which are the wrapped fairing improvement, the floating body shape improvement and the cable vibration reduction treatment. Through the improvement analysis and experimental comparison, the final treatments are proposed, namely the multilayer wrapped fairing structure with composite materials, the floating body with NACA0024 airfoil section and X-shape tail spoiler, as well as the brush cable. The sea test is carried out to evaluate the vibration restraint effect. Through comparison of the responses to the ocean ambient noise and the direction of arrival (DOA) estimations with the same underwater transmitting transducer, the results indicate that the horizontal floating platform with vibration restraint treatment has obvious flow resisting effect especially in low frequency range and more accurate DOA estimation.


Introduction
In order to evaluate the ocean ambient noise and detect the different underwater objects, the long-term and continuous observation of the underwater sound field is required (Mavrakos and Chatjigeorgiou, 1997). Nevertheless, the platforms are easily affected by the surrounding flow, generally in the form of vortex-induced vibration. In particular, resonance vibration will occur when the vibration frequency coincides with the natural frequency of the platform, which will have a serious impact on the accuracy of hydroacoustic measurement. Therefore, it is necessary to study the hydrodynamic characteristics of the platform and the vibration isolation measures of the sensor. This paper will introduce several vibration restraint methods for the floating underwater acoustic monitoring platform including the mooring cable and floating body, which have the advantages of better stealthiness, stability and maneuverability. Besides, there are also disadvantages such as the larger influences from the current and the cables (Shchurov et al., 1991(Shchurov et al., , 2010(Shchurov et al., , 2011. The shape design of existing underwater vehicles such as AUV and UUV can also be utilized to improve the hydrodynamic characteristics. It is indicated from several sea tests that the relevance is very obvious between the acoustic signal and ocean tide especially when the frequency is smaller than 100 Hz. Since the dimensions and the hydrodynamic characteristics of floating body and cable are different, there is coupling vibration between the cable and the body, and then the signal to noise ratio (SNR) will be obviously reduced. Furthermore, due to the military sensibility of the underwater acoustics, the available relevant literature is rare and the recent studies mainly focus on the signal processing and recognition method. The relevant research works can refer to Sarpkaya (2004), Kandasamy et al. (2016), and Sarpkaya (2004).
The underwater floating platform is single point mooring system. Hence, the floating body can be designed with the airfoil shape design method for AUV or underwater glider. Meanwhile, the method of damping tube design in offshore platform can also be applied to the cable. In view of the above consideration, the airfoil shape design is mainly based on the underwater navigation object consisting of the rotating body shell, the tail wing and the control surface. Besides, the hydrodynamic improvements are carried out through the modification of downstream profile with the dimension parameters in airfoil library (Ahmed and Duan, 2016;Sherman et al., 2001). Since the underwater navigation objects generally need to travel efficiently over long distances, a higher lift-drag ratio is usually required in shape design. However, to maintain the stable floating working condition and attitude, the profile should have lower values of drag force and lift force.
The existing investigations on marine risers are mainly achieved through modifying the outer surface shape or adding fairing devices to suppress the vortex shedding. Zdravkovich (1981) conducted the detailed study of such methods and divided them into the following three categories.
(1) Surface protrusions. To change the separation line or separate shear layers, such as threads, lines, fins, etc; (2) Wrap. To change the section layer, such as perforated sleeves, wire mesh, axial slats, etc; (3) Near wake stabilizer. To supply the entrainment layer interactions, such as streamers, fairings, dividers, guide wings, bottom rows and slits, etc.
In addition, Sakamoto and Haniu (1994), Dalton et al. (2001), and Lee et al. (2004) studied the influence of control rods on the drag characteristics and flow field structure of cylinder, concluding that the size and placement of control rods in different flow field environments influenced the vibration suppression effect. For the flexible and slender structure of the underwater measuring platform, the towing force is limited. Although the cable is always in a stressed condition, its rigidity is still smaller than that of the riser structure.
This paper will analyze the major interference sources that affect the localization accuracy and SNR. Moreover, several effective suppression methods including streamline design, low-drag cable and fairing structure are proposed to reduce the hydrodynamic influences. Finally, some sea tests are carried out to evaluate the vibration restraint effect of the new floating platform through the comparison of the responses to the ambient sea noise and the DOA estimation of the sound source target.

Influence analysis of floating platform
Two main parts of the floating platform are the cable and floating body. And the fluctuations generated from the unsteady current will transmit to the internal sensors through the platform support.
The hydrophone is quite sensitive to the flow-induced noise and the interference level in velocity channel will be up to 30−40 dB which is higher than the output level of the sound pressure channel (Hawkes and Nehorai, 2001). The vibration transmission along the cable and floating body will induce the interference to the sensor inside the platform, and the flow past blunt body will also generate hydrodynamic noise. In order to analyze the influence from the platform, the vibration responses from different test points on the table and floating body are measured through the sea test. The configuration is given in Fig. 1.
As shown in Fig. 1, three test points are respectively fixed on the vertical cable, the horizontal cable and the floating body. The accelerometers are used to measure the acceleration response in each point.
The vibration responses from the three points are shown in Fig. 2.
It is indicated from Figs. 2a−2c that the maximum responses of X-axis and Z-axis are in test point-1, while that of the Y-axis is in test point-2. Hence, the fluctuations generated in the radial direction of both vertical and horizontal cables are larger. Moreover, Figs. 2d−2f indicate that the floating body and cables share many coincident frequency values at the peak magnitude, which means that the fluctuation from the cables have the great vibration influence on the stability of the platform.

Vibration restraint solutions
The flow-induced noise is the main influence and related to the surface roughness, shape type and structural resonance vibration of the platform body (Dewi et al., 1999 ;Ghasemloonia et al., 2015). Moreover, as the section shapes of floating body are different and the current velocity is time-varying, the fluctuation frequency will fluctuate accordingly. Therefore, the vibration restraint methods are studied from three aspects, namely the wrapped fairing improvement, the floating body shape improvement and the cable vibration reduction treatment.

Fairing improvement in material and configuration
Fairing can efficiently reduce the turbulent flow influence on the wrapped structure according to its excellent performances of flow restraint and sound transmission (Wang et al., 2018). Hence, the material and structure of the fairing are invetigated and different fairing configurations are compared. WANG Zhen et al. China Ocean Eng., 2020, Vol. 34, No. 5, P. 718-729 719 (1) Flow restraint test The comparison test of three materials which are oxford cloth, felt and filtration fabric is carried out in the offshore area. The current velocities inside and outside the wrapped body are measured by two current meters respectively. To evaluate the different flow restraint effect, the current velocity ratio which is the internal current velocity from current meter-1 compared with the external current velocity from current meter-2 is defined. Hence, the fairing with better flow restraint effect has the lower current velocity ratio which means that it can reduce the turbulent flow influence from outside.
The test configuration is shown in Fig. 3 and the comparison of current velocity ratio with different materials is shown in Fig. 4, with the mean values of the velocity ratio being plotted as well. The fairing with oxford cloth has the minimum value of current velocity ratio which means that it has better flow restraint effect than the fairing with felt and filtration fabric.
(2) Sound transmission test The sound transmission effect is mainly evaluated by the sound insertion loss.
where I i is the sound insertion loss, p 0 is the sound pressure in test point without fairing, and p i is the sound pressure in test point with fairing. The sound pressure can be expressed as: where M is the hydrophone sensitivity and U is the output voltage. The configuration of sound transmission test is shown in Fig. 5. During the test, both the buoy and sound source are fixed in the same horizontal plane. The sound pressure responses with and without fairing are compared to evaluate the sound transmission effect.   Four fairing configurations of Oxford cloth, metal net, coarse spongy fabric and smooth spongy fabric are tested. The test bracket is shown in Fig. 6, and the configurations are listed in Table 1.
As seen from Table 1, the fairing with oxford cloth has the highest compactness while the metal net is sparest. The comparison of sound transmission test results is shown in Fig. 7.
It is indicated from Fig. 7 that the sound pressure inside the fairing wrapped body is smaller than the outside. And all four fairing types show similar tendency. The material with the best sound transmission effect is the metal net while the worst is Oxford cloth. Consequently, the density of the fairing material is tightly correlated with its sound transmission ability, i.e. the sparser the material, the better the sound transmits. Meanwhile, the sound insertion loss becomes larger when the frequency increases.
From the above analysis, variation tendencies between the flow restraint and sound transmission are opposite. Thus, both characteristics should be considered in the configuration of fairing.
A novel fairing structure that consists of two materials with different densities is proposed. There are seven layers for the fairing, of which the inner six layers are made of oxford fabric cloth and the outer layer is metal net. The configuration is shown in Fig. 8.
The comparison of floating bodies with and without de-signed fairing is carried out in the coastal test station, and the hydrophone responses are shown in Fig. 9, where P is the sound pressure; V is the velocity; ΔP and ΔV are the differences of the sound pressure and the velocity with and without fairing. Considering that applicable frequency range of the hydrophone is above 10 Hz, the results are analyzed from 10 Hz to 100 Hz. As shown in Fig. 10, the magnitude in the velocity channel is larger than that in the sound pressure channel when the floating body is not wrapped with fairing. Furthermore, the responses from both channels decrease obviously after wrapped with fairing especially for the velocity when the frequency is smaller than 30 Hz.

Floating body shape improvement
Because of the positive buoyancy, the platform can float without the lift force. To maintain stability, several parameters such as the fluctuation magnitude, the drag force, the lift force and the pitch moment should all be small enough. Meanwhile, the flow-induced noise around the platform also  WANG Zhen et al. China Ocean Eng., 2020, Vol. 34, No. 5, P. 718-729 721 needs to be minimal in consideration of the high-sensitive sensors such as hydrophone. NACA airfoil section is mostly adopted as the structure outline in the pre-existing vehicles to decrease the flow resistance (Ahmed and Duan, 2016). Therefore, this paper, based on the previous experience of AUV structure design, tries to improve the symmetric airfoil section with small lift coefficient. And then the velocity, pressure and flow-induced noise are compared to evaluate the hydrodynamic performance of different structures.
In order to fix the instruments, the structure dimension should be larger than that of a cylinder with the axial length being 1.2 m and basal diameter being 0.6 m. Therefore, four configurations are drafted based on NACA airfoil section. The configurations including the dimensions of radial section diameter (R d ) and axial length (L a ) are shown in Fig. 10. The dimensions of the computation flow field are shown in Fig. 11.
Three calculation conditions with the flow velocity being 0.54 m/s, 0.3 m/s and 0.1 m/s respectively are adopted according to the statistical results of the sea trial over years.
The settings of the computational fluid dynamics (CFD) analysis are as follows.
(1) Separate implicit solver is adopted, and the govern-  ing equation is discretized with finite volume method. The discretization scheme is second-order differential and the pressure-velocity coupling scheme adopts SIMPLE method. RMS residual monitoring type is adopted and the residual value is 10 −6 . Through repeated computations, it is completed convergence after 200 iterations.
(2) Inlet boundary condition: velocity-inlet type is adopted. Outlet boundary condition: outflow type is adopted. Side and top boundary condition: moving wall type is adopted and the speed is the same as the flow velocity. Bottom and fluid−structure interaction boundary condition: stationary wall type is adopted.
The fluid area is meshed by ICEM software and the number of meshes is more than 200 000 to ensure the calculation precision. Structured hexahedral grid is adopted on account of its convenient adjustment of the mesh density and higher accuracy.
The variations of drag, lift and pitch moment coefficients to the flow velocity are shown in Fig. 12. It is illustrated from Fig. 12 that the drag coefficient will decrease and the pitch moment coefficient will increase as the lengthdiameter ratio increasing, while the lift coefficient changes little. Meanwhile, the drag and pitch moment coefficients are directly proportional to the flow velocity, which means that the different platforms of this paper have obvious hydrodynamic difference in drag and pitch moment.
In order to study the flow-induced noise in the surrounding fluid field, FW-H acoustics solver is used. The time step size is 0.5 ms and the calculation steps are 1 000, thus the total calculation time is 0.5 s and the highest possible frequency the acoustic analysis can generate is 1/(2×(2×0.5 ms)) = 500 Hz. The number of time steps is 200, thus the frequency step is 1/(200×0.5 ms) = 10 Hz. Four test points which are 0.1 m away from the surface of the platform are adopted to analyze the flow-induced noise, as shown in Fig. 13.
The sound pressure levels of flow-induced noise in test points A, B, C and D are shown in Fig. 14. It is illustrated from Fig. 14 that the flow-induced noise will decrease with the increment of length-diameter ratio and is proportional to the flow velocity. The noise value achieves its maximum in the test points B and D but minimum in test point C.
It is concluded that the increasing of the length-diameter ratio of airfoil section can reduce the drag effect and then improve the floating body stability. Hence, NACA0024 airfoil section is adopted as the body shape.
Besides, the dimensions of the tail spoiler part needs to be calculated in order to compensate the stability decreasing caused by the large length-diameter ratio body. And the parameters such as top length, height and width of section are analyzed to obtain the optimized structure. The variations of flow-induced noise with the three parameters are shown in Fig. 15.
It is indicated that the flow-induced noise is directly proportional to the top length but inversely proportional to the height and width. Therefore, the top length is the most sensitive parameter to the hydrodynamic characteristic of the tail spoiler.
Consequently, the dimensions of the tail spoiler part are defined as below and shown in Fig. 16. The height is 150 mm, the top length is 18.75 mm and the sectional width is 15.6 mm.

Cable vibration reduction treatment
The vortex shedding is generated in the cable under the impact of the current. Especially when the vortex shedding frequencies are close to the inherent frequencies of the cables, the lock-in phenomenon of Vortex-Induced Vibration (VIV) will appear. Hence, the vibration magnitude and the phase angle increases obviously.   WANG Zhen et al. China Ocean Eng., 2020, Vol. 34, No. 5, P. 718-729 723 The brush cable is proposed to reduce the VIV along the cables, which are indicated in Fig. 17. The sea tests are carried out to compare the vibration restraint effect of brush cables with different configurations.
The cable length is 21 m, and the middle part of 15 m length is wrapped with tested brush. One side is fixed with weight while the other side fixed on the test boat. The structural parameters are listed in Table 2.
The sea test is carried out in clear day in the coastal area as the ocean condition is below the second level. In order to simulate the current velocity of 2 knots, the tested brush cables are dragged by the test boat, and the vibration responses are measured with accelerometers fixed on the cables.
The comparisons of vibration restraint effects with different structural parameters are conducted.
(1) Comparison of brush length As shown from Figs. 18−21, it is indicated that the cable wrapped with short brush has the lower responses in both axial and radial directions, which means that the increasing of the brush length will reduce the vibration restraint effect.
(2) Comparison of hardness As shown in Fig. 22 and Fig. 23, it is indicated that the cable wrapped with soft brush has the lower responses in both axial and radial directions.
(3) Comparison of spacing distance As shown in Fig. 24, it is indicated that the cable wrapped with large brush spacing distance has the lower responses in both axial and radial directions. Consequently, the cable wrapped with short, soft and large spacing distance brush has the better vibration restraint effect.

Experiment analysis
In order to evaluate the hydrodynamic characteristics of the platform after vibration reduction treatment, the sea test in the South China Sea was carried out between July 8, 2015 and July 20, 2015, during the fishing-off season, therefore the interference can be reduced. The sea floor in the test area is flat and composed of sediment, with the depth of 100 m. During the sea test, the ocean condition was below the third level.
The distance between two platforms is 1 km. As shown in Fig. 25, both platforms have similar floating connection structure, where the floating body is fixed by the horizontal cable and the vertical cable. In order to maintain the vertical posture of the vertical cable, a floating ball was tied at the top end. Both the vertical and horizontal cables can bend or stretch with the turbulent motion of the current, so as to re-       WANG Zhen et al. China Ocean Eng., 2020, Vol. 34, No. 5, P. 718-729 725 duce the vibration effect on the floating body. The floating body shape of Platform-1 is the structure of cylinder with four triangular tails, while Platform-2 has the same floating body shape as mentioned in Section 3.2. Besides, Platform-2 was treated by the vibration reduction methods as mentioned in Sections 3.2 and 3.3. The brush cables and the fairing adopted in Platform-2 are shown in Fig. 26.
The monitoring time lasted for more than 24 hours, and the tide influences can thereby be recorded. The flow velocity in the test period was recorded by the current meter fixed on the vertical cable and shown in Fig. 27. The maximum velocity was 0.41 m/s and the average value was     Fig. 27. Current velocity from July 9 to July 10.

726
WANG Zhen et al. China Ocean Eng., 2020, Vol. 34, No. 5, P. 718-729 0.24 m/s. The experimental analyses include the comparison of responses to the ocean ambient noise and the DOA estimation to the underwater transmitting transducer in the fixed point mode and navigation mode.

Ocean ambient noise test
Two test periods are analyzed. The first test period was from July 10 15:00 to 17:00 and the current velocity was 0.4 m/s, while the other was July 9 22:00 to July 10 00:00 and the current velocity was 0.04 m/s.
(1) Test period with current velocity being 0.4 m/s The responses are given in Fig. 28. As shown in Figs. 28a−28d, higher responses were obtained from both channels of Platform-1 especially when the frequency was smaller than 50 Hz. Meanwhile, the responses of both platforms above 50 Hz were close as seen from Fig. 28e. It is indicated that Platform-1 suffers more obvious flow influence. Nevertheless, a peak appeared in 32 Hz in both channels from Fig. 28e, and it was more obvious for Platform-2. Meanwhile, this phenomenon also exists in Fig. 29. This anomalous peak was caused by other simultaneous testing in the adjacent area. However, it is still illustrated that the vibration restraint treatments can effectively reduce the flow influence for the reason that the responses to this peak signal from Platform-2 were more obvious than that to Plat- Fig. 28. Time-frequency spectrums from July 10 15:00 to 17:00. Fig. 29. Time-frequency spectrums from July 9 22:00 to July 10 00:00.
(2) Test period with current velocity being 0.04 m/s The responses from both channels are shown in Fig. 29. As shown in Fig. 29e, the response magnitudes from both channels of two platforms were reduced by 12 dB with the decreasing of current velocity, while Platform-1 was still obviously influenced when the frequency was less than 50 Hz. Overall the flow influence on Platform-2 was significantly less than that on Platform-1. It is because that the current velocity value is small, while the magnitude of the abnormal response at the frequency of 32 Hz is higher than that of the ambient noise, so that both platforms have quite obvious responses on the pressure and velocity channels.
Consequently, it is concluded that the platform after vibration restraint treatment suffers from less flow influence.

DOA estimation test
The positions of two platforms and the sound source of transmitting transducer are shown in Fig. 30, where the sound source in fixed point mode is expressed by sound source-1, the track of navigation mode is expressed by sound source-2, and the sailing direction is along the black arrow.
The comparison of DOA estimation results from two platforms and the GPS data are shown in Fig. 31 and Fig. 32, where the orange line is the track of sound source, and green line is the track of test ship. It is indicated from Fig. 31 and Fig. 32 that both platforms could correctly estimate the sailing track, while Platform-1 was influenced when the current velocity increased to 0.4 m/s. Meanwhile, the estimation result of Platform-2 was less affected by the current velocity. Hence, the vibration restraint treatment this paper proposed can efficiently improve the measurement accuracy and stability of the underwater monitoring platform.

Conclusions
A novel horizontal floating measurement platform for underwater acoustic monitoring in very-low-frequency is proposed. Several conclusions can be drawn through the analysis of vibration restraint treatment and the sea test.
(1) The main hydrodynamic influence sources are the VIV from the cables and floating body as well as the cur-   rent fluctuation. The vibration in the radial direction of the cables is larger and there are several coincidence frequencies existing both in the cable and floating body.
(2) The vibration restraint solutions are concluded below. From the fairing improvement, it is indicated that the fairing with sparser material has better sound transmission but worse flow restraint and hence the bi-material multilayer fairing is proposed. From the body shape improvement, it is indicated that the hydrodynamic stability will increase and the flow-induced noise will decrease as the length-diameter ratio of body increases. From the cable treatment, it is indicated that the cable wrapped with soft, short and large spacing distance brush has the lower vibration disturbances.
(3) Through analyses of the suppression methods, the multilayer wrapped fairing structure, the floating body with NACA0024 airfoil section and X-shape tail spoiler, as well the brush cable are selected to restrain the hydrodynamic influences.
(4) As obtained from the sea test, the horizontal floating platform with vibration restraint treatment has obvious flow resisting effect especially in low frequency range and more accurate DOA estimation.