Composite material detection robot control method and system

Abstract

The invention relates to the field of composite material detection, and particularly provides a composite material detection robot control method and system.The composite material detection robot control method comprises the following steps that a control instruction is generated according to a preset scanning track so as to control a damage scanning device of a composite material detection robot to drive an ultrasonic phased array flaw detection device to scan a current composite material structural part; according to scanning data collected by the ultrasonic phased array flaw detection device, whether the current composite material structural part is damaged or not is judged; in the scanning process, whether a collision risk exists or not is predicted in real time according to the environmental data monitored by the sensing device; and when the collision risk exists, a posture adjustment instruction is generated according to the risk information of the obstacle, and the composite material detection robot is controlled to change the current posture. The device and the method are used for carrying out nondestructive testing on the large-size composite structural member.

Description

Technical Field

[0001] This invention relates to the field of composite material testing technology, and in particular to a control method and system for a composite material testing robot. Background Technology

[0002] Composite materials are made by combining two or more materials with different properties through physical or chemical methods to create a material with novel properties on a macroscopic scale. Composite materials are widely used in many fields due to their advantages such as light weight, high strength, corrosion resistance, good thermal stability, and high designability. However, during the production of composite materials, defects may exist on the surface or inside due to production processes, environmental control, and some random factors. During manufacturing, assembly, and service, damage may occur due to machining, external impacts, collisions, and scratches. Therefore, non-destructive testing of composite components is particularly important.

[0003] Currently, non-destructive testing (NDT) of composite structural components commonly employs techniques such as radiographic testing, ultrasonic testing, infrared thermography, acoustic-ultrasonic testing, eddy current testing, microwave testing, and laser holographic imaging. In NDT, a robotic arm typically moves the inspection probe, and software is used for identification, thus enabling the inspection of the composite structural component. For large-sized composite structural components, which have significant surface areas and volumes, and often complex dimensions and shapes, precise control of their motion trajectory and attitude is required to ensure accurate acquisition of surface damage information. However, in practice, the inspection robot may encounter various obstacles and unknown environmental factors, easily leading to collisions that can damage the robot or result in inaccurate scanning results.

[0004] A non-destructive testing device for large composite laminate structures, patent number CN201621234651.8, employs the following technical solution: The non-destructive testing device includes a worktable for placing the large composite material to be tested, with a drive device at the lower end of the worktable capable of rotating it; a robotic arm positioned on one side of the worktable, with a probe at its end for detecting the internal quality of the large composite laminate structure, and a corresponding nozzle at the end of the robotic arm. The nozzle is connected to a coupling water storage device via a water spray pipe, and a power pump on the water spray pipe allows coupling water to be sprayed onto the surface of the large composite material to be tested; and a PLC control system, which is communicatively connected to the drive device, robotic arm, probe, and power pump. This achieves the goal of accurately determining the internal quality of large composite laminate structures and reduces the labor intensity of testing personnel. However, the above invention ignores the potential collision risk. When a collision occurs, the posture and trajectory of the robotic arm will deviate, thereby reducing the accuracy of the robotic arm's operation.

[0005] Therefore, the present invention provides a control method and system for a composite material inspection robot. Summary of the Invention

[0006] This invention provides a control method and system for a composite material inspection robot. By controlling a damage scanning device, the robot uses scanning data collected by an ultrasonic phased array flaw detector to perform non-destructive testing on composite material structural components. It also uses environmental data collected by a sensor device to predict whether there is a collision risk, thus achieving efficient and accurate inspection of composite material structural components.

[0007] In a first aspect, this application provides a control method for a composite material inspection robot, the method comprising the following steps:

[0008] Control commands are generated based on the preset scanning trajectory to control the damage scanning device of the composite material inspection robot to drive the ultrasonic phased array flaw detector to scan the current composite material structure.

[0009] Determine whether there is damage to the current composite structure based on the scanning data collected by the ultrasonic phased array flaw detector;

[0010] During the scanning process, the presence of collision risk is predicted in real time based on the environmental data monitored by the sensing device.

[0011] When there is a risk of collision, the robot generates attitude adjustment instructions based on the risk information of the obstacle, and controls the composite material inspection robot to change its current attitude.

[0012] In one feasible embodiment, the damage scanning device includes a first six-axis robot and a second six-axis robot, which are located on both sides of the composite structure, and both the first six-axis robot and the second six-axis robot are equipped with an ultrasonic phased array flaw detector.

[0013] The steps of controlling the damage scanning device of the composite material inspection robot according to the scanning trajectory specifically include:

[0014] The first scanning command is generated based on the scanning trajectory and sent to the first six-axis robot. The first six-axis robot moves according to the scanning command to inspect the front and side of the composite structure.

[0015] Once the inspection is complete, a flipping command is sent to the power mechanisms located at both ends of the composite structure to drive the composite structure to rotate.

[0016] A second scanning command is generated based on the scanning trajectory and sent to the second six-axis robot. The second six-axis robot moves according to the scanning command to detect the reverse side and the opposite side of the composite structure, thus completing the scan.

[0017] In one implementable embodiment, prior to the step of generating control commands based on a preset scan trajectory, the method further includes the step of:

[0018] Generate a transport instruction and send it to the loading robotic arm to move the composite structural component to the inspection area;

[0019] A fixing instruction is generated and sent to the mobile platform, which then fixes the composite structural component.

[0020] Determine the scanning trajectory required for inspecting the current composite structural component;

[0021] After the inspection is completed, the method also includes the step of generating a transfer instruction and sending it to the unloading robot arm to remove the composite structure from the inspection area.

[0022] In one feasible approach, the step of determining whether the current composite structural component is damaged based on the scanning data acquired by the ultrasonic phased array flaw detector specifically includes:

[0023] The scanning data acquired by the two ultrasonic phased array flaw detectors are denoised to obtain the data to be tested;

[0024] The two scan data are compared, and the obtained detection results are mutually verified.

[0025] After successful verification, relevant feature values ​​are extracted from the data to be detected;

[0026] Based on the feature values, an image imaging algorithm is used to convert the feature values ​​into ultrasound images;

[0027] The ultrasonic images are analyzed using a defect identification model to determine whether there is damage to the current composite structure. When damage is found, damage information is generated.

[0028] In one feasible approach, environmental data is analyzed to obtain the position, velocity, and acceleration of obstacles within a preset range of the damage scanning device over a certain time period.

[0029] Draw motion trajectory diagrams of the damage scanning device and the obstacle based on their positions, velocities, and accelerations;

[0030] The motion trajectory diagrams of the damage scanning device and the obstacle are compared to determine whether the motion trajectory of the object and the scanning trajectory of the damage scanning device overlap or intersect.

[0031] There is a risk of collision when the trajectory of an object overlaps or intersects with the scanning trajectory of a damage scanning device.

[0032] In one feasible approach, during the scanning process, the method further includes: fault monitoring of the damage scanning device based on real-time motion data of various parts of the composite material inspection robot, specifically including:

[0033] The acquired motion signals are preprocessed;

[0034] Wavelet transform is performed on the preprocessed motion data using wavelet basis functions to decompose the signal into components of different frequencies and time scales, thereby obtaining multiple target features.

[0035] Extract key features that can characterize the fault from multiple target features;

[0036] Based on the pre-established mapping relationship between fault modes and key features, it is determined whether the current damage scanning device has malfunctioned and the corresponding fault type.

[0037] In one implementable embodiment, after obtaining the ultrasound image, the method further includes the step of:

[0038] The ultrasound images are used to extract features, and the extracted features are visualized and mapped, converting the features into color features.

[0039] Based on the results of the visualization mapping, the ultrasound images are displayed in graphical form;

[0040] Receive adjustment instructions from users and adjust the displayed content.

[0041] In one feasible approach, the step of receiving the user's adjustment instruction and adjusting the displayed content specifically includes:

[0042] The adjustment instructions are analyzed to obtain the data to be identified;

[0043] Based on the intent recognition model, the data to be identified is analyzed to predict the user's interaction patterns and preference settings;

[0044] Personalized recommendations and optimizations are provided based on users' interaction patterns and preferences.

[0045] On the other hand, the present invention also provides a composite material inspection robot control system, including an instruction generation module, a control module, a damage recognition module, a collision prediction module, and a pose adjustment module, wherein:

[0046] The instruction generation module is used to generate control instructions based on a preset scanning trajectory;

[0047] The control module is used to control the damage scanning device of the composite material inspection robot to drive the ultrasonic phased array flaw detector to scan the current composite material structure.

[0048] The damage identification module is used to determine whether there is damage to the current composite structure based on the scanning data collected by the ultrasonic phased array flaw detector.

[0049] The collision prediction module is used to predict whether there is a collision risk in real time based on the environmental data monitored by the sensing device during the scanning process.

[0050] The pose adjustment module is used to generate pose adjustment instructions based on the obstacle risk information when there is a collision risk, and control the composite material inspection robot to change its current pose.

[0051] In one feasible embodiment, the damage scanning device includes a first six-axis robot and a second six-axis robot, which are respectively located on both sides of the composite structure. Both the first and second six-axis robots are equipped with ultrasonic phased array flaw detectors.

[0052] The control commands include a first scan command and a second scan command;

[0053] The first scanning command is used to control the first six-axis robot and send it to the first six-axis robot. The first six-axis robot moves according to the scanning command to inspect the front and side of the composite structure.

[0054] The second scanning command is used to generate a second scanning command based on the scanning trajectory and send it to the second six-axis robot. The second six-axis robot moves according to the scanning command to detect the reverse side and the other side opposite to the side of the composite structure.

[0055] The instruction generation module is also used to send a flipping instruction to the power mechanism located at both ends of the composite structure after the first six-axis robot completes the scanning operation, so as to drive the composite structure to rotate.

[0056] The beneficial effects of this application are as follows:

[0057] This application achieves efficient and accurate inspection of composite structural components by controlling a damage scanning device and using scanning data collected by an ultrasonic phased array flaw detector. It also uses environmental data collected by a sensor to predict whether there is a collision risk, thereby improving the safety of the composite inspection robot, reducing downtime, and increasing working accuracy.

[0058] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings.

[0059] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0060] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0061] In the attached diagram:

[0062] Figure 1 This is a flowchart of a composite material inspection robot control method according to an embodiment of the present invention;

[0063] Figure 2 This is a system structure diagram of a composite material inspection robot control system according to an embodiment of the present invention;

[0064] Figure 3 This is a diagram showing the instruction topology of the damage scanning device in an embodiment of the present invention. Detailed Implementation

[0065] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0066] Composite material testing is the process of evaluating and verifying the performance, structure, and quality of composite materials. Existing composite material testing robots still have significant limitations. In non-destructive testing, a robotic arm is used to move the testing probe, and the probe is identified by testing software, thus enabling the testing of composite structural components.

[0067] However, in the existing composite material inspection robots, the robotic arms mostly adopt an articulated serial structure. When scanning and inspecting composite components with large curvature, such as wind turbine blades, due to the cumulative error of the joints and the insufficient rigidity of the end effector, the actual scanning trajectory deviates from the preset path, resulting in poor coupling between the ultrasonic probe and the workpiece, and cracks cannot be detected.

[0068] In addition, in terms of visual judgment, the robotic arm relies on a single vision sensor or lidar. In environments with strong light, dust, or electromagnetic interference, it cannot identify metal tooling fixtures on the ground, which may cause the robotic arm to collide with the ultrasonic probe, resulting in damage to the probe array elements.

[0069] In the coordination of ultrasonic testing and robotic arms, the lack of real-time closed-loop feedback in ultrasonic phased array data acquisition and robot motion control leads to increased defect location errors when the deviation between the probe and the workpiece surface normal exceeds a preset value, resulting in a shift in the acoustic wave incident angle. This can cause defects to go undetected during layered defect detection due to decreased sensitivity.

[0070] In terms of human-machine collaboration and rapid deployment, most existing robots adopt offline programming mode, which can only perform fixed path detection. For non-standard composite materials, human-machine interaction is required for detection.

[0071] To address the aforementioned issues, this application proposes a scheme for controlling a composite material inspection robot to perform composite material inspection. The composite material inspection robot includes a damage scanning device and an ultrasonic phased array flaw detector. The ultrasonic phased array flaw detector is mounted on a six-axis robot and moves with it, thereby achieving a complete scan of the composite material structure. The ultrasonic phased array flaw detector includes a phased array main unit and a phased array probe. The phased array probe is composed of multiple crystals, each forming an independent transmitting / receiving unit. By controlling the excitation delay time of each crystal and changing the phase relationship of the emitted or received ultrasonic waves, the desired sound beam is obtained, thereby achieving control over the direction and depth of focus of the ultrasonic waves.

[0072] The solutions in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0073] Example 1:

[0074] See Figure 1 The control method includes the following steps:

[0075] Step S100: Generate control instructions according to the preset scanning trajectory to control the damage scanning device of the composite material inspection robot to drive the ultrasonic phased array flaw detector to scan the current composite material structure.

[0076] Step S200: Determine whether there is damage to the current composite structure based on the scanning data collected by the ultrasonic phased array flaw detector;

[0077] Step S300: During the scanning process, predict whether there is a collision risk in real time based on the environmental data monitored by the sensing device;

[0078] Step S400: When there is a collision risk, generate an attitude adjustment command based on the obstacle risk information to control the composite material inspection robot to change its current attitude.

[0079] This application first generates control commands based on a preset scanning trajectory to control the damage scanning device to drive the ultrasonic phased array flaw detector to inspect the current composite structural component. The scanning trajectory can be obtained by analyzing the 3D model of the composite structural component or determined based on a preset trajectory template. Then, during the scanning process, the damage scanning device drives the ultrasonic phased array flaw detector to move, and the ultrasonic phased array flaw detector emits ultrasonic signals and receives echo signals, thereby realizing the scanning of the composite structural component. Finally, the scanning data is used to determine whether there is damage to the current composite structural component. This step can first preprocess the scanning data to improve signal quality, and then analyze it through detection software to obtain the damage type, location, and size. At the same time, the sensing device will acquire the environmental data sent by the sensing device in real time to predict collisions; when a collision risk is predicted, an attitude adjustment command is generated based on the obstacle risk information to control the composite inspection robot to change its current attitude.

[0080] In one embodiment, during the scanning process of the composite material inspection robot according to a preset trajectory, if there is an angular deviation between the ultrasonic phased array probe and the composite material surface, the driving torque of each axis of the robotic arm is adjusted in real time to eliminate the poor coupling between the ultrasonic probe and the workpiece caused by the angular deviation.

[0081] In one embodiment, during the scanning process of the composite material inspection robot, the detection trajectory of the composite material inspection robot is continuously changed by fusing multi-sensor data and obstacle motion models through real-time environmental prediction to prevent collisions.

[0082] In one embodiment, during the scanning and inspection process of the composite material inspection robot, the probe posture is dynamically corrected by coupling the scanning data and motion control, combined with the ultrasonic echo signal intensity, to ensure that the incident angle of the sound wave does not deviate.

[0083] In one embodiment, the composite material inspection robot can automatically determine the composite material structure to be scanned based on a preset scanning trajectory, automatically plan the detection path, and achieve automated detection.

[0084] This application achieves efficient and accurate inspection of composite structural components by controlling a damage scanning device and using scanning data collected by an ultrasonic phased array flaw detector. It also uses environmental data collected by a sensor to predict whether there is a collision risk, thereby improving the safety of the composite inspection robot, reducing downtime, and increasing work accuracy.

[0085] Example 2:

[0086] Based on Example 1, see [link / reference] Figure 3 The damage scanning device includes a first six-axis robot and a second six-axis robot, which are located on both sides of the composite structure. Both the first six-axis robot and the second six-axis robot are equipped with an ultrasonic phased array flaw detection device.

[0087] Structurally, the damage scanning device employs a dual six-axis robot design with a symmetrical layout on both sides. The first and second six-axis robots are deployed on opposite sides of the composite structural component, each equipped with an ultrasonic phased array flaw detector. The six-axis robots possess six degrees of freedom, capable of translation along the X, Y, and Z axes and rotation around them. During operation, inverse kinematics algorithms are used to adjust the end effector's pose in three-dimensional space. The symmetrical layout allows the robots to simultaneously or sequentially inspect both sides of the composite structural component.

[0088] In one embodiment, for curved and irregularly shaped composite structural components, the flexibility of a six-axis robot enables automatic adaptation of the detection trajectory.

[0089] The specific steps of controlling the damage scanning device of the composite material inspection robot according to the scanning trajectory include:

[0090] The first scanning command is generated based on the scanning trajectory and sent to the first six-axis robot. The first six-axis robot moves according to the scanning command to inspect the front and side of the composite structure.

[0091] In one embodiment, based on the geometry of the composite structure, a scanning trajectory is pre-determined by setting a helix, grating line, or custom path, and a first scanning command is generated by the robot control system. The first control command includes the joint angle, velocity, and acceleration of each axis. The operating parameters of the ultrasonic probe include frequency, focal length, and array element activation sequence. After receiving the command, the first six-axis robot drives the probe along the trajectory along the front and each side of the composite structure. Many structures have special sides, such as the leading edge of an airfoil component, simultaneously acquiring ultrasonic echo data.

[0092] Once the inspection is complete, a flip command is sent to the power mechanisms located at both ends of the composite structure to drive the composite structure to rotate. When the first robot completes the front and side inspections, the system controls the power mechanisms at both ends of the composite structure to rotate.

[0093] After the first robot completes the front and side inspections, the system sends a reverse command to the power mechanisms at both ends of the composite structure, driving the rotating fixture of the composite structure to rotate. Encoder feedback controls the rotation of the structure around its axis, adjusting the angle as needed to ensure the uninspected reverse side and the other side face the second robot. During the flipping process, the structure's posture is monitored in real time, and gyroscopes or vision positioning are used to prevent the workpiece from slipping or colliding.

[0094] In one embodiment, the power mechanism and the composite material inspection robot communicate in real time via an industrial bus, and the flipping action is seamlessly integrated with the inspection process, reducing the waiting time between processes.

[0095] A second scanning command is generated based on the scanning trajectory and sent to the second six-axis robot. The second six-axis robot moves according to the scanning command to detect the reverse side and the opposite side of the composite structure, thus completing the scan.

[0096] After the structural component is flipped, the system generates a second scanning command based on a preset trajectory, driving the second six-axis robot to scan the reverse side and the opposite side. The second robot shares a coordinate system with the first robot, and a calibration algorithm ensures the spatial consistency of the detection data from both sides, ultimately stitching them together to form a complete three-dimensional detection map.

[0097] In one embodiment, if one robot malfunctions, another robot takes over all the detection tasks and automatically generates the entire detection trajectory.

[0098] In practical implementation, the damage scanning device includes a first six-axis robot and a second six-axis robot located on both sides of the composite structure. The first and second six-axis robots slide left and right along the length of the composite structure using ground tracks, achieving 7-axis linkage. When inspecting the composite structure, a first scanning command is generated based on the scanning trajectory and sent to the first six-axis robot. The first six-axis robot moves according to the scanning command to inspect the front and sides of the composite structure. After the front and side inspections are completed, a flipping command is sent to two power components located at both ends of the composite structure to flip it. Finally, a second scanning command is generated based on the scanning trajectory and sent to the second six-axis robot. The first six-axis robot, according to the scanning command, inspects the reverse side and the opposite side of the composite structure, completing the scan. This application can perform comprehensive inspection of the front, reverse, and two sides of a composite structure.

[0099] Example 3:

[0100] Based on Example 1, before the step of generating control commands according to a preset scanning trajectory, the method further includes the following steps:

[0101] Generate a transport instruction and send it to the loading robotic arm to move the composite structural component to the inspection area;

[0102] The composite material inspection robot generates transport instructions through its control center. These instructions include the 3D positioning information of the composite material component to be inspected, the optimal trajectory for obstacle avoidance by the robot and its probes, and the opening and closing sequence of the pneumatic grippers of the loading robotic arm. The loading robotic arm, typically a four-axis or six-axis robot, identifies the location of the composite material component using vision sensors after receiving the transport instructions, grasps the workpiece, and moves it to the inspection area. During the process, force sensors provide feedback to ensure that the grasping force does not cause surface damage to the composite material component.

[0103] A fixing instruction is generated and sent to the mobile platform, which then fixes the composite structural component.

[0104] The mobile platform is equipped with adjustable clamps, vacuum suction cups, or electromagnetic positioning devices. After receiving a fixing command, it automatically adjusts the clamp spacing or suction cup layout based on visual observation of the composite structure's geometric information, and applies a preset clamping force through a force control system. A displacement sensor determines whether the workpiece is completely fixed, ensuring that the composite structure does not vibrate or shift due to vibration. At this point, the coupling gap between the ultrasonic probe and the workpiece surface is stable, resulting in more accurate detection results.

[0105] Determine the scanning trajectory required for inspecting the current composite structural component;

[0106] The scanning trajectory is automatically generated using a trajectory planning algorithm based on the 3D model of the composite structural component or real-time visual scanning data. In one possible embodiment, the planned path also needs to consider: the integrity of the coverage area without omissions; the shortest path with the highest motion efficiency; the probe posture adaptation without angular deviation; and avoiding the edges or weak areas of the component.

[0107] In one embodiment, for curved and variable thickness composite structural components, a smooth trajectory is generated by B-spline curve interpolation to ensure that the probe is always perpendicular to the detection surface, thereby improving the defect detection rate.

[0108] After the inspection is completed, the method also includes the step of generating a transfer instruction and sending it to the unloading robot arm to remove the composite structure from the inspection area.

[0109] After the inspection is completed, the workpiece is judged as qualified or unqualified, and a transfer instruction is generated and sent to the unloading robotic arm. The instruction includes the target position and handling posture to prevent the workpiece from colliding with the material rack. The robotic arm confirms the status of the workpiece through RFID or visual recognition, moves it to the corresponding area, and updates the workpiece inspection data in the MES system.

[0110] In practice, the workpiece ID, inspection results, and operator information may also be automatically recorded during the material unloading process. This data is then uploaded to the cloud in real time to meet the quality traceability requirements of industries such as aerospace and automotive.

[0111] In practice, this application controls the loading robotic arm, moving platform, and unloading robotic arm by sending instructions, thereby automating the loading, unloading, and fixing processes of composite structural components. The entire non-destructive testing process is completed automatically, improving production efficiency and reducing labor costs.

[0112] Example 4:

[0113] Based on Example 1, the step of determining whether there is damage to the current composite structure based on the scanning data collected by the ultrasonic phased array flaw detector specifically includes: denoising the scanning data obtained by the two ultrasonic phased array flaw detectors to obtain the data to be tested;

[0114] When the ultrasonic phased array flaw detector acquires raw scanning data, there may be interference signals such as electromagnetic interference, coupling noise, and grain scattering, so noise reduction processing is performed.

[0115] In one embodiment, dual-probe data is collected synchronously, and noise reduction ensures the reliability of the comparison, which is consistent with the processing of strong scattering materials such as carbon fiber composites.

[0116] The two scan data are compared, and the obtained detection results are mutually verified.

[0117] This application aligns the spatial coordinates of the scanning data collected by the first and second six-axis robots—that is, the front and back data of the same detection area—and calculates the data similarity using a cross-correlation algorithm. The similarity is determined by the amplitude of the reflected wave and the time difference of arrival. If the similarity does not reach a preset value, a rescan or manual review is triggered. This two-way verification of critical areas ensures the authenticity and validity of defect signals.

[0118] After successful verification, relevant feature values ​​are extracted from the data to be detected;

[0119] The key features include reflected wave amplitude, defect depth, and defect size.

[0120] Based on the feature values, an image imaging algorithm is used to convert the feature values ​​into ultrasound images;

[0121] In practice, synthetic aperture focusing technology or full focusing method converts feature values ​​into two-dimensional / three-dimensional ultrasonic images, which intuitively display the location, shape and distribution of defects.

[0122] The ultrasonic images are analyzed using a defect identification model to determine whether there is damage to the current composite structure. When damage is found, damage information is generated.

[0123] The defect identification model is a pre-trained model used for defect analysis and outputting defect coordinates.

[0124] In practice, this application first performs denoising on the two acquired scan data to obtain the data to be detected. Denoising can improve the accuracy of subsequent feature extraction and matching. Then, the two scan data are compared and the detection results are mutually verified. During the comparison, the similarity index can be used to match the two scan data, which can eliminate the influence of noise and non-uniformity on the matching to a certain extent and improve the reliability of the results.

[0125] After successful verification, relevant feature values ​​are extracted from the data to be tested. These feature values ​​serve as the basis for damage information determination. Then, based on these feature values, an image imaging algorithm is used to convert them into ultrasound images. Finally, a defect recognition model is used to analyze the ultrasound images to obtain damage information. The defect recognition model is a pre-trained model based on a convolutional neural network, used for feature extraction and classification of the two-dimensional grayscale image of each feature point. The defect recognition model can learn richer information from the high-level features of the image, thereby improving the accuracy of damage recognition.

[0126] Example 5:

[0127] Based on Example 1, the step of predicting the existence of collision risk in real time based on the environmental data monitored by the sensing device specifically includes:

[0128] The environmental data is analyzed to obtain the position, velocity, and acceleration of obstacles within a preset range of the damage scanning device over a certain period of time;

[0129] The analysis of environmental data mainly involves analyzing it into point cloud data, images, and distance information. Through spatiotemporal registration, the real-time status parameters of obstacles in the working area of ​​the composite material inspection robot are extracted.

[0130] In one embodiment, the sensor coordinate system data can be converted into the robot base coordinate system through coordinate transformation to determine the coordinate position of the obstacle;

[0131] In one embodiment, Kalman filtering can be used to predict the motion trend of obstacles, calculate linear velocity, angular velocity and acceleration, and distinguish between static and dynamic obstacles.

[0132] Draw motion trajectory diagrams of the damage scanning device and the obstacle based on their positions, velocities, and accelerations;

[0133] The motion trajectory graph is drawn using forward kinematics based on the preset scanning trajectory of the composite material inspection robot and the real-time state parameters of obstacles. The robot's trajectory calculates the position sequence over a period of time based on the preset scanning path and kinematic models for each axis. For dynamic obstacles, a uniform / uniformly accelerated motion model is used; for static obstacles, a fixed set of coordinate points is used. The trajectory graph uses time as the horizontal axis and position as the vertical axis, visually displaying the spatial relationship between the two over time. Potential collision points can be quickly identified through this intuitive trajectory curve.

[0134] The motion trajectory diagrams of the damage scanning device and the obstacle are compared to determine whether the motion trajectory of the object and the scanning trajectory of the damage scanning device overlap or intersect.

[0135] The comparison is mainly done by comparing the trajectory maps of the robot and the obstacle through the spatial geometric intersection algorithm: the judgment of spatial overlap is to calculate the position distance between the two at each moment within a certain period of time. If the distance is less than the safety threshold, it means that the buffer distance after the combination of the robot end diameter and the obstacle radius is large, and it is judged as trajectory overlap.

[0136] There is a risk of collision when the trajectory of an object overlaps or intersects with the scanning trajectory of a damage scanning device.

[0137] When a trajectory overlap / intersection is detected, the system immediately triggers a collision risk signal, interrupts the current scanning task using a priority scheduling algorithm, and generates an emergency response command. Furthermore, it generates interference commands before a collision occurs, reducing detection efficiency loss due to frequent interruptions.

[0138] Example 6:

[0139] Based on Example 1, fault monitoring of the damage scanning device is performed according to the motion data of various parts of the composite material inspection robot under real-time monitoring. Specifically, this includes:

[0140] The acquired motion signals are preprocessed;

[0141] Motion signals are collected by sensors installed at key locations on the composite material inspection robot. These key locations include, but are not limited to, the robotic arm joints and the ultrasonic probe drive shaft. Sensors include, but are not limited to, encoders, accelerometers, and torque sensors. Motion data primarily consists of joint angles, rotational speeds, and vibration amplitudes. Preprocessing mainly involves using the 3σ criterion to remove jumps in data caused by momentary sensor malfunctions; eliminating high-frequency jitter noise through moving average filtering; and converting the dimensional data from different sensors into normalized values ​​in the 0-1 range to standardize subsequent analysis.

[0142] The above preprocessing can prevent the vibration sensor's burr signal from being misjudged as a mechanical fault, which is caused by raw data noise.

[0143] Wavelet transform is performed on the preprocessed motion data using wavelet basis functions to decompose the signal into components of different frequencies and time scales, thereby obtaining multiple target features.

[0144] Wavelet transform decomposes preprocessed motion data into three levels using wavelet basis functions. This decomposes the original signal into: high-frequency detail components (corresponding to transient signals such as mechanical vibration and impact, and addressing high-frequency noise generated by bearing wear); and low-frequency approximation components (corresponding to stable signals of uniform robot motion, addressing the velocity curves of normal joint rotation). By calculating the energy proportion and singularity locations of each component, fault-related time-frequency features are extracted.

[0145] The above methods can be used to identify the initial factors of a fault that exhibit weak characteristics.

[0146] Extract key features that can characterize the fault from multiple target features;

[0147] Key features are those that can represent faults, such as: using peak value, root mean square (RMS), and kurtosis to determine the degree of signal impact; and using the dominant frequency amplitude and frequency centroid to determine the distribution of vibration energy.

[0148] Based on the pre-established mapping relationship between fault modes and key features, it is determined whether the current damage scanning device has malfunctioned and the corresponding fault type.

[0149] The system has a pre-set fault database. It determines the fault type by mapping key features to fault modes. Only by matching the fault database with the fault type and fault features can the system determine what kind of fault exists in the scanned composite structural component.

[0150] Specifically, in practical implementation, this application first preprocesses the acquired motion signal, including noise reduction and filtering, to reduce the impact of noise on fault identification. Then, wavelet transform is performed on the preprocessed motion data using wavelet basis functions to decompose the signal into components of different frequencies and time scales, obtaining multiple target features. Through wavelet transform, a series of features related to signal strength, amplitude, and time domain can be obtained; and key features closely related to known fault modes are selected from these multiple target features. During selection, key features are those that are relatively few in number under normal conditions but increase or decrease sharply when a fault occurs. Finally, based on the pre-established mapping relationship between fault modes and key features, it is determined whether the current damage scanning device has malfunctioned. When a fault occurs, a reminder message can be generated according to the corresponding fault type for timely notification. By monitoring the operating status of each joint of the damage scanning device, this application can identify potential problems before a fault occurs, thereby taking preventive measures to avoid downtime and increased maintenance costs caused by equipment failure.

[0151] Example 7:

[0152] Based on Example 1, after obtaining the ultrasound image, the method further includes the following steps:

[0153] The ultrasound images are used to extract features, and the extracted features are visualized and mapped, converting the features into color features.

[0154] Feature extraction from ultrasound images primarily relies on grayscale image feature extraction. Image segmentation algorithms, based on thresholding and edge detection, are used to extract the geometric features of defects, determining their area, perimeter, and circularity. Physical features, such as reflection amplitude and depth, are also extracted, establishing a mapping relationship between feature values ​​and color space. In feature mapping, defect depth is mapped to a gradient color spectrum of blue or red, and reflection amplitude is mapped to color saturation. Qualitatively, in practice, layered defects are marked yellow, void defects green, and cracks red, with contour lines overlaid to facilitate rapid identification of key defect parameters through color.

[0155] Based on the results of the visualization mapping, the ultrasound images are displayed in graphical form;

[0156] During the display process, the ultrasound image is presented in a multimodal format using an OpenGL graphics rendering engine. In 2D mode, a planar distribution map of the defect is displayed, including color coding, dimension annotations, and position coordinates. In 3D mode, the three-dimensional shape of the defect is reconstructed using a volume rendering algorithm. During the display, the original grayscale image and the color-mapped image are simultaneously displayed in a contrast mode, allowing for one-click switching.

[0157] In one embodiment, internal defects of complex components are visually presented through 3D rendering, allowing engineers to observe the spatial distribution of defects from any angle. During cross-platform operation, the system meets the collaborative interpretation needs of on-site workshop personnel and remote experts, displaying information on the robot control cabinet touchscreen, PC client, and mobile terminal.

[0158] Receive adjustment instructions from users and adjust the displayed content.

[0159] During the adjustment of the displayed content, the user inputs adjustment commands through a graphical interface and then adjusts the color mapping threshold, feature annotation information, and contrast by using mouse zooming, panning, and scrolling. Combined with the method of selecting defective areas, the data is transmitted in real time to the image processing module via TCP / IP protocol to achieve image processing.

[0160] Specifically, in practical implementation, this application first extracts features from the ultrasound image and then performs a visual mapping of the extracted features, converting them into color features. Next, based on the results of the visual mapping, the ultrasound image is displayed graphically. Finally, user adjustment commands are received to adjust the displayed content. This application improves information readability and enhances data interactivity, thereby increasing data analysis efficiency.

[0161] Example 8:

[0162] Based on Example 1, the steps of receiving the user's adjustment command and adjusting the displayed content specifically include:

[0163] The adjustment instructions are analyzed to obtain the data to be identified;

[0164] User-inputted adjustment commands via mouse, touchscreen, or voice are converted into structured data for recognition by a multimodal command parser. These commands include various types such as view control (zoom / rotation), parameter adjustment (color / threshold), and mode switching (2D / 3D). During operation, key numerical values, spatial coordinates, and temporal features are extracted from the commands to determine the associated data for recognition.

[0165] Based on the intent recognition model, the data to be identified is analyzed to predict the user's interaction patterns and preference settings;

[0166] During the analysis and prediction process, an intent recognition model based on a bidirectional LSTM + attention mechanism performs time-series analysis on the structured data to be recognized and outputs prediction results. During intent recognition, the intent is set by selecting a region and pre-loading the measurement tool panel.

[0167] In one embodiment, the prediction results include identifying the user's work scenario through a sequence of user actions.

[0168] In one embodiment, user habits are mined from historical data to form a preference vector.

[0169] Personalized recommendations and optimizations are provided based on users' interaction patterns and preferences.

[0170] In practice, based on the predicted interaction patterns and preferences, optimization actions are generated through reinforcement learning strategies: the frequently used functions and low-frequency functions of the user are listed separately, and the current user's operation behavior is quickly determined by the selection of interaction patterns and habitual preferences.

[0171] Specifically, in actual implementation, this application is able to parse received adjustment instructions; analyze the data to be identified based on the intent recognition model, predict the user's interaction patterns and preference settings; and provide personalized recommendations and optimizations for the user based on the user's interaction patterns and preference settings.

[0172] Example 9:

[0173] This embodiment 2 provides a control system for a composite material inspection robot, see below. Figure 2 It includes an instruction generation module, a control module, a damage recognition module, a collision prediction module, and a pose adjustment module, wherein:

[0174] The instruction generation module is used to generate control instructions based on a preset scanning trajectory;

[0175] The control module is used to control the damage scanning device of the composite material inspection robot to drive the ultrasonic phased array flaw detector to scan the current composite material structure.

[0176] The damage identification module is used to determine whether there is damage to the current composite structure based on the scanning data collected by the ultrasonic phased array flaw detector.

[0177] The collision prediction module is used to predict the existence of collision risk in real time based on the environmental data monitored by the sensing device during the scanning process;

[0178] The pose adjustment module is used to generate pose adjustment commands based on obstacle risk information when there is a collision risk, thereby controlling the composite material inspection robot to change its current pose.

[0179] Specifically, in practical implementation, this application generates control commands through an instruction generation module, and controls the damage scanning device to perform a comprehensive scan of the composite structural component through the control module. Then, the damage identification module performs non-destructive testing of the composite structural component based on the scan data collected by the ultrasonic phased array flaw detector. During the scanning process, the collision prediction module predicts the existence of collision risks based on environmental data monitored by the sensor device. This application enables efficient and accurate testing of composite structural components, which not only improves the efficiency and safety of the composite inspection robot, but also makes it more accurate, thus contributing to the improvement of the quality and safety of industrial products.

[0180] Example 10:

[0181] Based on Example 2, the damage scanning device includes a first six-axis robot and a second six-axis robot, which are located on both sides of the composite structure. Both the first six-axis robot and the second six-axis robot are equipped with an ultrasonic phased array flaw detection device.

[0182] The specific steps of controlling the damage scanning device of the composite material inspection robot according to the scanning trajectory include:

[0183] The first scanning command is generated based on the scanning trajectory and sent to the first six-axis robot. The first six-axis robot moves according to the scanning command to inspect the front and side of the composite structure.

[0184] Once the inspection is complete, a flipping command is sent to the power mechanisms located at both ends of the composite structure to drive the composite structure to rotate.

[0185] A second scanning command is generated based on the scanning trajectory and sent to the second six-axis robot. The second six-axis robot moves according to the scanning command to detect the reverse side and the opposite side of the composite structure, thus completing the scan.

[0186] Specifically, in actual implementation, the damage scanning device includes a first six-axis robot and a second six-axis robot located on either side of the composite structural component. The first and second six-axis robots move left and right along ground tracks, achieving 7-axis linkage. When inspecting the composite structural component, a first scanning command is generated based on the scanning trajectory and sent to the first six-axis robot. The first six-axis robot moves according to the scanning command to inspect the front and sides of the composite structural component. After the front and side inspections are completed, a flipping command is sent to two power components located at both ends of the composite structural component to flip it. Finally, a second scanning command is generated based on the scanning trajectory and sent to the second six-axis robot to inspect the reverse side and the opposite side of the composite structural component, completing the scan. This achieves comprehensive inspection of the composite structural component.

[0187] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A control method for a composite material inspection robot, characterized in that, The method includes the following steps: Control commands are generated based on the preset scanning trajectory to control the damage scanning device of the composite material inspection robot to drive the ultrasonic phased array flaw detector to scan the current composite material structure. Determine whether there is damage to the current composite structure based on the scanning data collected by the ultrasonic phased array flaw detector; During the scanning process, the presence of collision risk is predicted in real time based on the environmental data monitored by the sensing device. When there is a risk of collision, the robot generates attitude adjustment instructions based on the risk information of the obstacle, and controls the composite material inspection robot to change its current attitude.

2. The control method for a composite material inspection robot as described in claim 1, characterized in that, The damage scanning device includes a first six-axis robot and a second six-axis robot, which are located on both sides of the composite structure. Both the first six-axis robot and the second six-axis robot are equipped with an ultrasonic phased array flaw detection device. The steps of controlling the damage scanning device of the composite material inspection robot according to the scanning trajectory specifically include: The first scanning command is generated based on the scanning trajectory and sent to the first six-axis robot. The first six-axis robot moves according to the scanning command to inspect the front and side of the composite structure. Once the inspection is complete, a flipping command is sent to the power mechanisms located at both ends of the composite structure to drive the composite structure to rotate. A second scanning command is generated based on the scanning trajectory and sent to the second six-axis robot. The second six-axis robot moves according to the scanning command to detect the reverse side and the opposite side of the composite structure, thus completing the scan.

3. The composite material inspection robot control method as described in claim 2, characterized in that, Before the step of generating control commands based on a preset scan trajectory, the method further includes the following step: Generate a transport instruction and send it to the loading robotic arm to move the composite structural component to the inspection area; A fixing instruction is generated and sent to the mobile platform, which then fixes the composite structural component. Determine the scanning trajectory required for inspecting the current composite structural component; After the inspection is completed, the method also includes the step of generating a transfer instruction and sending it to the unloading robot arm to remove the composite structure from the inspection area.

4. The control method for a composite material inspection robot as described in claim 2, characterized in that, The step of determining whether there is damage to the current composite structure based on the scanning data collected by the ultrasonic phased array flaw detector specifically includes: The scanning data acquired by the two ultrasonic phased array flaw detectors are denoised to obtain the data to be tested; The two scan data are compared, and the obtained detection results are mutually verified. After successful verification, relevant feature values ​​are extracted from the data to be detected; based on the feature values, an image imaging algorithm is used to convert the feature values ​​into an ultrasound image; The ultrasonic images are analyzed using a defect identification model to determine whether there is damage to the current composite structure. When damage is found, damage information is generated.

5. The control method for a composite material inspection robot as described in claim 1, characterized in that, The step of predicting the existence of collision risk in real time based on environmental data monitored by the sensing device specifically includes: The environmental data is analyzed to obtain the obstacles, speeds, and accelerations within a preset range of the damage scanning device over a certain time period. Draw motion trajectory diagrams of the damage scanning device and the obstacle based on their positions, velocities, and accelerations; The motion trajectory diagrams of the damage scanning device and the obstacle are compared to determine whether the motion trajectory of the object and the scanning trajectory of the damage scanning device overlap or intersect. There is a risk of collision when the trajectory of an object overlaps or intersects with the scanning trajectory of a damage scanning device.

6. The control method for a composite material inspection robot as described in claim 4, characterized in that, During the scanning process, the method also includes: monitoring the damage scanning device for faults based on the real-time motion data of various parts of the composite material inspection robot, specifically including: The acquired motion signals are preprocessed; Wavelet transform is performed on the preprocessed motion data using wavelet basis functions to decompose the signal into components of different frequencies and time scales, thereby obtaining multiple target features. Extract key features that can characterize the fault from multiple target features; Based on the pre-established mapping relationship between fault modes and key features, it is determined whether the current damage scanning device has malfunctioned and the corresponding fault type.

7. The composite material inspection robot control method as described in claim 4, characterized in that, After obtaining the ultrasound image, the method further includes the following steps: The ultrasound images are used to extract features, and the extracted features are visualized and mapped, converting the features into color features. Based on the results of the visualization mapping, the ultrasound images are displayed in graphical form; Receive adjustment instructions from users and adjust the displayed content.

8. The control method for a composite material inspection robot as described in claim 7, characterized in that, The step of receiving the user's adjustment instruction and adjusting the displayed content specifically includes: The adjustment instructions are analyzed to obtain the data to be identified; Based on the intent recognition model, the data to be identified is analyzed to predict the user's interaction patterns and preference settings; Personalized recommendations and optimizations are provided based on users' interaction patterns and preferences.

9. A control system for a composite material inspection robot, characterized in that, The system includes an instruction generation module, a control module, a damage recognition module, a collision prediction module, and a pose adjustment module, wherein: The instruction generation module is used to generate control instructions based on a preset scanning trajectory; The control module is used to control the damage scanning device of the composite material inspection robot to drive the ultrasonic phased array flaw detector to scan the current composite material structure. The damage identification module is used to determine whether there is damage to the current composite structure based on the scanning data collected by the ultrasonic phased array flaw detector. The collision prediction module is used to predict whether there is a collision risk in real time based on the environmental data monitored by the sensing device during the scanning process. The pose adjustment module is used to generate pose adjustment instructions based on the obstacle risk information when there is a collision risk, and control the composite material inspection robot to change its current pose.

10. The composite material inspection robot control system as described in claim 9, characterized in that, The damage scanning device includes a first six-axis robot and a second six-axis robot, which are located on both sides of the composite structure. Both the first and second six-axis robots are equipped with ultrasonic phased array flaw detection devices. The control commands include a first scan command and a second scan command; The first scanning command is used to control the first six-axis robot and send it to the first six-axis robot. The first six-axis robot moves according to the scanning command to inspect the front and side of the composite structure. The second scanning command is used to generate a second scanning command based on the scanning trajectory and send it to the second six-axis robot. The second six-axis robot moves according to the scanning command to detect the reverse side and the other side opposite to the side of the composite structure. The instruction generation module is also used to send a flipping instruction to the power mechanism located at both ends of the composite structure after the first six-axis robot completes the scanning operation, so as to drive the composite structure to rotate.

Images