A curved surface component ultrasonic self-adaptive detection method and system

Abstract

The present application relates to the technical field of nondestructive testing, and discloses a kind of curved surface component ultrasonic adaptive detection method and system, comprising the following steps: three-dimensional curved surface online reverse and point cloud data processing;Detection area definition and adaptive detection path generation;Probe posture optimization and interference collision avoidance;Ultrasonic detection parameter adaptive matching;Closed loop adaptive execution and data fusion;The present application first online three-dimensional reverse component curved surface, generates triangular mesh model, then plans initial detection path and posture on model, further with optimal incidence and coupling as target optimization probe posture, and avoids movement collision, simultaneously according to local curvature adaptive adjustment ultrasonic parameter, finally by robot executes adaptive scanning under constant force tracking, synchronously collects spatial pose and ultrasonic data, finally realizes the visualization of detection result on three-dimensional model, realizes the full-automatic, high-precision, adaptive ultrasonic detection to complex curved surface component.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, specifically to an ultrasonic adaptive testing method and system for curved surface components. Background Technology

[0002] Ultrasonic testing is a non-destructive testing technique that uses high-frequency sound waves to detect internal defects in materials or to evaluate material properties. Ultrasonic testing of curved components is based on the propagation characteristics of ultrasonic waves in a medium. By emitting ultrasonic waves and receiving their echo signals, the characteristics of defect reflection and scattering are analyzed to ultimately form an image and achieve non-destructive testing. Key technologies include beam focusing, signal processing, and curved surface modeling. It is widely used in aerospace (engine blades, aircraft structures), automobile manufacturing (body, cylinder block), precision instruments, etc.

[0003] Currently, ultrasonic testing is a crucial non-destructive testing method in the industrial field. However, for critical components with complex free-form surfaces, its testing faces the following challenges:

[0004] Model dependency: Traditional offline programming inspection methods rely heavily on the accurate CAD model of the component. However, actual workpieces often have manufacturing errors, service deformation, or no CAD model at all (such as reverse engineering parts or repair parts), which causes the preset inspection path to fail, resulting in missed detections or poor coupling.

[0005] Fixed path and attitude: Existing methods often fix the detection path planning and probe attitude (sway angle, roll angle), which cannot adapt to the continuous changes in the local curvature of the surface and makes it difficult to ensure that the ultrasonic beam always enters the inspected area at the optimal angle (usually the normal), affecting the defect detection rate and quantitative accuracy.

[0006] Insufficient automation and adaptability: Lacking closed-loop adaptive capability of perception-planning-execution, the system cannot adjust in real time when the actual curved surface does not match the expectation, requiring manual intervention, which is inefficient. Therefore, there is an urgent need for a fully automated ultrasonic testing method that does not rely on a preset CAD model, can adapt to the actual curved surface geometric features, and optimize the detection parameters in real time. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides an ultrasonic adaptive inspection method and system for curved surface components. It solves the problems of traditional offline programming inspection methods that heavily rely on the accurate CAD model of the component, and existing methods that fix the detection path planning and probe posture (sway angle, roll angle), which cannot adapt to the continuous changes in the local curvature of the curved surface. It is difficult to ensure that the ultrasonic beam always enters the inspected area at the optimal angle (usually the normal), which affects the defect detection rate and quantitative accuracy, and lacks the closed-loop adaptive capability of perception-planning-execution.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an ultrasonic adaptive testing method for curved surface components. This testing method does not rely on a preset CAD model, can adapt to the actual geometric features of the curved surface, and optimize the testing parameters in real time to achieve fully automatic ultrasonic testing, comprising the following steps:

[0009] S1. Online inverse 3D surface calculation and point cloud data processing;

[0010] S2. Definition of detection area and generation of adaptive detection path;

[0011] S3. Probe attitude optimization and interference collision avoidance;

[0012] S4. Adaptive matching of ultrasound detection parameters;

[0013] S5, closed-loop adaptive execution and data fusion;

[0014] S6. Three-dimensional visualization and evaluation of test results.

[0015] As a preferred technical solution, in step S1, real-time three-dimensional geometric information of the surface component to be inspected is obtained, the clamped surface component to be inspected is scanned online in three dimensions to obtain its surface point cloud data, and a triangular mesh surface model is reconstructed.

[0016] Online 3D scanning utilizes a line laser scanner or a structured light 3D scanner mounted at the end of an inspection robot to perform high-precision scanning of the curved surface component to be inspected, which is clamped in place. Among them, the line laser scanner is suitable for rapid measurement of small and medium-sized complex curved surface components, while the structured light 3D scanner is suitable for comprehensive measurement of complex shapes.

[0017] The surface point cloud data is obtained by scanning, and a series of post-processing operations are performed on the obtained point cloud data, including denoising, filtering and simplification.

[0018] After data processing is completed, the moving least squares method or Poisson surface reconstruction algorithm is used to construct a triangular mesh surface model of the surface component to be inspected based on the processed point cloud data, which serves as the real-time geometric reference for inspection.

[0019] Among them, the moving least squares method can preserve the detailed features of the surface through local weighted fitting, while the Poisson surface reconstruction algorithm can generate a smoother and more complete surface model.

[0020] As a preferred technical solution, S2 is to define the detection area on the triangular mesh surface model constructed in step S1, and generate the initial ultrasonic probe movement path points and theoretical probe posture. Specifically, on the triangular mesh surface model, the area to be detected is defined, and an adaptive ultrasonic probe movement path is generated according to the detection requirements. This area can be determined based on the failure mode analysis of the component, historical detection records, and design requirements.

[0021] After the detection area is determined, the initial ultrasonic probe movement path points are generated within the detection area based on the isoparametric method or the equal bow height error method. The isoparametric method distributes the path points uniformly along the parameterization direction of the surface, which is suitable for surfaces with regular parameter domains. The equal bow height error method dynamically determines the spacing between path points according to the bow height error threshold, which is particularly suitable for complex surfaces with large curvature changes.

[0022] For each initial path point, the theoretical normal orientation of the probe is initially calculated based on the normal vector of its local triangular facet. The theoretical normal orientation refers to the orientation when the probe axis coincides with the surface normal vector, that is, when the probe axis coincides with the surface normal vector. The theoretical probe orientation is determined by the normal vector of the triangular facet at the path point.

[0023] As a preferred technical solution, S3 optimizes the probe posture at each path point based on the optimal incident and coupling conditions of the ultrasonic beam, and performs robot motion collision detection and avoidance planning.

[0024] In actual testing, it is necessary to consider the physical dimensions of the ultrasonic probe, the characteristics of the ultrasonic beam, and the kinematic constraints of the robot, and optimize the probe posture at each path point:

[0025] During attitude optimization, a search and optimization is performed within a spatial cone-shaped range near the theoretical normal attitude. The dual objective functions are minimizing the normal incident deviation of the ultrasonic beam centerline to the measured area and optimizing the ultrasonic coupling state. The optimal probe yaw angle and roll angle at this point are determined by optimization. Specifically, the probe attitude optimization uses minimizing the normal incident deviation of the ultrasonic beam centerline to the measured area and optimizing the ultrasonic coupling state as joint optimization objectives. The search is performed in the neighborhood space of the theoretical normal attitude to obtain the optimized probe yaw angle and roll angle.

[0026] During collision detection, the robot's kinematic model and the 3D models of the probe and tooling are used to continuously detect collisions throughout the entire motion process from the current point to the next path point. If a collision is predicted to occur during the motion, an avoidance point is inserted in the robot's joint configuration space using the artificial potential field method or random sampling method to generate a collision-free motion trajectory.

[0027] As a preferred technical solution, in step S4, the selection of ultrasonic testing parameters has a direct impact on the testing results. Different curved surfaces require different testing parameters to obtain the best testing effect. Based on the local curvature of the surface at the testing point, the emission and scanning parameters of the ultrasonic testing equipment are adaptively adjusted. The adaptive adjustment specifically involves: establishing a mapping relationship library between surface curvature and ultrasonic testing parameters; querying and calling the corresponding pulse repetition frequency, scanning speed, and probe selection strategy based on the real-time calculated curvature of the testing point.

[0028] Based on the local Gaussian curvature and average curvature of the surface at the current detection point, as well as the acoustic properties of the component material, the emission parameters of the ultrasonic testing instrument are adaptively adjusted. Curvature is a geometric quantity describing the degree of surface bending. The curvature value can be used to determine whether the current area is a flat region or a high-curvature region, as detailed below:

[0029] For high curvature regions, the repetition frequency is automatically increased to increase the number of sampling points per unit time, the scanning speed is reduced to increase the detection coverage time, and a smaller probe and a higher frequency are selected to focus the sound beam to improve resolution.

[0030] For flat areas, the scanning step length can be increased and the scanning speed can be improved, thereby increasing detection efficiency while ensuring detection quality. At the same time, a mapping database of curvature-sound beam diffusion-detection parameters can be established for real-time query and retrieval.

[0031] As a preferred technical solution, in the actual detection execution stage, S5 controls the robot to carry the ultrasonic probe along the optimized path and posture, performs ultrasonic detection under constant force control, and simultaneously collects spatial pose data and ultrasonic signals. The spatial pose data includes the three-dimensional coordinates of the probe center point and the optimized posture angle.

[0032] During the movement, the contact force between the probe and the curved surface is monitored in real time by the six-dimensional force / torque sensor at the end of the robot, and constant force tracking control is performed to ensure stable coupling. This achieves constant force tracking control between the probe and the curved surface. By monitoring the magnitude of the contact force in real time, the robot's movement position is dynamically adjusted to keep the contact force within the set target range.

[0033] Furthermore, high-precision detection is achieved through data fusion. The spatial coordinates (X,Y,Z), optimized attitude angles (A,B), corresponding ultrasonic A-scan signals, and processed C-scan image data of each detection point are strictly synchronized and bound. The ultrasonic A-scan signal is the original echo signal of the ultrasonic probe, which contains the temporal information of the defect. The C-scan image is a two-dimensional projection image after spatial mapping of multiple A-scan signals, which can intuitively display the distribution of defects on the scanning plane.

[0034] The synchronized data binding forms a full matrix detection dataset with three-dimensional spatial location information.

[0035] As a preferred technical solution, S6 integrates ultrasonic test data with a triangular mesh surface model to perform three-dimensional visualization and evaluation, presenting the test data to the user in an intuitive way.

[0036] The full matrix detection data obtained in step S5 is mapped back onto the triangular mesh surface model constructed in step S1 to generate a 3D color image and isosurface map for visualization.

[0037] In the 3D visualization process, based on the ultrasonic signal characteristics of each detection point, the corresponding defect indication value is calculated and these values ​​are mapped to the corresponding positions on the surface. The precise location, depth and equivalent size of the defect on the component surface are displayed intuitively by the shades of color and the density of contour lines.

[0038] The burial depth information is calculated using the flight time of the ultrasonic signal with an accuracy of 0.1 mm. The equivalent size is obtained by comparing the echo amplitude of the artificial injury of the standard defect.

[0039] The 3D visualization of the inspection results provides a highly intuitive basis for analysis. Operators can rotate, scale, and section the 3D model to observe the distribution and severity of defects from any angle, providing reliable support for subsequent maintenance decisions.

[0040] An adaptive ultrasonic testing system for curved surface components enables adaptive ultrasonic testing of curved surface components. The system includes a motion actuator, an adaptive clamping mechanism, a composite end effector, an ultrasonic transmitter / receiver and data acquisition unit, a central control and processing unit, and a visualization and human-machine interaction unit.

[0041] As a preferred technical solution, the motion actuator is the execution layer of the system, responsible for driving the ultrasonic probe and the three-dimensional scanner to move precisely in three-dimensional space, and adopts a six-axis industrial robot or a three-axis gantry scanner.

[0042] Six-axis industrial robots can realize various motion trajectories required for the inspection of complex curved surfaces. In inspection applications, a composite end effector is installed at the end of the robot, which is suitable for the inspection of small and medium-sized complex curved surface components.

[0043] The three-axis gantry scanner adopts a linear guide and slider structure, and realizes spatial movement through three mutually perpendicular linear axes. It is suitable for the inspection of large components. When inspecting large workpieces, it uses its built-in adaptive clamping mechanism to adaptively center and limit the workpiece. The adaptive clamping mechanism will stably limit the workpiece to be inspected.

[0044] The composite end effector is the core execution component of the system. It is installed at the end of the motion actuator and directly contacts the curved surface component to be inspected to perform the inspection task. It integrates an ultrasonic probe, a 3D scanner and a force sensor.

[0045] The ultrasonic probe is a key component for ultrasonic testing. It can be coupled by water immersion or water spray. Water immersion coupling involves completely immersing the probe in water, while water spray coupling uses a nozzle to spray a water film between the probe and the workpiece to achieve coupling. The choice of probe needs to be determined based on the thickness of the material being tested, the type of defect, and the testing requirements.

[0046] 3D scanners are used to acquire three-dimensional geometric information of workpiece surfaces online. 3D scanners can be either line laser scanners or structured light 3D scanners. Line laser scanners acquire point cloud data by emitting laser lines and capturing reflected light, while structured light 3D scanners acquire rich information by projecting gratings with specific patterns onto the object surface and calculating three-dimensional coordinates based on the deformation of the gratings. The choice of scanner depends on the characteristics of the workpiece and the inspection requirements.

[0047] The force sensor is a six-dimensional force / torque sensor, which can simultaneously monitor forces and torques in three directions. It is installed between the probe and the robot end effector and can sense the contact state between the probe and the curved surface in real time. By monitoring the magnitude of the contact force, constant force tracking control can be achieved.

[0048] As a preferred technical solution, the ultrasonic transmitting / receiving and data acquisition unit is responsible for generating, receiving and acquiring ultrasonic signals, and working in conjunction with the robot control system and the computer system to achieve accurate acquisition and processing of ultrasonic detection signals.

[0049] The central control and processing unit is responsible for coordinating the work of each part and executing the core algorithm functions. This unit is used to run the algorithm of steps S1-S6 and control the robot, scanner and ultrasonic equipment to work together. It adopts a high-performance industrial computer and has powerful computing power and real-time performance.

[0050] The visualization and human-computer interaction unit provides a three-dimensional visualization and human-computer interaction interface, enabling operators to intuitively monitor the system's operating status, view detection results, and perform necessary intervention operations.

[0051] Compared with the prior art, the present invention provides an ultrasonic adaptive detection method and system for curved surface components, which has the following beneficial effects:

[0052] 1. This invention first performs online 3D reverse engineering of the component surface to generate a triangular mesh model. Then, it plans the initial detection path and posture on the model, optimizes the probe posture with the goal of optimal acoustic beam incidence and coupling, and avoids motion collisions. At the same time, it adaptively adjusts the ultrasonic parameters according to the local curvature. Finally, the robot performs adaptive scanning under constant force tracking, and simultaneously collects spatial pose and ultrasonic data. Ultimately, the detection results are visualized on the 3D model, completely eliminating the dependence on prefabricated CAD models. This invention achieves fully automatic, high-precision, and adaptive ultrasonic detection of unknown or complex curved surface components, significantly improving the reliability, efficiency, and intelligence level of the detection.

[0053] A method and system that does not rely on precise prior CAD models and can automatically generate and optimize ultrasonic testing paths and probe postures online based on the measured three-dimensional shape of components, enabling efficient, reliable, and fully automated adaptive ultrasonic testing of complex curved surface components.

[0054] 2. By adopting a scan-then-detection mode, planning is performed directly based on the real-time 3D reverse model of the workpiece, eliminating model dependence and completely solving the detection failure problem caused by missing or inaccurate models. Through local normal vector calculation and attitude optimization, it ensures that the ultrasonic beam is always incident at the optimal angle, maximizing the reception of defect echoes, improving the detection rate and quantitative accuracy of defects, especially directional defects. The detection accuracy is high. The entire process, from surface reverse engineering, path planning, attitude optimization, parameter matching to closed-loop force control, is completed automatically. It can also adaptively handle unknown surfaces, deformable surfaces, and complex geometric features, with full automation and strong adaptability.

[0055] Furthermore, by integrating kinematic collision detection and avoidance, the safety of expensive equipment and workpieces is ensured, guaranteeing safety. Through three-dimensional visualization processing of the detection results, defect information is accurately associated with the workpiece entity, facilitating intuitive analysis and precise positioning by engineers, making the detection results more intuitive and reliable.

[0056] 3. The longitudinal, transverse, and vertical electric slide rails facilitate multi-dimensional movement of the composite end effector for scanning and detection according to actual needs. The drive motor drives the drive disc to rotate at the bottom of the platform, and in conjunction with the limiting arc groove and drive column, the slide rod slides along the limiting groove in the guide rail. The slide rod drives the limiting fixture to move. Through the synchronous movement of the four sets of limiting fixtures, the workpiece placed on the platform is stably limited and clamped. Flexible airbags and anti-slip strips ensure the stability of the limiting fixture in clamping the workpiece. The adaptive clamping mechanism achieves adaptive centering and limiting of the workpiece, ensuring the stability of the workpiece during the detection process and guaranteeing detection accuracy. Attached Figure Description

[0057] Figure 1 This is a flowchart of the detection steps of the adaptive detection method of the present invention.

[0058] Figure 2 This is a schematic diagram of the adaptive detection system of the present invention.

[0059] Figure 3 This is a schematic diagram of the structure of the ultrasonic scanning frame of the present invention.

[0060] Figure 4 This is a schematic diagram of the structure of the water tank of the present invention.

[0061] Figure 5 This is a schematic diagram of the structure of the frustum-shaped object carrier of the present invention.

[0062] Figure 6 This is a schematic diagram of the limiting clamp of the present invention.

[0063] Figure 7This is a schematic diagram of the slide bar of the present invention.

[0064] In the diagram: 1. Ultrasonic scanning frame; 2. Water tank; 3. Longitudinal electric slide rail; 4. Transverse electric slide rail; 5. Vertical electric slide rail; 6. Composite end effector; 7. Loading frustum; 8. Support bar; 9. Guide rail; 10. Limiting groove; 11. Slide rod; 12. Limiting clamp; 13. Drive disc; 14. Drive motor; 15. Limiting arc groove; 16. Drive column; 17. Flexible airbag; 18. Anti-slip strip. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. However, it should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0066] In the description of this invention, it should be noted that when an element is referred to as being "fixed to" or "set on" another element, it can be directly on or indirectly on the other element. When an element is referred to as being "connected to" another element, it can be directly connected to or indirectly connected to the other element.

[0067] In the description of this invention, it should be noted that the terms "center," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this invention is in use. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. "Several" means one or more, unless otherwise explicitly specified.

[0068] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0069] Example: Please refer to Figure 1 The present invention provides the following technical solution: an ultrasonic adaptive detection method for curved surface components, which is for the ultrasonic automated detection of components with complex curved surfaces. This embodiment takes aero-engine turbine blades as the detection object. Aero-engine turbine blades are one of the most critical load-bearing components in aero-engines. They operate in extremely harsh environments and need to run for a long time under high temperature, high pressure and high speed conditions.

[0070] This detection method and system do not rely on precise prior CAD models and can automatically generate and optimize the ultrasonic testing path and probe posture online based on the measured three-dimensional shape of the component. It achieves efficient, reliable, and fully automated adaptive ultrasonic testing of complex curved surface components, including the following steps:

[0071] S1. Online inverse 3D surface calculation and point cloud data processing;

[0072] S2. Definition of detection area and generation of adaptive detection path;

[0073] S3. Probe attitude optimization and interference collision avoidance;

[0074] S4. Adaptive matching of ultrasound detection parameters;

[0075] S5, closed-loop adaptive execution and data fusion;

[0076] S6. Three-dimensional visualization and evaluation of test results.

[0077] S1, as the starting point of the entire inspection process, aims to obtain the real-time three-dimensional geometric information of the surface component to be inspected.

[0078] The clamped turbine blades to be inspected are subjected to online 3D scanning to obtain their surface point cloud data and reconstruct a triangular mesh surface model.

[0079] Online 3D scanning utilizes a line laser scanner mounted at the end of an inspection robot to perform high-precision scanning of a clamped turbine blade. Line laser scanning technology features high scanning speed and high accuracy, making it suitable for rapid measurement of small to medium-sized complex curved surface components. During the scanning process, the robot carries a line laser sensor and moves along the spanwise direction of the blade, while the laser line emitted by the sensor itself scans along the blade height. Through this combined motion, complete point cloud data of the blade surface can be obtained.

[0080] The surface point cloud data is obtained by scanning, and a series of post-processing operations are performed on the obtained point cloud data, including denoising, filtering and simplification. In the denoising process, environmental noise and outliers introduced during the scanning process are removed. In the filtering process, the point cloud data is smoothed and noise interference is reduced. In the point cloud simplification process, the amount of data is reduced while ensuring the geometric features of the surface, thereby improving the efficiency of subsequent processing.

[0081] After data processing is completed, the Moving Least Squares (MLS) algorithm is used to construct a triangular mesh surface model of the surface component to be inspected based on the processed point cloud data, which serves as the real-time geometric reference for inspection. The MLS algorithm is a method based on local fitting, which can well preserve the detailed features of the surface, and is especially important for turbine blades with complex freeform surfaces.

[0082] Among them, the moving least squares method can preserve the detailed features of the surface well through local weighted fitting, while the Poisson surface reconstruction algorithm can generate a smoother and more complete surface model. In specific implementation, the scanning accuracy can reach ±0.05mm, which fully meets the accuracy requirements of subsequent turbine blade inspection.

[0083] S2 is to define the detection area on the triangular mesh surface model constructed in step S1, and generate the initial ultrasonic probe movement path points and theoretical probe posture. Specifically, on the triangular mesh surface model, the area to be detected is defined, and an adaptive ultrasonic probe movement path is generated according to the detection requirements. This area can be determined based on the failure mode analysis of the component, historical detection records, and design requirements.

[0084] After the detection area is determined, for turbine blades with complex curved surfaces, the equal bow height error method is used to generate initial ultrasonic probe movement path points within the detection area. The equal bow height error method dynamically determines the path point spacing based on the bow height error threshold, which can optimize path efficiency while ensuring detection coverage. It is particularly suitable for complex curved surfaces with large curvature changes and meets the detection requirements of turbine blades.

[0085] For each initial path point, the theoretical normal orientation of the probe is initially calculated based on the normal vector of its local triangular facet. The theoretical normal orientation refers to the orientation when the probe axis coincides with the normal vector of the surface, that is, when the probe axis coincides with the normal vector of the surface. The theoretical probe orientation is determined by the normal vector of the triangular facet at the path point, which serves as the most ideal incident angle in ultrasonic testing, because normal incidence can maximize the intensity of the defect echo signal and minimize the sound beam propagation path error.

[0086] S3 optimizes the probe posture at each path point based on the optimal incident and coupling conditions of the ultrasonic beam, and performs robot motion collision detection and avoidance planning. Collision detection and avoidance are important aspects of ensuring safe operation of the detection.

[0087] The theoretical normal orientation is only an ideal reference orientation. In actual testing, it is necessary to consider the physical size of the ultrasonic probe, the characteristics of the ultrasonic beam, and the kinematic constraints of the robot, and optimize the probe orientation at each path point.

[0088] During attitude optimization, a search and optimization is performed within a spatial cone-shaped range near the theoretical normal attitude. The half-apex angle of this spatial cone-shaped range is set to 15-20° to cover the optimal attitude under various curvature conditions. The dual objective functions are minimizing the normal incident deviation of the ultrasonic beam centerline to the measured area and optimizing the ultrasonic coupling state. The optimal probe yaw angle and roll angle at this point are then determined. Specifically, probe attitude optimization uses minimizing the normal incident deviation of the ultrasonic beam centerline to the measured area and optimizing the ultrasonic coupling state as joint optimization objectives. The search is performed in the neighborhood space of the theoretical normal attitude to obtain the optimized probe yaw angle and roll angle.

[0089] During collision detection, the robot's kinematic model and the 3D models of the probe and tooling are used to continuously detect collisions throughout the entire motion process from the current point to the next path point. If a collision is predicted to occur during the motion, the Random Sampling Method (RRT) is used to insert avoidance points in the robot's joint configuration space to generate a collision-free motion trajectory. In this way, the Fast Extended Random Tree (FAST) algorithm is used to perform collision detection and avoidance planning in the robot's joint space.

[0090] The fast expanding random tree algorithm can quickly search for collision-free paths in a large space, with a higher planning success rate, and is suitable for obstacle avoidance planning in complex environments.

[0091] S4. The selection of ultrasonic testing parameters has a direct impact on the testing results. Different surface curvatures require different testing parameters to obtain the best testing effect. Based on the local surface curvature of the testing point, the emission and scanning parameters of the ultrasonic testing equipment are adaptively adjusted. Turbine blades are usually made of titanium alloy or high-temperature alloy materials, which have high acoustic impedance and sound velocity. The propagation characteristics of sound waves in the material are significantly different from those of aluminum alloy materials. The curvature of different regions of the blade varies greatly. Among them, the main body of the blade is relatively flat, while the leading edge, trailing edge and tenon transition area have large curvature. The adaptive adjustment is specifically as follows: establish a mapping relationship library between surface curvature and ultrasonic testing parameters, and query and call the corresponding pulse repetition frequency, scanning speed and probe selection strategy based on the real-time calculated curvature of the testing point.

[0092] Based on the local Gaussian curvature and average curvature of the surface at the current detection point, as well as the acoustic properties of the component material, the emission parameters of the ultrasonic testing instrument are adaptively adjusted. Curvature is a geometric quantity describing the degree of surface bending. The curvature value can be used to determine whether the current area is a flat region or a high-curvature region, as detailed below:

[0093] For high curvature areas, including the leading edge, trailing edge, and tenon transition area of ​​turbine blades, the repetition frequency is automatically increased to increase the number of sampling points per unit time, the scanning speed is reduced to increase the detection coverage time, and a smaller probe and a higher frequency are selected to focus the sound beam to improve resolution.

[0094] For flat areas, including the middle of turbine blades, the scanning step length and scanning speed can be increased to improve detection efficiency while ensuring detection quality. At the same time, a mapping database of curvature-sound beam diffusion-detection parameters can be established for real-time query and retrieval. This database contains a large amount of experimental data and engineering experience summaries, which can quickly match the optimal detection parameters according to different curvature conditions. When selecting parameters, the acoustic properties of the component materials are also considered, and the ultrasonic penetration rate and attenuation coefficient compensation are adjusted for different material characteristics to ensure the effectiveness and reliability of the detection signal.

[0095] S5, during the actual detection execution phase, the robot is controlled to carry the ultrasonic probe along the optimized path and posture, and ultrasonic detection is performed under constant force control. Spatial pose data and ultrasonic signals are collected simultaneously. The spatial pose data includes the three-dimensional coordinates of the probe center point and the optimized posture angle.

[0096] During the movement, the contact force between the probe and the curved surface is monitored in real time by the six-dimensional force / torque sensor at the end of the robot. Constant force tracking control is performed to ensure stable coupling and achieve constant force tracking control between the probe and the curved surface. Constant force control ensures the detection quality. If the pressure of the probe on the curved surface is too small, it will lead to poor coupling and signal loss. If the pressure is too large, it will damage the workpiece and cause the probe to be damaged. By monitoring the magnitude of the contact force in real time, the movement position of the robot is dynamically adjusted to keep the contact force within the set target range.

[0097] Furthermore, high-precision detection is achieved through data fusion. The spatial coordinates (X,Y,Z), optimized attitude angles (A,B), corresponding ultrasonic A-scan signals, and processed C-scan image data of each detection point are strictly synchronized and bound. The ultrasonic A-scan signal is the original echo signal of the ultrasonic probe, which contains the temporal information of the defect. The C-scan image is a two-dimensional projection image after spatial mapping of multiple A-scan signals, which can intuitively display the distribution of defects on the scanning plane.

[0098] The synchronized data forms a full-matrix detection dataset with three-dimensional spatial location information, laying the foundation for subsequent three-dimensional visualization evaluation.

[0099] The key technical parameters of the aero-engine turbine blade are as follows: the ultrasonic probe frequency is 5MHz, the water immersion focusing method is adopted, the point cloud scanning accuracy reaches ±0.05mm, the pulse repetition frequency in the high curvature area is set to 5kHz, the scanning speed is set to 10mm / s, the scanning speed in the flat area is increased to 50mm / s, and the constant force control accuracy is 2N±0.5N.

[0100] S6, the ultrasonic test data is fused with the triangular mesh surface model for three-dimensional visualization and evaluation, presenting the test data to the user in an intuitive way. Example 1 successfully detected a small fatigue crack located 1.2 mm from the surface on the blade back, with a length of 3 mm, verifying the detection efficiency of the method.

[0101] The full matrix detection data obtained in step S5 is mapped back onto the triangular mesh surface model constructed in step S1 to generate a 3D color image and isosurface map for visualization.

[0102] In the 3D visualization process, based on the ultrasonic signal characteristics of each detection point, including echo amplitude, phase, and spectrum, the corresponding defect indication value is calculated and these values ​​are mapped to the corresponding positions on the surface. Through the depth of color and the density of contour lines, the precise location, burial depth, and equivalent size of the defect on the component surface are displayed intuitively.

[0103] The burial depth information is calculated using the flight time of the ultrasonic signal with an accuracy of 0.1 mm. The equivalent size is obtained by comparing the echo amplitude of the artificial injury of the standard defect.

[0104] The 3D visualization of the inspection results provides a highly intuitive basis for analysis. Operators can rotate, scale, and section the 3D model to observe the distribution and severity of defects from any angle, providing reliable support for subsequent maintenance decisions.

[0105] In summary, this detection method can detect microcracks with a depth of 1.2 mm and a length of only 3 mm, indicating that it has high detection sensitivity when applied to the inspection of aero-engine turbine blades. This is due to the following factors:

[0106] First, high-precision 3D scanning and surface reconstruction ensure the accuracy of path planning. Second, attitude optimization ensures that the ultrasonic beam is incident at the best angle, maximizing the defect echo signal. Third, adaptive parameter adjustment ensures that the best detection effect can be obtained in different curvature areas. Finally, constant force control ensures the stability and repeatability of the detection signal.

[0107] Please see Figures 2-7 An adaptive ultrasonic testing system for curved surface components is disclosed, which enables adaptive ultrasonic testing of curved surface components. The system includes a motion actuator, an adaptive clamping mechanism, a composite end effector, an ultrasonic transmitter / receiver and data acquisition unit, a central control and processing unit, and a visualization and human-computer interaction unit.

[0108] The motion actuator is the execution layer of the system, responsible for driving the ultrasonic probe and the 3D scanner to move precisely in three-dimensional space. It adopts a three-axis gantry scanner.

[0109] The three-axis gantry scanner adopts a linear guide and slider structure, and realizes spatial movement through three mutually perpendicular linear axes. It is suitable for the inspection of large components. When inspecting large workpieces, it uses its built-in adaptive clamping mechanism to adaptively center and limit the workpiece, ensuring the stability and accuracy of the workpiece during the inspection process. The adaptive clamping mechanism stably limits the workpiece.

[0110] The adaptive clamping mechanism includes an ultrasonic scanning frame 1, a water tank 2, a longitudinal electric slide rail 3, a transverse electric slide rail 4, a vertical electric slide rail 5, a composite end effector 6, a loading frustum 7, a support bar 8, a guide rail 9, a limiting slide groove 10, a slide rod 11, a limiting clamp 12, a drive disc 13, a drive motor 14, a limiting arc groove 15, a drive column 16, a flexible airbag 17, and an anti-slip strip 18.

[0111] A water tank 2 is embedded in the inner side of the ultrasonic scanning frame 1. The four corners of the bottom of the ultrasonic scanning frame 1 are supported by adjustable supports. The top two sides of the ultrasonic scanning frame 1 are symmetrically equipped with longitudinal electric slide rails 3. The top of the longitudinal electric slide rail 3 is slidably connected to a transverse electric slide rail 4. The side of the transverse electric slide rail 4 is slidably connected to a vertical electric slide rail 5. A composite end effector 6 is installed at the bottom sliding end of the vertical electric slide rail 5.

[0112] A truncated cone 7 is installed on the inner bottom of the water tank 2. Support bars 8 are evenly arranged on the top surface of the truncated cone 7. The edge of the truncated cone 7 is fixedly installed on the inner wall of the water tank 2 via a guide rail 9. A limiting groove 10 is opened inside the guide rail 9. A sliding rod 11 is slidably connected in the limiting groove 10. A limiting clamp 12 is fixedly connected to the top of one end of the sliding rod 11.

[0113] The bottom of the loading platform 7 is rotatably connected to a drive disk 13. The bottom of the drive disk 13 is fixedly connected to the output end of the drive motor 14 via a transmission shaft. The drive disk 13 has four sets of limiting arc grooves 15 inside. The bottom end of the slide rod 11 slides along the limiting arc grooves 15 via a drive column 16. A flexible airbag 17 is provided in the middle of the inner side of the limiting clamp 12, and an anti-slip strip 18 is provided on the inner side edge of the limiting clamp 12.

[0114] The composite end effector is the core execution component of the system. It is installed at the end of the motion actuator and directly contacts the curved surface component to be inspected to perform the inspection task. It integrates an ultrasonic probe, a 3D scanner and a force sensor.

[0115] The ultrasonic probe is a key component for ultrasonic testing. It is a water immersion type, in which the probe is completely immersed in water. The selection of the probe needs to be determined according to the thickness of the material being tested, the type of defect, and the testing requirements. Common types include straight probes, angle probes, and focusing probes.

[0116] 3D scanners are used to acquire three-dimensional geometric information of workpiece surfaces online. 3D scanners are line laser scanners, which acquire point cloud data by emitting laser lines and capturing reflected light. They have a fast scanning speed and high accuracy.

[0117] The force sensor adopts a six-dimensional force / torque sensor, which can simultaneously monitor forces and torques in three directions. It is installed between the probe and the robot end effector and can sense the contact state between the probe and the curved surface in real time. By monitoring the magnitude of the contact force, constant force tracking control is achieved, ensuring coupling stability while avoiding damage to the workpiece.

[0118] The ultrasonic transmitting / receiving and data acquisition unit is responsible for generating, receiving, and acquiring ultrasonic signals. It works in conjunction with the robot control system and computer system to achieve accurate acquisition and processing of ultrasonic detection signals.

[0119] The central control and processing unit is responsible for coordinating the work of each part and executing the core algorithm functions. This unit is used to run the algorithm of steps S1-S6 and control the robot, scanner and ultrasonic equipment to work together. It adopts a high-performance industrial computer and has powerful computing power and real-time performance.

[0120] In terms of data processing, it is responsible for processing point cloud data collected by 3D scanners in real time and executing surface reconstruction algorithms to build triangular mesh models;

[0121] In terms of path planning, an initial detection path is generated based on the detection area definition and surface model, and then the path is optimized.

[0122] In terms of attitude optimization, the optimal probe yaw angle and roll angle are calculated to ensure optimal sound beam incidence;

[0123] In terms of parameter matching, the parameter database is queried based on curvature information to perform adaptive parameter adjustments.

[0124] In terms of motion control, the planning results are converted into robot motion commands, and the execution process is monitored.

[0125] In terms of data fusion, spatial pose data and ultrasonic signal data are synchronously bound to form a complete dataset;

[0126] The visualization and human-computer interaction unit provides a 3D visualization and human-computer interaction interface, enabling operators to intuitively monitor the system's operating status, view test results, and perform necessary intervention operations;

[0127] The monitoring interface displays key information such as the robot's motion status, the probe's current position and posture, the magnitude of the contact force, and the ultrasonic signal waveform in real time, allowing operators to monitor the detection process at any time.

[0128] In terms of result presentation, the detection data will be displayed in various forms such as 3D color pictures, isosurface maps, and animations to help operators intuitively understand the distribution and severity of defects;

[0129] In terms of interactive operation, it provides functions such as path point editing, parameter adjustment, and region redefinition, enabling operators to optimize and adjust the detection scheme according to the actual situation.

[0130] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An ultrasonic adaptive testing method for curved surface components, characterized in that: Includes the following steps: S1. Online inverse 3D surface calculation and point cloud data processing; The clamped curved surface component to be inspected is scanned online to obtain its surface point cloud data. The 3D point cloud data of its surface is obtained by scanning, and a series of post-processing operations are performed on the obtained point cloud data. After data processing is completed, the moving least squares method or Poisson surface reconstruction algorithm is used to construct a triangular mesh surface model of the surface component to be inspected based on the processed point cloud data, which serves as the real-time geometric reference for inspection. Online 3D scanning utilizes a line laser scanner or structured light 3D scanner installed at the end of an inspection robot to perform high-precision scanning of the curved surface component to be inspected, which is clamped in place. S2. Definition of detection area and generation of adaptive detection path; The detection area is defined on the triangular mesh surface model constructed in step S1, and the initial ultrasonic probe movement path points and theoretical probe posture are generated. After the detection area is determined, the initial ultrasonic probe movement path points are generated within the detection area based on the isoparametric method or the equal bow height error method. For each initial path point, the theoretical normal orientation of the probe is initially calculated based on the normal vector of its local triangular facet. S3. Probe attitude optimization and interference collision avoidance; Based on the optimal incident and coupling conditions of the ultrasonic beam, the probe posture at each path point is optimized, and robot motion collision detection and avoidance planning are performed. During attitude optimization, the search optimization is performed within a spatial cone-shaped range near the theoretical normal attitude. The dual objective functions are to minimize the normal incident deviation of the ultrasonic beam centerline to the measured area and to optimize the ultrasonic coupling state. The optimal probe yaw angle and roll angle at this point are then determined. During collision detection, the robot's kinematic model and the 3D models of the probe and tooling are used to continuously detect collisions throughout the entire motion process from the current point to the next path point. If a collision is predicted to occur during the motion, an avoidance point is inserted in the robot's joint configuration space using the artificial potential field method or random sampling method to generate a collision-free motion trajectory. S4. Adaptive matching of ultrasound detection parameters; Based on the local Gaussian curvature and average curvature of the surface at the current detection point, as well as the acoustic properties of the component material, the emission parameters of the ultrasonic testing instrument are adaptively adjusted. Establish a mapping relationship library between surface curvature and ultrasonic testing parameters. Based on the real-time calculated curvature of the testing point, query and call the corresponding pulse repetition frequency, scanning speed and probe selection strategy. At the same time, establish a mapping relationship database between curvature-beam diffusion-test parameters for real-time query and call. S5, closed-loop adaptive execution and data fusion; The robot is controlled to carry an ultrasonic probe along an optimized path and posture, and performs ultrasonic testing under constant force control. Spatial pose data and ultrasonic signals are collected simultaneously. The spatial pose data includes the three-dimensional coordinates of the probe center point and the optimized posture angle. High-precision detection is achieved through data fusion. The spatial coordinates (X,Y,Z), optimized attitude angles (A,B), corresponding ultrasonic A-scan signals, and processed C-scan image data of each detection point are strictly synchronized and bound. The synchronized and bound data form a full matrix detection dataset with three-dimensional spatial position information. S6. Three-dimensional visualization and evaluation of test results; The full matrix detection data obtained in step S5 is mapped back onto the triangular mesh surface model constructed in step S1 to generate a 3D color image and isosurface map for visualization.

2. The ultrasonic adaptive testing method for curved surface components according to claim 1, characterized in that: S1 involves acquiring the real-time three-dimensional geometric information of the component to be inspected and reconstructing a triangular mesh surface model. Specific operations include denoising, filtering, and simplification. The line laser scanner is suitable for rapid measurement of small and medium-sized complex curved surface components, while the structured light 3D scanner is suitable for comprehensive measurement of complex shapes. Moving least squares method preserves the detailed features of the surface through local weighted fitting, while Poisson surface reconstruction algorithm can generate a smoother and more complete surface model.

3. The ultrasonic adaptive testing method for curved surface components according to claim 2, characterized in that: Specifically, S2 defines the area to be detected on the triangular mesh surface model and generates an adaptive ultrasonic probe movement path according to the detection requirements. This area is determined based on the failure mode analysis of the component, historical detection records, and design requirements. The isoparametric method evenly distributes path points along the parameterization direction of the surface, which is suitable for surfaces with regular parameter domains. The equal bow height error method dynamically determines the spacing between path points based on the bow height error threshold, which is suitable for complex surfaces with large curvature variations. The theoretical normal orientation refers to the orientation when the probe axis coincides with the surface normal vector. That is, the probe axis coincides with the surface normal vector. The theoretical probe orientation is determined by the normal vector of the triangular facet at the path point.

4. The ultrasonic adaptive testing method for curved surface components according to claim 1, characterized in that: In the actual detection process, the physical size of the ultrasonic probe, the characteristics of the ultrasonic beam, and the kinematic constraints of the robot need to be considered to optimize the probe posture at each path point. Specifically, probe attitude optimization takes minimizing the normal incident deviation of the ultrasonic beam centerline to the measured area and optimizing the ultrasonic coupling state as the joint optimization objectives. It searches within the neighborhood space of the theoretical normal attitude to obtain the optimized probe yaw angle and roll angle.

5. The ultrasonic adaptive testing method for curved surface components according to claim 1, characterized in that: In S4, the selection of ultrasonic testing parameters has a direct impact on the testing results. Different surface regions with different curvatures require different testing parameters to obtain the best testing effect. The emission and scanning parameters of the ultrasonic testing equipment are adaptively adjusted according to the local surface curvature of the testing point. For high curvature regions, the repetition frequency is automatically increased to increase the number of sampling points per unit time, the scanning speed is reduced to increase the detection coverage time, and a smaller probe and a higher frequency are selected to focus the sound beam to improve resolution. For flat areas, increasing the scanning step length and scanning speed can improve detection efficiency while ensuring detection quality. Curvature is a geometric quantity that describes the degree of curvature of a surface. By using the curvature value, we can determine whether the current region is a flat region or a region with high curvature.

6. The ultrasonic adaptive testing method for curved surface components according to claim 1, characterized in that: S5, the ultrasonic A-scan signal is the original echo signal of the ultrasonic probe, which contains the time domain information of the defect, and the C-scan image is a two-dimensional projection image after spatial mapping of multiple A-scan signals, which can intuitively display the distribution of the defect on the scanning plane. By using a six-dimensional force / torque sensor at the robot's end effector to monitor the contact force between the probe and the curved surface in real time, constant force tracking control is achieved. By monitoring the magnitude of the contact force in real time, the robot's movement position is dynamically adjusted to keep the contact force within the set target range.

7. The ultrasonic adaptive testing method for curved surface components according to claim 6, characterized in that: S6 integrates ultrasonic testing data with a triangular mesh surface model for three-dimensional visualization and evaluation, presenting the testing data to the user in an intuitive way. In the 3D visualization process, based on the ultrasonic signal characteristics of each detection point, the corresponding defect indication value is calculated and these values ​​are mapped to the corresponding positions on the surface. The precise location, depth and equivalent size of the defect on the component surface are displayed intuitively by the shades of color and the density of contour lines. The burial depth information is calculated using the flight time of the ultrasonic signal, while the equivalent size is obtained by comparing the echo amplitude of the artificial injury of the standard defect. The 3D visualization of the inspection results provides a highly intuitive basis for analysis. Operators can rotate, scale, and section the 3D model to observe the distribution and severity of defects from any angle, providing reliable support for subsequent maintenance decisions.

8. An ultrasonic adaptive testing system for curved surface components for implementing the method according to any one of claims 1-7, characterized in that: The system enables adaptive ultrasonic testing of curved components, including a motion actuator, an adaptive clamping mechanism, a composite end effector, an ultrasonic transmitter / receiver and data acquisition unit, a central control and processing unit, and a visualization and human-machine interaction unit.

9. The ultrasonic adaptive testing system for curved surface components according to claim 8, characterized in that: The motion actuator is the execution layer of the system, responsible for driving the ultrasonic probe and the 3D scanner to move precisely in three-dimensional space, using a six-axis industrial robot or a three-axis gantry scanner; Six-axis industrial robots can realize various motion trajectories required for the inspection of complex curved surfaces. In inspection applications, a composite end effector is installed at the end of the robot, which is suitable for the inspection of small and medium-sized complex curved surface components. The three-axis gantry scanner adopts a linear guide and slider structure, and realizes spatial movement through three mutually perpendicular linear axes. It is suitable for the inspection of large components. When inspecting large workpieces, it uses its built-in adaptive clamping mechanism to adaptively center and limit the workpiece. The adaptive clamping mechanism will stably limit the workpiece being inspected. The composite end effector is the core execution component of the system. It is installed at the end of the motion actuator and directly contacts the component to be inspected to perform the inspection task. It integrates an ultrasonic probe, a 3D scanner and a force sensor. The ultrasonic probe is a key component for ultrasonic testing. It can be coupled by water immersion or water spray. Water immersion coupling involves completely immersing the probe in water, while water spray coupling uses a nozzle to spray a water film between the probe and the workpiece to achieve coupling. The choice of probe needs to be determined based on the thickness of the material being tested, the type of defect, and the testing requirements. 3D scanners are used to acquire three-dimensional geometric information of workpiece surfaces online. 3D scanners can be either line laser scanners or structured light 3D scanners. Line laser scanners acquire point cloud data by emitting laser lines and capturing reflected light, while structured light 3D scanners acquire rich information by projecting gratings with specific patterns onto the object surface and calculating three-dimensional coordinates based on the deformation of the gratings. The choice of scanner depends on the characteristics of the workpiece and the inspection requirements. The force sensor is a six-dimensional force / torque sensor, which can simultaneously monitor forces and torques in three directions. It is installed between the probe and the robot end effector and can sense the contact state between the probe and the curved surface in real time. By monitoring the magnitude of the contact force, constant force tracking control can be achieved.

10. The ultrasonic adaptive testing system for curved surface components according to claim 8, characterized in that: The ultrasonic transmitting / receiving and data acquisition unit is responsible for generating, receiving, and acquiring ultrasonic signals. It works in conjunction with the robot control system and the computer system to achieve accurate acquisition and processing of ultrasonic detection signals. The central control and processing unit is responsible for coordinating the work of each part and executing the core algorithm functions. This unit is used to run the algorithm of steps S1-S6 and control the robot, scanner and ultrasonic equipment to work together. It adopts a high-performance industrial computer and has powerful computing power and real-time performance. The visualization and human-computer interaction unit provides a three-dimensional visualization and human-computer interaction interface, enabling operators to intuitively monitor the system's operating status, view detection results, and perform necessary intervention operations.

Images