Wall-climbing robot device carrying ultrasonic nondestructive testing system

Abstract

The invention provides a wall-climbing robot device carrying an ultrasonic nondestructive testing system, and relates to the technical field of testing equipment, and the wall-climbing robot device is technically characterized by comprising a main body frame, a power supply system, a magnetic chuck, a power device, a clamping device and an ultrasonic acquisition device; wherein a transmitting probe in the ultrasonic wave collecting device transmits ultrasonic waves, and an ultrasonic wave receiving probe receives transmitted sound signals. Ultrasonic signals of ship surface material samples which are not subjected to rust removal and subjected to ultrahigh-pressure water jet rust removal are collected respectively, a deep learning model is established and trained, and intelligent nondestructive testing of the ship surface rust removal effect is achieved on the basis of the deep learning model. According to the wall-climbing robot device, stable adsorption and flexible movement on the surface of a complex ship are achieved, and the detection efficiency and reliability are remarkably improved by combining an ultrasonic nondestructive detection system; meanwhile, the identification method based on the LSTM deep learning model solves the problems of low efficiency and high misjudgment rate of a traditional detection method through high-precision signal processing and a cross validation strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of detection equipment, and particularly relates to a wall-climbing robot device equipped with an ultrasonic non-destructive testing system. Background Art

[0002] With the advancement of economic globalization, ships play an irreplaceable role in many fields such as global trade, transportation, and ocean development. A large amount of goods rely on ships for cross-sea and cross-ocean transportation, making ships an important artery for the global economic operation. At present, ship enterprises have carried out research on ultra-high pressure water rust removal technology, designed a complete process plan for intelligent ultra-high pressure pure water jet spraying rust removal and new surface treatment recycling, developed ultra-high pressure pump units and ultra-high pressure water rust removal integrated systems, formulated technical standards for ultra-high pressure water jet ship rust removal, and completely replaced the traditional and backward sandblasting process with pure water jet spraying rust removal, realizing the demonstration application of ultra-high pressure water rust removal complete sets of equipment. On the basis of ultra-high pressure water jet ship rust removal, it is necessary to develop intelligent detection technology to evaluate the effect of ship rust removal. Summary of the Invention

[0003] The purpose of the present invention is to solve the problem of the lack of intelligent detection technology for ship rust removal in the prior art.

[0004] To achieve the above purpose, the present invention adopts the following technical solutions:

[0005] A wall-climbing robot device equipped with an ultrasonic non-destructive testing system, including a main body frame, a power supply system, a magnetic suction cup, a power device, a clamping device, and an ultrasonic acquisition device;

[0006] The main body frame is spliced into a square structure by transverse rods and longitudinal rods. The transverse rods include short rods and long rods, the longitudinal rod is a middle rod. The short rod is located at one end of the power supply system, the long rod is located at the other end of the main body frame. The two ends of the middle rod are respectively connected to the short rod and the long rod. The short rod and the long rod are respectively connected to the middle rod through connecting angle plates;

[0007] The power supply system is fixed on the outside of the short rod, and the magnetic suction cup is fixed on the inside of the main body frame; the power devices are symmetrically arranged at both ends of the long rod;

[0008] The clamping devices are symmetrically arranged on both sides of the midline of the long rod and are located inside the power devices. The ultrasonic acquisition device is installed below the clamping device.

[0009] Preferably, the connecting angle plate connecting the middle rod and the short rod is located inside the frame, and the angle plate connecting the middle rod and the long rod is located outside the frame;

[0010] An outer connecting angle plate is provided at the middle position of the middle rod, on which an ultra-short rod is fixed, and the end of the ultra-short rod is coplanar with the long rod and connected to the power device.

[0011] Preferably, the power supply system includes:

[0012] An L-shaped power supply bracket, the vertical side of which is fixed to the outside of the short rod, and the horizontal side extends outward and is located at the bottom of the vertical side;

[0013] Three groups of power supply support rods, which are respectively the first support rod, the second support rod and the third support rod. Among them, the first support rod is vertically fixed on the upper side of the short rod, the second support rod is vertically fixed on the first support rod and extends outward, and the third support rod is parallel to the short rod and connects the two second support rods;

[0014] An electric box is placed between the second support rod and the horizontal side of the power supply bracket.

[0015] Preferably, two groups of fixing rods extend respectively on the upper and lower sides of the magnetic chuck. Each group of the fixing rods is provided with four. The ends of the two groups of fixing rods are fixedly connected to the upper and lower sides of the middle rod through connecting rods.

[0016] Preferably, the power device includes a driving wheel, a driven wheel, a wheel shaft fixing member, a motor outer box and a crawler, wherein:

[0017] The driving wheel is installed at the end of the long rod through the wheel shaft fixing member, and the driven wheel is installed at the end of the ultra-short rod through the wheel shaft fixing member;

[0018] The motor outer box is fixed to the outside of the long rod to drive the driving wheel to rotate, and the crawler is sleeved on the outer circumferences of the driving wheel and the driven wheel.

[0019] This application also provides a ship surface rust removal recognition method based on ultrasonic non-destructive testing, including the following steps:

[0020] S1: Excite ultrasonic Rayleigh waves on the surfaces of specimens without rust removal and with ultra-high pressure water jet rust removal, and receive the transmitted signals;

[0021] S2: Collect 200 groups of rust removal and non-rust removal signal samples respectively, and each sample is 1 signal segment;

[0022] S3: Construct an LSTM deep learning model, and the input gate, forget gate and output gate control the information flow through the sigmoid function;

[0023] S4: Use 10-fold cross-validation to train the model, and use the classification accuracy of the test set as the recognition rate.

[0024] Preferably, a 1 MHz sine tone burst with a Hanning window is used, with a peak-to-peak value of 10 Vp-p and amplified 40 times; the angle of the plexiglass inclined block is 64.5°, and the pressure between the transducer and the specimen is maintained at 75 N - 80 N.

[0025] Preferably, the update formula for the memory unit of the LSTM model is:

[0026] c t = f t ⊙ c t-1 + i t ⊙ tanh(W xc x t + W hc h t-1 + b c

[0027] In the above formula, f t is the forget gate; i t is the input gate, and c t-1 is the value of the memory unit at the previous moment, and the weight matrix and bias term are determined through training.

[0028] Preferably, the 10-fold cross-validation is specifically as follows:

[0029] Divide 200 samples of each class into 10 folds, with 20 samples in each fold;

[0030] Each time, take 1 fold as the test set, and the remaining 9 folds as the training set, and repeat 10 times to take the average recognition rate.

[0031] Preferably, the parameters of the ultra-high pressure water jet rust removal include: the target distance is 25 mm, the pressure is 100 MPa, the diversion device uses a rotary nozzle, the rotary nozzle consists of 4 nozzles with a pore diameter of 0.3 mm, the erosion speed is 240 mm / min, and the flushing time is 1 minute and 20 seconds.

[0032] Beneficial effects:

[0033] 1. Improvement in structural stability and adaptability: The main frame adopts a combined design of short rods, long rods and connecting angle plates, combined with the U-shaped connection structure of the magnetic chuck and the T-shaped wheel shaft fixing part of the power device, effectively dispersing stress and enhancing the anti-deformation ability, ensuring that the robot adsorbs firmly and moves smoothly on the ship surface; the modular layout of the power system through bolt connection and layered rack design takes into account both the installation convenience and the power supply stability.

[0034] 2. Optimization of Detection Accuracy and Adaptive Ability: The clamping device adopts a hinge structure and a spring clamping plate, combined with an adjustable ultrasonic probe, which can adapt to different surface curvatures and buffer the impact force to ensure the accuracy of ultrasonic signal acquisition. In addition, by optimizing the ultrasonic excitation parameters (peak-to-peak value of 10Vp-p, inclined block angle of 64.5°) and pressure control (75N - 80N), signal interference is further reduced and the detection sensitivity is improved.

[0035] 3. Enhancement of Model Efficiency and Robustness: The LSTM model dynamically updates the memory unit through the forget gate and input gate, combined with a 10-fold cross-validation strategy, effectively avoiding overfitting and enhancing the generalization ability of the model. And by setting the ultra-high pressure water jet parameters (target distance of 25mm, pressure of 100MPa, etc.), the erosion conditions are standardized to provide a high-quality data basis for model training, and finally a 100% recognition accuracy is achieved.

[0036] 4. Engineering Applicability and Maintenance Convenience: In the overall design, bolt-nut connections, modular motor outer boxes, and detachable power brackets significantly reduce manufacturing and maintenance costs. The combined design of the rotating nozzle and the motor-controlled cross-movement speed makes the erosion operation range controllable and the efficiency improved, meeting the engineering requirements of ship detection and maintenance. Brief Description of the Drawings

[0037] Figure 1 It is a schematic diagram of the overall structure of the wall-climbing robot device equipped with an ultrasonic non-destructive testing system in an embodiment of the present invention;

[0038] Figure 2 It is a schematic diagram of the overall structure of the wall-climbing robot device equipped with an ultrasonic non-destructive testing system in another angle in an embodiment of the present invention;

[0039] Figure 3 It is a schematic diagram of the power system of the wall-climbing robot device equipped with an ultrasonic non-destructive testing system in an embodiment of the present invention;

[0040] Figure 4 It is a schematic diagram of the overall structure of the clamping device of the wall-climbing robot device equipped with an ultrasonic non-destructive testing system in an embodiment of the present invention;

[0041] Figure 5 It is a schematic diagram of another angle of the clamping device of the wall-climbing robot device equipped with an ultrasonic non-destructive testing system in an embodiment of the present invention, aiming to show the ultrasonic acquisition device;

[0042] Figure 6 It is a diagram showing the primary ultrasonic signal of an untreated rust-removed ship surface specimen in an embodiment of the present invention;

[0043] Figure 7This is a display diagram of the primary ultrasonic signal for the ultra-high pressure water jet rust removal of the ship surface sample in an embodiment of the present invention.

[0044] Legend:

[0045] 1. Main frame; 11. Short rod; 12. Long rod; 13. Medium rod; 14. Connecting angle plate; 15. Ultra-short rod; 2. Power supply system; 21. Power supply bracket; 22. Power supply pole; 221. First pole; 2211. Angle seat; 222. Second pole; 2221. Connector; 223. Third pole; 3. Magnetic chuck; 31. Fixed rod; 32. Connecting rod; 4. Power device; 41. Crawler; 42. Driving wheel; 43. Driven wheel; 44. Motor outer box; 45. Wheel shaft fixing part; 451. T-shaped plate; 452. Wheel shaft rotating shaft; 5. Clamping device; 51. Top plate; 52. Guide rail fixing plate; 521. Longitudinal steering projection; 522. Longitudinal rotating shaft; 53. Hinge structure; 531. Rotating plate; 5311. Limit block; 5312. Transverse steering projection; 5313. Transverse rotating shaft; 532. Rotating block; 54. Spring; 55. Clamping plate; 551. Fixed bolt; 6. Ultrasonic acquisition device. Detailed implementation manners

[0046] The present invention will be further described in detail below in conjunction with specific embodiments.

[0047] Please refer to Figure 1 and Figure 2 , a wall-climbing robot device equipped with an ultrasonic non-destructive testing system, including a main frame 1, a power supply system 2, a magnetic chuck 3, a power device 4, a clamping device 5 and an ultrasonic acquisition device 6. The main frame 1 is arranged as a square frame. The power supply system 2 and the clamping device 5 are located at both ends of the main frame 1 in the traveling direction. The power supply system 2 is used to provide current and control for the magnetic chuck 3, the power device 4 and the ultrasonic acquisition device 6. The clamping device 5 is used to fix the ultrasonic acquisition device 6. The magnetic chuck 3 is located at the middle position of the main frame 1. The power devices 4 are located on both sides of the main frame 1 in the traveling direction. The power devices 4 are used to drive the movement of the wall-climbing robot.

[0048] Please refer to Figure 1 and Figure 2 , the main frame 1 is constructed by splicing various profiles of different specifications. In one embodiment, the main frame 1 is formed into a square frame by transverse rods and longitudinal rods. There are two transverse rods and two longitudinal rods respectively. The transverse rods are located at both ends of the wall-climbing robot in the traveling direction, and the longitudinal rods extend along the traveling direction of the wall-climbing robot.

[0049] In one embodiment, one of the transverse rods is a short rod 11, and the other transverse rod is a long rod 12, the short rod 11 is located at one end of the main frame 1 pointing to the power system 2, and the two longitudinal rods are middle rods 13, one end of the two middle rods 13 is connected to the end of the short rod 11, and the other end is fixed to the long rod 12. In the present application, the short rod 11, the middle rod 13 and the long rod 12 are only different in length, and their materials are the same.

[0050] The connection between the longitudinal rod and the transverse rod is fixed by a connecting angle plate 14. In one embodiment, the connecting angle plate 14 connecting the middle rod 13 and the short rod 11 is located on the inner side of the square formed by the main frame 1, and the connecting angle plate 14 connecting the long rod 12 and the middle rod 13 is located on the outer side of the main frame 1; a connecting angle plate 14 is also provided at the middle position of the two middle rods 13, and the connecting angle plate 14 is located on the outer side of the main frame 1 and is symmetrically arranged with the connecting angle plate 14 connecting the long rod 12 and the middle rod 13. An ultra-short rod 1511 is also provided on the connecting angle plate 14 located at the middle position of the middle rod 13, and the ultra-short rod 1511 is fixedly connected to the middle rod 13 through the connecting angle plate 14, and the outer end surface of the ultra-short rod 1511 is arranged coplanar with the outer end surface of the long rod 12.

[0051] In the present application, the long rod 12 is provided to provide the main frame 1 with the necessary length extension and stable longitudinal structural foundation. The shape and stability of the frame are precisely adjusted at a specific position through the cooperation between the short rod 11 and the long rod 12. The long rod 12 is evenly distributed in the main frame 1 to further enhance the lateral stability and overall structural strength of the main frame 1, so that the main frame 1 can maintain a good state when subjected to lateral forces. The ultra-short rod 1511 is used to strengthen and improve the local structure of the main frame 1 to ensure the stability of each connection part. In one embodiment, the ultra-short rod 1511 and the end of the long rod 12 are used to connect with the power device 4.

[0052] In addition, in the present application, each profile is tightly and firmly connected by connecting the angle plate 14. The connecting angle plate 14 can not only effectively and tightly fix each profile, but also disperse the stress to a certain extent, thereby improving the deformation resistance of the entire frame.

[0053] See also Figure 1 and Figure 3, the power supply system 2 includes a power supply bracket 21 and a power supply pole 22. The power supply bracket 21 is provided as an L-shaped plate. The vertical side of the power supply bracket 21 is connected to the outer side of the short rod 11 and is directly fixed by bolts and nuts. The horizontal side of the power supply bracket 21 is located below its vertical side and extends away from the main body frame 1. Through the fixation of bolts and nuts, on the one hand, it provides a strong fastening force to ensure a tight connection between the power supply bracket 21 and the main body frame 1. During the operation of the robot, even under the impact and vibration of various complex external forces, it can ensure that the power supply bracket 21 always firmly adheres to the main body frame 1 and will not easily loosen or shift; on the other hand, the fixation method of bolts and nuts has good versatility and operability, which is convenient for the staff to install, disassemble and adjust during both the production and manufacturing stage and the later maintenance and repair process.

[0054] There are three groups of the power supply poles 22. The three groups of the power supply poles 22 are respectively a first pole 221, a second pole 222 and a third pole 223. There are two of the first poles 221 and two of the second poles 222 respectively. The two first poles 221 are fixed on the upper side of the short rod 11 and are perpendicular to the upper side of the short rod 11. In an embodiment, the first pole 221 is fixed on the upper side of the short rod 11 through an angle seat 2211. Through the design of the angle seat 2211, various forces borne by the first pole 221 during operation can be effectively dispersed, including gravity, inertia force, and impact force generated during the movement of the robot, etc.

[0055] The two second poles 222 are respectively fixed on the two first poles 221 through connectors 2221. The two second poles 222 are perpendicular to the outer side of the short rod 11 and extend away from the main body frame 1. An electrical box is placed between the second pole 222 and the horizontal side of the power supply bracket 21.

[0056] There is one third pole 223. The third pole 223 is used to connect the two second poles 222. In an embodiment, the third pole 223 is located at the end faces of the two second poles 222 away from the first pole 221. The third pole 223 is arranged parallel to the short rod 11. The third pole 223 and the second pole 222 are connected and fixed through a connector 2221. In an embodiment, the connector 2221 is provided in an L shape, and the two sides of the L shape are respectively connected to two different power supply poles 22 to achieve the connection and fixation of each power supply pole 22.

[0057] Through the setting of the connecting piece 2221, not only is the connection between the power rack poles 22 ensured to be tight and firm, preventing loosening or disconnection during the movement of the robot, but it also has a certain flexibility and adaptability to cope with the slight deformation and relative displacement of various components of the robot under different movement postures. Through the connection of the connecting piece 2221, the power rack poles 22 can work together, further optimizing the structural layout of the electrical box placement device, so that the entire device can meet the power installation requirements while better adapting to the integrated movement requirements of the robot.

[0058] The third frame rod 223 is connected to the third frame rod 223 by the connecting piece 2221, so that the force borne on the third frame rod 223 can be evenly dispersed in the entire supporting structure, so that the power frame rods 22 together build a hierarchical, compact, stable and reliable support system, which can accurately locate and firmly support the power supply, ensure that the power supply always remains stable during the integrated movement of the robot, avoid the normal power supply of the power supply due to shaking, bumps or collisions, and thus ensure the stable operation of the robot.

[0059] In order to ensure that the robot can be reliably adsorbed on the surface of the ship and meet the needs of flexible movement and continuous work in different parts of the ship, a key component, the magnetic chuck 3, is designed. The magnetic chuck 3 plays a pivotal role in the operation of the entire robot. It is the core device to achieve a close connection between the robot and the surface of the ship.

[0060] In one embodiment, see Figure 1 and Figure 2 In order to ensure that the magnetic suction cup 3 can be firmly integrated into the overall structure of the robot, a fixing rod 31 is provided on the magnetic suction cup 3 to extend outwardly around it. The fixing rod 31 is divided into two groups. The two groups of fixing rods 31 are respectively located on the upper and lower sides of the magnetic suction cup 3. A connecting rod 32 is provided at one end of the fixing rod 31 away from the magnetic suction cup 3. The connecting rod 32 is located on the side of the fixing rod 31 away from each other. The connecting rod 32 is perpendicular to the middle rod 13 and fixed to the upper and lower sides of the middle rod 13. The two ends of the connecting rod are respectively fixed to the fixing rod 31 and the middle rod 13 by bolts.

[0061] In this way, the combination of the fixed rod 31 and the connecting rod 32 forms a U-shaped connection, enabling the magnetic chuck 3 to be connected to the main body frame 1. The U-shaped connection can tightly surround and hold the connection part between the magnetic chuck 3 and the main body frame 1, providing a large contact area and stable support. This connection method not only ensures the firmness of the connection between the magnetic chuck 3 and the main body frame 1, but also effectively buffers the vibration and impact generated due to the unevenness of the ship's surface or the movement of the robot during the operation of the robot, preventing the loosening of the connection between the magnetic chuck 3 and the main body frame 1 and ensuring the stability of the entire adsorption system.

[0062] A magnet (not shown in the figure) is provided at the bottom of the magnetic chuck 3. When the robot approaches the ship's surface, a strong magnetic attraction force will be generated between the magnet and the ship. This magnetic attraction force can overcome the gravity of the robot itself and various external force interferences during the operation process, firmly adsorbing the robot on the ship's surface. Whether it is on the horizontal ship deck, the vertical ship's side, or even some complex curved parts of the ship, the magnetic attraction force generated by the magnet can ensure that the robot is stably attached, enabling the robot to crawl freely on the ship's surface like a gecko, providing a solid guarantee for it to successfully complete various tasks such as ship inspection, rust removal, and painting.

[0063] The design of the robot for ship operations needs to meet the complexity and particularity of the ship surface operation environment. In order to enable the robot to have a powerful and reliable wall-climbing ability to successfully complete the inspection task, please refer to Figure 1 Two sets of the power devices 4 are provided, and the two sets of the power devices 4 are symmetrically arranged along the midline of the long rod 12.

[0064] Each set of the power devices 4 respectively includes a crawler 41, a driving wheel 42, a driven wheel 43, a motor outer box 44, and a wheel shaft fixing part 45; wherein, the crawler 41, the driving wheel 42, and the driven wheel 43 are located outside the middle rod 13, and the motor outer box 44 is fixed outside the long rod 12 for driving the driving wheel 42. By installing the motor outer box 44 at the front position of the robot, on the one hand, it is beneficial to optimize the center of gravity distribution of the robot, making the robot more balanced and stable during the climbing process; on the other hand, it is convenient for the motor to be efficiently connected to the subsequent transmission components, reducing the energy loss during the power transmission process.

[0065] Moreover, the outer motor case 44 is directly connected to the long rod 12, which provides stable support and reliable protection for the motor. The outer motor case 44 can not only prevent the motor from being eroded by external environmental factors such as dust and water vapor, but also buffer the vibration generated during the operation of the motor to a certain extent, avoiding the impact of vibration on other components of the robot. At the same time, the direct connection between the outer motor case 44 and the frame enables the power generated by the motor to be transmitted to the robot's walking system efficiently and directly, reducing energy loss and transmission error in the intermediate links and improving the power transmission efficiency.

[0066] The wheel shaft fixing member 45 is fixed to the end portions of the ultra-short rod 1511 and the long rod 12. There are two wheel shaft fixing members 45 provided on each power device 4. The two wheel shaft fixing members 45 are respectively used to fix the driving wheel 42 and the driven wheel 43. In one embodiment, the wheel shaft fixing member 45 includes a T-shaped plate 451 and a wheel shaft rotating shaft 452. There are two T-shaped plates 451 provided on each wheel shaft fixing member 45. The horizontal sections of the T-shaped plates 451 on the two wheel shaft fixing members 45 are respectively fixed to the opposite sides of the ultra-short rod 1511 and the long rod 12. And the vertical sections of the T-shaped plates 451 on the two wheel shaft fixing members 45 extend in the direction away from each other. The wheel shaft rotating shaft 452 is fixed between the two T-shaped plates 451 of the same wheel shaft fixing member 45. The driving wheel 42 and the driven wheel 43 are sleeved on the wheel shaft rotating shaft 452 and are respectively located between the two T-shaped plates 451 of the two wheel shaft fixing members 45. The crawler 41 is sleeved on the outer peripheries of the driving wheel 42 and the driven wheel 43.

[0067] Through the arrangement of the wheel shaft fixing member 45, not only can the wheel shafts (the driving wheel 42 and the driven wheel 43) be accurately positioned and fixed to ensure that the wheel shafts maintain a stable position during the operation of the robot, but also it can bear various forces transmitted from the wheel shafts and the crawler 41, including gravity, friction force, driving force, etc., to ensure the structural integrity and reliability of the robot's walking system. In addition, the design and manufacture of the wheel shaft fixing member 45 strictly follow high-precision standards to ensure the fitting accuracy between it and the wheel shafts, reducing abnormal wear or jamming phenomena caused by too large or too small gaps, thereby extending the service life of the entire walking system.

[0068] The driving wheel 42 is located at the end of the long rod 12. A motor (not shown in the figure) is provided in the outer motor case 44, and the driving wheel 42 is driven by the motor.

[0069] When the robot starts to work, the motor is powered on and starts. The rotor of the motor rotates at high speed, driving the driving wheel 42 connected to it to rotate. There is a tight meshing between the driving wheel 42 and the crawler 41 through a specific tooth shape or pattern structure. The rotation of the driving wheel 42 is converted into the continuous movement of the crawler 41 through this meshing relationship. Due to the frictional force between the crawler 41 and the ship surface, as the crawler 41 continuously rolls, the robot can steadily move forward along the ship wall under the push of this frictional force, thus realizing the autonomous walking of the robot on the ship.

[0070] Please refer to Figure 4 and Figure 5 The clamping devices 5 are provided in two. The two clamping devices 5 are located between the outer boxes 44 of the motor and are symmetrically arranged along the midline of the long rod 12.

[0071] In one embodiment, the two clamping devices 5 respectively include a top plate 51, a guide rail fixing plate 52, a hinge structure 53, a spring 54 and a clamping plate 55. Among them, the top plate 51 is used to connect the clamping device 5 and the long rod 12. The guide rail fixing plate 52 is located between the top plate 51 and the hinge structure 53 and is used to fix the guide rail. The hinge structure 53 is used to realize the rotation between various components. The clamping plate 55 is located below the hinge structure 53. There are two clamping plates 55, and the spring 54 is located between the two clamping plates 55.

[0072] In one embodiment, the top plate 51 is L-shaped. The vertical side of the top plate 51 is fixed on the long rod 12 by bolts. The horizontal side is above the vertical side and perpendicular to the outer side of the long rod 12. The guide rail fixing plate 52, the hinge device, the spring 54 and the clamping plate 55 are all located below the horizontal side of the top plate 51. In one embodiment, the top plate 51 is made of a high-strength and lightweight material to ensure that while ensuring the structural strength, it will not add too much burden to the entire robot.

[0073] The guide rail fixing plate 52 is located below the horizontal side of the top plate 51 and is used to connect the top plate 51 and the hinge device. In one embodiment, two longitudinal turning protrusions 521 extend downward at one end of the guide rail fixing plate 52 pointing away from the horizontal side of the top plate 51, and a longitudinal rotating shaft 522 is arranged between the longitudinal turning protrusions 521.

[0074] The hinge device includes a rotating plate 531 and a rotating block 532. Among them, the rotating plate 531 rotates around the longitudinal rotating shaft 522 between the longitudinal turning protrusions 521. The rotating block 532 is located below the rotating plate 531. A limiting block 5311 is provided above the rotating plate 531. The limiting plate is arranged parallel to the two longitudinal turning protrusions 521 and is located between the two longitudinal turning protrusions 521. The limiting plate is also sleeved between the longitudinal rotating shafts 522. Through the arrangement of the limiting plate and the longitudinal rotating shaft 522, the longitudinal rotation of the clamping device 5 is realized. A transverse turning protrusion 5312 is provided below the rotating plate 531. There are two transverse turning protrusions 5312 on each rotating plate 531. The projections of the two transverse turning protrusions 5312 in the vertical direction are perpendicular to the projection of the longitudinal protrusion in the vertical direction. A transverse rotating shaft 5313 is also provided between the two transverse turning protrusions 5312. The rotating block 532 is located between the two transverse turning protrusions 5312 and is rotatably arranged around the transverse rotating shaft 5313. The lower part of the rotating block 532 is connected to one of the clamping plates 55. Through the arrangement of the hinge device, relative rotation can be achieved between various components, so that the clamping mechanism can be adaptively adjusted according to the shape and position of the object to be clamped. This adaptability greatly improves the versatility and applicability of the clamping mechanism, enabling it to cope with various complex working scenarios.

[0075] There are two clamping plates 55, and the two clamping plates 55 are arranged parallel to each other. Among them, the upper clamping plate 55 is fixedly arranged with the rotating block 532. Fixing bolts 551 pointing to the lower clamping plate 55 are arranged at the four corners of the upper clamping plate 55. The fixing bolts 551 pass through the lower clamping plate 55 and are fixed by nuts. A spring 54 is sleeved outside the fixing bolts 551 and is located between the two clamping plates 55. The maximum distance between the two clamping plates 55 is limited by the fixing bolts 551 and the nuts. Through the arrangement of the spring 54, the necessary elastic force is provided during the clamping process. When clamping an object, the spring 54 can automatically adjust the clamping force according to the size and shape of the object, ensuring that the object can be firmly clamped without being damaged due to excessive clamping force. At the same time, the elasticity of the spring 54 can also buffer the impact force during the clamping process to a certain extent, protecting the object to be clamped and the clamping mechanism itself. In one embodiment, the components of the upper and lower clamping plates 55 that directly contact the object to be clamped usually undergo special treatment to increase friction and improve the stability of clamping. Through the coordinated action of the above-mentioned various components of the upper and lower clamping plates 55, flexible opening and closing actions can be realized, so as to complete the clamping task of different objects.

[0076] Please refer to Figure 5, the ultrasonic acquisition device 6 is arranged below the clamping device 5. The ultrasonic acquisition device 6 includes two ultrasonic probes, which are respectively a transmitting probe and a receiving probe. Among them, the transmitting probe is used to emit ultrasonic waves to detect the surface of the ship, and the receiving probe receives the transmitted acoustic signal, and judges whether there is damage to the ship at that place by analyzing the acoustic signal. For the convenience of installation, the ultrasonic probe uses a threaded connection.

[0077] In the field of ship rust removal, ultrasonic testing, as a crucial and effective non-destructive testing method, is a method of detecting material defects by utilizing the characteristics of ultrasonic waves propagating in materials. Ultrasonic waves, as a mechanical wave with a frequency higher than the upper limit of human hearing, have many unique physical properties, which lay the foundation for ultrasonic testing technology. Its detection principle is based on the reflection, refraction, etc. of ultrasonic waves in materials. When ultrasonic waves enter the material to be detected at a specific angle and frequency, they propagate inside the material. The transmission method uses a transmitting and receiving dual-probe placed on both sides of the object to be detected, and the material is detected according to the change in acoustic wave energy after penetrating the object to be detected. When ultrasonic waves propagate in ship materials and encounter the surface of the ship without rust removal or the surface of ship rust removal by high-pressure water jet, a part of the ultrasonic waves will be affected, and thus a difference in acoustic wave energy can be detected in the received signal.

[0078] The wave equation of ultrasonic Rayleigh waves in an isotropic medium can be expressed as

[0079]

[0080] where λ and μ are Lame constants, u is the displacement vector, and ρ is the density.

[0081] The stress-free boundary condition on the surface can be written as

[0082]

[0083] where x is the propagation direction of the ultrasonic Rayleigh wave, z is the direction perpendicular to the surface, and σ zz and σ xz are the normal and tangential stresses and are the normalized amplitudes of the displacement components in the x and z directions respectively, equal to

[0084]

[0085] where k is the wave number of the Rayleigh wave, and q, s, and r are equal to and c T and c R are the velocities of the shear wave and the Rayleigh wave respectively.

[0086] In addition, the tangential stress the normal stress and the amplitude of and are normalized by and they are respectively equal to

[0087]

[0088] It is found that most of the energy of Rayleigh waves is in the region near the surface, and its depth is less than one wavelength. Therefore, Rayleigh waves show higher sensitivity to surface and subsurface defects and damages.

[0089] c R The expression of can be written as

[0090]

[0091] where ν is the Poisson's ratio, and ξ and Δ are respectively equal to c T / c L and

[0092] (64 / 27)(11 - 62ξ 2 + 107ξ 4 - 64ξ 6 ).

[0093] Rayleigh waves are usually excited by an acrylic wedge, and its angle can be determined by Snell's law

[0094]

[0095] where c R and are respectively the velocities of Rayleigh waves in the medium and longitudinal waves in the acrylic wedge.

[0096] The ultrasonic signals of specimens without derusting and after ultra-high pressure water jet derusting are collected, and each collected ultrasonic signal segment is taken as a sample. On the basis of collecting a certain number of samples, a deep learning model is built and trained using a long short-term memory network (LSTM). The LSTM model uses three components: an input gate, a forget gate, and an output gate. The output h t- of the previous recurrent layer is used to calculate the output h t of the next recurrent layer. The memory cell of the LSTM model maintains a memory value c t at time t, and the output value calculation formula of the recurrent layer state is

[0097] h t = o t ⊙ tanh(c t )

[0098] Among them, o t is the output gate, which is a vector and is calculated according to the following formula;

[0099] o t = σ(W xo x t + W ho h t-1 + b o )

[0100] Among them, σ is the sigmoid function. The output gate determines what proportion of the memory value stored in the memory cell can be output. The sigmoid function is used because its value range is (0, 1), so that the values of all components of o t are in the range of 0 to 1, and they are multiplied by the components of another vector respectively, which can control the output proportion of another vector. W xo , W ho , b o are the weight matrix and bias term of the output gate, and these parameters are obtained through training.

[0101] The memory value c t is the state value of the previous moment remembered by the neurons in the recurrent layer. It is weighted and updated over time, and its update formula is

[0102] c t = f t ⊙ c t-1 + i t ⊙ tanh(W xc x t + W hc h t-1 + b c

[0103] Among them, f t is the forget gate; c t-1 is the value of the memory cell at the previous moment. The forget gate determines the situation where the value of the memory cell at the previous moment is passed to the current moment. The current value of the memory cell is the weighted sum of the value at the previous moment and the current input value, and the memory value is just an intermediate value. The calculation formula of the forget gate is f t = σ(W xf x t + W hf h t-1 + b f )

[0104] i t is the input gate, which controls how much of the current input can enter the memory cell, and its calculation formula

[0105] i t = σ(W xi xt +W hi h t-1 +b i )

[0106] The calculation formulas for these three gates are the same. Each uses its own weight matrix and bias vector, and the calculation of these three values all involves x t and h t-1 , which play a role in controlling the information flow.

[0107] The state value of the hidden layer is jointly determined by the forget gate, the value of the memory cell at the previous moment, as well as the input gate and the output gate. Apart from the three gates, what truly determines h t is only x t and h t-1 . To sum up, the calculation idea of LSTM is as follows: The input gate acts on the input value at the current moment, and the forget gate acts on the previous memory value. The weighted sum of the two obtains the aggregated information; finally, the output gate determines the output value. During the process of detecting the rust removal of ships by ultra-high pressure water jets, the ultrasonic detection device moves on the ship surface and collects signals, and the effect of ultra-high pressure water jet ship rust removal is evaluated through the output result of the model for the signals.

[0108] The following are the specific application implementations:

[0109] As the core equipment of the entire experiment, the ultra-high pressure water jet platform needs to be preheated before being officially put into use. This is because the equipment has been in a stationary state for a long time, and various components and fluid systems inside it need a certain amount of time to reach a stable working temperature and performance state. The preheating process can gradually increase and stabilize the oil temperature, water temperature, and the temperature of each mechanical component inside the equipment, reduce equipment wear and performance fluctuations caused by temperature differences, and provide a good equipment foundation for subsequent precise operations.

[0110] While the ultra-high pressure water jet platform is being preheated, a fixture needs to be used to clamp the sample block. As the research object of this experiment, the fixing stability of the sample block directly affects the erosion effect and the accuracy of the test data. Ensure that the fixture can provide sufficient clamping force to firmly fix the sample block in the specified position and prevent the sample block from shifting or shaking under the impact of the ultra-high pressure water jet. Only when the sample block is stably clamped can the consistency of the conditions for each erosion test be guaranteed, so that the test results are comparable and reliable.

[0111] The target distance is the distance between the ultra-high pressure water jet nozzle and the surface of the sample block. An appropriate target distance can enable the water jet to maintain the best energy distribution and impact intensity when reaching the surface of the sample block. After the sample block is fixed, adjust the target distance to 25 mm. Next, use an ultra-high pressure piston pump to increase the pressure to 100 MPa. As a key device for generating ultra-high pressure water jets, the pressure boosting ability of the ultra-high pressure piston pump directly determines the impact intensity of the water jet. Precisely boosting the pressure to 100 MPa is obtained through repeated demonstration based on test requirements and material properties. At this pressure value, the water jet can have sufficient energy to erode the sample block, simulating the high-pressure impact situation that may be encountered in the actual engineering environment, so as to more realistically study the performance changes of materials under the action of ultra-high pressure water jets.

[0112] The diversion device selects a rotary nozzle, which consists of 4 nozzles with a diameter of 0.3 mm. This specially designed rotary nozzle has unique advantages. The 4 nozzles with a diameter of 0.3 mm can precisely control the jet direction and flow distribution of the water jet, enabling the water jet to impact the surface of the sample block in a uniform and dense manner. At the same time, the rotation function of the rotary nozzle can expand the erosion range, avoiding excessive local erosion or uneven erosion. The working diameter of the rotary nozzle is 40 mm. This parameter determines the size of the erosion area. An appropriate working diameter can ensure that the surface of the sample block is uniformly eroded within the specified range, providing a strong guarantee for accurately evaluating the erosion performance of materials.

[0113] During the entire erosion process, use a motor to control the transverse movement speed and direction of the nozzle. The precise control of the motor can ensure that the nozzle moves according to the predetermined path and speed, thereby achieving comprehensive and uniform erosion of the surface of the sample block. The erosion speed is set to 240 mm / min. This speed has been optimized through multiple tests. It can not only ensure sufficient action time of the water jet on the sample block but also complete the erosion test within a reasonable time, improving the test efficiency. The one-time flushing time is 1 minute and 20 seconds. This setting is determined by comprehensively considering various factors such as the material properties, the energy of the water jet, and the erosion effect. During this time period, the water jet can have an effective erosion effect on the surface of the sample block, causing obvious changes on the material surface, which is convenient for subsequent analysis of the erosion performance of the material.

[0114] The ultrasonic Rayleigh wave is excited on the left side of the specimens without rust removal and with ultra-high pressure water jet rust removal, and the acoustic signals are received on the right side of the specimens without rust removal and with ultra-high pressure water jet rust removal respectively. The excitation signal is a 1 MHz sinusoidal tone burst with a Hanning window and a period of 10, a peak-to-peak value of 10 Vp-p, and an amplification factor of 40 times. The angle of the Plexiglas inclined block is selected as 64.5°. The inclined block transducers with the same angle are used to excite the ultrasonic Rayleigh wave and receive the transmitted wave. To reduce the volatility of the measured data, in this study, the specimen and the inclined block transducer are fixed with the clamping device 5 and the pressure sensor, and the pressure between the two is maintained between 75 N and 80 N. The Rayleigh wave obtained from the ship surface specimen without rust removal is as Figure 1 shown, and the Rayleigh wave obtained from the specimen with ultra-high pressure water jet ship rust removal is as Figure 2 shown. The Rayleigh wave is collected 200 times for the ship specimen without rust removal, and each 1 signal segment is used as 1 sample, obtaining 200 non-rust removal signal samples. The Rayleigh wave is collected 200 times for the ship specimen with ultra-high pressure water jet rust removal, and each 1 signal end is used as 1 sample, obtaining 200 rust removal signal samples.

[0115] The 10-fold cross-validation method is adopted to evaluate the accuracy of the model for ultrasonic signal classification, that is, 200 signal samples of each type are divided into 10 folds, and each fold has 20 samples. Each time, 1 fold of rust removal signal samples and non-rust removal signal samples are taken as the test set, and the remaining 9 folds of non-rust removal signal samples and rust removal signal samples are taken as the training set. An LSTM deep learning model is established, and the parameters of the model are adjusted through the training set to realize the training of the model. The classification accuracy obtained by the trained model on the test set is used as the recognition rate of whether the ship surface has been rust removed by ultra-high pressure water jet. The average value of the recognition rates obtained on the 10 test sets is used as the final verification result. The experimental results designed in the present invention show that the recognition rate of ultra-high pressure water jet ship rust removal reaches 100%.

[0116] The wall-climbing robot device of the present invention realizes stable adsorption and flexible movement on the complex ship surface through optimizing the structural design and the layout of the power system. Combined with the ultrasonic non-destructive testing system, the detection efficiency and reliability are significantly improved; at the same time, the recognition method based on the LSTM deep learning model enables the recognition accuracy of the rust removal state of the ship surface to reach 100% through high-precision signal processing and cross-validation strategies, solving the problems of low efficiency and high misjudgment rate of the traditional detection methods.

Claims

1. A wall-climbing robot device equipped with an ultrasonic nondestructive testing system, characterized in that: It includes a main frame, a power system, a magnetic chuck, a power device, a clamping device and an ultrasonic collection device; The main frame is spliced ​​into a square structure by transverse rods and longitudinal rods, the transverse rods include short rods and long rods, the longitudinal rods are middle rods, the short rods are located at one end of the power system, the long rods are located at the other end of the main frame, the two ends of the middle rods are respectively connected to the short rods and the long rods, and the short rods and the long rods are respectively connected to the middle rods through connecting angle plates; The power supply system is fixed on the outside of the short rod, and the magnetic chuck is fixed on the inside of the main frame; the power device is symmetrically arranged at both ends of the long rod; The clamping device is symmetrically arranged on both sides of the midline of the long rod and located on the inner side of the power device, and the ultrasonic collecting device is installed below the clamping device.

2. The wall-climbing robot device according to claim 1, characterized in that: The connecting angle plate connecting the middle rod and the short rod is located on the inner side of the frame, and the angle plate connecting the middle rod and the long rod is located on the outer side of the frame; An outer connecting angle plate is provided in the middle of the middle pole, on which an ultra-short pole is fixed, and an end of the ultra-short pole is coplanar with the long pole and connected to the power device.

3. The wall-climbing robot device according to claim 1, characterized in that: The power supply system comprises: An L-shaped power supply bracket, with a vertical side fixed to the outside of the short rod and a transverse side extending outward and located at the bottom of the vertical side; Three groups of power supply racks, the three groups of power supply racks are respectively a first rack, a second rack and a third rack, wherein the first rack is vertically fixed to the upper side of the short rack, the second rack is vertically fixed to the first rack and extends outward, and the third rack is parallel to the short rack and connects the two second racks; An electrical box is placed between the second rack rod and the lateral edge of the power supply bracket.

4. The wall-climbing robot device according to claim 1, characterized in that: Two groups of fixing rods are respectively extended from the upper and lower sides of the magnetic suction cup, and each group of the fixing rods is respectively provided with four fixing rods. The ends of the two groups of fixing rods are fixedly connected to the upper and lower side surfaces of the middle rod through connecting rods.

5. The wall-climbing robot device according to claim 1, characterized in that: The power device includes a driving wheel, a driven wheel, a wheel axle fixing member, a motor outer box and a crawler track, wherein: The driving wheel is mounted on the end of the long rod through a wheel axle fixing piece, and the driven wheel is mounted on the end of the super short rod through a wheel axle fixing piece; The motor outer box is fixed on the outer side of the long rod to drive the driving wheel to rotate, and the crawler belt is sleeved on the outer periphery of the driving wheel and the driven wheel.

6. A method for identifying rust removal on ship surface based on ultrasonic nondestructive testing, characterized in that: The following steps are involved: S1: Excite ultrasonic Rayleigh waves on the surface of the sample without rust removal and the sample with ultra-high pressure water jet rust removal, and receive the transmission signal; S2: Collect 200 groups of un-rusted and rusted signal samples, each sample is a signal segment; S3: Build an LSTM deep learning model, where the input gate, forget gate, and output gate control the information flow through the sigmoid function; S4: The model was trained using 10-fold cross validation, and the classification accuracy of the test set was used as the recognition rate.

7. The identification method according to claim 6, characterized in that: The ultrasonic signal excitation includes: using a 1MHz short sine pure tone with a Hanning window, a peak-to-peak value of 10Vp-p, and an amplification of 40 times; the angle of the organic glass bevel is 64.5°, and the pressure between the transducer and the sample is maintained at 75N-80N.

8. The identification method according to claim 6, characterized in that: The memory unit update formula of the LSTM model is: c t =f t ☉c t-1 +i t ⊙tanh(W xc x t +W hc h t-1 +b c In the above formula, f t is the forget gate; i t is the input gate, c t-1 It is the value of the memory unit at the previous moment, and the weight matrix and bias term are determined through training.

9. The identification method according to claim 6, characterized in that: The 10-fold cross validation is specifically as follows: Divide the 200 samples of each category into 10 folds, with 20 samples in each fold; Each time, 1 fold is taken as the test set and the remaining 9 folds are taken as the training set. This is repeated 10 times and the average recognition rate is taken.

10. The identification method according to claim 6, characterized in that: The parameters of the ultra-high pressure water jet rust removal include: target distance 25mm, pressure 100MPa, a rotary nozzle is selected as the diversion device, the rotary nozzle is composed of 4 nozzles with an aperture of 0.3mm, an erosion speed of 240mm / min, and a flushing time of 1 minute and 20 seconds.

Images