Steel dock door twisting opening and closing system for immersed tunnel dry dock

Abstract

The invention relates to the field of immersed tube tunnel engineering, in particular to an immersed tube tunnel dry dock steel dock door twisting opening and closing system which comprises a traction module, an intelligent connecting and disconnecting module and an intelligent control module which are in wireless communication connection, and the traction module is used for controlling opening, stopping, closing and fixing of a steel dock door through a wireless remote control winch; the intelligent connecting and separating module is used for connecting and separating the mooring rope and the steel dock gate through matching of the mechanical arm and the unmanned aerial vehicle; and the intelligent control module is used for acquiring sensing data through a remote control winch, a cable, the steel dock gate and a sensor on the unmanned aerial vehicle, analyzing each state of the steel dock gate after processing, and issuing a control instruction. The steel dock gate can be opened, stopped and closed through remote control, and data is obtained through real-time monitoring in the operation process, so that possible risks are pre-judged, and suggestions are given in time for adjustment.

Description

Technical Field

[0001] The invention relates to the field of immersed tube tunnel engineering, in particular to a hinged opening and closing system for a steel dock door of an immersed tube tunnel dry dock. Background Art

[0002] With the development, more and more immersed tube tunnels are being planned and constructed. Among them, the dry dock is an important site for the prefabrication, storage and outfitting of immersed tube sections. It often needs to be opened and closed many times, and even serves multiple immersed tube tunnel projects. Before the immersed tube section is undocking, the dry dock door is closed; when the immersed tube section is undocking, the dry dock door is opened.

[0003] The existing steel dock door winch pulling system has some shortcomings: the operation of the winch requires workers to reach the winch position, and the winches are scattered, so it is often necessary for one worker to go to the locations of multiple winches or to assign one worker to each winch, which is inefficient and there is a risk of miscommunication of instructions; workers directly observe the position and posture of the steel dock door on site, and often can only take care of the local but not the overall situation, and are prone to making wrong judgments and instructions; when the steel dock door is docked near one side of the pier, the cable between the winch of the opposite pier and the steel dock door almost spans the entire harbor, creating obstacles for the submerged tube to be undocking, and it is impossible to simultaneously take into account the stable docking of the steel dock door and the smooth undocking of the submerged tube. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a hinged opening and closing system for the steel dock door of an immersed tube tunnel dry dock. Through remote control, the steel dock door can be opened, docked and closed, and real-time monitoring and data acquisition are carried out during the operation to predict possible risks and give timely suggestions for adjustment.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] A hinged opening and closing system for a steel dock door of an immersed tunnel dry dock includes: a traction module, an intelligent connection and separation module and an intelligent control module connected by wireless communication, wherein:

[0007] The traction module is used to control the opening, docking, closing and fixing of the steel dock door through a wireless remote control winch;

[0008] The intelligent connection and separation module is used to connect and separate the cable and the steel dock door through the combination of the robotic arm and the drone;

[0009] The intelligent control module is used to obtain sensor data through remote control winches, cables, dock doors and sensors on drones. After processing, it analyzes the various states of the dock doors and issues control instructions.

[0010] Furthermore, a first remote-controlled winch and a second remote-controlled winch are respectively provided on the dock gate piers on both sides, a third remote-controlled winch and a fourth remote-controlled winch are provided on the pier on one side of the steel dock gate, and a fifth remote-controlled winch is provided on the pier on the other side; mooring posts are respectively provided at the four corners W, X, Y and Z above the steel dock gate, mooring posts w1, w2, x1 and x2 are respectively provided at the two corners W and X close to the dock gate piers, and a mooring post y and z are respectively provided at the other two corners Y and Z, and the remote-controlled winch is connected to the steel dock gate by cables and mooring posts; a ballast water tank and a water pump are provided inside the steel dock gate.

[0011] Furthermore, all remote-controlled winches are provided with inertial measurement units, proximity switches and tension sensors; the wading area of ​​the steel dock door is provided with position sensors, angle sensors, water pressure sensors and water flow velocity sensors; the ballast water tank of the steel dock door is provided with a pressure-type liquid level sensor; a drone G is provided above the steel dock door, and an industrial camera is provided on the drone G; the cable connection end between the cable and the steel dock door is an alloy component made of aluminum, cobalt and nickel.

[0012] Furthermore, the method of opening, docking, closing and fixing the steel dock door by wireless remote control winch comprises the following steps:

[0013] A1: Pump out the water from the ballast tank of the steel dock door through a water pump, connect the first remote control winch to the mooring post w1, the second remote control winch to the mooring post w2, the third remote control winch to the mooring post x1, the fourth remote control winch to the mooring post y, the fifth remote control winch to the mooring post z, the first remote control winch, the second remote control winch and the third remote control winch release the cable, the fourth remote control winch and the fifth remote control winch pull the cable, and the steel dock door is opened and moved to the center of the harbor;

[0014] A2: The first remote-controlled winch, the second remote-controlled winch and the fifth remote-controlled winch release the cable, and the third remote-controlled winch and the fourth remote-controlled winch pull the cable to move the steel dock door to the vicinity of the jetty, and during the process, the cable release speed of the fifth remote-controlled winch is lower than that of the first remote-controlled winch and the second remote-controlled winch;

[0015] A3: Untie the connection between the cable and the mooring bollard w1 and the mooring bollard z, and tighten the cable with the second remote control winch, the third remote control winch and the fourth remote control winch to fix the steel dock door near the pier;

[0016] A4: Untie the cable from the mooring bollard w2, reconnect the first remote-controlled winch to the mooring bollard w1, the fifth remote-controlled winch to the mooring bollard z, and the second remote-controlled winch to the mooring bollard x2, the third remote-controlled winch and the fourth remote-controlled winch release the cable, and the first remote-controlled winch, the second remote-controlled winch and the fifth remote-controlled winch pull the cable to move the steel dock door to the center of the harbor;

[0017] A5: The third remote-controlled winch, the fourth remote-controlled winch and the fifth remote-controlled winch release the cable, the first remote-controlled winch and the second remote-controlled winch pull the cable to move the steel dock door from the center of the harbor to the dock entrance, and the first remote-controlled winch and the second remote-controlled winch continue to pull the cable to fix the steel dock door at the dock entrance.

[0018] Furthermore, a mechanical arm is fixedly arranged in the middle platform area of ​​each mooring post above the steel dock door, and the grabbing component of the mechanical arm is provided with a roller and an angle sensor.

[0019] Furthermore, an industrial drone H is arranged above the mechanical arm, an electromagnetic suction cup is connected to the bottom of the drone H, a steering shaft is arranged at the upper end of the electromagnetic suction cup, and an electromagnetic suction cup is arranged at the lower end of each remote control winch.

[0020] Furthermore, a mechanical arm is fixedly arranged in the middle platform area of ​​each mooring post above the steel dock door, and the grabbing component of the mechanical arm is provided with a roller and an angle sensor, comprising the following steps:

[0021] B1: The captured cable connection end is close to the threaded connection end of the corresponding mooring bitt;

[0022] B2: Use the improved least square method combined with the data collected by the angle sensor to align the angles of the cable connection end and the threaded connection end of the bollard;

[0023] B3: The cable connection end aligned by the robot arm is parallel to the threaded connection end of the mooring post. It is translated toward the connection port and pressed against it, and then the roller is controlled to rotate, thereby driving the cable connection end to rotate in a specific direction, and the thread is rotated and inserted into the mooring post for connection.

[0024] Furthermore, the sensor data is obtained by remotely controlling the winch, the cable, the steel dock door and the sensors on the drone, and after processing, the analysis of each state of the steel dock door is completed to issue control instructions, including the following steps:

[0025] C1: According to the inertial measurement unit, proximity switch and tension sensor set on the remote control winch, the position sensor, angle sensor, water pressure sensor and water flow velocity sensor set at the wading position of the steel dock door, the sensor data of the steel dock door status and cable condition are obtained, and the collected data are cleaned, outliers and noise data are removed, and normalized;

[0026] C2: Design a neural network model. The input layer receives sensor data as input, selects a multi-layer fully connected neural network as a hidden layer, and the output layer outputs control instructions for the remote control winch to retract and release the cable.

[0027] C3: Use mean squared error as the loss function to measure the difference between the model's predicted value and the actual value, and use stochastic gradient descent and its variants to optimize the parameters of the neural network to minimize the loss function;

[0028] C4: Use root mean square error and mean absolute error to evaluate model performance, and adjust the hyperparameters of the neural network through cross-validation method;

[0029] C5: Deploy the trained model to the actual control system, receive sensor data in real time, and output control instructions for the cable retracting and discharging motor in the remote control winch to achieve opening and closing control of the steel dock door.

[0030] Furthermore, it also includes a warning alarm module, which obtains data on cable tension value, cable retraction and release speed, cable retraction and release length, steel dock door position, steel dock door angle and steel dock door water conditions through sensors, and sets a safety threshold to determine whether to issue a warning alarm based on the threshold range comparison.

[0031] Furthermore, the drone G is used to determine the risks and provide control instructions for the steel dock door during the process of opening, docking and closing, including the following steps:

[0032] D1: The industrial camera of UAV G is used to collect the motion image of the steel dock door in real time in a specific area at high altitude, and the image is pre-processed by denoising and enhancing.

[0033] D2: Use the deep learning YOLOv5 target detection algorithm to detect the steel dock door. In the detection stage, the preprocessed image is input into the trained YOLOv5 model, and the model outputs the bounding box coordinates of the steel dock door and the corresponding category confidence. Based on the detection results, the DeepSORT target tracking algorithm is used to track the steel dock door. For each frame of the image, the DeepSORT algorithm outputs the tracking ID of the steel dock door and its position in the image.

[0034] D3: Determine the normal motion trajectory of the steel dock door, calculate the vertical distance to the normal motion trajectory, and if it is greater than the preset position deviation threshold, it is judged that there is a position deviation risk; calculate the instantaneous speed of the steel dock door according to the position and time interval of the steel dock door in two adjacent frames of images. If the instantaneous speed is greater than the preset minimum or maximum speed threshold, it is judged that there is a speed abnormality risk; perform edge detection on the detected steel dock door area, calculate the perimeter and area of ​​the edge contour and define the shape coefficient of the steel dock door. If the change of the shape coefficient exceeds a certain threshold in multiple consecutive frames of images, it is judged that there is a shape change risk;

[0035] D4: If there is a risk of position deviation, calculate the length of the cable that needs to be adjusted according to the direction and size of the deviation; according to the safety threshold set for the speed of change of the cable length, when the speed of change of the cable length exceeds the safety threshold, adjust it by reducing the speed of the motor driving the cable or increasing the speed of the motor; if an abnormal movement of the steel dock door is detected, stop the movement of the steel dock door immediately, and send an alarm to notify the maintenance personnel to check the structure of the steel dock door and the tension of the cable;

[0036] D5: The generated control suggestions are converted into control instructions and sent to the control system of the steel dock door through the communication protocol. After receiving the instructions, the control system adjusts the corresponding motor speed or cable length to adjust the motion state of the steel dock door; the camera continuously collects the motion image of the steel dock door, repeats the target detection, tracking, risk judgment and control suggestion generation process, forms a closed-loop feedback control system, and continuously optimizes the control strategy based on the real-time feedback results.

[0037] The beneficial effects of the present invention are:

[0038] 1. The traction module is equipped with a minimum number of remote-controlled winches and mooring posts, which can achieve the purpose of opening, docking and closing the steel dock door while reducing the amount of renovation work as much as possible and reducing the renovation cost of the system;

[0039] 2. Through the intelligent connection and disconnection module, the connection and disconnection between the mooring post and the cable can be completed without manual work under the cooperation of the robotic arm and the drone, eliminating the process of manually going to the top of the steel dock door by boat to connect or disconnect the cable, saving the time for the state change of the steel dock door, reducing the tediousness and danger of manual operation, and improving the overall completion efficiency;

[0040] 3. In the intelligent docking module, the disconnected cables are connected by drones and transmitted to the magnetic suction cups set under the corresponding remote-controlled winches, and the steel dock door is docked on one side of the jetty, ensuring that there are no obstacles of cables and steel dock doors in the route of the immersed tube undocking, thus improving the efficiency of dry docking operations;

[0041] 4. The intelligent control module obtains data and makes calculations in real time through algorithms and sensors, avoiding the erroneous judgments and instructions made by workers in the traditional mode who directly observe the position and posture of the steel dock door on site without considering the overall situation, thus improving the accuracy and efficiency of control;

[0042] 5. The risk status of the steel dock door, remote-controlled winch and cable is monitored by both the early warning alarm module and the UAV G, with multiple safeguards to reduce the probability of risk. In addition, the UAV G can give timely control instructions in the face of risk situations, and directly respond to the instructions to reduce the occurrence of risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a top view of the steel dock door opening;

[0044] Figure 2 This is a top-down diagram of the steel dock door docking;

[0045] Figure 3 This is a top view of the steel dock door being closed;

[0046] Figure 4 This is a schematic diagram of the mechanical arm structure on the steel dock door.

[0047] Among them: 1. The first remote-controlled winch; 2. The second remote-controlled winch; 3. The third remote-controlled winch; 4. The fourth remote-controlled winch; 5. The fifth remote-controlled winch; 6. The W angle of the steel dock door; 7. The X angle of the steel dock door; 8. The Y angle of the steel dock door; 9. The Z angle of the steel dock door; 10: Jetty; 11. Cable; 12. Harbor; 13. Dock door pier; 14. Roller. DETAILED DESCRIPTION

[0048] See also Figure 1 , Figure 2 and Figure 3 As shown, the present invention relates to a hinged opening and closing system for a steel dock door of an immersed tunnel dry dock, comprising: a traction module, an intelligent connection and disconnection module and an intelligent control module connected by wireless communication, wherein:

[0049] The traction module is used to control the opening, docking, closing and fixing of the steel dock door through a wireless remote control winch;

[0050] The intelligent connection and separation module is used to connect and separate the cable and the steel dock door through the combination of the robotic arm and the drone;

[0051] The intelligent control module is used to obtain sensor data through remote control winches, cables, dock doors and sensors on drones. After processing, it analyzes the various states of the dock doors and issues control instructions.

[0052] Furthermore, a first remote-controlled winch 1 and a second remote-controlled winch 2 are respectively arranged on the dock gate piers 13 on both sides, a third remote-controlled winch 3 and a fourth remote-controlled winch 4 are arranged on the pier 10 on one side of the steel dock gate, and a fifth remote-controlled winch 5 is arranged on the pier 10 on the other side; mooring posts are respectively arranged at the four corners W, X, Y and Z above the steel dock gate, mooring posts w1, w2, x1 and x2 are respectively arranged at the two corners W and X close to the dock gate pier 13, and a mooring post y and z are respectively arranged at the other two corners Y and Z, and the remote-controlled winch is connected to the steel dock gate through a cable 11 and a mooring post; a ballast water tank and a water pump are arranged inside the steel dock gate.

[0053] Specifically, there is a large amount of water in the harbor 12, and a ballast tank and a pump are provided inside the steel dock gate. When the water in the ballast tank is released by the pump, the weight of the steel dock gate is reduced, and the buoyancy generated by the water in the harbor 12 facilitates pulling the steel dock gate through the cable 11. In the setting of the remote control winch, since the steel dock gate needs to reach a docked state, in order to save the average tension of the cable 11 and play a role in guiding the moving direction of the steel dock gate, one more remote control winch is provided on the pier 10 on the side of the docked steel dock gate. As for the setting of the mooring post, since the tensile force that a single mooring post can bear is limited, it is required that a single mooring post can only bind one cable 11. Since the moving direction of the steel dock gate needs to be changed by the cable 11 on the side of the steel dock gate close to the dock gate pier 13, two mooring posts are provided at the W and X angles respectively.

[0054] Furthermore, all remote-controlled winches are provided with inertial measurement units, proximity switches and tension sensors; the wading area of ​​the steel dock door is provided with position sensors, angle sensors, water pressure sensors and water flow velocity sensors; the ballast water tank of the steel dock door is provided with a pressure-type liquid level sensor; a drone G is provided above the steel dock door, and an industrial camera is provided on the drone G; the cable connection end between the cable 11 and the steel dock door is an alloy component made of aluminum, cobalt and nickel.

[0055] Specifically, in the remote control winch, the inertial measurement unit can measure the linear acceleration of the cable 11 during the process of cable winding and releasing, and the speed of the cable 11 can be adjusted by regulating the motor in the remote control cable; the proximity switch is a non-contact position detection device, which is installed in the remote control winch to measure the length data of the cable 11 during the process of cable winding and releasing; the tension sensor is used to measure the tension exerted on the cable 11 during operation. These data are very important for ensuring the smooth movement of the steel dock door, preventing the cable 11 from breaking due to excessive tension, and ensuring the tension balance between the cables 11 to maintain the correct posture of the steel dock door, and dynamically adjusting the speed and strength of cable 11 winding and releasing.

[0056] The steel dock door is equipped with a position sensor to accurately understand the real-time position of the steel dock door, determine whether it has reached the target open or closed position, and provide position feedback for the control algorithm to adjust the retraction and release speed and length of the cable 11. The angle sensor is set to ensure that the steel dock door does not tilt or rotate excessively during movement, maintain its correct posture, and prevent problems such as jamming and damage due to abnormal posture. Water pressure sensor, considering the effect and influence of water pressure on the steel dock door when working underwater, these data can assist in adjusting the tension and retraction and release speed of the cable 11 to cope with the interference of pressure changes at different water depths on the movement of the steel dock door. Water flow velocity sensor, the water flow velocity will produce impact force and resistance on the steel dock door. Through the data of this sensor, the influence of these forces on the movement of the steel dock door can be predicted and compensated in advance, and the control strategy of the cable 11 can be optimized.

[0057] The pressure level sensor installed in the ballast water tank of the steel dock door can obtain the displacement of the ballast water tank, judge the weight of the steel dock door, and determine the change of each tension value. The cable connection end of the cable 11 and the steel dock door is an alloy component made of aluminum, cobalt and nickel. The alloy made of aluminum, cobalt and nickel has stronger magnetism and is often used as a key component of the electromagnetic crane after power is turned on. Since the cable 11 used in this solution is heavy, in order to facilitate the connection end of the cable 11 to the electromagnetic suction cup to be more tightly connected and not easy to fall off, this kind of alloy is used.

[0058] Furthermore, the method of opening, docking, closing and fixing the steel dock door by wireless remote control winch comprises the following steps:

[0059] A1: Pump out the water from the ballast tank of the steel dock door through a water pump, connect the first remote control winch 1 to the mooring post w1 through a cable 11, connect the second remote control winch 2 to the mooring post w2, connect the third remote control winch 3 to the mooring post x1, connect the fourth remote control winch 4 to the mooring post y, and connect the fifth remote control winch 5 to the mooring post z. The first remote control winch 1, the second remote control winch 2 and the third remote control winch 3 release the cable, and the fourth remote control winch 4 and the fifth remote control winch 5 pull the cable. The steel dock door is opened and moved to the center of the harbor 12.

[0060] A2: The first remote-controlled winch 1, the second remote-controlled winch 2 and the fifth remote-controlled winch 5 release the cable, and the third remote-controlled winch 3 and the fourth remote-controlled winch 4 pull the cable to move the steel dock door to the vicinity of the jetty 10. In the process, the cable release speed of the fifth remote-controlled winch 5 is lower than that of the first remote-controlled winch 1 and the second remote-controlled winch 2;

[0061] A3: Untie the connection between the cable 11 and the mooring bollard w1 and the mooring bollard z, and tighten the cable 11 with the second remote-controlled winch 2, the third remote-controlled winch 3 and the fourth remote-controlled winch 4 to fix the steel dock door near the jetty 10;

[0062] A4: Untie the connection between the cable 11 and the mooring bollard w2, reconnect the first remote control winch 1 and the mooring bollard w1, the fifth remote control winch 5 and the mooring bollard z, and the second remote control winch 2 and the mooring bollard x2, the third remote control winch 3 and the fourth remote control winch 4 release the cable, and the first remote control winch 1, the second remote control winch 2 and the fifth remote control winch 5 pull the cable to move the steel dock door to the center of the harbor 12;

[0063] A5: The third remote-controlled winch 3, the fourth remote-controlled winch 4 and the fifth remote-controlled winch 5 release the cable, the first remote-controlled winch 1 and the second remote-controlled winch 2 pull the cable, and the steel dock door is moved from the center of the harbor 12 to the dock entrance, and the first remote-controlled winch 1 and the second remote-controlled winch 2 continue to pull the cable to fix the steel dock door at the dock entrance.

[0064] Specifically, in step A1, the first remote control winch 1, the second remote control winch 2 and the third remote control winch 3 release the cable, and the fourth remote control winch 4 and the fifth remote control winch 5 pull the cable, so that the distance between the steel dock door and the dock door pier 13 is increased and the steel dock door is moved to the center of the harbor 12. In step A2, the first remote control winch 1, the second remote control winch 2 and the fifth remote control winch 5 release the cable, and the third remote control winch 3 and the fourth remote control winch 4 pull the cable. Figure 1 As shown, the steel dock door is translated to the right to approach the right jetty 10. During the process, the cable release speed of the fifth remote-controlled winch 5 is reduced and the corresponding cable 11 is in a tensioned state to maintain the stability of the steel dock door, so as to avoid the danger and loss caused by the steel dock door tipping due to the cable release too fast. Figure 2 As shown, the connection between the cable 11 and the mooring post w1 and the mooring post z is untied, and the connecting end of the cable 11 is connected by a drone to connect the cable 11 to the electromagnetic suction cup below the pier 10 to avoid other hazards caused by the random placement of the cable 11. The second remote control winch 2, the third remote control winch 3 and the fourth remote control winch 4 tighten the cable 11 to make the steel dock door close to one side of the pier 10, thereby improving the stability of the steel dock door and not affecting the docking port to the greatest extent, thereby ensuring smooth docking of the immersed tube.

[0065] In step A4, the third remote-controlled winch 3 and the fourth remote-controlled winch 4 are changed from cable reeling to cable unreeling, and the second remote-controlled winch 2 is changed to be connected to the mooring post x2, thereby changing the traction direction and moving the steel dock door to the center of the harbor 12. In step A5, if Figure 3 As shown, the third remote control winch 3, the fourth remote control winch 4 and the fifth remote control winch 5 release the cable, and the first remote control winch 1 and the second remote control winch 2 pull the cable to move the steel dock door from the center of the harbor 12 to the dock port to close the dock port. The first remote control winch 1 and the second remote control winch 2 continuously pull the cable to make the steel dock door stably stagnant at the dock port and not easily tipped over by external forces.

[0066] Furthermore, a mechanical arm is fixedly arranged in the middle platform area of ​​each mooring post above the steel dock door, and the grabbing component of the mechanical arm is provided with a roller 14 and an angle sensor.

[0067] Specifically, the mechanical arm is fixedly connected to the upper surface platform area of ​​the steel dock door, and the operating direction of the mechanical arm is controlled by the mechanical arm's own multiple steering axes. According to the layout of the mechanical arm, the six mooring posts can be flexibly operated one by one to connect or disconnect the cable connection end with the mooring post; Figure 4As shown, the gripping component of the manipulator is provided with a roller 14 and an angle sensor, wherein the rotation of the roller 14 will not affect the gripping operation of the gripping component, the rotation direction of the roller 14 is determined according to the thread direction of the mooring post, and the movement direction of the transmission motor of the roller 14 is adjusted according to different operation requirements. When the roller 14 is operating, the gripping component will relax the gripping force on the cable connection end, only maintain the gripping direction to prevent the cable connection end from leaving the gripping range, and make the roller 14 close to the cable connection end, ensuring that the rotation of the roller 14 can effectively drive the cable connection end to rotate, thereby achieving the operation purpose.

[0068] Furthermore, an industrial drone H is arranged above the mechanical arm, an electromagnetic suction cup is connected to the bottom of the drone H, a steering shaft is arranged at the upper end of the electromagnetic suction cup, and an electromagnetic suction cup is arranged at the lower end of each remote control winch.

[0069] Specifically, the drone H is suspended above the mechanical arm during operation. Since the current industrial drone has a flight time of about one hour, it is necessary to control the operation time of the drone. An electromagnetic suction cup is connected below the drone H, and the distance between the drone H and the electromagnetic suction cup can also be extended through a pipeline or other device. When the drone H operates to separate the cable 11 from the steel dock door, the flight posture is adjusted to make the electromagnetic suction cup close to the designated cable connection end. When the cable connection end is separated from the mooring column, the approaching electromagnetic suction cup is energized to generate magnetic attraction, so that the cable connection end is adsorbed on the electromagnetic suction cup. The flight of the drone H drives the cable connection end to fly to the vicinity of the remote control winch to which the cable 11 belongs. The electromagnetic suction cup set below the corresponding remote control winch is energized, so that the cable connection end and the electromagnetic suction cup set below the remote control winch and the electromagnetic suction cup set below the drone H are adsorbed together. After determining that the electromagnetic suction cup set below the remote control winch is firmly adsorbed, the drone H cuts off the power to the electromagnetic suction cup connected to the fuselage and flies to the next operation point. At this time, the cable connection end is adsorbed on the electromagnetic suction cup set below the corresponding remote control winch, meeting the operation requirements of the steel dock door, and does not affect the route of the submerged tube out of the dock.

[0070] When the drone H is operating to connect the cable 11 to the steel dock door, the drone H flies close to the designated cable connection end, and energizes the electromagnetic suction cup under the fuselage, so that the cable connection end is adsorbed together with the electromagnetic suction cup set under the remote control winch and the electromagnetic suction cup set under the drone H. After confirming that the electromagnetic suction cup under the fuselage is firmly adsorbed, the electromagnetic suction cup under the remote control winch is powered off, and the drone H flies away from the cable connection end to the designated position; the steering shaft at the upper end of the electromagnetic suction cup of the drone H is adjusted to adjust the angle of the electromagnetic suction cup, thereby driving the adjustment of the angle of the cable connection end, so that the mechanical arm can grasp the cable connection end at the correct angle. After the grasping component of the mechanical arm grasps the cable connection end, the electromagnetic suction cup under the drone is powered off and flies away to the next operation area.

[0071] Furthermore, a mechanical arm is fixedly arranged in the middle platform area of ​​each mooring post above the steel dock door, and the grabbing component of the mechanical arm is provided with a roller 14 and an angle sensor, comprising the following steps:

[0072] B1: The captured cable connection end is close to the threaded connection end of the corresponding mooring bitt;

[0073] B2: Use the improved least square method combined with the data collected by the angle sensor to align the angles of the cable connection end and the threaded connection end of the bollard;

[0074] B3: The cable connection end aligned by the robot arm is parallel to the threaded connection end of the mooring post. After being translated toward the connection port and pressed against it, the control roller 14 is rotated, thereby driving the cable connection end to rotate in a specific direction, and the thread is rotated and inserted into the mooring post for connection.

[0075] Specifically, the specific implementation steps of the least squares method for angle correction are as follows:

[0076] 1. Use the angle sensor to measure the actual angle of the cable connection end at multiple positions to obtain a series of angle measurement values ​​yi, and record the corresponding measurement point serial number or other identification i;

[0077] 2. Assuming that there is a linear relationship between the cable connection angle and a potential factor (such as position, time, etc.), the model can be expressed as y = a + bx, where a and b are the parameters to be estimated;

[0078] 3. The goal of the least squares method is to make the actual measurement value y i and the model predicted value a+bx i The sum of squared errors is minimized, that is, solving the following optimization problem:

[0079]

[0080] Taking partial derivatives of the formula with respect to a and b respectively, and setting the partial derivatives to 0, we can get the following system of equations:

[0081]

[0082] By solving this system of equations, we can get the estimated values ​​of parameters a and b;

[0083] 4. Use another set of independent validation data, substitute it into the estimated model, calculate the difference between the predicted angle and the actual angle, and evaluate the accuracy of the model;

[0084] 5. Based on the estimated model, calculate the angle value that the pipeline should have at the target position or state. If the current angle deviates from the target angle, it can be calibrated by adjusting the gripping components of the robot arm and the support structure of the steering axis of the robot arm;

[0085] 6. Install the pipe accurately according to the calibrated angle requirements, and after ensuring that its angle meets the expected requirements, start installing the cable connection end to the mooring post.

[0086] Furthermore, the sensor data is obtained by remotely controlling the winch, the cable 11, the steel dock door and the sensors on the drone, and after processing, the analysis of each state of the steel dock door is completed to issue control instructions, including the following steps:

[0087] C1: According to the inertial measurement unit, proximity switch and tension sensor set on the remote control winch, the position sensor, angle sensor, water pressure sensor and water flow velocity sensor set at the wading position of the steel dock door, the sensor data of the state of the steel dock door and the condition of the cable 11 are obtained, and the collected data are cleaned, outliers and noise data are removed, and normalized;

[0088] C2: Design a neural network model, where the input layer receives sensor data as input, a multi-layer fully connected neural network is selected as a hidden layer, and the output layer outputs control instructions for the remote control winch to retract and release the cable 11;

[0089] C3: Use mean squared error as the loss function to measure the difference between the model's predicted value and the actual value, and use stochastic gradient descent and its variants to optimize the parameters of the neural network to minimize the loss function;

[0090] C4: Use root mean square error and mean absolute error to evaluate model performance, and adjust the hyperparameters of the neural network through cross-validation method;

[0091] C5: Deploy the trained model to the actual control system, receive sensor data in real time, and output control instructions for the cable 11 retractable motor in the remote control winch to achieve opening and closing control of the steel dock door.

[0092] Specifically, in step C1, a sensor network is formed according to various installed sensors, and the steel dock door is operated under various working conditions (different water flow speeds, different loads, different opening and closing speed requirements, etc.), and sensor data is collected. At the same time, the control instructions of the corresponding cable 11 receiving and discharging motor (such as motor speed, direction and torque, etc.) are recorded, and the collected data is cleaned to remove abnormal values ​​and noise data. Statistical methods can be used to remove data points that obviously deviate from the normal range, and the data is normalized to map data of different dimensions and ranges into a unified range, such as [0, 1]. For the position data x, the following formula can be used for normalization:

[0093] X normalized =(xx min ) / (x max -x min )

[0094] Among them, x min and x max are the maximum and minimum values ​​of the position data respectively.

[0095] In step C2, the neural network model is designed, and the input layer receives sensor data as input. Assume that we have n sensors, and the data collected by each sensor at a time is a vector S = [s1, s2, ..., s n ], where s i The data collected by the i-th sensor at the time. A multi-layer fully connected neural network (FCN) can be selected as the hidden layer. Assume that the l-th layer has m l neurons, for the input of the jth neuron in the lth layer for:

[0096]

[0097] in, is the connection weight between the ith neuron in the l-1th layer and the ith neuron in the lth layer, It is the l-1 The input of the i-th neuron in the layer out, is the bias term of the i-th neuron in the l-th layer. The output of the neuron is obtained through the activation function f:

[0098] The commonly used activation function is the ReLU function: f(x) = max(0,x). The output layer outputs the control instructions of the cable 11 for receiving and discharging the motor, such as the motor speed v, the direction d (1 for forward direction, -1 for reverse direction), the torque T, etc.

[0099] In step C3, the model is trained and the mean square error (MSE) is used as the loss function to measure the difference between the model prediction value and the actual value. Assume that the model's prediction output is The actual target output is y = [v, d, T], then the loss function L is:

[0100] Here, m is the number of training samples.

[0101] Use stochastic gradient descent (SGD) and its variant (Adam) to optimize the parameters of the neural network (connection weights w and bias terms b) to minimize the loss function. Taking the Adam algorithm as an example, the parameter update formula is:

[0102] Calculate the gradient g: Calculate the first moment estimate m t and the second moment estimate v t :

[0103] m t =β1m t-1 +(1-β1)gv t =β2v t-1 +(1-β2)g 2 Among them, β1 and β2 are hyperparameters, usually set to 0.9 and 0.999 respectively. Bias correction is performed on the first-order moment estimate and the second-order moment estimate:

[0104] Update parameter w:

[0105] Here, α is the learning rate and ∈ is a small constant to prevent the denominator from being zero.

[0106] In step C4, the model is evaluated and adjusted, and the performance of the model is evaluated using indicators such as root mean square error (RMSE) and mean absolute error (MAE):

[0107] Root Mean Square Error (RMSE):

[0108] Mean Absolute Error (MAE):

[0109] The hyperparameters of the neural network (such as the number of layers, the number of neurons in each layer, the learning rate, the regularization parameters, etc.) are adjusted through methods such as cross-validation to improve the performance of the model.

[0110] In step C5, model deployment and real-time control are performed to deploy the trained model into the actual control system, receive sensor data in real time, and output control instructions for the cable 11 to retract and discharge the motor, thereby realizing the opening and closing control of the steel dock door.

[0111] Furthermore, it also includes a warning alarm module, which obtains data on cable tension value, cable retraction and release speed, cable retraction and release length, steel dock door position, steel dock door angle and steel dock door water conditions through sensors, and sets a safety threshold to determine whether to issue a warning alarm based on the threshold range comparison.

[0112] Specifically, the early warning alarm module is activated based on the threshold comparison algorithm of the sensors set on the remote control winch, cable 11 and steel dock door, which can effectively and timely convey risk situation prompts and improve the safety of the system.

[0113] Furthermore, the drone G is used to determine the risks and provide control instructions for the steel dock door during the process of opening, docking and closing, including the following steps:

[0114] D1: The industrial camera of UAV G is used to collect the motion image of the steel dock door in real time in a specific area at high altitude, and the image is pre-processed by denoising and enhancing.

[0115] D2: Use the deep learning YOLOv5 target detection algorithm to detect the steel dock door. In the detection stage, the preprocessed image is input into the trained YOLOv5 model, and the model outputs the bounding box coordinates of the steel dock door and the corresponding category confidence. Based on the detection results, the DeepSORT target tracking algorithm is used to track the steel dock door. For each frame of the image, the DeepSORT algorithm outputs the tracking ID of the steel dock door and its position in the image.

[0116] D3: Determine the normal motion trajectory of the steel dock door, calculate the vertical distance to the normal motion trajectory, and if it is greater than the preset position deviation threshold, it is judged that there is a position deviation risk; calculate the instantaneous speed of the steel dock door according to the position and time interval of the steel dock door in two adjacent frames of images. If the instantaneous speed is greater than the preset minimum or maximum speed threshold, it is judged that there is a speed abnormality risk; perform edge detection on the detected steel dock door area, calculate the perimeter and area of ​​the edge contour and define the shape coefficient of the steel dock door. If the change of the shape coefficient exceeds a certain threshold in multiple consecutive frames of images, it is judged that there is a shape change risk;

[0117] D4: If there is a risk of position deviation, calculate the length of the cable that needs to be adjusted according to the direction and size of the deviation; according to the safety threshold set for the speed of change of the cable length, when the speed of change of the cable length exceeds the safety threshold, adjust it by reducing the speed of the motor driving the cable or increasing the speed of the motor; if an abnormal movement of the steel dock door is detected, stop the movement of the steel dock door immediately, and send an alarm to notify the maintenance personnel to check the structure of the steel dock door and the tension of the cable;

[0118] D5: The generated control suggestions are converted into control instructions and sent to the control system of the steel dock door through the communication protocol. After receiving the instructions, the control system adjusts the corresponding motor speed or cable length to adjust the motion state of the steel dock door; the camera continuously collects the motion image of the steel dock door, repeats the target detection, tracking, risk judgment and control suggestion generation process, forms a closed-loop feedback control system, and continuously optimizes the control strategy based on the real-time feedback results.

[0119] Specifically, in step D1, image acquisition and preprocessing are performed. The image sequence of the steel dock door movement process is collected by the drone G. The noise in the image is removed by using methods such as median filtering and Gaussian filtering. The contrast and clarity of the image are enhanced by operations such as histogram equalization and contrast stretching.

[0120] In step D2, target detection and tracking are performed. First, a large amount of image data containing steel dock doors is collected and the steel dock doors are labeled. Then, the YOLOv5 model is trained using the labeled data. In the detection stage, the preprocessed image is input into the trained YOLOv5 model, and the model outputs the bounding box coordinates of the steel dock door and the corresponding category confidence. Based on the detection results, the DeepSORT target tracking algorithm is used to track the steel dock door. The DeepSORT algorithm combines target detection results with Kalman filter prediction and appearance feature matching to achieve continuous tracking of the target. For each frame of the image, the DeepSORT algorithm outputs the tracking identification ID of the steel dock door and its position in the image (x t ,y t ).

[0121] In step D3, the risk judgment algorithm is performed. First, the normal motion trajectory of the steel dock door is determined, assuming it is a function f(x). For the tracked steel dock door position (x t ,y t ), calculate the vertical distance d from the normal motion trajectory, d = |y t -f(x t )|If d>d threshold (d threshold is the preset position deviation threshold), it is judged that there is a position deviation risk. t-1 ,y t-1 ) and (x t ,y t ), and the time interval Δt between the two frames, calculate the instantaneous velocity v of the steel dock door:

[0122]

[0123] If v <v min or v>v max (v min and v max The speed is determined to be abnormal if the speed is greater than the preset minimum and maximum speed thresholds. Perform edge detection on the detected steel dock door area, for example, use the Canny edge detection algorithm to obtain the edge contour of the steel dock door. Calculate the perimeter C and area A of the edge contour. Define the shape coefficient of the steel dock door

[0124] If the change of the morphological coefficient F in multiple consecutive frames of images exceeds a certain threshold, it is determined that there is a risk of morphological change.

[0125] In step D4, control suggestions are generated. If there is a risk of position deviation, the length of the cable 11 that needs to be adjusted is calculated based on the direction and size of the deviation. Assuming that the steel dock door is controlled by n cables 11, the adjustment length ΔL corresponding to the i-th cable 11 is i Calculated by the following formula: ΔL i =k i ×d, where k i is the influence coefficient of the i-th cable 11 on the position of the steel dock door, which can be obtained through the mechanical model of the system or actual calibration. If the speed is too fast, reduce the speed of the motor driving the cable 11; if the speed is too slow, increase the speed of the motor. The adjustment amount of the motor speed Δω can be calculated by the following formula Δω=λ×(vv target ), where λ is the speed adjustment factor, v target If an abnormal shape is detected, the movement of the steel dock door is stopped immediately, and an alarm is sent to notify the maintenance personnel to check the structure of the steel dock door and the tension of the cable 11.

[0126] In step D5, the system responds and provides feedback in real time, converts the generated control suggestions into control instructions, and sends them to the control system of the steel dock door through a communication protocol (such as TCP / IP protocol). After receiving the instructions, the control system adjusts the corresponding motor speed or cable 11 length to adjust the motion state of the steel dock door. The camera of the drone G continuously collects the motion image of the steel dock door, repeats the above target detection, tracking, risk judgment and control suggestion generation process, and forms a closed-loop feedback control system. According to the results of real-time feedback, the control strategy is continuously optimized to improve the safety and accuracy of the steel dock door movement.

[0127] When the early warning alarm module and the UAV G work together, the normal and safe operation of the steel dock door and various equipment is guaranteed, reducing the operational risks in all aspects.

[0128] The above implementation modes are merely descriptions of the preferred implementation modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering and technical personnel in the field shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A hinged opening and closing system for steel dock doors of immersed tunnel dry dock, characterized in that: include: Traction module, intelligent pick-up and drop-off module and intelligent control module connected by wireless communication, including: The traction module is used to control the opening, docking, closing and fixing of the steel dock door through a wireless remote control winch; The intelligent connection and separation module is used to connect and separate the cable and the steel dock door through the combination of the robotic arm and the drone; The intelligent control module is used to obtain sensor data through remote control winches, cables, dock doors and sensors on drones. After processing, it analyzes the various states of the dock doors and issues control instructions.

2. The hinged opening and closing system for the steel dock door of the immersed tube tunnel dry dock according to claim 1 is characterized in that: The first and second remote-controlled winches are respectively provided at the dock gate piers on both sides, the third and fourth remote-controlled winches are provided at the pier on one side of the steel dock gate, and the fifth remote-controlled winch is provided at the pier on the other side; mooring posts are respectively provided at the four corners W, X, Y and Z above the steel dock gate, the two corners W and X close to the dock gate piers are respectively provided with mooring posts w1, w2, x1 and x2, and the other two corners Y and Z are respectively provided with a mooring post y and z, and the remote-controlled winch is connected to the steel dock gate by cables and mooring posts; a ballast water tank and a water pump are provided inside the steel dock gate.

3. The hinged opening and closing system for the steel dock door of the immersed tube tunnel dry dock according to claim 2 is characterized in that: All remote-controlled winches are equipped with inertial measurement units, proximity switches and tension sensors. The wading area of ​​the steel dock door is equipped with position sensors, angle sensors, water pressure sensors and water flow velocity sensors. The ballast water tank of the steel dock door is equipped with a pressure liquid level sensor. A drone G is installed above the steel dock door, and an industrial camera is installed on the drone G. The cable connection end between the cable and the steel dock door is an alloy component made of aluminum, cobalt and nickel.

4. The hinged opening and closing system for the steel dock door of the immersed tube tunnel dry dock according to claim 1 is characterized in that: The method of realizing the opening, docking, closing and fixing of the steel dock door by a wireless remote control winch comprises the following steps: A1: Pump out the water from the ballast tank of the steel dock door through a water pump, connect the first remote control winch to the mooring post w1, the second remote control winch to the mooring post w2, the third remote control winch to the mooring post x1, the fourth remote control winch to the mooring post y, the fifth remote control winch to the mooring post z, the first remote control winch, the second remote control winch and the third remote control winch release the cable, the fourth remote control winch and the fifth remote control winch pull the cable, and the steel dock door is opened and moved to the center of the harbor; A2: The first remote-controlled winch, the second remote-controlled winch and the fifth remote-controlled winch release the cable, and the third remote-controlled winch and the fourth remote-controlled winch pull the cable to move the steel dock door to the vicinity of the jetty, and during the process, the cable release speed of the fifth remote-controlled winch is lower than that of the first remote-controlled winch and the second remote-controlled winch; A3: Untie the connection between the cable and the mooring bollard w1 and the mooring bollard z, and tighten the cable with the second remote control winch, the third remote control winch and the fourth remote control winch to fix the steel dock door near the pier; A4: Untie the cable from the mooring bollard w2, reconnect the first remote-controlled winch to the mooring bollard w1, the fifth remote-controlled winch to the mooring bollard z, and the second remote-controlled winch to the mooring bollard x2, the third remote-controlled winch and the fourth remote-controlled winch release the cable, and the first remote-controlled winch, the second remote-controlled winch and the fifth remote-controlled winch pull the cable to move the steel dock door to the center of the harbor; A5: The third remote-controlled winch, the fourth remote-controlled winch and the fifth remote-controlled winch release the cable, the first remote-controlled winch and the second remote-controlled winch pull the cable to move the steel dock door from the center of the harbor to the dock entrance, and the first remote-controlled winch and the second remote-controlled winch continue to pull the cable to fix the steel dock door at the dock entrance.

5. The hinged opening and closing system for the steel dock door of the immersed tube tunnel dry dock according to claim 1 is characterized in that: A robotic arm is fixedly installed in the middle platform area of ​​each mooring post above the steel dock door, and the grabbing part of the robotic arm is provided with rollers and angle sensors.

6. The hinged opening and closing system for the steel dock door of the immersed tube tunnel dry dock according to claim 1 is characterized in that: An industrial drone H is arranged above the mechanical arm, an electromagnetic suction cup is connected to the bottom of the drone H, a steering shaft is arranged at the upper end of the electromagnetic suction cup, and an electromagnetic suction cup is arranged at the lower end of each remote-controlled winch.

7. The hinged opening and closing system for the steel dock door of the immersed tube tunnel dry dock according to claim 5 is characterized in that: A mechanical arm is fixedly arranged in the middle platform area of ​​each mooring post above the steel dock door, and the grabbing part of the mechanical arm is provided with a roller and an angle sensor. The following steps are involved: B1: The captured cable connection end is close to the threaded connection end of the corresponding mooring bitt; B2: Use the improved least square method combined with the data collected by the angle sensor to align the angles of the cable connection end and the threaded connection end of the bollard; B3: The cable connection end aligned by the robot arm is parallel to the threaded connection end of the mooring post. It is translated toward the connection port and pressed against it, and then the roller is controlled to rotate, thereby driving the cable connection end to rotate in a specific direction, and the thread is rotated and inserted into the mooring post for connection.

8. The hinged opening and closing system for the steel dock door of the immersed tube tunnel dry dock according to claim 1 is characterized in that: The sensor data is obtained by remotely controlling the winch, the cable, the steel dock door and the sensors on the drone, and after processing, the analysis of each state of the steel dock door is completed and the control instructions are issued, including the following steps: C1: According to the inertial measurement unit, proximity switch and tension sensor set on the remote control winch, the position sensor, angle sensor, water pressure sensor and water flow velocity sensor set at the wading position of the steel dock door, the sensor data of the steel dock door status and cable condition are obtained, and the collected data are cleaned, outliers and noise data are removed, and normalized; C2: Design a neural network model. The input layer receives sensor data as input, selects a multi-layer fully connected neural network as a hidden layer, and the output layer outputs control instructions for the remote control winch to retract and release the cable. C3: Use mean squared error as the loss function to measure the difference between the model's predicted value and the actual value, and use stochastic gradient descent and its variants to optimize the parameters of the neural network to minimize the loss function; C4: Use root mean square error and mean absolute error to evaluate model performance, and adjust the hyperparameters of the neural network through cross-validation method; C5: Deploy the trained model to the actual control system, receive sensor data in real time, and output control instructions for the cable retracting and discharging motor in the remote control winch to achieve opening and closing control of the steel dock door.

9. The hinged opening and closing system for the steel dock door of the immersed tube tunnel dry dock according to claim 1 is characterized in that: It also includes a warning alarm module, which uses sensors to obtain data on cable tension value, cable retraction and release speed, cable retraction and release length, dock door position, dock door angle and dock door water conditions. A safety threshold is set and whether to issue a warning alarm is determined based on the threshold range comparison.

10. The hinged opening and closing system for the steel dock door of the immersed tube tunnel dry dock according to claim 3 is characterized in that: The drone G is used to determine the risks and provide control instructions for the steel dock door during the process of opening, docking and closing, and includes the following steps: D1: The industrial camera of UAV G is used to collect the motion image of the steel dock door in real time in a specific area at high altitude, and the image is pre-processed by denoising and enhancing. D2: Use the deep learning YOLOv5 target detection algorithm to detect the steel dock door. In the detection stage, the preprocessed image is input into the trained YOLOv5 model, and the model outputs the bounding box coordinates of the steel dock door and the corresponding category confidence. Based on the detection results, the DeepSORT target tracking algorithm is used to track the steel dock door. For each frame of the image, the DeepSORT algorithm outputs the tracking ID of the steel dock door and its position in the image. D3: Determine the normal motion trajectory of the steel dock door, calculate the vertical distance to the normal motion trajectory, and if it is greater than the preset position deviation threshold, it is judged that there is a position deviation risk; calculate the instantaneous speed of the steel dock door according to the position and time interval of the steel dock door in two adjacent frames of images. If the instantaneous speed is greater than the preset minimum or maximum speed threshold, it is judged that there is a speed abnormality risk; perform edge detection on the detected steel dock door area, calculate the perimeter and area of ​​the edge contour and define the shape coefficient of the steel dock door. If the change of the shape coefficient exceeds a certain threshold in multiple consecutive frames of images, it is judged that there is a shape change risk; D4: If there is a risk of position deviation, calculate the length of the cable that needs to be adjusted according to the direction and size of the deviation; according to the safety threshold set for the speed of change of the cable length, when the speed of change of the cable length exceeds the safety threshold, adjust it by reducing the speed of the motor driving the cable or increasing the speed of the motor; if an abnormal movement of the steel dock door is detected, stop the movement of the steel dock door immediately, and send an alarm to notify the maintenance personnel to check the structure of the steel dock door and the tension of the cable; D5: The generated control suggestions are converted into control instructions and sent to the control system of the steel dock door through the communication protocol. After receiving the instructions, the control system adjusts the corresponding motor speed or cable length to adjust the motion state of the steel dock door; the camera continuously collects the motion image of the steel dock door, repeats the target detection, tracking, risk judgment and control suggestion generation process, forms a closed-loop feedback control system, and continuously optimizes the control strategy based on the real-time feedback results.

Images