Research on Erosion Effect of Various Submerged Cavitating Jet Nozzles and Design of Self-Rotating Cleaning Device

Abstract

Submerged cavitating jets can effectively remove marine organisms from ship hulls without damaging the surface paint. To enhance the cleaning efficiency of cavitating jets, the selection of an appropriate nozzle structure and the design of an efficient cleaning device are crucial. In this study, the submerged cavitation effect of different nozzles was analyzed by numerical simulation. The actual cleaning efficacy of the nozzles was confirmed through erosion experiments as well. The simulation and experiment showed that the shear nozzle, absent of a pre-shrinking section and featuring a spherical outlet connected to a diffusion cylindrical section, maintained stable erosion performance at a standoff distance of 30–50 mm. This erosion was primarily attributed to denudation caused by bubble collapse. Based on this shear nozzle, a self-rotating cleaning device was designed and manufactured. A test rig was also established to test the cleaning effect and some parameters of the cleaning device.

1. Introduction

Marine biofouling, which is caused by the settlement and accumulation of marine organisms on hull surface during sailing, can significantly impact the navigation efficiency and energy efficiency of ships [1,2]. The additional fuel consumption caused by biofouling on global vessels can result in an estimated annual cost of USD 22–204 billion [3]. Moreover, biofouling can result in species invasion [4,5,6] and increased maintenance costs associated with artificial structures [7,8]. Consequently, it is imperative to investigate cleaning technologies capable of eliminating biofouling.

Cavitating jet technology emerges as a viable solution for in situ removal of biofouling without damaging the hull surface paint [9,10], rendering it a preferable choice for practical applications. Cavitation is a phenomenon occurring when the local pressure falls below the saturated vapor pressure, vapor bubbles arise in water. The bubbles break at a solid surface in a submerged environment, generate high-pressure micro-shock jets, causing damage of the solid surface [11]. Thus, cavitation jets can be utilized for biofouling cleaning.

Generally, the cavitation jet is produced in water, especially for biofouling cleaning, so the submerged cavitation jet was mainly researched in this paper. A cavitating nozzle is the main actuator for generating a cavitating jet in biofouling cleaning. The nozzle geometry has a significant impact on the intensity of the cavitation jet [12]. A lot of work has been performed on the design of cavitation nozzles to optimize the cavitation characteristics. Chen et al. [13] proposed a cavitating nozzle optimization algorithm to improve axial vapor volume fraction and a 9.41% increment was realized. Wu et al. [14] found that increasing the length and diameter of the resonator of the organ-pipe self-resonating cavitating nozzles contributed to the occurrence of cavitation. Liu et al. [15] experimentally studied the influence of chamber dimensions on aggressive intensity of the cavitating jet. Deng et al. [16] proposed a dual cavitating waterjet, which proved to be more efficient than the traditional cavitation nozzle. Liu et al. [17] investigated a double-hole nozzle and the aggressive ability of cavitating jets was significantly improved in contrast to the single-hole nozzle. Cai et al. [18,19,20] extensively researched the cavitation characteristics of a self-resonance pulsed organ pipe nozzle. They examined the influence of nozzle geometry, including the nozzle lip and upstream and downstream contraction ratios, on these characteristics. Wang et al. [21] further conducted experimental studies on the impact of the downstream contraction ratio of the nozzle on its resonant chamber. Recent studies [22,23,24,25,26] have also investigated the erosion behavior and flow regime of jets generated by double-hole nozzles, revealing a significant enhancement in aggressiveness due to interactions between adjacent jets. In addition, surface roughness [27,28], standoff [29,30], and pressure [31] can also influence the cavitation characteristics.

Computational fluid dynamics (CFD) and erosion test are two common methods for studying cavitating jet. Fluid simulation helps a lot in the visualization of flow structures and capturing intricate flow details, while the erosion test can be used to study the damage pattern and failure mechanisms of cavitating jet [32]. Apte et al. [33] simulated cavitating flow in a venturi nozzle by several turbulence models and found that the models can reproduce the cavitating process well. Peng et al. [34] provided a comprehensive explanation of the underlying mechanism responsible for the diverse distribution patterns observed on the impingement surface. They employed an inverse finite element method grounded in pitting analysis. The hybrid model was utilized due to its accurate depiction of both cavitation instability and periodic shedding [35,36,37,38,39]. Liu et al. [17] employed a model that combined RANS and LES methodologies to represent the instability and periodic shedding of cavitation around the hydrofoil. Zhu et al. [40] further investigated the different stages of cavitation life using the Stress-Blended Eddy Simulation (SBES) turbulent model, which yielded results that closely matched experimental data. Some new models based on the existing ones were also proposed recently [41,42].

In this paper, cavitation nozzles with different geometries were studied by combining erosion experiment and CFD technology, and a device that can achieve multi-nozzle collaborative work was designed to improve cleaning efficiency. The paper is organized as follows. In Section 2, cavitation characteristics of nozzles with different structures and geometric parameters were simulated and analyzed. In Section 3, erosion tests were carried out, and the erosion effect of different cavitation nozzles were compared, and the optimal nozzle was selected. In Section 4, a self-rotating cleaning device was designed and manufactured, and an experiment was carried out to test the cleaning performance. Some conclusions are drawn in Section 5. The above studies have improved the erosion ability of single nozzle. In order to improve the cleaning efficiency of cavitation water jets, it can also be considered from the design of actuators. The Cybernetix Company has designed an underwater magnetic adsorption cleaning vehicle. The array arrangement of cavitation nozzles is used to increase the jet cleaning area [31]. Kalumuck et al. [7] determined the feasibility of installing cavitation nozzles on the rotating mechanism and obtained cleaning equipment with a cleaning efficiency of 130 m²/h/kw.

Under controlled input pressure, nozzle standoff distance, and working water depth conditions, this paper examines the impact of factors such as different throttling orifice diameters, the presence of a pre-shrinking section, the existence of an expanded diffusion cylindrical section, and different outlet shapes on the cleaning efficacy of shear-type nozzles. Additionally, various shear-type nozzles were compared with self-resonance pulsed Helmholtz and organ pipe nozzles to identify the most suitable nozzle structure for marine organism cleaning. Subsequently, an efficient cleaning device with a rotating mechanism was designed and manufactured, and the selected nozzle was installed on the device. Finally, an experimental device was developed to measure the cleaning efficiency of the rotating device.

2. Numerical Simulation of Submerged Cavitating Jet

2.1. Nozzle Structure

In this section, five shear nozzles with different configurations were investigated. Figure 1 illustrates the common structure of the shear nozzle, which consists of six segments, i.e., the inlet and outlet of the nozzle, the pre-shrinking and shrinking segment, the throttling orifice, and the expand cylinder segment.

Two self-resonance pulsed nozzles, one Helmholtz nozzle, and one organ-pipe nozzle were also simulated for comparison. The configurations of the five shear nozzles and two self-resonance pulsed nozzles are shown in

Table 1. Configurations of the seven shear nozzles and self-resonance pulsed nozzle.

Nozzle Number	Structure Type	Outlet Shape	Profile
1	Shear nozzle with a pre-shrinking segment	Cone
2	Shear nozzle with a pre-shrinking segment	Cone
3	Shear nozzle with a pre-shrinking segment	Conical outlet with expand cylinder segment
4	Shear nozzle with no pre-shrinking segment	Conical outlet with expand cylinder segment
5	Shear nozzle with no pre-shrinking segment	Spherical outlet with expand cylinder
6	Self-resonance pulsed nozzle Helmholtz type	/
7	Self-resonance pulsed nozzle organ pipe type	/

. The red arrows indicate the direction of water flow inside the nozzle. Shear nozzle No.1 and No.2 all had a pre-shrinking segment and conical outlet. Their difference lay in the diameter of the throttling orifice, with the former being 1 mm and the latter being 1.5 mm. Nozzle No.3 differed from No.2 in that it had a conical outlet with expand cylinder segment. Nozzle No.4 had no pre-shrinking segment compared with No.3. Unlike the four nozzles above, nozzle No.5 had a spherical outlet. Nozzles No.6 and No.7 were two self-resonance nozzles which were the optimums of Zhang’s [43] and Yu’s [44] research, respectively.

2.2. Numerical Model

The software Ansys Fluent (the version is 19.0) was utilized to simulate the submerged cavitation jet which was a two-phase flow. The mixture model was adopted here to simulate the cavitation jet due to its superior precision and computational efficiency. The continuity equation for the mixture model is:

$$\frac{\partial }{\partial t}\left({\rho }_m\right)+\nabla ·\left({\rho }_m{\overline{v}}_m\right)=\dot{m},$$

(1)

The momentum equation of the mixture model is:

$$\frac{\partial }{\partial t}\left({\rho }_m{\overrightarrow{v}}_m\right)+\nabla ·\left({\rho }_m{\overrightarrow{v}}_m⨂{\overrightarrow{v}}_m\right)=-\nabla p+\nabla ·\left[{\mu }_m\left(\nabla {\overrightarrow{v}}_m+\nabla {\overrightarrow{v}}_m^T\right)\right]+{\rho }_m\overrightarrow{g}+\overrightarrow{F}+\nabla ·\left(\sum _{K=1}^n{\alpha }_k{\rho }_k{\overrightarrow{v}}_{dr,k}⨂{\overrightarrow{v}}_{dr,k}\right),$$

(2)

where $\dot{m}$ represents the mass source, ${\rho }_m$ denotes the mixing density, ${\overrightarrow{v}}_m$ signifies the average mass velocity, $\overrightarrow{F}$ stands for the volume force, ${\alpha }_k$ represents the volume fraction of the k phase, and ${\overrightarrow{v}}_{dr,k}$ is identified as the drift velocity of the phase k.

Water and water vapor were selected as the flow field materials for this study. The density of the water vapor was set as $0.02558 \mathrm{k}\mathrm{g}/{\mathrm{m}}^3$, with a viscosity of $1.26\times {10}^{-6 }\mathrm{k}\mathrm{g}/(\mathrm{m}·\mathrm{s})$. For water, the density was set as $1000 \mathrm{k}\mathrm{g}/{\mathrm{m}}^3$, with a viscosity of $0.001 \mathrm{k}\mathrm{g}/(\mathrm{m}·\mathrm{s})$. The conversion between the liquid phase (water) and vapor phase (water vapor) was simulated using the Schnerr and Sauer cavitation model. The transport equation of this model is as follows:

$$\frac{\partial }{\partial t}\left(\alpha .{\rho }_v\right)+\nabla ·\left(\alpha .{\rho }_v.{\overrightarrow{V}}_V\right)=\frac{{\rho }_v.{\rho }_l}{\rho }\frac{d\alpha }{dt}$$

(3)

In Equation (3), the density of vapor is denoted as ${\rho }_v$, while the density of liquid is represented by ${\rho }_l$. The volume fraction of vapor in the system is defined as $\alpha$, and the velocity of the vapor phase is denoted as ${\overrightarrow{V}}_V$. Additionally, time is represented by the variable $t$.

The viscous model was selected as the Realizable model in the k-ε model. The Realizable k-ε model adheres to the constraint condition of Reynolds stress, ensuring consistency with real turbulence on this scale and providing a more accurate description of the diffusion velocity in circular jets. This model is comprised of two equations: the turbulent kinetic energy (k) equation and the dissipation rate (ε) equation.

The turbulent kinetic energy k equation is as follows:

$$\frac{\partial \left(\rho k\right)}{\partial t}+\frac{\partial \left(\rho k{u}_j\right)}{\partial x_j}=\frac{\partial }{\partial x_j}\left[\left(\mu +\frac{{\mu }_l}{{\sigma }_k}\right)\frac{\partial k}{\partial x_j}\right]+G_k+G_b-\rho \epsilon -Y_M+S_k$$

(4)

The dissipation rate ε equation is given below:

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

(5)

Figure 2 shows the numerical model of a submerged cavitating jet. The simulation area included the internal flow field of the nozzle and the external water environment domain. Since the nozzle itself is a rotational structure, a two-dimensional axisymmetric model was used to save the calculation time under the premise of ensuring the accuracy of the simulation. The pre-processing software Mesh 19.0 was used to mesh the model. The pressure-inlet was selected, and the gauge total pressure was established at 30 MPa. By utilizing the Bernoulli equation (Equation (6)), the Reynolds number equation (Equation (7)), and the turbulence intensity calculation equation (Equation (8)), the turbulence intensity at the inlet can be determined as 5%.

$$P_G=P_S+\frac{1}{2}\rho U^2$$

(6)

$$R_e=\frac{\rho UL}{\mu }$$

(7)

$$I\cong 0.16{\left(R_e\right)}^{-0.125}$$

(8)

In the equation, $P_G$ represents the total pressure at the inlet, $P_S$ represents the static pressure, $\rho$ represents the density of water, $U$ represents the velocity at the inlet, $L$ represents the diameter of the inlet, and $\mu$ is the dynamic viscosity coefficient.

The outlet was assigned to the pressure-outlet, with the gauge pressure at 0.1 MPa. The pressure difference was measured according to the principles of pressure gauge measurement, implying that the actual pressure is the atmospheric pressure exerted on the gauge pressure. Consequently, the actual pressure at the outlet corresponds to the pressure when the water depth is ten meters. To expedite convergence, the pressure correction method employed the SIMPLEC algorithm. The discrete format adopted the second-order upwind scheme, hybrid initialization was adopted, the number of iterations was set to 500, and residual accuracy was maintained at 10⁻⁵.

To ensure that the grid quality does not affect the simulation results, a grid dependence validation is necessary. Three different grid element sizes were selected, as shown in

Table 2. The error in pressure drop for different grid element sizes.

Grid Element Size [mm]	Pressure Drop [Pa]	Error [%]
0.1	30,000,010.409	0
0.07	29,985,590.485	−0.0481
0.05	29,998,177.102	−0.0061

. The pressure drop was calculated based on the average total pressure at the inlet and outlet. By comparing the pressure drops obtained from the second and third grid sizes with that from the first grid size, it was observed that when the grid size reached 0.1 mm, further reducing the grid size resulted in an error of less than 1%. This indicates that decreasing the grid size further has a minimal impact on the simulation results. Therefore, a grid size of 0.1 mm can be used for meshing the model.

2.3. Analysis of Simulation Results

Similar to the submerged free jet which was identified to have 3 distinct regions of flow: the undisturbed region of flow, the potential core region and the fully developed region [45,46,47], the cavitation jet also has 3 regions: the growing region, the shedding region, and the collapsing and diffusing region [48]. The damage to the target is mainly caused by the last region. The contours of pressure (mixture), volume fraction (vapor), and velocity for each nozzle, obtained through simulation results, are presented in Figure 3. The cavitating jet’s cleaning effect primarily manifests in two forms: erosion due to water flow impact and denudation resulting from bubbles collapse. The velocity contours and pressure (mixture) contours in the simulation results indicate the extent of erosion caused by water flow impact, while the volume fraction (vapor) contours reveal the degree of denudation induced by bubbles collapse.

Nozzle No.2, which has a larger orifice diameter than nozzle No.1, demonstrated a larger flow velocity coverage and a larger area after the jet reached the wall. This suggests that increasing the orifice diameter can enlarge the cleaning area. Nozzle No.3 featured an expanding cylinder segment at the outlet in comparison to nozzle No.2. The simulation results showed that nozzle No.3 had a superior large pressure holding distance and a larger area of the water vapor phase than nozzle No.2. Therefore, configuring the expanding cylinder segment can mitigate the decrease in vapor volume fraction and extend the outlet pressure holding distance. The pre-shrinking segment was removed in nozzle No.4 compared with nozzle No.3, and the simulation results indicated that although the pressure holding distance of nozzle No.4 decreased, the continuity of the vapor volume fraction increased, as well as the coverage area. The removal of the pre-shrinking segment can improve the cavitation effect of the nozzle and increase the bubble content of the jet. The outlet in nozzle No.5 was changed from conical in nozzle No.4 to spherical. Through the simulation results, it was found that the continuity of the vapor volume fraction was further improved, and the vapor showed a noticeable aggregation tendency. Consequently, it can be inferred that converting the conical outlet into a spherical one can facilitate the cavitation bubbles gathering together, thereby allowing more bubbles to reach the wall surface and enhance the bubbles denudation effect. When compared to the shear nozzle, the self-resonance pulsed Helmholtz nozzle exhibited a longer outlet velocity retention distance, but a lower vapor volume fraction. The organ pipe nozzle did not perform as effectively as the Helmholtz nozzle in terms of flow velocity and vapor volume distribution.

3. Submerged Cavitating Jet Experiment

Numerical simulation can only serve as a reference for trends, and actual cleaning effects must be verified through experiments. The seven nozzles described above were manufactured and tested in this section.

3.1. Erosion Experiment

In order to verify the practical performance of each nozzle, an erosion experiment was designed, and pure aluminum blocks were used as the objects for erosion, as shown in Figure 4. The nozzle was connected to the high-pressure hose, and the high-pressure water impacted the aluminum block after cavitation through the nozzle. The nozzle was fixed on the height adjustment clamp through which the height of the nozzle could be adjusted to achieve different standoff distances. The aluminum block was fixed by a clamp to avoid moving. Both the aluminum block and the nozzle were submerged in the water tank during the experiment. The output pressure of the pump was constant at 30 MPa during the experiment; that is, the water flow into the water tank was constant. By adjusting the tank outlet valve, the water outflow could be adjusted to be the same as the water inflow, therefore all nozzles were always at the same water depth, removing the influence of water pressure difference caused by different water depths on the actual effect of cavitating jet. The optimal standoff distance of the nozzle is between 30 mm and 50 mm [49,50]. In order to compare the performance under different standoff distances, two groups of experiments were designed. For the first group, the aluminum blocks were eroded by the seven nozzles listed in Table 1 for ten minutes at a standoff distance of 30 mm, and three nozzles with better erosion effect were selected. The selected three nozzles form the second group, and the experiment was conducted at a standoff distance of 50 mm. The aluminum block was weighed before and after the experiment and the masses were recorded.

3.2. Experimental Results and Analysis

Seven aluminum blocks sized 50 mm × 50 mm × 10 mm were eroded, and the morphology before and after experiments is shown in Figure 5. It can be seen that except for nozzle No.6, the Helmholtz nozzle, the cavitation jets generated by other nozzles caused serious damage to the surface of the aluminum blocks. The damage manifests as a central dent with pitting around it, but the pitting caused by nozzle No.7 was almost negligible. It is not appropriate to determine the erosion performance of the nozzles solely based on the erosion morphology of aluminum blocks, thus the mass loss is listed in

Table 3. Mass of aluminum blocks before and after erosion in the first group (standoff distance of 30 mm).

Nozzle Number	Initial Mass [g]	Final Mass [g]	Mass Loss [mg]
1	60.8800	60.8666	13.4
2	63.6957	63.6782	17.5
3	64.9366	64.8898	46.8
4	63.1979	63.1706	27.3
5	63.2803	63.2600	20.3
6	62.2402	62.2386	1.6
7	66.5985	66.5869	11.6

. It can be seen that the nozzles No.3, No.4, and No.5 caused the greatest mass loss to the aluminum blocks, so these three nozzles were selected to form the second group to test the erosion performance under a standoff distance of 50 mm.

Figure 6 shows the morphology of the aluminum blocks before and after erosion in the second group. It can be seen that the damage of the three aluminum blocks all includes central dent and spitting around, wherein the dent caused by nozzle No.3 seems the deepest and the spitting caused by nozzle No.5 has the widest range. The mass loss of aluminum block caused by nozzle No.5 was the largest, as seen from

Table 4. Mass of the aluminum blocks before and after erosion in the second group (standoff distance of 50 mm).

Nozzle Number	Initial Mass [g]	Final Mass [g]	Mass Loss [mg]
3	62.5468	62.5261	20.7
4	63.1108	63.0954	15.4
5	64.4566	64.4234	33.2

.

The central dent was attributed to high-pressure water flow, while the pitting around was caused by bubbles collapse. Therefore, the shallower the central dent and the wider the pitting range, the better the cavitation effect. In the first group, the nozzles No.3, No.4, and No.5 which have the expanded cylinder segment had better a erosion effect than the other nozzles. The expanded cylinder segment can prolong the lifespan of the cavitation bubbles, making more bubbles reach and collapse at the target surface, ultimately improving the erosion effect. Under the standoff distance of 30 mm, the predominant erosion effect was due to high-pressure water flow impact. However, the denudation caused by bubbles collapse became more pronounced under the standoff distance of 50 mm. Overall, nozzle No.5 in second group had the best erosion effect. This can be attributed to the spherical outlet, which facilitates the cavitation bubbles gathering together and collapsing at the target surface.

When cavitation jet is applied to cleaning, e.g., ship hull cleaning, the erosion effect of nozzle No.5 was more favored, because the shallow central dent and wide pitting range mean less damage to the surface of the hull paint layer and larger cleaning range.

4. Design and Experiment of Self-Rotating Cleaning Device

Cavitation jet cleaning is one of the main methods of ship hull cleaning. It is required that the ship’s paint layer surface remains undamaged when removing biofouling. Nozzle No.5 was selected as the cleaning component in this section. However, the cleaning area of a single nozzle is very limited, so a cleaning device with multiple cavitation nozzles is proposed to enhance cleaning efficiency.

4.1. Cleaning Device Design

The device was designed in such a way that the cavitating jet nozzle can provide recoil force to drive the rotation of the device, thereby forming a rotating water jet for cleaning purposes. This method of cleaning not only expands the cleaning area but also improves efficiency by transitioning from constant scouring to pulse scouring. The self-rotating cleaning device is shown in Figure 7, and the main internal structure of the rotation axis is depicted in Figure 8. The cleaning device includes four cavitation nozzles, which are symmetrically installed on both sides of the rotation axis. All the four nozzles have an angle of 30° with the rotation axis. Nozzle 1 and nozzle 4 have an angle of 60° with the water outlet axis. Nozzle 2 and nozzle 3 are perpendicular to the water outlet axis. The nozzles are fixedly connected to the mounting holder (5 in Figure 8) and can rotate around the rotating spindle (number 4 in Figure 8).

While in operation, the cavitation nozzle bears the reaction force generated by the jet, and the force analysis is shown in Figure 9. From the force balance equations in the X, Y, and Z directions, it can be concluded that:

$$\left\{\begin{array}{l}{F}_{R1x}+F_{R2x}=F_{L1x}+F_{L2x}\\ F_{R1y}=F_{L1y}\end{array}\right.$$

(9)

From the torque balance equation, it can be concluded that:

$$\left\{\begin{array}{l}{F}_{R1z}\cdot a=F_{L1z}\cdot a\\ F_{R1y}\cdot b=F_{L1y}\cdot b\\ F_{R1z}\cdot c+F_{R2z}\cdot d=F_{L1z}\cdot c+F_{L2z}\cdot d\end{array}\right.$$

(10)

The rotational torque of the self-rotating cleaning device is:

$$T=F_{L1x}\cdot b+F_{L2x}\cdot e+F_{R1x}\cdot b+F_{L2x}\cdot e$$

(11)

4.2. Self-Rotating Device Test and Result Analysis

The self-rotating cleaning device was manufactured and assembled, as shown in Figure 10. To evaluate the actual cleaning effect of the device, a test rig was established, as shown in Figure 11. The recoil force and the radius of cleaning area were also measured. The test rig was fixed on the pool by clamping to the wall with fixtures. The self-rotating cleaning device was horizontally mounted on the support rod which could rotate around the revolution axis. During the experiment, the self-rotating cleaning device scoured the test plate and generated a right thrust, causing the support rod to rotate counterclockwise at a small angle, thereby pulling the electronic tension meter. Then, the tension could be calculated according to principle of torque balance as:

$$Thrust=\frac{Tension\cdot Length of tension arm}{Length of thrust arm}$$

(12)

The experimental results showed that when the input water pressure was 30 Mpa, the reading of the electronic tension meter was 303.31 N (30.95 Kgf). Considering the thrust arm length to be in a 2:1 ratio with the tension arm length, the recoil force of a single rotating device was $F_{anti}=151.7 \mathrm{N}$. According to Equation (12), the actual reaction force F provided by a single nozzle can be calculated as follows:

$$\begin{array}{c}F\times {\left(\mathit{cos}30°\right)}^2\times 2+F\times \mathit{cos}30°\times 2=F_{anti}\end{array}$$

(13)

The recoil force value of a single nozzle was calculated to be 47 N.

In the experiment, the test plate was manufactured by pasting shells on a painted iron plate with waterproof glue to simulate the ship hull with biofouling, as shown in Figure 12a. The peel strength of the glue was about 1 MPa, which is equivalent to the peel strength of marine organisms such as barnacles. The test plate after cleaning is shown in Figure 12b,c. It can be seen that both the shells and glue were complete cleared. The cleaning area is a ring with an outer diameter of 270 mm and a width of 50 mm.

In addition to laboratory testing, the self-rotating cleaning device was also tested in actual engineering applications. The self-rotating cleaning device was installed onto the underwater cleaning robot for in situ cleaning of ships at sea, as shown in Figure 13. According to the images transmitted back by the underwater cleaning robot, it can be observed that the ship’s hull, previously covered by marine biofouling, was thoroughly cleaned. This test also demonstrated the effectiveness of the self-rotating cleaning device.

5. Conclusions

In this paper, the erosion (or cleaning) effect of cavitation nozzles was investigated. Five shear nozzles and two self-excited oscillating nozzles were researched through both simulation analysis and experimental study. The results showed that the shear nozzle had a better cleaning effect than the two self-excited oscillating nozzles. For shear nozzles, increasing the orifice diameter can enlarge the cleaning area. The expanded cylinder segment can mitigate the decrease in vapor volume fraction and extend the outlet pressure holding distance. A spherical outlet can facilitate the cavitation bubbles gathering together, thereby allowing more bubbles to reach and collapse at the wall surface. Therefore, the shear nozzle with no pre-shrinking section, a spherical outlet connecting to a diffusion cylindrical section outlet shape and orifice diameter of 1.5 mm had the best cleaning effect. This nozzle primarily capitalizes on denudation induced by bubble collapse during cleaning, thereby ensuring that the surface paint layer remains undamaged when cleaning ship hull. Furthermore, the nozzle maintained consistent erosion ability at a standoff distance of 30–50 mm, enabling it to adapt to changes in the cleaning standoff distance triggered by waves. A self-rotating cleaning device based on the optimum nozzle was designed and manufactured, and a test rig was established to test the cleaning effect. The cleaning area, and reaction force of one nozzle were obtained as well, which provides a design basis for the cleaning module of a cleaning robot.