Ships that sail for extended periods often experience harsh marine environments, leading to extensive corrosion and rust on their cargo hold surfaces. This corrosion can contaminate the cargo inside. Therefore, when cargo ships dock to change cargo, especially when carrying clean goods like grain, large-scale cleaning of the cargo holds is required each time, resulting in a significant workload. Currently, cargo hold cleaning primarily involves workers using handheld cleaning tools at heights, a highly dangerous method. Furthermore, many shipyards use large quantities of chemicals to ensure cleaning quality and effectiveness; statistics show that a single shipyard can use up to 300,000 tons of chemicals annually for cargo hold cleaning, causing severe environmental pollution. To improve the current state of cargo hold cleaning, one solution is to develop wall-climbing robots for interior cleaning, thereby achieving automated cleaning operations.
Currently, in the field of ship cleaning, wall-climbing robots are mainly divided into two types: tracked and wheeled. Tracked wall-climbing robots are widely used in cleaning operations on large surfaces of ships due to their advantages such as good wall adaptability and high load capacity. Ding Wensi et al. developed a permanent magnet adsorption type ship wall-climbing robot. This tracked robot is designed with a permanent magnet omnidirectional wheel mechanism with auxiliary adsorption and driven floating functions, which greatly improves the robot's load and obstacle-crossing ability. However, this robot has the disadvantage of inflexible steering. Wang Mingqiang et al. designed a multi-tracked omnidirectional wall-climbing robot, which solved the problem of low steering efficiency of tracked robots. It uses four tracks and unidirectional wheels on the tracks to achieve precise rotation of the robot. However, this robot is relatively heavy, which poses challenges in operation.
Installation, disassembly, and maintenance are difficult during on-site ship operations. To address the challenge of adapting to curved surfaces, existing wall-climbing robots face difficulties. Wang et al. designed two four-degree-of-freedom articulated mechanisms for the existing tracked magnetic adsorption wall-climbing robot's walking mechanism, enabling the robot to walk on walls with smaller curvatures and exhibiting better surface adaptability. However, the addition of multi-degree-of-freedom articulated mechanisms increases the difficulty of controlling the robot's movement on walls, making precise position control challenging. Compared to tracked wall-climbing robots, wheeled wall-climbing robots offer advantages such as lighter weight, more flexible movement, and easier installation and maintenance. Song Wei et al. developed a magnetic adsorption wall-climbing robot, improving the magnetic mass ratio and load capacity of the wheeled robot by optimizing the magnetic adsorption components. However, this robot is currently only suitable for walking on uniform ship walls, lacking sufficient ability to walk on curved surfaces or overcome obstacles. Jiang Aimin et al. developed a dual-joint wheeled robot with wall transition capabilities, capable of traversing large obstacles on walls, but its control is complex, making it difficult to achieve the control stability performance required for industrial applications. To improve the wall adaptability of wheeled wall-climbing robots, Jiang Yong et al. designed a two-wheeled transition wall-climbing robot. This wall-climbing robot adopts a bipedal wheel hybrid motion mechanism, which can move between two walls with a certain angle or even verticality. However, the robot has insufficient load capacity and cannot carry marine cleaning equipment to complete the cleaning operation of ship walls.
The interior walls of ship cargo holds are mostly intersecting, with transition angles between them. This necessitates that wall-climbing robots possess the ability to navigate these transitions. This paper leverages the agility of wheeled magnetic adsorption wall-climbing robots, optimizes the existing magnetic adsorption mechanism, and enhances the adaptability of the cleaning mechanism to different wall surfaces. A wall-adaptive wall-climbing robot is designed, and experiments verify its feasibility for movement on ship cargo hold walls, achieving green, efficient, and safe cleaning of ship cargo holds.
2. Overall Design of the Wall-Climbing Robot
The interior walls of bulk carrier holds are uneven, and the various bulkheads are at angles, with the smallest angle reaching 120°. Figure 1 shows the internal structure of a bulk carrier hold, which mainly includes bottom side sloping plates, side vertical plates, top side sloping plates, hatch coamings, and side ribs. While the outer surface of the ship is uniformly distributed, the internal wall structures of the cargo holds are complex, with low continuity between surfaces, posing a challenge to automated cleaning of the cargo holds using robots. To address this issue, this paper develops a wall-climbing robot capable of carrying cleaning equipment to complete the transitional movement across multiple walls within the cargo hold. This mainly involves transitioning from the bottom side sloping plates to the side vertical plates, and from the side vertical plates to the top side sloping plates. Through this transitional movement across multiple walls within the cargo hold, a single robot can complete the cleaning of a large area of the hold's walls to the greatest extent possible.
Figure 1 Internal structure diagram of bulk carrier hold
The overall structure of the wall-adaptive climbing robot designed in this paper is shown in Figure 2. It includes a walking mechanism, front and rear magnetic adsorption mechanisms, an adaptive cleaning mechanism, and a control box. The robot's walking mechanism (divided into left and right walking mechanisms) uses the powerful driving force generated by the motor and reducer to rotate the rubber wheels, thus enabling the robot to move. The robot's steering is achieved by utilizing the speed difference between the left and right walking mechanisms. A special arc-shaped magnetic adsorption mechanism is installed inside the two walking mechanisms. The magnetic adsorption force of this arc-shaped magnetic adsorption mechanism allows for transitions across multi-angle walls. Simultaneously, this magnetic adsorption force ensures the robot's safety, preventing slippage and tipping during movement. The front end of the wall-climbing robot is equipped with a wall-adaptive cleaning mechanism. This mechanism uses high-pressure water to clean rust and other deposits from the inner wall of the chamber and is equipped with a vacuum recovery pipe for wastewater recycling. The high-pressure water pipe and vacuum recovery pipe on the cleaning mechanism are secured to a pipe fixator, improving the stability of the robot's movement. The entire mechanism has a certain degree of flexibility, allowing it to transition along the angle of the wall to another surface when encountering a transition surface. The robot's control box is installed at the rear of the machine and contains a wireless communication device, allowing for remote control operation of the robot.
3. Design and Simulation of Magnetic Adsorption Mechanism
3.1 Mechanical Model of Wall Transition
To ensure the wall-climbing robot's ability to transition across walls, the magnetic force distribution of the magnetic adsorption mechanism needs to be adjusted, especially the front-end magnetic adsorption mechanism. When the robot transitions across a wall, a sufficiently large magnetic adsorption force is required on the front transition surface so that the resulting frictional driving force can guarantee the successful transition. The primary objective of this study is to determine the mechanical relationship of this magnetic adsorption force. Therefore, a force analysis is performed on the robot during wall transition, and the results are shown in Figure 3. In the figure, Fmag1 is the magnetic adsorption force of the robot on the front wall, Fmag2 is the magnetic adsorption force of the front magnetic adsorption structure on the bottom contact surface, Fmag3 is the magnetic adsorption force of the rear magnetic adsorption structure on the bottom contact surface, Ff1, Ff2, and Ff3 are the frictional forces at the contact points between the robot wheels and the wall, FN1, FN2, and FN3 are the supporting forces at the contact points between the robot wheels and the wall, and G is the robot's gravity.
Assuming the maximum inclination angle between the interior walls of the cargo hold is 60°, when the robot transitions from the bottom of the hold to the ramp at the bottom edge, Fmag1 needs to be large enough so that the frictional force Ff1 it provides can drive the robot to transition to the other surface. At this point, the mechanical relationship that needs to be satisfied along the x-axis is:
(1)
Along the y-axis:
(2)
Figure 3. Transitional force on the robot wall.
From the critical condition of the transition, we know that , and , if the coefficient of friction between the wheel and the wall is 0.7, the following relationship can be obtained according to formula (1) and formula (2):
(3)
3.2 Design of Magnetic Adsorption Mechanism
According to the relationship between Fmag1 and Fmag2 in formula (3), the magnetic adsorption force of the front magnetic adsorption mechanism on the front transition surface is much greater than that on the bottom contact surface. Moreover, this mechanical relationship remains stable throughout the entire wall transition process. Therefore, this paper designs an arc-shaped magnetic adsorption mechanism, which is installed on the inner side of the left and right walking mechanisms respectively to satisfy the magnetic adsorption force relationship, as shown in Figure 4.
Figure 4 Front magnetic adsorption mechanism
The arc-shaped magnetic adsorption mechanism consists of nine neodymium iron boron permanent magnets arranged in a row. To optimize the magnetic adsorption force, i.e., to increase the ratio of magnetic adsorption force to magnet mass (magnetic mass ratio), this study utilizes the characteristics of the Halbach array to concentrate the magnetic flux density at the wall contact end, greatly improving the magnetic energy utilization rate. Furthermore, the magnet has an arc-shaped structure with an arc angle of 120°, ensuring that the mechanism maintains a magnetic adsorption force on the contact surface during the wall transition, preventing the robot from falling due to loss of magnetism. The magnetic adsorption force decreases as the distance between the magnet and the wall increases. To enable the robot to transition, i.e., to satisfy the relationship of magnetic adsorption force in formula (3), it is necessary to determine the distance d1 between the arc-shaped magnet and the front transition surface and the distance d2 between the bottom contact surface. When d1 < d2, the magnetic adsorption force on the entire arc-shaped magnetic adsorption mechanism decreases from top to bottom during the transition.
3.3 Simulation of Magnetic Adsorption Mechanism
Based on the characteristics of the magnetization direction of the Halbach array magnets, this paper presents two different magnet array arrangements and conducts simulation analysis on the magnetic attraction force of the two arrangements at d2=8mm. The simulation analysis results are shown in Figure 5, where Figure 5(a) shows the arrangement with the magnets horizontally magnetized at the bottom, and Figure 5(b) shows the arrangement with the magnets vertically magnetized at the bottom. The simulation results show that the magnetic flux density at the wall surface is greater in the first arrangement than in the second arrangement. The magnetic attraction forces on the wall surface for both arrangements are Fm1=3587N and Fm2=3242N, respectively. In summary, the first arrangement has higher magnetic energy utilization and better magnetic attraction performance. Therefore, the first magnet array is selected for the magnetization direction in the front-end magnetic attraction mechanism designed in this paper.
( a ) First arrangement (b) Second arrangement
Figure 5 Simulation results of two Halbach arrays
To determine the relationship between the adsorption force of the magnetic adsorption mechanism and the wall spacing d1 and d2, the adsorption force of the transversely magnetized magnetic adsorption mechanism was simulated, and two relationship curves, Fmag1-d1 and Fmag2-d2, were obtained, as shown in Figure 6.
(a) Relationship curve between Fmag1 and d1 (b) Relationship curve between Fmag2 and d2
Figure 6 shows the relationship between magnetic attraction forces (Fmag1, Fmag2) and wall distances (d1, d2).
Figure 7. Transition process of the wall surface of the adaptive cleaning mechanism
Based on the mechanical model for preventing tipping and slipping of the robot, assuming the robot weighs 100kg, the magnetic attraction force of the robot on the bottom contact surface must satisfy Fmag2≥3000N. To ensure that the robot has a certain obstacle-crossing ability, especially the ability to cross a 6mm high weld seam, a certain distance must be maintained between the magnetic attraction mechanism and the wall surface. Therefore, d2=8.5mm is taken, and from curve (b) in Figure 6, we can obtain Fmag2=3000N at this time. According to formula (3), the magnetic attraction force of the robot on the front transition surface must satisfy Fmag1≥5834N. To satisfy this relationship, according to curve (a), d1=2.5mm is taken, and Fmag1=5987N is obtained. In summary, the installation position of the front magnetic attraction mechanism relative to the wall surface is d1=2.5mm and d2=8.5mm.
4 Adaptive Cleaning Mechanism Design
The magnetic adsorption mechanism enables the robot to transition across walls. However, in actual operation, the robot needs a cleaning mechanism to complete the wall cleaning task; therefore, the cleaning mechanism also needs to have wall transition capabilities. Currently, ship cleaning wall-climbing robots mainly use a disc-type cleaning mechanism, which is equipped with a high-pressure cleaning device and a vacuum recovery device, providing both cleaning and recovery functions. Based on the disc-type structure, this paper designs a cleaning mechanism with an added three-wheel transition mechanism. This mechanism consists of three Mecanum wheels symmetrically distributed at 120°, enabling the cleaning disc to transition across walls at a certain angle. The end of the cleaning mechanism is connected to the robot body via a rotating arm, which can rotate around the axis at the connection point. The main principle of the adaptive cleaning mechanism's wall transition is shown in Figure 7.
Figure 8 Experiment of the three-wheel transition mechanism
When the cleaning mechanism reaches the transition surface, under the action of the horizontal thrust Fp, the Mecanum wheel in the three-wheel transition mechanism that is in contact with the transition surface will make close contact with the transition surface, while the other two Mecanum wheels will rotate around the contact wheel. Then, the top wheel will make contact with the transition surface after rotating. During the entire wall transition process, the three-wheel transition mechanism drives the cleaning disc to rotate at a certain angle, causing the thrust Fp to generate a component force along the direction of the transition surface. Under the action of this force, the entire cleaning mechanism begins to move along the transition surface, finally completing the wall transition.
To verify the feasibility of the three-wheel transition mechanism, a simplified model of the mechanism was constructed, and a wall transition experiment was conducted, as shown in Figure 8. During the experiment, a thrust was applied to the simplified model along the walking wall, causing it to move towards the transition surface. The experiment showed that the three-wheel transition mechanism can complete actions such as rotation, lifting, and walking along the transition surface at the transition wall. Therefore, mounting this three-wheel transition mechanism on a cleaning mechanism can give the cleaning mechanism a certain degree of wall self-adaptation capability.
5. Prototype Experiment and Result Analysis
5.1 Prototype Testing
The prototype obtained after assembling the various modules of the robot is shown in Figure 9.
It includes an adaptive cleaning mechanism, a magnetic adsorption mechanism, and a walking mechanism. To verify the prototype's wall transition capability and its stability in moving within a ship's cargo hold, this paper conducted relevant experiments by simulating the environment inside the cargo hold, including bottom side tank walking experiments and top side tank walking experiments.
(1) Bottom side cabin walking experiment
To simulate the actual conditions at the bottom of the cargo hold, an experimental platform with a 120° wall angle was constructed, allowing the robot to perform transition tests along this wall, as shown in Figure 10. The robot travels at a constant speed of 2 m/min from the bottom side ramp. When it encounters the front transition surface, the cleaning disc equipped with a three-wheeled transition mechanism slowly moves along the front transition surface. During the movement, the entire cleaning mechanism rotates relative to the robot body towards the transition surface. When the cleaning disc is fully in contact with the transition surface, the robot's front wheels have just reached the transition surface. At this point, the magnetic adsorption mechanism acts on the transition surface and the walking surface. Under the action of the magnetic adsorption force, the robot detaches from the bottom side ramp and begins to walk along the vertical transition surface. When the rear wheels reach the transition surface, the robot completes the process of moving from the bottom side ramp to the side vertical plate.
(2) Top Side Cabin Walking Experiment
The maximum angle between the ramp of the top side of the cargo hold and the top surface is 120°. To verify that the wall-climbing robot developed in this paper has the ability to transition from the ramp of the top side of the cargo hold to the top surface, two steel plates with an angle of 120° were constructed in this experiment to simulate the characteristics of the top wall of the cargo hold. The robot's walking experiment process in the top side of the cargo hold is shown in Figure 11. During the experiment, the robot starts moving from the ramp of the top side of the cargo hold. When the adaptive cleaning mechanism moves to the top transition surface, the transition surface will give the cleaning mechanism a reaction force. At this time, the telescopic push rod connecting the robot body and the cleaning mechanism starts to act, and the telescopic push rod gives the cleaning mechanism a thrust along the normal direction of the wall, preventing the cleaning mechanism from overturning under the action of the wall reaction force and its own weight. As shown in Figure 11, after the cleaning mechanism transitions to the top wall of the cargo hold, the front wheel walking mechanism of the entire robot also completes the transition action under the action of the magnetic adsorption mechanism. During the entire top side of the cargo hold walking experiment, the robot walks stably without overturning or falling off.
5.2 Results Analysis
Figure 9 Robot Prototype
Figure 10 Bottom Side Cabin Walking Experiment
To address the limitation of existing wall-climbing robots in navigating multiple walls within ship cargo holds, this paper develops a wall-adaptive wall-climbing robot. Simulation analysis and prototype experiments verify that this robot possesses a certain degree of wall-transition capability. Two experimental platforms were established—a bottom platform and a top platform—with a 120° angle between the transition walls. Wall-walking experiments demonstrate that the magnetic adsorption mechanism provides the necessary magnetic force for the transition, and the three-wheeled transition mechanism within the cleaning system can perform rotation, lifting, and transition actions. In conclusion, the wall-climbing robot developed in this paper can complete the journey from the bottom to the top of the cargo hold, meeting the requirements for cargo hold cleaning.
This paper also discusses relevant magnetic adsorption wall-climbing robots both domestically and internationally, finding that numerous studies have conducted simulation analyses and prototype tests on the wall-transition capabilities of wall-climbing robots. Compared to the adaptive variable curvature facade split-type flexible wall-climbing robot developed by Wang Yang et al., the wall-climbing robot developed in this paper can adapt not only to concave and convex surfaces with certain curvatures, but also to walls with certain angles, making its operation and application scenarios more extensive. The wall-adaptive wheeled wall-climbing robot designed by Eto et al. can climb...
While capable of climbing multi-angled walls, this robot exhibits weak magnetic adsorption capabilities compared to the one designed in this paper, making it unsuitable for carrying heavy cleaning mechanisms. Tche et al. designed a magnetic wheel-type wall-climbing robot for vertical transitions and curved surface walking—utilizing the support of two front and rear magnetic wheels and auxiliary wheels to complete wall transitions. This robot's main function is surface inspection of various irregular walls, but it cannot be applied to cleaning cargo hold walls, and the structural stability of the two wheels is poor. The magnetic adsorption mechanism designed in this paper employs a Halbach array, which offers higher magnetic energy utilization compared to existing type A and type B magnetic circuit magnet arrangements. Liu Feng et al. applied a Halbach array to the permanent magnet adsorption unit of a tracked wall-climbing robot. Based on a traditional Halbach permanent magnet array, this research added a yoke on the weak magnetic side, effectively reducing magnetic field leakage on that side. In addition, Liu Feng et al. studied the effects of the width, height, thickness of the magnet array, and the height of the yoke on the adsorption force and adsorption efficiency. This provided a reference for the optimization of the magnetic adsorption mechanism of the robot in this paper. However, the permanent magnet adsorption unit only generates magnetic adsorption force on the walking surface, which cannot meet the adsorption force requirements for wall transitions. Given that the application of this magnet array in wheeled robots can greatly reduce the robot's weight, An Lei et al. designed a wheeled wall-climbing robot that uses an arc-shaped magnetic adsorption mechanism. Based on this magnetic mechanism, this paper adds the magnetization direction of the Halbach array and optimizes the magnetic force distribution, providing sufficient magnetic adsorption force for the robot's stable walking, while ensuring that the robot can complete the transition of walls with a minimum angle of 120°.
The wall-adaptive climbing robot developed in this paper has dual functions of cargo compartment cleaning and rust removal. The robot walks stably and can be equipped with a vacuum recovery device to prevent wastewater from polluting the environment, truly achieving efficient, safe, and environmentally friendly cleaning operations. However, the robot currently has certain limitations in its application. First, the minimum wall transition angle for the robot is 120°, making it unable to adapt to smaller wall angles or even vertical surfaces. Second, the robot is primarily suited to walking on concave walls; when walking on convex surfaces with very small angles, it is prone to losing its magnetic force and falling off. Future work will focus on further optimizing the robot's magnetic adsorption mechanism to address the challenge of vertical surface transitions and enhancing the strength of the robot's cleaning mechanism to improve the efficiency of cleaning operations.
6 Conclusions
This paper studies the wall transition characteristics of a cargo hold cleaning robot through magnetic simulation analysis and prototype experiments, and designs an arc-shaped magnetic adsorption mechanism and an adaptive cleaning mechanism. Addressing the insufficient adsorption force of conventional magnetic adsorption mechanisms, this paper uses ANSYS Maxwell 3D simulation software to optimize the magnetic force simulation, improving the magnetic flux distribution around the magnetic adsorption mechanism so that the magnetic adsorption force on the wall satisfies the mechanical relationship for wall transition. After solving the transition problem of the robot body, a three-wheel transition mechanism is designed for the wall transition of the cleaning mechanism. This mechanism, installed on the cleaning tray, enables the wall cleaning mechanism to have a certain degree of adaptability to inclined walls.
Currently, this paper has conducted relevant experiments and obtained suitable solutions to the problem of wall-climbing robots transitioning between multiple walls inside ship cargo holds. However, experiments on the robot's long-term operation at the work site have not been conducted, so it cannot be concluded that the robot can work continuously at the work site. When the robot performs cleaning operations, it generates a large amount of water mist, which poses a challenge to the robot's waterproofing. In addition, the wall-climbing robot developed in this paper is mainly designed for cleaning large-scale walls such as the bottom, sides, and top of the hold. The problem of cleaning the sides of the ribs and the narrow areas inside the cargo hold remains unsolved. Future research will focus on various wall cleaning mechanisms and integrate them onto the robot to achieve multi-robot collaborative operation, which can greatly improve the cleaning efficiency of the cargo hold and truly realize automated cleaning operations in the shipbuilding industry.
Chen Jiankun 1,2, He Kai 1*, Fang Haitao 1
1. Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences
2. University of Chinese Academy of Sciences
Reprinted from "Integration Technology"
