Open Access
Add to BibSonomy
Abstract
In real-life marine environments, the composition and grain size of suspended sediments and the resuspension and sedimentation of sediments caused by turbulence may have a significant impact on underwater wireless optical communication (UWOC). However, to date, researchers have not conducted quantitative research on this issue. To this end, we innovatively study the effects of different compositions and grain sizes of suspended sediments on UWOC and the effects of turbulence-induced sediment resuspension and sedimentation on UWOC in this paper. Quartz and kaolin with different grain sizes are used to simulate sediments in seawater. An oscillating grid that can vary frequency and stroke is used to generate turbulence of different intensities. By comparing the turbidity and optical power density of different simulated sediments with different grain sizes, we find that the smaller the grain size of the simulated sediments, the higher the bit error rate (BER) under the same turbidity. But different simulated sediments with different grain sizes have similar effects on BER performance under the same optical power density. Therefore, turbidity can be used to characterize the changes of underwater channels, and optical power density can be used to evaluate the attenuation of light at the receiving end after transmission through the underwater channel. By continuously changing the frequency of the grid to cause the sediments to resuspend and sink, we prove that the process of turbulence-induced sediment resuspension and sedimentation can seriously affect the BER performance. The larger the frequency of the grid, the greater the turbulence intensity and the worse the BER performance. This study lays a foundation for the practical application of UWOC in mobile ocean observation networks.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Ocean observation networks composed of various platforms (e.g., buoys, submersibles, bottom-sitting platforms, etc.) and instruments (e.g., conductivity-temperature-depth sensor, seismograph, camera, etc.), play an important role in marine environmental monitoring, marine resource exploration, and marine disaster warning. The massive data collected by the instruments is mainly transmitted to the shore through the submarine photoelectric composite cable. Nowadays, with more and more underwater robots (e.g., remote operated vehicles, autonomous underwater vehicles, gliders, etc.) applied in ocean observation networks, underwater acoustic communication has become an effective complementary technology for submarine cable communication, providing an important guarantee for underwater mobile platforms to obtain and transmit information [1]. It can support long-distance (∼km) wireless data transmission, but it has low data rate (∼kbit/s) and high latency. In the future, underwater wireless optical communication (UWOC) technology with high bandwidth and low latency has great potential in mobile ocean observation networks. It can support real-time and high-speed (∼Gbit/s) data transmission over short distances (∼100 m) [2].
One of China’s ocean observation networks is built in the East China Sea. The East China Sea is extremely turbid. Because the Yangtze and the Qiantang River carry hundreds of millions of tons of sediments to the East China Sea every year. Moreover, the movement of sediments in the East China Sea is extremely complicated due to the mixed effects of wind waves, ocean currents, estuarine freshwater, and seawater [3]. The suspended sediments and complex hydrodynamic factors in the East China Sea can easily lead to the degradation of UWOC performance and even the failure of UWOC. In this context, it is of great significance to quantitatively study the sediments in the East China Sea and the effects of sediment resuspension and sedimentation under hydrodynamic factors on UWOC.
Table 1 lists typical work on turbidity and turbulence. We summarize that (1) most researchers use Maalox [4] to study the effect of turbidity on UWOC. A few researchers use iron oxide [5], kaolin [6], Mg(OH)2 powder [7], and dirt [8] to study the effect of turbidity on UWOC. However, none of them consider the effect of composition and grain size of suspended sediments on UWOC. (2) Some researchers use optical power/optical power density [9,10] to characterize the attenuation of the optical path, while others use turbidity to characterize the degree of light blocking in water. However, turbidity is not only related to the content of suspended substances in water, but also to their size, shape, refraction coefficient, etc. (3) Researchers commonly study the effect of temperature [11] or salinity [12] gradients-induced turbulence on UWOC. Some researchers also study the effect of turbulence caused by bubbles [13] and mechanical methods (e.g., propellers [14], water jets [15]), and pumps [16]) on UWOC. However, the above temperature or salinity gradients-induced turbulence is difficult to achieve the resuspension of sediments on the laboratory testbed. The above mechanical methods-induced turbulence is not stable, uniform, and easy to control, which poses a great challenge to study the effects of turbulence-induced sediment resuspension and sedimentation processes on UWOC.
Table 1. Typical work on turbidity and turbulencea
In this paper, we first innovatively study the effects of different compositions and grain sizes of suspended sediments on UWOC. Quartz and kaolin with different grain sizes are proposed as the main representative of sediments. An oscillating grid is used to generate stable and uniform isotropic turbulence with the same intensity [17]. By studying the relationship of turbidity and optical power density, we find that different kinds of sediments with different grain sizes have different effects on bit error rate (BER) performance under the same turbidity, but have similar effects on BER performance under the same optical power density. Turbidity increases with the deterioration of channel conditions. Moreover, the smaller the grain size, the higher the BER under the same turbidity. Therefore, it is reasonable to use turbidity to characterize variations in underwater channel and optical power density to assess light attenuation at the receiver after transmitting through the underwater channel. When using turbidity to characterize UWOC channels, the composition and grain size of the suspended sediments must be considered. We then innovatively study the effects of sediment resuspension and sedimentation on UWOC under different turbulence intensities. The experimental results show that the larger the frequency and stroke of the grid, the greater the turbulence intensity and the worse the BER performance. In the case of continuous change of grid frequency, the process of sediment resuspension and sedimentation caused by turbulence can seriously affect the BER performance.
2. Experimental setup
2.1 UWOC system
Figure 1 shows the schematic diagram of the UWOC system. In this work, we used 4-quadrature amplitude modulation (QAM) orthogonal frequency division multiplexing (OFDM) with high spectral efficiency to effectively improve the data rate. The specific parameters of the 4-QAM OFDM signals are listed in Table 2. It is worth noting that cyclic prefix was used to effectively resist inter-symbol interference, and the frequency gap near direct current (DC) was used to effectively resist low-frequency noise. At the transmitter side, 4-QAM OFDM signals generated by MATLAB were first loaded into an arbitrary waveform generator (AWG, Siglent SDG7102A), where the sampling rate was set at 5 MSa/s and the peak-to-peak voltage (Vpp) was set at 0.5 V. After transmitting through an amplifier (AMP, Mini-Circuits ZHL-6A-S+) with a fixed gain of 25 dB and a key-press variable electrical attenuator (ATT, KT3.0-90/1S-2S) with an attenuation of 14 dB, the signals were superimposed on a blue light-emitting diode (LED, Cree XLamp XB-D) via a bias tee (Bias-T, Mini-Circuits ZFBT-4R2GW+). The AMP and the ATT were used to adjust the amplitudes of the signals in the linear operating region of the LED. A planoconvex lens with a focal length of 5.08 cm was placed in front of the LED to collimate light.
Table 2. Parameter values of the 4-QAM OFDM signals
After transmitting through a 46-cm long water tank, the optical signals were detected by an avalanche photodiode (APD, Thorlabs APD430A2/M). The water tank was made of highly transparent acrylic sheets. In the experiment, we characterize the channel in terms of turbidity and received optical power density, respectively. Turbidity was measured by a recording turbidity monitor (Campbell Scientific OBS-3A). Received optical power density was measured by an optical power and energy meters (Thorlabs, PM100). A planoconvex lens with a focal length of 3 cm was fixed to the front of the APD to make the beam converge to the 0.2-mm detection active area of the APD. A DC block was used to remove the DC noise. Finally, the output signals were captured by a mixed signal oscilloscope (MSO, Tektronix MSO64) and processed offline. The sampling rate and sampling length of the MSO were set at 25 MSa/s and 2 M, respectively.
2.2 Oscillating grid turbulence system
Figures 2(a) and 2(b) show a main view and 45° view of the schematic diagram of the oscillating grid turbulence (OGT) system, respectively. In the OGT system, we used a customized alternating current asynchronous motor with 750-W power. The motor was equipped with a speed controller to ensure precise control of the rotation speed. The motor spindle drives an adjustable eccentric wheel that reciprocates the connected rods and drives the grid reciprocating. The oscillation frequency and stroke of the grid were adjusted by controlling the motor speed and the eccentricity of the adjustable eccentric wheel. The grid was composed of crossed metal rods and was driven by a motor that vibrated in a direction perpendicular to the plane of the grid. The length, width, and thickness of the grid were 40 cm, 40 cm, and 0.4 cm, respectively. The grid was placed vertically in the middle of the water tank (Length: 46 cm, width: 46 cm, and height: 73 cm). The edge of the grid was 3 cm away from the side wall of the water tank and 5 cm away from the bottom of the water tank, as shown in Fig. 2(a). The height of the light path from the bottom of the water tank (H) was set at 10 cm, 20 cm, and 30 cm in the experiment. The center of the oscillation stroke of the grid was at the middle of the water tank (i.e., 23 cm away from the side wall of the water tank), as shown in Fig. 2(b). In the experiment, we first put the sediments evenly at the bottom of the empty water tank, and then slowly injected water from the bottom of the water tank outlet until the water flooded the top of the grid, so as to avoid the resuspension of sediments.
Because sediments are mainly composed of sand, silt, and clay. The main components of sand and silt are quartz, feldspar, mica and so on. The main components of clay minerals are kaolinite, montmorillonite, hydromica, glauconite and so on. Among them, the chemical properties of quartz and kaolin are relatively stable, which are not prone to aggregation and sedimentation. We select quartz with different grain sizes as the main representative of sand and silt, and kaolin with different grain sizes as the main representative of clay minerals, respectively. Figures 3(a) and 3(b) show the Sediment-A and Sediment-B collected from the East China Sea. Sediment-A was collected by a homemade multi-corer. The longitude and latitude coordinates of the collected Sediment-A are 122°19.256E, 29°54.413N, and longitude and latitude coordinates of the collected Sediment-B are 121°55.227E, 30°51.469N, as shown in Fig. 3(c).
Fig. 3. (a) Sediment-A collected from the East China Sea, (b) Sediment-B collected from the East China Sea, and (c) longitude and latitude coordinates of Sediment-A and Sediment-B.
During the oscillation of the grid, jets were formed at the openings of the grid and wakes were formed at the shields of the metal rods. The combination of jets and wakes resulted in isotropic and zero-mean turbulence. Equation (1) is the empirical equation for the root mean square of the flow velocity of the OGT [18].
where, C is a constant, f is the oscillation frequency of the grid, S is the oscillation stroke of the grid, M is the size of the inner side of each grid, z is the vertical distance from the plane of the grid which was set at 5 cm in this work. It can be seen that the quantitative control of turbulence can be easily achieved by changing the frequency, stroke, and size of the grid.3. Experimental results
3.1 UWOC system demonstration
We first measured the optical spectrum of the blue LED under a drive voltage of 2.75 V. It has a peak wavelength of 477 nm, as shown in Fig. 4(a). We then measured the current versus voltage curve of the blue LED. It has good linearity, as shown in Fig. 4(b). After that, we measured the frequency response of the system consisting of the blue LED and the APD over a 46-cm air channel. The -3 dB bandwidth of the system is not more than 2 MHz due to the limited bandwidth of the LED, as shown in Fig. 4(c). Finally, we measured BER performance of the 4-QAM OFDM signals with different data rates in clean water. The maximum data rate is 5.01 Mbit/s with a BER of 3.3 × 10−3, which is below the forward error correction (FEC) limit of 3.8 × 10−3, as shown in Fig. 4(d). In the following experiments, we used an error-free data rate of 4.88 Mbit/s in clean water to study the effect of underwater channels on the performance of the UWOC system.
Fig. 4. (a) Optical spectrum of the blue LED under a drive voltage of 2.75 V, (b) current versus voltage curve of the blue LED, (c) frequency response of the system consisting of the LED and the APD over a 46-cm air channel, and (d) BER performance of the 4-QAM OFDM signals with different data rates in clean water.
3.2 Test results of grain size distribution
We measured the grain size distribution of Sediment-A and Sediment-B, so as to simulate them using the quartz of similar grain size distribution in the subsequent experiments. A laser particle size analyzer (Coulter LS-230) was used to measure the grain size distribution of Sediment-A, Sediment-B, and 110-160 mesh, 160-200 mesh, and 200-300 mesh quartz. As shown in Fig. 5, the grain size distribution of Sediment-A is close to that of the 200-300 mesh quartz, but it is slightly finer. The grain size distribution of Sediment-B is between two kinds of quartz (i.e., 110-160 mesh and 160-200 mesh). Figure 5 also shows that 200-300 mesh quartz contains many finer grains. 200-300 mesh refers to the main grain sizes. The vertical axis indicates the volume percent content. When the components of the substance to be measured are the same, the volume percentage content is the mass percentage content.
Fig. 5. Measured grain size distribution of Sediment-A, Sediment-B, and 110-160 mesh, 160-200 mesh, and 200-300 mesh quartz.
3.3 Test results of turbulence in tap water
We studied the effect of turbulence on the BER performance of the UWOC system in the water tank filled with tap water. S was set to 7 cm, which was the maximum oscillation stroke of the grid. Figure 6 shows the BER performance of the OFDM signals and the corresponding received optical power density varying with the frequency of the grid and the height of the light path from the bottom of the water tank. The BERs fluctuated within a range of 0∼2.05 × 10−5 (see Fig. 6(a)) and the corresponding received optical power density fluctuated within a range of 3.41 $\textrm{mW}/\textrm{c}{\textrm{m}^2}$∼3.74 $\textrm{mW}/\textrm{c}{\textrm{m}^2}$ (see Fig. 6(b)), both of which were within a reasonable error range. It indicates that turbulence in 46 cm tap water has little effect on the BER performance.
Fig. 6. (a) BER performance of the OFDM signals and (b) received optical power density varying with the frequency of the grid and the height of the optical path from the bottom of the water tank.
3.4 Test results of turbidity and optical power density of different simulated sediments
We tested the relationship between turbidity of different sediments and BER, and compared it with the relationship between optical power density of different sediments and BER. As shown in Fig. 7(a), Mg(OH)2, three grain sizes of kaolin, 200-300 mesh quartz, Sediment-A, and Sediment-B have different BERs under the same turbidity. It indicates that different sediments with different grain sizes have different impact on BER performance under the same turbidity. When the sediments were 325-mesh kaolin with a turbidity of 15 NTU, the BER of the OFDM signals was 5.34 × 10−4. When the suspended sediments were finer kaolin (1250-mesh) with a turbidity of 15, the BER of the OFDM signals was as high as 5.17 × 10−2. It indicates that the smaller the grain size of the kaolin is, the higher the BER is, when the turbidity is the same. Therefore, turbidity can only be used as a reference when a single type of sediment is considered. From Fig. 7(b), we can see those different sediments with different grain sizes have almost the similar effect on BER performance under the same optical power density. When the optical power density at the received end was higher than 0.35 $\textrm{mW}/\textrm{c}{\textrm{m}^2}$, BERs of different sediments were below the FEC limit of 3.8 × 10−3. Since the turbidity measured by the recorded turbidity monitor we adopted characterizes the backscattering (90°∼180°) ability of suspended particles and dissolved substances in the liquid medium recording turbidity monitor, it can indirectly estimate the concentration and distribution of these particles and substances in the liquid, and thus characterize the change of the channel. Since optical power density characterizes the transmittance of light, it can assess the attenuation of different sediments to the optical path.
Fig. 7. BER performance of the OFDM signals at different (a) turbidity and (b) received optical power density for different sediments with different grain sizes.
We then studied the effect of 110-160 mesh quartz on the BER performance of the OFDM signals at different time under the conditions of varying the frequency and stroke of the grid and the height of the light path from the bottom of the water tank. f took values of 1 Hz, 1.5 Hz, and 2 Hz. S took values of 3 cm, 5 cm, and 7 cm. H took values of 10 cm, 20 cm, and 30 cm.
3.4.1 Test results on changing the OGT system parameters
We set S to 7 cm and H to 10 cm, and measured the turbidity, optical power density, and BER at different f. Figures 8(a) and 8(b) show that the turbidity and received optical power density remain in a relatively stable state after 5 mins of grating oscillation when f are 1 Hz and 1.5 Hz. The corresponding BERs of the OFDM signals fluctuate in a small range from 0 to 1.37 × 10−5, as shown in Fig. 8(c). Therefore, the 110-160 mesh quartz had little effect on the BER performance of the OFDM signals under the circumstances. When f increased to 2 Hz, the concentration of resuspended quartz increased, resulting in an obvious increase in turbidity and an obvious decrease in received optical power density. After 5 mins of oscillation, BERs showed an obvious upward trend, which fluctuated in the range from 2.05 × 10−5 to 1.71 × 10−4 with the increase of time. It indicates that the greater the frequency of the grid, the greater the turbulence intensity, resulting in a higher content of resuspended quartz and the greater the impact on BER.
Fig. 8. (a) Turbidity, (b) received optical power density, and (c) BER performance of the OFDM signals measured at different time and different frequencies of the grid using 110-160 mesh quartz.
We set f to 2 Hz and H to 10 cm, and measured the turbidity, optical power density, and BER at different S. As shown in Figs. 9(a) and 9(b), turbidity and optical power density reach a relatively stable state after 2.5 mins of oscillation. When S is 3 cm and 5 cm, the BERs of the OFDM signals fluctuate in the range from 0 to 3.42 × 10−5, as shown in Fig. 9(c). When S was 7 cm, 110-160 mesh quartz had a greater impact on the BER performance of the OFDM signals, which was the same experiment as in Fig. 8 (c) where f was 2 Hz. It indicates that the greater the stroke of the grid, the greater the turbulence intensity, resulting in a higher content of resuspended quartz and the greater the impact on BER.
Fig. 9. (a) Turbidity, (b) received optical power, and (c) BER performance of the OFDM signals measured at different time and different strokes of the grid using 110-160 mesh quartz.
3.4.2 Test results on changing the frequency and the height of the optical path
We set S to 7 cm, and measured the turbidity, optical power density, and BER at different H and f using 110-160 mesh quartz, as shown in Figs. 10–12. When the grid frequency was 1 Hz, the turbidity measured at different H fluctuated in a small range from 0.9 NTU to 2.5 NTU. The received optical power density decreased with the increase of H. The BERs of the OFDM signals fluctuated in a small range from 6.85 × 10−6 to 2.05 × 10−5. When the grid frequency was 1.5 Hz, the turbidity measured at 30 cm was higher than that measured at 10 cm and 20 cm after 5 mins of oscillation with a turbidity of up to 4.5 NTU. The received optical power density still decreased with the increase of heights. The BERs of the OFDM signals fluctuated in a small range from 6.85 × 10−6 to 5.84 × 10−5. When the grid frequency was 2 Hz, the turbidities measured at 10cm and 30 cm were higher than those measured at 20 cm with a turbidity of up to 7.2 NTU. The received optical power densities measured at 20 cm were higher than those measured at 10 cm and 30 cm. The BERs of the OFDM signals at 20 cm were higher than those at 10 cm and 30 cm. Because the resuspended quartz was enriched near the water surface (at 30 cm), the turbidities and BERs at 30 cm were high, and the received optical power at 30 cm was low, when the grid frequency was 2 Hz.
Fig. 10. Turbidity measured at different time, H, and f of (a) 1 Hz, (b) 1.5 Hz, and (c) 2 Hz using 110-160 mesh quartz.
Fig. 11. Optical power density measured at different time, H, and f of (a) 1 Hz, (b) 1.5 Hz, and (c) 2 Hz using 110-160 mesh quartz.
Fig. 12. BER performance of the OFDM signals at different H and f of (a) 1 Hz, (b) 1.5 Hz, and (c) 2 Hz using 110-160 mesh quartz.
3.4.3 Test results on changing the turbulence intensity
We studied the influence of hydrodynamic changes on the UWOC performance. We set f to 1 Hz, 1.5 Hz, 2 Hz, 1.5 Hz, and 1 Hz every 5 mins, and then stopped oscillating and sunk for 30 mins. S was set at 7 cm. Figure 13 shows the turbidity, received optical power density, and BER performance of the OFDM signals at different H using 110-160 mesh quartz. We can see that the turbidities and BERs of the OFDM signals measured at 30 cm were higher than those measured at 10 cm and 20 cm and the received optical power densities measured at 30 cm were smaller than those measured at 10 cm and 20 cm. Although there were differences in the measured values at different optical path heights, the upward and downward trends were similar. In the first 15 mins, turbidities showed an obvious upward trend with the increase of turbulence intensity, and the received optical power densities have an obvious downward trend. At 15 to 25 mins, turbidities showed a significant downward trend with the decrease of turbulence intensity, and the received optical power densities showed a significant upward trend. It indicates that the suspension state of 110-160 mesh quartz can no longer be maintained due to the decrease of hydrodynamic force, and the quartz immediately began to sink. After 25 mins, the grid stopped oscillating, and the resuspended quartz gradually sunk. The turbidities and received optical power densities remained in a relatively stable state. When H was 10 cm and 20 cm, the BERs of the OFDM signals fluctuated in a small range from 0 to 2.74 × 10−5. When H was 30 cm. the BERs of the OFDM signals showed an obvious upward/downward trend with the increase/decrease of turbulence intensity.
Fig. 13. (a) Turbidity, (b) received optical power density, and (c) BER performance of the OFDM signals measured at different H using 110-160 mesh quartz, when f increases from 1 Hz to 1.5 Hz and 2 Hz, and then decreases from 2 Hz to 1.5 Hz and 1 Hz every 5 mins, and finally sinks for 30 mins without turbulence.
We continued to test the effects of finer-grained quartz (i.e., 160-200 mesh and 200-300 mesh) on the UWOC performance under varying turbulence intensities. We still set the grid frequency to 1 Hz, 1.5 Hz, 2 Hz, 1.5 Hz, and 1 Hz every 5 mins, and then stopped oscillating and settling for 30 mins. S was set at 7 cm and H was set at 10 cm. Figure 14 shows that the finer the grain size, the higher the turbidity and BER, and the smaller the received optical power density. For example, compared with the turbidities, BERs, and received optical power densities measured using 110-160 mesh and 160-200 mesh quartz, the turbidities and BERs measured using 200-300 mesh quartz were higher, and the received optical power densities were lower. When the frequency increased from 1 Hz to 1.5 Hz (i.e., after 7.5-mins oscillation), BERs directly increased to 1.49 × 10−1. Moreover, 200-300 mesh quartz was difficult to sink after 5 mins of resuspension. Upon the attenuation of hydrodynamic forces (after 15 mins), the settling time of 160-200 mesh quartz lagged behind that of the 110-160 mesh quartz by 5 mins. It indicates that once the finer sediments in the channel were activated and resuspended, their effect on the optical path was severe, manifested as significant deterioration of BER performance. From the 200-300 mesh quartz data in Figs. 14(a) and 14(b), we can see that when the channel becomes turbid enough that the receiver cannot detect the light spot, the received optical power density cannot characterize the channel. However, turbidity can still be used as an indicator of changes occurring within the channel.
Fig. 14. (a) Turbidity, (b) received optical power density, and (c) BER performance of the OFDM signals measured with different quartz, when f increases from 1 Hz to 1.5 Hz and 2 Hz, and then decreases from 2 Hz to 1.5 Hz and 1 Hz every 5 mins, and finally sinks for 30 mins without turbulence.
The previous experimental results show that the sediments with smaller grain size have a greater impact on the UWOC performance. Therefore, we selected four kinds of kaolin with different grain sizes to simulate the finer sediments (e.g., clay minerals) in the real-life seawater environments. In the experiment, S was set at 7 cm, f was set at 1 Hz, and H was set at 10 cm. As shown in Fig. 15, it basically conforms to the law that the finer the grain size of kaolin, the higher the BER.
Fig. 15. (a) Optical power density and (b) BER performance of the OFDM signals at different time using kaolin with different grain sizes.
3.5 Test results on Sediment-B
Further tests were conducted with Sediment-B from the East China Sea. Sediment-B was sampled on a beach. Active hydrodynamic factors, such as tides and waves, make it difficult for finer sediments to sink. Therefore, Sediment-B was mainly composed of coarser sediments. Active hydrodynamic environments also cause the sediments to be less reliably drained, compacted, and consolidated. On the contrary, the sedimentary environment of Sediment-A was relatively stable. Fine sediments will be sunk, drained, compacted, and consolidated, resulting in strong adhesion between sediments. The experiments of quartz and kaolin also did not have the process of drainage, compaction, and consolidation, which was similar to that of Sediment-B. Therefore, we tested Sediment-B using the same experimental parameters as in Section 3.4.3 to vary the turbulence intensity.
In the first 20 mins, turbidities showed an obvious upward trend with the increase of turbulence intensity, and the received optical power densities have an obvious downward trend. At 20 to 30 mins, turbidities showed a significant downward trend with the decrease of turbulence intensity, and the received optical power densities showed a significant upward trend. After 30 mins, the grid stopped oscillating, and the resuspended quartz gradually sunk. The turbidities and received optical power densities remained in a relatively stable state. Overall, the upward and downward trends with increased and decreased turbulence were delayed by 5 mins. This may be due to the complex composition of Sediment-B. Even though Sediment-B was mainly composed of coarse sediments, there were still some fine sediments that were not easy to sink. A small number of fine sediments were enriched near the water surface, and a large number of coarse sediments sunk at the bottom of the water tank. Therefore, turbidity and BER at 20 cm were successively higher than those at 30 cm and 10 cm during the process of hydrodynamic weakening, as shown in Figs. 16(a) and 16(c).
Fig. 16. (a) Turbidity, (b) received optical power measured, and (c) BER performance of the OFDM signals at different H using Sediment-B, when the frequency of the grid increases from 1 Hz to 1.5 Hz and 2 Hz, and then decreases from 2 Hz to 1.5 Hz and 1 Hz every 5 mins, and finally sinks for 30 mins without turbulence.
The grain size distribution of Sediment-B was distributed between 110-160 mesh quartz and 160-200 mesh quartz. Therefore, we compared the turbidity, optical power density, and BER performance of Sediment-B, 110-160 mesh quartz, and 160-200 mesh quartz. Except that H was set to 10 cm, other parameters were the same as the experimental parameters of Section 3.4.4 to vary the turbulence intensity. Figure 17 shows that the turbidity, optical power density, and BER performance of Sediment-B are between those of two kinds of quartz, which proves the feasibility of using quartz to simulate real-life sediments.
Fig. 17. (a) Turbidity, (b) received optical power measured, and (c) BER performance of the OFDM signals at different time using 110-160 mesh quartz sand, 160-200 mesh quartz sand, and Sediment-B, respectively, when the frequency of grid increases from 1 Hz to 1.5 Hz and 2 Hz, and then decreases from 2 Hz to 1.5 Hz and 1 Hz every 5 mins, and finally sinks for 30 mins without turbulence.
4. Conclusions
In this paper, we propose using different grain sizes of quartz and kaolin to simulate sediments in seawater and using an oscillating grid to generate turbulence of different intensities, and thus quantitatively study the effects of turbulence-induced sediment resuspension and sedimentation on UWOC. Firstly, we innovatively study the effects of different compositions and grain sizes of suspended sediments on UWOC. The experimental results show that different kinds of sediments with different grain sizes have similar effects on BER performance under the same optical power density, but the smaller the grain size, the higher the BER under the same turbidity. It indicates that it is valid to use turbidity to represent changes in the underwater channel and received optical power density to assess the attenuation of light after transmission through the underwater channel. Then, we innovatively study the effects of sediment resuspension and sedimentation on UWOC under continuous varying turbulence intensity. The experimental results show that turbulence has little effect on BER performance in the case of tap water, but the process of sediments resuspension and sedimentation caused by turbulence can seriously affect the BER performance. The larger the frequency of the grid, the greater the turbulence intensity and the worse the BER performance. In the future, a wavelength division multiplexing UWOC system with adaptive modulation, channel estimation, and channel equalization techniques can be studied to solve the time-varying channel issues and improve the data rate in turbid and dynamic underwater environment.
Funding
Fundamental Research Funds for the Central Universities (13502150069); Interdisciplinary Project in Ocean Research of Tongji University (1350121004/016).
Acknowledgments
The authors acknowledge Zhifei Liu, Fuwu Ji, Wei Liu, and Junbiao Tu for their help in improving the experimental program.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. Y. Guo, M. Kong, O. Alkhazragi, et al., “Current Trend in Optical Internet of Underwater Things,” IEEE Photonics J. 14(6), 1–8 (2022). [CrossRef]
2. Z. Xu, Z. Luo, X. Lin, et al., “15.26Gb/s Si-substrate GaN high-speed visible light photodetector with super-lattice structure,” Opt. Express 31(20), 33064–33076 (2023). [CrossRef]
3. Y. Zhu and R. Chang, “Preliminary study of the dynamic origin of the distribution pattern of bottom sediments on the continental shelves of the Bohai Sea, Yellow Sea and East China Sea,” Estuarine, Coastal Shelf Sci. 51(5), 663–680 (2000). [CrossRef]
4. L. Liu, T. Wang, R. Tian, et al., “Underwater Channel Characteristics and Wireless Optical Communications Using a Blue Mode- Locked Semiconductor Disk Laser,” J. Lightwave Technol. 41(16), 5340–5349 (2023). [CrossRef]
5. E. G. Mital, V. C. Michael, Olarte, et al., “Characterization of underwater optical data transmission parameters under varying conditions of turbidity and water movement,” in 2017 5th International Conference on Information and Communication Technology (ICoIC7) (IEEE, 2017), pp. 1–6
6. P. N. Ramavath, S. A. Udupi, and P. Krishnan, “Experimental demonstration and analysis of underwater wireless optical communication link: Design, BCH coded receiver diversity over the turbid and turbulent seawater channels,” Microw. Opt. Technol. Lett. 62(6), 2207–2216 (2020). [CrossRef]
7. Y. Song, W. Lu, B. Sun, et al., “Experimental demonstration of MIMO-OFDM underwater wireless optical communication,” Opt. Commun. 403, 205–210 (2017). [CrossRef]
8. A. S. Alatawi, “A Testbed for Investigating the Effect of Salinity and Turbidity in the Red Sea on White-LED-Based Underwater Wireless Communication,” Appl. Sci. 12(18), 9266 (2022). [CrossRef]
9. C. Shen, Y. Guo, H. M. Oubei, et al., “20-meter underwater wireless optical communication link with 1.5 Gbps data rate,” Opt. Express 24(22), 25502–25509 (2016). [CrossRef]
10. C. Fei, X. Hong, G. Zhang, et al., “Improving the Performance of Long Reach UOWC With Multiband DFT-Spread DMT,” IEEE Photonics Technol. Lett. 31(16), 1315–1318 (2019). [CrossRef]
11. W. Liu, Z. Ye, N. Huang, et al., “Multilevel polarization shift keying modulation for turbulence-robust underwater optical wireless communication,” Opt. Express 31(5), 8400–8413 (2023). [CrossRef]
12. Y. Guo, M. Kong, M. Sait, et al., “Compact scintillating-fiber/450-nm-laser transceiver for full-duplex underwater wireless optical communication system under turbulence,” Opt. Express 30(1), 53–69 (2022). [CrossRef]
13. E. Zedini, H. M. Oubei, A. Kammoun, et al., “Unified Statistical Channel Model for Turbulence-Induced Fading in Underwater Wireless Optical Communication Systems,” IEEE Trans. Commun. 67(4), 2893–2907 (2019). [CrossRef]
14. M. Khodadadi Karahroudi, S. A. Moosavi, A. Mobashery, et al., “Performance evaluation of perfect optical vortices transmission in an underwater optical communication system,” Appl. Opt. 57(30), 9148–9154 (2018). [CrossRef]
15. Z. Vali, A. Gholami, Z. Ghassemlooy, et al., “Experimental study of the turbulence effect on underwater optical wireless communications,” Appl. Opt. 57(28), 8314–8319 (2018). [CrossRef]
16. D.-C. Li, C.-C. Chen, S.-K. Liaw, et al., “Performance Evaluation of Underwater Wireless Optical Communication System by Varying the Environmental Parameters,” Photonics 8(3), 74 (2021). [CrossRef]
17. R. Huang, Q. Zhang, W. Zhang, et al., “Experimental research on the effect of suspended sediment stratification on turbulence characteristics,” Estuarine, Coastal Shelf Sci. 278, 108128 (2022). [CrossRef]
18. N.-S. Cheng and A. W.-K. Law, “Measurements of Turbulence Generated by Oscillating Grid,” J. Hydraul. Eng. 127(3), 201–208 (2001). [CrossRef]
