The experiment on the effect of the inlet pipe design on waste collection in the aquaculture tank was conducted at the National Engineering Research Center for Marine Aquaculture, Zhejiang Ocean University. The aquaculture tank is connected to the conical barrel through a wire pipe, and the water in the tank forms a circulation through the pump. As indicated in Figure 1, the blue arrow represents the direction of water flowing out of the aquaculture tank, and the red arrow represents the direction of water flowing into the aquaculture tank. The circular tank is made of transparent acrylic sheets, with a height of 50 cm, an internal diameter of 98 cm, a flat bottom, and a drainage opening right in the center with a diameter of 5 cm. The circular tank is placed on an operating table made of aluminum profiles. The inlet pipe is a rigid PVC pipe with an external diameter of 20 mm, which is set up on an aluminum operating table, and the distance of the inlet pipe is adjusted by changing the length of the horizontal PVC pipe above the aquaculture tank. A dial is placed at the connection of the inlet pipe and the bracket, a pointer is set on the inlet pipe in the same direction as the inlet, and the pointer is kept relatively stationary with the inlet pipe. When the inlet pipe turns, the pointer points out the corresponding angle on the dial, thus adjusting the angle of the inlet pipe. The flow meter measurement system consists of a valve and two flow meters (Keyence FD-Q20C) mounted on the inlet pipe. The valve regulates the volume of the water flow in the inlet pipe, and the flow meter monitors the flow rate (L/min) of the water input from the inlet pipe to the aquaculture tank in real time, thus achieving precise control of the inlet pipe flow. The inlet pipe has three uniform round holes of 5 mm at an interval of 3 cm from the bottom so that water enters the aquaculture tank evenly. The image acquisition system consists of a high-definition camera (VA-200PRO) placed directly above the tank and the corresponding control software, which mainly records the waste particle's movement process. Besides, a white film is laid on the bottom of the aquaculture tank to more clearly capture the waste distributed at the bottom of the tank.

The effect of the deployment and the angle of the inlet pipe driven by the jet on the waste collection characteristics in the circular aquaculture tank is mainly studied in this research. The experimental design water depth for measuring the waste collection characteristics was 20 cm, the diameter-to-depth ratio (diameter/water depth) was about 5:1, the inlet flow rate was 5.2 L/min (the flow velocity at each inlet was about 0.51 m/s in the dual inlet experiment and about 1.02 m/s in the single inlet experiment, and the hydraulic retention time is about 30 minutes. Based on the previous studies Hu et al. (2021) and the inlet pipe design in aquaculture practices, we totally designed 54 inlet pipe design cases in this experiment, including two inlet pipe designs of a single inlet pipe and dual inlet pipes (Figure 2) three groups of inlet pipe layout distance of d=1/50 r, 1/4 r, and 1/2 r, and nine groups of inlet pipe design angles: α=0°, 10°, 20°, 30°, 40°, 45°, 50°, 60°, 70°. The angles above 70° are not considered in this paper since the self-cleaning ability is obviously poor in such conditions and 45° is included for its wide use in aquaculture practice Gorle et al. (2018). In which before the experiment was started, the angle and distance of the inlet pipe were first adjusted to the preset condition, the pump and the flow meter were turned on, and the valve regulated the flow of the inlet pipe. Then, the flow of water through the inlet pipe into the aquaculture tank was monitored in real-time (L/min) by the flow meter, and the flow was regulated to the required value of the experiment. After waiting about 30 min, the flow field in the tank tended to be stable (the pre-test results showed that the flow field in the aquaculture tank was almost unchanged when the system was operated for 30 minutes). At this time, the high-definition camera was turned on, and 10 g shrimp feed with a diameter of 1.6 mm and a length of 1.0-2.0 mm were sprayed quickly and evenly into the tank, and the waste collection situation and pattern in the tank were observed from this time (t=0 min at this point). Based on the previous studies Summerfelt et al. (2016), Xue et al. (2021), the maximum monitoring time was set to 30 min. That is to say, we will turn off the recording camera and start the next cases even though not all the waste particles were completely discharged.

In the experiment of waste collection under the dual inlet pipe design, it was found that the waste residuals could not completely discharge within 30 min. To quantitatively compare the self-cleaning ability, MATLAB R2020b software was used to process the raw image of the waste residuals in the tank. (1) After the raw image (Figure 3A) of the waste in the aquaculture tank was input into MATLAB software, the rgb2gray function was used to convert it into a grayscale image. (2) The grayscale image of the waste in the aquaculture tank was binarized, and the binarized image of the aquaculture tank was extracted by deploying the grayscale threshold after processing. In this study, the gray index value threshold was set to 200, i.e., when the gray value of a pixel point is greater than 200, the color of the pixel point is set to black; otherwise, it is set to white, and the black pixel point represents the waste residual in the aquaculture tank. (Figure 3B) (3) The calculation range in the binarized image was set, and all the black pixels outside the bottom of the aquaculture tank were removed to obtain the binarized bottom waste image. (Figure 3C) (4) After binarization, the number of black pixels in the bottom waste image of the aquaculture tank was calculated, and the self-cleaning ability of the aquaculture tank was quantified by comparing the number of pixels in the binarized image of the bottom waste residuals under different experimental working conditions.

The flow field measurement experiment in the aquaculture tank was conducted in the marine measurement laboratory of QD GL Technology Co, Ltd. The flow field measurement setup (Figure 4) was similar to the waste collection experiments except for the flow field measurement system. In this test, the PIV technique was employed to measure the flow field distribution characteristics in the aquaculture tank. The circular aquaculture tank was embedded in a square acrylic tank to avoid refraction of the laser by the circular tank wall. The waste residuals have a small effect on the flow velocity in the aquaculture tank. The water clarity will be affected if the water is in the aquaculture tank for a long time, thus weakening the laser penetration and affecting the PIV experiment results. Considering this, no waste residuals were spread into the aquaculture tank in the flow field distribution characteristic measurement experiments.

The PIV system combines tracer particles, a laser transmitter, high-speed cameras, control systems, and data processing software. Measurements were conducted by scattering tracer particles in the aquaculture tank, and the velocity of the fluid in the flow field was represented with particle velocity. Meanwhile, a laser was applied to irradiate a test plane, and the position of particles exposed twice was recorded with an imaging method. An image analysis technique obtained the displacement of the particle swarm, and the flow velocity vector at each point was obtained by the time interval of displacement and exposure.

50-μm polystyrene tracer particles were used in this experiment (the white particles in Figure 4B), with a density of 1.3 kg/m3 (close to the density of water) and good following and light diffusion characteristics. The laser is an ideal light source for the PIV system, which can ensure good collimation and uniformity of the slice beam. In the experiment, a continuous laser plane was generated by the GL-532-1045 model from QD GL Technology Co, Ltd. Besides, the GL-JR-20P high-speed camera was used, with a maximum frame rate of 16 fps and an image resolution of 5,120×3,800.

The SVMC capture software of QD GL Technology Co, Ltd. was used for image acquisition of the flow field (Figure 5). The camera shooting frame rate was 16 fps, the image frequency was set to 8, and the number of images collected in the experiment was 96 (with every two images forming a group, there were 48 groups in total). The experimental acquisition duration was 6 s; dt is the acquisition time interval of two images in the same group, and the dt was set to 15000 μs. The high-speed camera automatically stopped shooting after 96 images were acquired, and the captured photos were stored automatically. The flow pattern tracking master V50 software performed preliminary flow field processing on the captured images and automatically saved the flow field image and flow velocity data.

The PIV images were analyzed by a direct cross-correlation method to obtain the flow field under different inlet pipe design cases. The removal efficiency of fish wastes is evaluated in a comprehensive analysis. For this purpose, the EQ method was deployed to determine the inlet setup considering the uniformity of water flow velocity (1) Masaló and Oca et al. (2016). To systematically compare the hydrodynamic characteristics in the aquaculture tank under different working conditions, the flow field was quantitatively analyzed based on the hydrodynamic characteristic quantities such as the average flow velocity (v_avg ) in the aquaculture tank, the flow uniformity coefficient (DU₅₀ ), and the resistance coefficient (C_t ) of the aquaculture tank. v_avg can be expressed mathematically in the following form Venegas, et al. (2014):

where,

The average flow velocity (v_avg ) is the average flow velocity of each monitoring point in the aquaculture tank. In formula (1), v is the average weighted velocity (m/s); v_i is the velocity of the monitoring point (m/s); r_i is the distance of the monitoring point from the center of the tank (m); DU₅₀ is the uniformity coefficient of the water flow in the aquaculture tank; v₅₀ is the average value of the first 50% of the velocity at each point of a depth section (m/s); DU₅₀ is larger than 0 and smaller than 100, and the closer to 100 represents better velocity uniformity. In formula (1), the aquaculture tank resistance coefficient (C_t ) refers to the quantified expression of the aquaculture tank on the energy obstruction of the water flow in the tank, and a larger aquaculture tank resistance coefficient represents more energy loss of the inlet pipe input to the aquaculture tank; Q is the inlet flow rate (m³/s); v₁ is the inlet velocity (m/s).

Figure 6 shows the waste residuals distribution driven by dual inlet pipe at t=30 min, and the inlet pipe design angles of 0°, 10°, 20°, 30°, 40°, 45°, 50°, 60°, and 70° and distance of 1/50 r, 1/4 r, and 1/2 r are represented in horizontal and vertical axis represents, respectively. Obviously, it can be observed that the angle and distance of the inlet pipe significantly affect the waste collecting process in the aquaculture tank.

When the distance of the inlet pipe is at d=1/50 r and the angle is at α=0°~70°, the waste collection ability of the aquaculture tank shows an overall trend of first strengthening and then weakening with the increase of the angle of the inlet pipe, and the highest ability was achieved when the angle of the inlet pipe was at α=50°; when the waste residual in the tank was distributed in a hollow oval shape, the overall waste collection ability was low. When the distance at d=1/4 r and the angle at α=0°~10°, some waste remained in the aquaculture tank, and the waste collection ability was low; when the angle of the inlet pipe was at α=20°~40°, the waste remained in the area near the outside of the aquaculture tank, and the outlet was reduced; meanwhile, most of the waste was gathered in the middle area of the aquaculture tank in a hollow elliptical distribution, and the waste collection ability was poor; when the angle of the inlet pipe was at α=50°~70°, a small amount of circular waste remained at the wall of the aquaculture tank, and the waste collection ability was low.

When the distance of the inlet pipe was at d=1/2 r, and the angle was at α=0°~20°, the waste in the aquaculture tank was basically discharged out of the tank, and the waste collection effect was the best; when the angle of the inlet pipe was at α=30°~70°, a small amount of circular waste remained at the wall of the aquaculture tank, and the waste collection effect was poor.

To further measure the effect of the distance and angle of the inlet pipe on the hydrodynamic characteristics of the waste collection of the aquaculture tank, the images in the aquaculture tank were quantitatively analyzed at t=30 min, and the results are illustrated in Figure 7. It can be observed from the figure that the overall amount of waste residual (pixel numbers) decreased with the increase in the distance of the inlet pipe, and the difference of waste residual between d=1/4 r and 1/2 r was significantly smaller than that at d=1/50 r. When the angle of the inlet pipe was at α=0°~20°, the amount of waste residual gradually decreased with the increase of d; when the angle of the inlet pipe was at α=30°~50°, the difference of waste residual between d=1/4 r and 1/2 r was small but significantly smaller than that of d=1/50 r. When the angle of the inlet pipe was at α=50°~70°, there was no obvious pattern change in the waste residual.

The above analysis indicates that the aquaculture tank achieves the best effect of collecting waste at distance of d=1/4 r, angle of α=20°~40° or at distance of d=1/2 r, angle of α=10° and its self-cleaning ability is the highest.

The single inlet pipe is to remove one inlet pipe in the dual inlet pipe experiment, and the rest of the device is the same as that of the dual inlet pipe experiment. The inlet pipe flow rate was set to 5.2 L/min in the single inlet pipe experiment to ensure the same hydraulic retention time. The single inlet pipe experiment design was identical to the dual inlet pipe experiment. However, during the experiment, it was found that due to the large flow rate of the single inlet pipe, the waste in the aquaculture tank could be completely discharged within the hydraulic retention time of 30 minutes except for three groups of special conditions. Thus, in the cases of single inlet, the time consumed for completing the discharge of waste residuals in the aquaculture tank was used as an index to quantify the waste collection ability of the aquaculture tank, and the shorter the waste collection time, the stronger the waste collection ability.

The waste collection time under different inlet pipe design angles is indicated in Figure 8. Note that the data of waste collection time exceeding 30 min will not be displayed in the figure. In the case of d=1/50 r, the curve of waste collection time featured a parabolic shape, showing a bottom at the angle of α=45°. It decreased rapidly at α=0°~30°, slowing its reduction among angle of α=30°~45° and finally increased from 45°~70°. For the case the distance of d=1/4 r, the waste collection time was almost the same for the angles of α=0°~45° and depicted an increasing trend from 45°~70°. The waste collection time increased gradually with the inlet pipe angle increment, achieving the best waste collecting performance for α=0° at the distance of d=1/2. The waste in the aquaculture tank cannot be completely discharged at the angle of α=60°~70°.

The flow field in the near bottom (2 cm from the bottom) in the cases of dual inlet pipe design was indicated in Figure 9 with the variations of jetting angle and jetting distance in the horizontal and vertical axes, respectively. Two long black arrows (not so obvious) in each subpanel show the jetting angle and distance. Meanwhile, the small arrows and colormap indicated the velocity direction and magnitude, respectively. Not all the small arrows were displayed in the figure to enhance visibility.

As depicted in Figure 9, two kinds of high-speed circulation flow were formed in the tank at all the cases, one is around the center of the tank caused by the narrow outlet and one is around the inlets due to the high jetting flow. However, the location and area of the high-speed circulation flow due to the jetting varied with the jetting angle. The flow evolution process with jetting angle is similar in general and still several little inconsistencies exist and the details are described in the following. For the case of d=1/50 r, with the increase of the jetting angle, the flow velocity close to the wall of the tank gradually decreases and the high-speed circulation flow moves away from the tank wall. In addition, the area of the high-speed flow increases from α=0° ~ 45° and decreases after α=45°, indicating a uniform good flow condition for aquaculture. The area increment arises from the less water flow collision between the jetting flow and the tank wall. On the contrary, the area decrement is caused by energy loss due to the discrepancy in the directions of the general flow in the tank and the jetting. The similar phenomenon (flow evolution) also can be found in the case of d=1/4 r but it differs in the velocity magnitude with obviously larger velocities than these in the case of d=1/50 r owing to the longer distance between the jetting and tank wall. Meanwhile, a gradual decrease in the area of high-speed flow with increasing jetting angle is observed in the case of d=1/2 r, revealing that the jetting distance is far enough to avoid the flow collision in this case. Figure 10 shows the flow field measured in the cases of single inlet pipe design. Although the inlet pipe design was adjusted to single pipe, the characteristics of flow field are basically the same with those in Figure 9. Notably, the maximum value of the colorbar in Figure 10 is twice that in Figure 9, thus the velocity magnitude in single inlet pipe design is generally larger. Assessing from the views of high-speed flow area and the flow uniformity, the optimum flow environment can be achieved at the angel of α=40° ~ 50°, 20° ~ 30° and 10° ~ 20° for d=1/50 r, 1/4 r and 1/2 r, respectively.