Study on Collection Performance of Hydraulic Polymetallic Nodule Collector Based on Solid–Liquid Two-Phase Flow Numerical Simulation

Abstract

The hydraulic collector is an important device for collecting seafloor polymetallic nodules. In this study, a hydraulic polymetallic nodule collector with two acquisition nozzles and one transmission nozzle is described. The numerical model of the hydraulic collector is established based on the solid–liquid two-phase flow method, and it is verified by experimental tests. On this basis, the collection mechanism of the hydraulic collector is analyzed, and the effects of structural parameters and working parameters on its collection performance are explored. The results show that the collection height and slant angle of the acquisition nozzle are key factors for collection efficiency, with optimal heights below 150 mm and angles between 45 and 49${}^{\circ }$. The recommended range for the center distance between the two acquisition confluence tubes is 650–730 mm. Excessive acquisition and transmission flow rates make a negligible contribution to improving the collection efficiency, but can also cause a significant increase in energy consumption. Therefore, the recommended ranges for acquisition and transmission flow rates are 140–160 m${}^3$/h and less than 80 m${}^3$/h, respectively. All of the results indicated that the parameters of the developed hydraulic collector were set reasonably, which thus ensured a balance between the collection efficiency and energy consumption.

1. Introduction

With the increasing demand for mineral resources, the development of deep-sea mineral resources is attracting the interest of increasing numbers of countries and researchers, and it has become one of the hot spots for new resource development [1]. There are abundant resources such as cobalt-rich crusts, polymetallic nodules, and polymetallic sulfides in the vast deep sea [2], which can be employed in the research and development of energy batteries, special materials, superconductors, and other advanced technologies [3]. Polymetallic nodules are a mineral resource in the submarine plain at a depth of 3500–6500 m, which contain several kinds of metals such as copper, cobalt, nickel, and manganese [4]. At the same time, they have more abundant reserves and wider distributions on the seafloor, as well as simpler collection operations than other minerals [5,6]. Therefore, polymetallic nodules are currently considered the most promising mineral resource, and they hold great mining value (especially for commercial mining).

In general, polymetallic nodules are located in the sediment at a depth of 0–200 mm on the seafloor surface, and they are oval in shape at a size of 20–100 mm [7], as shown in Figure 1. The deep-sea miner in Figure 2 is the most common and practical equipment for collecting polymetallic nodules at present [8]. The front collector collects the nodules from the seafloor to the middle storage bin while the miner is moving on the seafloor, and then the collected nodules are transported to the surface mining vessel by the lifting pipeline system connected to the miner [9], thus completing the collection of the polymetallic nodules.

As a critical component of the miner, the front collector directly relates to the miner’s working performance. The collection form of the front collector includes the mechanical type [10], the hydraulic type [11] and the compound mechanical-hydraulic type [5]. Currently, the hydraulic type collector has been widely applied due to its advantages of high efficiency, great environmental protection, and low cost [12].

Researchers have carried out a large number of studies on the collection mechanism of polymetallic nodule hydraulic collectors through experimental tests and numerical simulations. Hong et al. [13] manufactured a hydraulic collector with water-jet piping and baffle plates. They also analyzed the collection efficiency of the hydraulic collector through experimental tests, and an operation condition with high collection efficiency was identified. Yang and Tang [14] designed a hydraulic collector consisting of two rows of jet nozzles and a conducting nozzle. The influence of the main parameters of this collector on the mining performance was explored through experimental tests. The geometric considerations and related power consumption were provided in their research. Lim et al. [15] analyzed the flow field characteristics with outflow discharge from a collecting device in deep seawater while gathering manganese nodules through computational fluid dynamics simulation. Xiong et al. [16] adopted the CFD-DEM method to establish a solid–liquid two-phase flow model in which a single particle leaves the ground and enters the lifting pipe under the action of suction from the vertical pipe. Through this, the flow field and motion state of the single particle was explored. Yue et al. [17] studied and analyzed the operating performance of three different types of hydraulic collectors, and they found that the Coandă-effect-based hydraulic collector has the most promising application due to its combination of low energy consumption and small flow field disturbances. Zhao et al. [18] conducted a study on the shape of ellipsoidal polymetallic nodules on the collection suction force. By adopting the particle image velocimetry (PIV) technique and numerical method, Yue et al. [19] investigated the flow velocity distribution of two polymetallic-nodule-collecting devices. Jia et al. [20] conducted a theoretical analysis and numerical simulation study of a Coandă effect-based hydraulic collector, and they found that the Coandă effect became stronger with increasing initial velocity, increasing non-dimensional jet slot height, and decreasing non-dimensional wall height. Jia et al. [21] also investigated the new Coandă effect-based hydraulic polymetallic nodule collector with a logarithmic spiral surface. Alhaddad and Helmons [22] studied the erosion phenomenon of deep-sea sediments during nodule collection by a Coandă-effect-based hydraulic collector. On this basis, Alhaddad et al. [23] developed and designed a Coandă-effect-based hydraulic polymetallic nodule collector, as well as conducted CFD numerical analysis and experimental tests. The results showed that the designed collector can realize collection operations with high collection efficiency and low deep-sea sediment disturbance. With the resolved computational fluid dynamics discrete element method, Ren et al. [24] investigated the motion characteristics of the coarse particles during the hydraulic collecting process.

The essential role of a hydraulic collector is to aggregate the polymetallic nodules located on the seabed and to store them in a container on the mining vehicle for transportation to the sea surface through a pipeline. Therefore, the hydraulic collector needs to perform the two functions of peeling and lifting the nodules from the seabed, as well as transporting the nodules to the container in the mining vehicle. However, in the existing research, the majority of scholars have concentrated on the peeling of the nodules from the seabed, with less focus given to the transportation process of the nodules. During the transportation process, the nodules may not be delivered to the desired location as expected, thus resulting in a significant decline in the hydraulic collector’s performance. Therefore, it is more accurate and comprehensive to evaluate the performance of the hydraulic collector as a whole by considering the two functions of collection and transportation at the same time. In addition, the performance of a hydraulic collector is influenced by numerous factors, including structural and working parameters. Therefore, a thorough and encompassing analysis is needed. Particularly, it is critical to determine which parameters deserve priority in the design and development process. The development of a new type of hydraulic collector with good collection performance is also of great research value.

In this paper, a hydraulic polymetallic nodule collector with two acquisition nozzles and one transmission nozzle is studied. Firstly, the structural composition and operating principle of the hydraulic collector are analyzed and introduced. Subsequently, based on the solid–liquid two-phase flow method, a numerical model of the hydraulic collector is established, and the correctness of the established model is verified by experimental tests. Based on the established numerical model, the acquisition and transportation processes of this hydraulic collector are analyzed and their working mechanisms are explored. Finally, the influence of various structural parameters and working parameters on the collection efficiency of the hydraulic collector is investigated, and a parameter importance analysis is conducted. The work conducted in this study can provide a theoretical basis and design reference for the development of the hydraulic collector.

2. Hydraulic Collector

2.1. Structural Composition

The structural components of the developed hydraulic collector are shown in Figure 3. The baffle plates are welded together to form a closed conveying channel. The pumps are connected to the confluence tubes via the flange connections. The bearings are welded to the lateral baffle plates and carry the transverse confluence tube. The middle confluence tube is fixed to the conveying channel by a bolt connection. The acquisition nozzles and transmission nozzles are connected with the confluence tubes at the bottom of the conveying channel. The acquisition nozzle comprises many single-round orifices, and the transmission nozzle has only a flat rectangular mouth. The splashboards are connected with the lateral baffle plate and are used to avoid nodules being scoured outside by the high-speed seawater jets. The terrain followers can stably support the hydraulic collector and adjust the distance between the seafloor and the acquisition nozzle to ensure an optimal collection height.

2.2. Operating Principle

The operating principle of the hydraulic collector in this study is described in Figure 4. The hydraulic collector performs the extraction and transport of the polymetallic nodules from the deep seafloor while moving forward. The pumps continuously supply seawater into the confluence tubes. Due to the minute calibers of the nozzles, the seawater in the confluence tubes sprays out as high-speed jets. The oblique jets ejected from the acquisition nozzles strongly scour the seafloor, thereby causing the polymetallic nodules to loosen and peel from the seafloor. At the same time, these acquisition jets are reflected by the seafloor and converge together, which further lift the polymetallic nodules into the bottom of the conveying channel. Finally, the polymetallic nodules in the conveying channel are transmitted to the top exit of the collector by the upward jets that are ejected from the transmission nozzles.

When the nodule is in the conveying channel, it will be mainly subjected to a combination of the buoyancy ($F_b$), gravity (G), and the drag force under the action of flowing seawater ($F_d$), as shown in Figure 5. The three forces are expressed as follows:

$$F_b={\rho }_s{V}_{n}g,$$

(1)

$$G={\rho }_n{V}_{n}g,$$

(2)

$$F_d=\frac{1}{2}{\rho }_s{C}_D{A}_n{v}_s^2.$$

(3)

where ${\rho }_s$ and ${\rho }_n$ are the densities of the seawater and nodules, respectively; g is the gravitational acceleration; $C_D$ is the drag coefficient; $v_s$ is the seawater flow velocity in the conveying channel; and $V_n$ and $A_n$ are the volume of the nodule and the projected area of the nodule in the direction of seawater flow, respectively. Assuming that the nodule is a standard sphere with a diameter of $d_n$, $V_n$ and $A_n$ can be expressed as follows:

$$V_n=\frac{1}{6}\pi d_n^3,$$

(4)

$$A_n=\frac{1}{4}\pi d_n^2.$$

(5)

To ensure that the nodules do not settle and fall under the pull of gravity, the sum of $F_b$ and the component of $F_d$ in the direction of gravity should be greater than G.

$$F_b+F_{d}sin\alpha \ge G.$$

(6)

Substituting Equations (1)–(5) into Equation (6), as well as transforming it, yields

$$v_s\ge \sqrt{\frac{4\left({\rho }_n-{\rho }_s\right)d_{n}g}{3{\rho }_s{C}_{D}sin\alpha }}.$$

(7)

Theoretically, the absence of the gravitational settling of nodules is ensured when the seawater flow velocity in the collector conveying channel, $v_s$, satisfies Equation (7). However, due to the narrowness of the conveying channel, the seawater flow rate in the channel should be increased from the minimum value shown in Equation (7) in order to avoid clogging due to the low transport velocity of the nodules in the channel. In this study, a factor of two was taken. As such, $v_s$ was found to be

$$v_s=\sqrt{\frac{16\left({\rho }_n-{\rho }_s\right)d_{n}g}{3{\rho }_s{C}_{D}sin\alpha }}.$$

(8)

3. Numerical Method and Validation

3.1. Governing Equation

For the calculation, which was based on a solid–liquid two-phase flow, the governing equations of the liquid phase and the solid phase were, respectively taken into account. For the fluid phase, which was treated as incompressible, the Reynolds-Averaged Navier–Stokes (RANS) was employed in the simulation. The mass equation and momentum Equation were as follows [25]:

$$\frac{\partial \rho }{\partial t}+\frac{\partial }{\partial x_i}\left(\rho \overline{u_i}\right)=0,$$

(9)

$$\frac{\partial }{\partial t}\left(\rho \overline{u_i}\right)+\frac{\partial }{\partial x_i}\left(\rho \overline{u_i{u}_j}\right)=\frac{\partial \rho }{\partial x_i}+\frac{\partial \rho }{\partial x_j}\left(\mu \frac{\partial \overline{u_i}}{\partial x_j}-\rho \overline{{u^{\prime }}_i{u^{\prime }}_j}\right)+F_i.$$

(10)

where $\rho$ and $\mu$ represent the density and dynamic viscosity of the fluid, respectively; $\overline{u_i}$ and $\overline{{u^{\prime }}_i}$ are the mean components of velocity and pulsation component of the velocity in the i-directions, respectively; p is the total pressure; $F_i$ denotes the external body force components in the i-direction; and $-\rho \overline{{u^{\prime }}_i}\overline{{u^{\prime }}_j}$ denotes the Reynolds stress.

To simulate the turbulent flow, the realizable $k-ϵ$ model was employed thanks to the stability and precision it demonstrates when dealing with pressure gradients [26]. The transport Equations were expressed as follows [25]:

$$\frac{\partial \left(\rho k\right)}{\partial t}+\frac{\partial \left(\rho k{u}_i\right)}{\partial x_i}=\frac{\partial }{\partial x_i}\left[\left(\mu +\frac{{\mu }_t}{{\sigma }_k}\right)\frac{\partial k}{\partial x_j}\right]+G_k+G_b-\rho \epsilon -Y_M,$$

(11)

$$\frac{\partial \left(\rho \epsilon \right)}{\partial t}+\frac{\partial \left(\rho \epsilon u_i\right)}{\partial x_i}=\frac{\partial }{\partial x_i}\left[\left(\mu +\frac{{\mu }_t}{{\sigma }_{\epsilon }}\right)\frac{\partial \epsilon }{\partial x_j}\right]+\rho C_{1}S\epsilon -\rho C_2\frac{{\epsilon }^2}{k+\sqrt{v\epsilon }}+C_{1\epsilon }\frac{\epsilon }{k}{C}_{3\epsilon }{G}_b.$$

(12)

where k is the turbulent kinetic energy; $ϵ$ is the dissipation rate; $G_k$ and $G_b$ represent the turbulent kinetic energy generated from the average speed gradients and buoyancy effect, respectively; S is the average strain rate; $Y_M$ is the overall dissipation rate due to the fluctuating dilatation; $C_1=\max \left[0.43,\eta /\left(\eta +5\right)\right]$ and $\eta =S·k/\epsilon$; $C_{3\epsilon }$ is a function representing the effect of buoyancy on the dissipation rate; $C_2$ and $C_{1\epsilon }$ are constants; ${\sigma }_k$ and ${\sigma }_{\epsilon }$ denote the turbulent Prandtl number for k and $ϵ$, respectively; and, generally, $C_{1\epsilon }$ = 1.44, $C_2$ = 1.9, ${\sigma }_k$ = 1.0 and ${\sigma }_{\epsilon }$ = 1.2 were the default values in the software.

For the governing equation of the discrete phase particle, the motion trajectory was predicted by integrating its force balance in a Lagrangian reference system. The force balance equation can be written as [27]:

$$m_p\frac{d{\overrightarrow{u}}_p}{dt}=m_p\frac{\overrightarrow{u}-{\overrightarrow{u}}_p}{{\tau }_r}+m_p\frac{\overrightarrow{g}\left({\rho }_p-\rho \right)}{{\rho }_p}+\overrightarrow{F}.$$

(13)

where $m_p$ is the mass of the discrete solid particle, $\overrightarrow{u}$ is the velocity of the fluid phase, ${\overrightarrow{u}}_p$ is the velocity of the discrete solid particle, ${\rho }_p$ and $\rho$ are the densities of discrete particles and continuous fluid, respectively, $m_p\frac{\overrightarrow{u}-{\overrightarrow{u}}_p}{{\tau }_r}$ is the drag force, and $\overrightarrow{F}$ is the additional force. $\overrightarrow{F}$ includes the virtual mass force, thermophoretic force, Brownian force, Saffman’s lift force, etc.

3.2. Modeling

3.2.1. Model Simplification

The hydraulic collector comprises many complex parts and components. To ensure calculation efficiency and convergence in simulation, the original model needs to be simplified. Since the nodules are extracted and transmitted in the interior of the hydraulic collector, certain outside components, such as the pumps, flanges, and terrain followers, are removed. The splashboards and the lateral baffle plate are incorporated together. For the confluence tubes, only the parts connected to the nozzles are retained. The simplified model is presented in the form of the inside geometric profile to satisfy the collection mechanism of the nodules, as shown in Figure 6. The specific structural parameters of the hydraulic collector are displayed in Figure 7. The definition and values of each parameter are listed in

Table 1. Definition and value of structure parameters.

Parameter	Definition	Value
$B_1$	Width of the entrance of the conveying channel	600 mm
$B_2$	Width of the exit of the conveying channel	400 mm
d	Inner diameter of the acquisition nozzle	12 mm
D	Diameter of the confluence tube	150 mm
H	Collection height	120 mm
$H_s$	Width of the conveying channel	115 mm
$H_d$	Height of the collector	1600 mm
$\alpha$	Slant angle of the conveying channel	45
$\beta$	Slant angle of the acquisition nozzle	45${}^{\circ }$
L	Center distance between the front and rear acquisition confluence tubes	690 mm
$L_1$	Horizontal center distance between the transmission and rear acquisition confluence tubes	390 mm

.

3.2.2. Model Setting

Considering the symmetry of the hydraulic collector, the half structure is intercepted to apply in the computational domain for less calculation time, as shown in Figure 8. In Figure 8, the circular end faces of the confluence tubes are defined as the velocity inlets. The inlets of No. 1 and No. 2 provide seawater to the acquisition nozzles at a velocity of 1.26 m/s (equivalent to 80 m${}^3$/h), and the inlet of No. 2 provides seawater to the transmission nozzles at a velocity of 0.55 m/s (equivalent to 35 m${}^3$/h). Note that the one-half model was employed in this study and that all the aforementioned volume flow rates were halved. Therefore, the simulated acquisition and transmission flow rates were 160 m${}^3$/h and 70 m${}^3$/h, respectively. The top exit of the conveying channel and the vertical surfaces under the acquisition nozzles, namely the surfaces numbered No. 3 to No. 6, were defined as the pressure outlets with a relative pressure of 0. The polymetallic nodules are distributed in large quantities on the seafloor at a depth of 3.5–6.5 km [4], with a hydrostatic pressure of 35–65 Mpa. The ambient pressure in this study was set as 60 MPa. The cross-section that is numbered No. 7 is the symmetrical plane of this computational domain. Other surfaces were set up as walls. The hydraulic collector was fixed, and the nodules flow was relative to it. The particles were continuously injected with a speed of 0.5 m/s on the ground and their abundance was set to 15 kg/m${}^2$. In addition, the nodules were modeled as spherical particles with the same diameter, and they were evenly arranged on the smooth ground. According to the properties of the polymetallic nodule, the density and diameter of the particle were set to 2000 kg/m${}^3$ and 0.05 m, respectively. In addition, the dynamic viscosity and density of seawater are 0.00161 kg/m·s and 1025 kg/m${}^3$, respectively, [28]. The collection duration was set to 30 s.

The whole computational domain was divided into unstructured grids. The regions near the wall of the hydraulic collector were refined by the prism layer grid to accurately simulate the flow state. The y+ value of 50 was chosen, and the corresponding thickness of the first prism layer grid was 1.5 mm. The layer number and growth rate of the prism grid were 5 and 1.2, respectively. The realizable $k-ϵ$ model was selected to simulate turbulence. Considering the low concentration of polymetallic nodules in the conveying channel, the interaction between nodules can be ignored, and particles in the collector were simulated by the DPM (discrete particle model ). The SIMPLE algorithm with the Eulerian model for bidirectional coupling calculations was used to realize the interaction between the particles and fluid.

3.2.3. Mesh Independence Study

The collection efficiencies of the hydraulic collector under four different mesh densities of coarse (0.51 $\times {10}^6$ elements), middle (0.98 $\times {10}^6$ elements), fine (1.77 $\times {10}^6$ elements), and finer (2.52 $\times {10}^6$ elements) were compared to ensure the mesh convergence. The collection efficiency of the hydraulic collector can be expressed as follows [17]:

$$\eta =\frac{M_1}{M_2}.$$

(14)

where $\eta$ denotes the collection efficiency, and $M_1$ and $M_2$ denote the mass of nodules lifted to the exit of the transport channel and the total nodule mass, respectively.

The collection efficiencies under the four mesh densities are listed in

Table 2. Collection efficiencies under different mesh densities.

Type	Number of Elements	Collection Efficiency	Error against Very Fine
Coarse	0.51 $\times {10}^6$	95.8%	−2.8%
Medium	0.98 $\times {10}^6$	97.1%	−1.5%
Fine	1.77 $\times {10}^6$	98.2%	−0.4%
Very Fine	2.52 $\times {10}^6$	98.6%	-

. As the number of elements increased, the simulation errors relative to the Very Fine case decreased gradually. The collection efficiencies for the Fine case and the Very Fine case were almost the same, i.e., less than 1%. Therefore, to balance the calculation accuracy and efficiency, the computational domain of the Fine case with 1.77 $\times {10}^6$ elements was chosen to conduct the subsequent study, which is shown in Figure 9.

3.3. Validation

3.3.1. Experimental Setting

In order to verify the accuracy of the simulation model, an experimental platform was established to perform the collection of the simulated polymetallic nodules by the hydraulic collector. As shown in Figure 10a, the experimental platform consists of the hydraulic system, the moving platform, and the hydraulic collector. According to the operating principle, when the moving platform carries the hydraulic collector forward, the hydraulic system inputs water into the confluence tubes to form jets for nodule acquisition and transmission. The simulated polymetallic nodules used in the experiments are shown in Figure 10b, and the density, diameter, and abundance are consistent with those in Section 3.2.2. The structural parameters of the hydraulic collector are consistent with those in Section 3.2.1 and the working parameters are consistent with those in Section 3.2.2.

3.3.2. Experimental validation

The experimental test scene and the simulated nodules extracted by this hydraulic polymetallic nodule collector are shown in Figure 11. The experimental and numerically simulated collection efficiencies of the collector at different collection heights are shown in

Table 3. Comparison of collection efficiency.

Collection Height H (mm)	Simulated Collection Efficiency	Experimental Collection Efficiency	Error
70	99.3%	100%	−0.7%
90	99.0%	96.0%	3.0%
110	97.3%	99.2%	−1.9%
150	90.7%	89.4%	1.3%
180	84.0%	88.3%	−4.3%
200	68.3%	72.1%	−3.8%

. As the collection height increased, the collection efficiencies in both the simulation and experiment decreased. The error between the two was small at all collection heights, with a maximum error of no more than 5%. Therefore, it could be verified that the established simulation model of the hydraulic collector, which was based on the solid–liquid two-phase flow method, has high accuracy.

4. Results and Discussions

4.1. Distribution of Flow Fields and Polymetallic Nodules

Figure 12 gives the flow field distributions of the hydraulic collector. As can be seen in Figure 12a,b, the front and rear acquisition nozzles produced two high-speed seawater jets at the bottom of the collector. The two sweater jets scoured the seafloor and the reflected seawater jets converged together. Meanwhile, the transmission nozzle ejected a high-speed seawater jet; thus, the seawater flow in the conveying channel was accelerated. The above flow velocity distribution was consistent with the operating principle in Section 2.2. As a result of the scouring of the seawater jets, there were significantly three high pressures on the seafloor and upper baffle, as shown in Figure 12c. The reflected seawater jet was blocked by the upper baffle and diverted to the two sides. In combination with the high-speed water jets from the acquisition nozzle, two vortices were formed, as shown in Figure 12b. Moreover, two low pressures in Figure 12c were generated in the vortex region. In addition, the static pressure gradient between the entrance of the conveying channel and the seafloor could further drive the polymetallic nodules to be lifted into the conveying channel.

By taking the drag coefficient $C_D$ as 0.44 [29,30] and substituting the relevant parameters in Section 3, the seawater flow velocity ($v_s$), which ensures that the nodules do not settle gravitationally in the conveying channel, was calculated from Equation (8) to be 2.85 m/s. The average velocity of the water flow at the outlet of the conveying channel in the numerical simulation was read to be 3.04 m/s, which is evidently greater than 2.85 m/s. This ensures that the nodules can be lifted and transported smoothly in the channel, as shown by Figure 13a. Thus, finally, a collection efficiency of 96.4% was achieved.

The 30 s distribution of the polymetallic nodules in the hydraulic collector is presented in Figure 13b. It can be seen that most of the polymetallic nodules were collected successfully. During the first 6 s, a small number of particles were collected due to the unstable flow field in the hydrodynamic collector. After the flow field was stabilized, a large number of polymetallic nodules were collected in an orderly manner during the period from the 12th to the 24th second. By the 30th second, the collection of polymetallic nodules in the collector was completed ahead of schedule.

In order to further investigate the collection performance of the hydraulic collector, the effects on the structural and working parameters are investigated in the following subsections based on the developed numerical model.

4.2. Influence of Structural Parameters

4.2.1. Collection Height

As shown in Figure 14, the velocity contours inside the hydraulic collector at different collection heights (H) were achieved by keeping other parameters constant. Furthermore, the extent to which the collection efficiency ($\eta$) and the average outlet flow velocity of the conveying channel ($v_{ao}$) varied with H is shown in Figure 15.

As can be seen from Figure 14, as the collection height (H) increased, the front and rear acquisition jets continued to attenuate, such that the upward flow formed by reflection after scouring the ground was also weakened. As a result, some of the nodules flew out of the back side of the collector before they were washed up to a sufficient height, and they then entered the conveying channel, as shown in Figure 16.

In addition, when H was less than 120 mm, the space in the lower part of the collector gradually increased as the H increased, the energy lost from the mixing of the acquisition jets decreased, and the average outlet flow velocity of the conveying channel ($v_{ao}$) increased. When H was greater than 120 mm, the H was greater than the non-attenuation length of the acquisition jet, and the water flow attenuation became serious. At the same time, more of the seawater output from the acquisition nozzles flowed out of the back side of the collector rather than into the conveying channel. Then, as can be seen from the red curve in Figure 15, when H was greater than 120 mm, $v_{ao}$ decreased rapidly with the increase in H. When H was 180 mm, $v_{ao}$ was less than 2.85 m/s, which is the required velocity for nodules to be exempted from gravitational settling according to Equation (8). Moreover, Figure 16 and Figure 17 clearly show that the nodule moved much slower than those represented in Figure 13a. In addition, as can be seen by comparing Figure 16 and Figure 17, the three nodules near the entrance of the conveying channel even moved downward.

Under the combined effect of the above influences on the acquisition and transmission capacity that were caused by the collection height (H), the collection efficiency ($\eta$) decreased continuously with increasing H, as shown by the black curve in Figure 15. In particular, $\eta$ decreased sharply when H was greater than 180 mm. In light of the above analysis, H should be no more than 150 mm to ensure good collection efficiency of the hydraulic collector. This can be achieved by the terrain follower in Figure 3, which can dynamically and adaptively adjust according to the desired settings and terrain conditions so as to keep H within a reasonable range.

4.2.2. Center Distance between the Front and Rear Acquisition Confluence Tubes

With the other parameters being set constant, the velocity contours inside the hydraulic collector at different center distances between the front and rear acquisition confluence tubes (L) are shown in Figure 18. Moreover, how the collection efficiency ($\eta$) and average outlet flow velocity of the conveying channel ($v_{ao}$) vary with L is shown in Figure 19.

Figure 18 together with Figure 12a indicate that, as the center distance between the front and rear acquisition confluence tubes (L) increased from 610 mm to 690 mm, the reflected upflow formed by the two acquisition jets scouring the ground increased, and the lifting effect on the nodules was enhanced. When L was greater than 690 mm, the upflow was slightly weakened with the increase in L. From the red curve in Figure 19, it can be seen that, with the increase of L, the average outlet flow velocity of the conveying channel ($v_{ao}$) keeps decreasing, which weakens the nodule transport capacity in the channel. However, at the L of 730 mm, $v_{ao}$ was still about 2.9 m/s, which was larger than the required 2.85 m/s.

The above phenomena eventually lead to the collection efficiency ($\eta$) of the hydraulic collector first increasing with L, but slightly decreasing after L was greater than 690 mm, as shown by the black curve in Figure 19. Therein, $\eta$ was found to be greater than 90% when the L was between 650 mm and 730 mm, which is the recommended range for designing the hydraulic collector.

4.2.3. Slant Angle of the Acquisition Nozzle

With the other parameters are constant, the velocity contours inside the hydraulic collector at different slant angles of the acquisition nozzle ($\beta$) are shown in Figure 20. When $\beta$ varied between 41${}^{\circ }$ and 53${}^{\circ }$, the average outlet flow velocity of the conveying channel ($v_{ao}$) varies less and was above 3 m/s. How the collection efficiency ($\eta$) varied with $\beta$ is plotted in Figure 21.

As can be seen from Figure 20, when $\beta$ was less than 45${}^{\circ }$, the two acquisition jets, in addition to forming an upward flow after scouring the ground, formed a backward flow at the bottom of the hydraulic collector. As a result, some of the nodules were washed away by this flow before they could be lifted to a sufficient height, which is similar to that shown in Figure 16. This also makes the collection efficiency ($\eta$) extremely low when the $\beta$ was less than 45${}^{\circ }$, as shown in Figure 21. Figure 20 also shows that, when $\beta$ was greater than 51${}^{\circ }$, the reflected upflow begins to move forward, i.e., it is no longer directly toward the conveying channel. This prevents some of the nodules from successfully entering the conveying channel, which, in turn, results in a decrease in $\eta$, as shown in Figure 21. According to Figure 21, the hydraulic collector has a high collection efficiency when $\beta$ is between 45${}^{\circ }$ and 49${}^{\circ }$.

4.3. Influence of Working Parameters

4.3.1. Acquisition Flow Rate

The velocity contours at different acquisition flow rates ($Q_a$) are shown in Figure 22. Furthermore, how the collection efficiency ($\eta$) and the power consumed by the collector (P) varied with L are plotted in Figure 23.

Figure 22 clearly shows that, as the acquisition flow rate ($Q_a$) increased, the strength of the two acquisition jets increased. This also applied to the strength of the updraft formed after scouring the ground, thus resulting in an increase in the ability to lift the nodules. Most of the $Q_a$ flowed into the conveying channel. Therefore, it can also be seen in Figure 22 that, as $Q_a$ increases, the flow velocity in the conveying channel also increases, thus resulting in a smoother transport of the nodules in the channel. These above effects ultimately lead to an increase in collection efficiency ($\eta$) with increasing $Q_a$, as shown by the black curve in Figure 23. Note that when $Q_a$ is greater than or equal to 160 m${}^3$/h, $\eta$ is already above 95% and the increase in $\eta$ is already minimal. In contrast, the power consumed by the collector (P) increases dramatically with the increase in $Q_a$, which is shown by the red curve in Figure 23. In light of the above discussion, it is more appropriate to set $Q_a$ at 140–160 m${}^3$/h in order to achieve a balance between the collection efficiency and power consumption.

4.3.2. Transmission Flow Rate

Figure 24 shows the velocity contours at different transmission flow rates ($Q_t$). It clearly shows that the variation of $Q_t$ has no effect on the flow velocity distribution in the lower part of the collector; therefore, it has no effect on the ability to peel and lift the nodules from the ground. Figure 24 also shows that, as $Q_t$ increases, the flow velocity in the conveying channel increases, thus allowing the nodules to be transported more smoothly and quickly. Most of the acquisition flow $Q_a$, which is 160 m${}^3$/h, also entered the conveying channel and mixed with the transmission flow to co-transport the nodules. Thus, even if $Q_t$ is only 40 m${}^3$/h, the total flow rate in the channel can still guarantee the smooth transport of the nodules. As a result, the collection efficiency ($\eta$) does not vary much at different $Q_t$, and it was all above 95%, as shown by the black curve in Figure 25. The increase in $Q_t$ will also inevitably make the power consumption (P) larger, as shown by the red curve in Figure 25. In addition, as $Q_t$ increases, the percentage of $Q_t$ compared to ($Q_a$ + $Q_t$) keeps increasing, thus making the power curve shown in Figure 25 increase faster with increasing $Q_t$. In particular, it was found that P is 21.21 kW at a $Q_t$ of 80 m${}^3$/h, which is only a 9% increase from the 19.55 kW at a $Q_t$ of 30 m${}^3$/h. Therefore, it is recommended that the transmission flow rate $Q_t$ of the hydraulic collector should not exceed 80 m${}^3$/h so as to avoid the unnecessary waste of energy.

4.4. Parameter Importance Analysis

The effects of various structural and working parameters on the collection efficiency of hydrodynamic collectors are analyzed in Section 4.2 and Section 4.3. The parameters that have a greater impact on the performance need to be prioritized in the design. In this study, the Variable Importance in the Projection (VIP) index [31] is utilized to quantitatively evaluate the extent to which each structural and working parameter affected the collection efficiency. After normalizing the numerical simulation results in Section 4.2 and Section 4.3, the calculated VIP index was determined, as shown in Figure 26. Figure 26 clearly shows that the collection height H had the greatest effect on the collection efficiency, and it was followed by the slant angle of the acquisition nozzle $\beta$. The remaining parameters, such as the center distance between the acquisition confluence tubes L, acquisition flow rate $Q_a$ and transmission flow rate $Q_t$, had relatively less influence. Therefore, when designing and developing a hydraulic collector, the focus needs to be on the H and $\beta$. In the hydraulic collector in this study, the terrain follower shown in Figure 3 was adopted for dynamic adaptive adjustment so as to ensure that the collection height could be kept near the optimal value during the walking collection process.

5. Conclusions

1. A hydraulic polymetallic nodule collector with two acquisition nozzles and one transmission nozzle is described in this study. The numerical model of the hydraulic collector was established based on the solid–liquid two-phase flow method. An experimental test platform was also constructed to verify the correctness of the model. The experimental and simulated collection efficiencies were found to be in good agreement at different collection heights, with a maximum error of no more than 5%.

2. The parameter importance analysis indicated that the collection height and slant angle of the acquisition nozzle were the two primary factors affecting the collection efficiency, and thus they should be prioritized in the design and development of hydraulic collectors. Collection efficiency decreases as the collection height increases; as such, the recommended setting is below 150 mm. Collection efficiency increases and then decreases slightly with the slant angle of the acquisition nozzle; thus, the recommended setting is 45${}^{\circ }$–49${}^{\circ }$. Collection efficiency also increases and then decreases slightly with the center distance between the two acquisition confluence tubes; as such, the recommended setting is 650–730 mm.

3. The collection efficiency increases with the acquisition flow rate, but the increase is particularly limited at >160 m${}^3$/h. At the same time, the power consumption of the collector keeps increasing with the acquisition flow rate. Therefore, it is recommended to set the acquisition flow rate between 140 and 160 m${}^3$/h to minimize the power consumption at higher collection efficiencies. A large amount of the acquisition flow was reflected into the conveying channel to ensure the smooth transportation of the polymetallic nodules. Thus, the transmission flow rate had little effect on the collection efficiency. However, the power consumption also kept increasing with the transmission flow rate. Therefore, it is recommended that the transmission flow rate be set as less than 80 m${}^3$/h to save energy.

To further optimize the collection performance and energy utilization of the hydraulic collector, further research should be conducted in the following aspects: exploring the influence of the diameter, shape, and arrangement of the acquisition nozzle on the acquisition and transmission flow field, as well as looking to the internal channel structure of the collector. Furthermore, the lightweight design of the collector, while ensuring strength and stiffness, can also reduce the total energy consumption of the entire mining vehicle and improve driving performance.