A method for trajectory planning in ultrasonic nondestructive testing of model-free components

Abstract

This invention discloses a method for trajectory planning in ultrasonic nondestructive testing of model-less components. The method acquires all surface point cloud information of the component under test through scanning. The point cloud information is preprocessed spatially to align the centroid of the component horizontally. The detection path of the component is pre-planned in the point cloud file, and the point cloud along the preset detection path is extracted. High-order curve fitting is performed in the software, and outliers are eliminated to generate a scanning path line on the surface of the component. According to preset scanning parameters such as the scanning interval, the scanning path line on the surface of the component is discretized. The path line is extended along the normal to the sampling point on the fitted curve to obtain the probe position corresponding to the sampling point. This position is used as the motion coordinate of the robotic arm flange holding the probe, generating a robotic arm flange motion and robotic arm pose matrix, thus realizing the generation of the robotic arm scanning trajectory. This invention enables automatic detection of model-less components by a robotic arm holding a probe.

Description

Technical Field

[0001] This invention relates to a trajectory planning method for nondestructive testing of model-free components, specifically a trajectory planning method for an ultrasonic testing system containing a multi-degree-of-freedom robotic arm, belonging to the field of nondestructive testing. Background Technology

[0002] The rapid development of the modern aviation industry has placed increasingly higher demands on the performance and shape of complex curved surface workpieces. The detection and characterization of internal defects in complex-shaped components has spurred the rapid development of ultrasonic non-destructive testing technology. High-strength, lightweight alloy components with complex curved surfaces are widely used in aerospace propulsion systems and load-bearing structures. Currently, the aerospace industry experiences numerous and long-duration flight missions, placing higher demands on the fatigue life of key components, including engines. Due to the complex and costly machining process of complex curved surface workpieces, they are widely used in critical load-bearing structures and core components of power systems, such as the upper wing skin of space shuttles, aircraft shells, and engine blades in the aerospace field. The inspection of aero-engine blades is highly complex, and effective detection and quantification of blade defects are difficult. Fatigue, cracks, and corrosion are several common and important defects that occur in aero-engine blades. Currently, non-destructive testing methods are mainly used for defects in in-service aero-engine blades.

[0003] Ultrasonic non-destructive testing (UNDT), as an important testing method in non-destructive testing, is widely used in various fields due to its advantages such as low testing cost, wide range of test objects, high sensitivity, and harmlessness to humans. Ultrasonic imaging technology, based on the directivity, transmission, and reflection characteristics of ultrasonic waves, characterizes the acoustic properties of the test piece based on parameters such as the time difference and intensity of the received signal, and indirectly achieves internal imaging visualization through post-processing. The directivity of ultrasonic waves requires a strict definition of the incident angle of the transducer, especially for complex curved surfaces, ensuring that the transducer's pose is aligned with the normal direction of the tested surface to obtain the best testing results. To ensure that the ultrasonic transducer's pose is perpendicular to the normal direction of the tested surface, high requirements are placed on the spatial motion capability of the transducer. Traditional three-axis and five-axis motion mechanisms are typically used for the inspection of plate-shaped components. When multi-degree-of-freedom robots carry ultrasonic transducers to complete automated inspection tasks, they can ensure that the ultrasonic transducer is perpendicular to the workpiece surface and the distance remains stable, but high precision is required for the robot's trajectory. Six-degree-of-freedom (DOF) industrial robotic arms possess high-precision repeatability and flexible pose transformation capabilities. Using a six-DOF industrial robotic arm as the main motion structure, ultrasonic non-destructive testing (NDT) is performed on the tested components by holding an ultrasonic probe. Applying multi-DOF robotic arms to NDT systems effectively leverages their flexibility and testing range. However, path planning for robotic arms in NDT systems has always been a key challenge. Most existing testing systems, both domestically and internationally, require a model of the tested component for path planning, and there is currently no effective method for testing components without a model. Addressing the current technological limitations, a trajectory planning algorithm for six-DOF robotic arms is needed to control the robotic arm's contour-following motion along the tested surface during testing, enabling high-precision ultrasonic NDT for complex curved surface components without a model. Summary of the Invention

[0004] This invention addresses the trajectory control of ultrasonic nondestructive testing for model-less components, and innovatively proposes a method for high-precision scanning by planning the motion trajectory of a robotic arm using spatial point clouds.

[0005] This invention provides a method for high-precision motion path planning in an ultrasonic testing system incorporating a multi-degree-of-freedom robotic arm. The method involves acquiring surface point cloud information of the tested environment and components using data scanning. The point cloud files are then aligned in space to align with the robotic arm in the same world coordinate system. Subsequently, based on the testing path requirements, the point cloud information of the components along the required path is extracted, and a high-order curve is fitted in space to generate the surface route of the tested component. Following the testing plan, the sampling points on the fitted curve are discretized to obtain the spatial coordinates of the discrete points. These discrete points are then extended along the normal to the fitted curve to the required testing distance, yielding the spatial coordinates of the six-degree-of-freedom robotic arm flange. Finally, the spatial coordinates of all sampling points along the single-line scanning path are extended to obtain the path point information of the flange during the six-degree-of-freedom robotic arm's testing process, thus achieving high-precision path planning for the robotic arm's scanning path.

[0006] To achieve the above objectives, the technical solution adopted in this invention is to use laser reverse engineering technology to obtain the surface point cloud information of the component under test, then import it into point cloud processing software to perform spatial correction on the component point cloud, perform path pre-planning according to actual detection requirements, import the point cloud file into MATLAB for data processing, perform high-order curve fitting on the point cloud of the pre-planned path to obtain a high-precision surface path line, then obtain each fitted scanning path line and the sampling points on each line according to the scanning interval and step distance of the automated detection plan, extend the normal of the sampling points to obtain the corresponding robotic arm flange coordinates, connect all the sampling point flange coordinates in space to obtain the robotic arm scanning path, and finally use a six-degree-of-freedom robotic arm to hold the ultrasonic probe and perform ultrasonic detection according to the planned path.

[0007] A method for trajectory planning in ultrasonic nondestructive testing of model-free components utilizes laser reverse engineering technology to obtain point cloud files, and performs multi-degree-of-freedom trajectory planning based on the point cloud files. The specific implementation steps of this method include:

[0008] S1. Obtain the point cloud file of the surface of the component under test, and perform spatial correction on the point cloud file to ensure that the center point of the component under test is in a horizontal position in space.

[0009] S2. Pre-plan the ultrasonic testing path, extract the point cloud along the pre-planned path, fit it with a high-order curve, and fit the component surface testing path corresponding to each scanning path in turn according to the pre-planned testing parameters to obtain the surface testing path of the component under test.

[0010] S3. Extend the normal vector of all sampling points on the fitted curve to obtain the trajectory coordinates of the robotic arm flange and the pose matrix of the robotic arm during the detection. Connect all the spatial motion points of the flange with the pose matrix of the robotic arm to finally guide the multi-degree-of-freedom robotic arm through the detection path.

[0011] In step S1, for components with complex shapes, manual inspection is often used in ultrasonic testing. For components without a model, commonly used automated inspection equipment requires path planning based on a model of the component under test. Therefore, before inspection, the component under test is laser-scanned to obtain its point cloud file. The point cloud file is then imported into point cloud processing software. The centroid of the component is obtained in space, and the point cloud is adjusted in the software to align its centroid with the spatial origin, ensuring that the point cloud coordinates of the component are consistent with the actual inspection position when placed horizontally or vertically.

[0012] In step S2, based on the actual testing requirements, the point cloud file is first pre-planned for testing. The scanning path is planned according to the commonly used "bow" pattern in ultrasonic C-scanning, and the scanning interval and step distance in ultrasonic testing are set. The point cloud along the actual testing path is extracted based on the step distance, and a high-order curve is fitted for the point cloud involved in each path. Points with excessive errors are deleted, and a high-precision curve is fitted to obtain the actual surface testing path of the component under test.

[0013] In step S3, based on the pre-planned path and pre-detection parameters, all pre-set detection sampling points are separated, and their spatial coordinates on the fitted curve are obtained. These coordinates are then extended along the normal of the fitted curve, with the extension distance including the distance between the ultrasonic water jet and the length of the ultrasonic probe clamp, resulting in the spatial coordinates of the multi-degree-of-freedom robotic arm flange. The robotic arm motion matrix is ​​derived based on the flange coordinates corresponding to all sampling points, yielding the pose of the robotic arm to each sampling point. Connecting the robotic arm poses corresponding to all sampling points provides the robotic arm motion trajectory, i.e., the detection path.

[0014] Compared with the prior art, the present invention has the following advantages.

[0015] 1. This invention utilizes point cloud information files, selects corresponding point clouds by pre-setting paths within the point cloud files, and fits surface scan lines using the point cloud information, thus realizing scan path planning for complex-shaped components in ultrasonic non-destructive testing. Currently, complex-shaped components are typically inspected using traditional handheld devices. This invention provides a novel detection path planning technology for automated and intelligent non-destructive testing.

[0016] 2. This invention, based on the scanning path of components in point cloud files, combines motion trajectory planning with a multi-degree-of-freedom robotic arm used in practical applications to achieve automated inspection of model-less components. In current technologies, automated non-destructive testing systems require a model of the component under test for trajectory planning. This invention enables trajectory planning for model-less components within a multi-degree-of-freedom testing system, overcoming the current bottleneck in the inspection of model-less components.

[0017] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention. Attached Figure Description

[0018] Figure 1 This is a point cloud file image obtained in an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of the high-order path fitting of the present invention;

[0020] Figure 3 This is a schematic diagram of the robotic arm trajectory matrix of the present invention; Detailed Implementation

[0021] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0022] At present, there are more and more new practical applications of components with complex shapes. The surface and interior of the components often suffer micro-damage during production and practical application. However, due to the service scenarios of the components, their inspection often requires the use of non-destructive testing methods.

[0023] This invention provides a method for planning the trajectory of nondestructive testing of model-less components. First, a point cloud scan file of the component under test is obtained using laser reverse engineering technology, containing information about the component's outer surface. Second, the point cloud scan file is imported into point cloud processing software, and its spatial orientation is adjusted to align its centroid with the horizontal plane. Then, a pre-planning of the scanning path is performed on the point cloud file, selecting the point clouds traversed by the planned path and performing high-order curve fitting to generate a high-precision path line on the surface of the component under test. Finally, based on the pre-planned scanning parameters, the generated surface path line is discretized and extended according to its normal extension line to obtain the coordinate matrix of the robotic arm flange corresponding to each scanning point. Connecting all flange coordinate matrices yields the complete detection trajectory of the model-less component.

[0024] The equipment used in this embodiment of the invention is a multi-degree-of-freedom ultrasonic non-destructive testing system, comprising a multi-degree-of-freedom robotic arm with an ultrasonic probe holder, an ultrasonic pulse exciter / receiver, a 25MHz ultrasonic immersion probe, a digital acquisition card, and an industrial control computer for system control. In actual testing, the industrial control computer controls the robotic arm, which, in conjunction with the ultrasonic probe holder, holds the immersion probe. A water spray structure within the holder utilizes ultrasonic water jet coupling to perform ultrasonic non-destructive testing on the tested component. The ultrasonic signal excitation and reception are achieved using the ultrasonic pulse exciter / receiver, and the echo signal is collected by the digital acquisition card and fed back to the industrial control computer. Finally, the collected data is analyzed and processed to realize ultrasonic non-destructive testing of the tested component. The specific implementation steps are as follows:

[0025] S1. Obtain the point cloud file of the surface of the component under test, and perform spatial correction on the point cloud file. The corrected point cloud file is shown as follows. Figure 1 As shown, the center point of the measured component is ensured to be in a horizontal position in space. In this embodiment, a handheld laser scanner is used to acquire the point cloud file, which can perform scanning of the measured component with an accuracy of 0.05mm and generate a point cloud file in .ply format. A retired blade from an aerospace engine is scanned; the blade is 4.5cm long, 2.2cm wide, and 7cm high. The entire blade is scanned using a handheld laser scanner, and information from each surface is converted into point cloud coordinates. The point cloud file is displayed, and subsequent calculations are performed in MATLAB.

[0026] S2. Process the point cloud file in MATLAB, pre-planning the ultrasound detection path. The scanning path is planned according to the commonly used "bow" pattern in ultrasound C-scans, setting the measurement range and the step interval and route during scanning. Then, extract the point cloud data along the measurement path, process each scanning path individually, and perform curve fitting. The fitting process is also performed in MATLAB. Note that large fluctuations in single-point errors may cause issues in the fitted screenshot. Figure 2 As shown, depending on the actual situation, high-order Gaussian fitting and Fourier fitting are usually used in the fitting process. In this embodiment, second-order Fourier fitting is used. Figure 2 The equation of the fitted curve shown is: f(x) = 5.501 - 6.111 * cos(x * 0.07474) - 0.3124 * sin(x * 0.07474) + 0.5684 * cos(2 * x * 0.07474) + 0.2987 * sin(2 * x * 0.07474). Its goodness-of-fit parameters are: SSE: 2464; R-squared: 0.7575; RMSE: 1.559.

[0027] S3. The path line generated after fitting is segmented according to the pre-designed scanning parameters to obtain the actual sampling points on the surface of the component being tested. The sampling points are extended in the reverse direction according to the normal of the fitted curve, that is, extended towards the water jet in the actual test; the position of the ultrasonic probe during the actual test is obtained, that is, the coordinate position of the robotic arm flange. Using all the coordinates of the robotic arm flange under the preset test path, and based on the robotic arm pose transformation matrix, the path file of the robotic arm is written to obtain the actual test path file. A partial screenshot of the path file generated in the embodiment is shown below. Figure 3 As shown, trajectory planning for a multi-degree-of-freedom ultrasonic testing system is realized.

[0028] This invention, based on curve fitting and utilizing the characteristics of point cloud files, proposes an ultrasonic non-destructive testing (NDT) path planning method for model-less components during the pre-scanning stage before NDT. This method ensures that the testing system can set a high-precision testing path when inspecting model-less components, guaranteeing that ultrasonic waves are incident perpendicularly to the surface of the tested component in real time, receiving high-quality ultrasonic signals, and acquiring surface and internal acoustic information of the tested component. Based on the robotic arm pose description and coordinate transformation matrix, after obtaining the apparent path information of the model-less tested component, this invention discretizes the sampling points and extends the normals to obtain the robotic arm flange coordinates during actual testing. Connecting all flange coordinates sequentially yields the robotic arm pose transformation matrix during testing, enabling automated ultrasonic NDT of model-less components.

[0029] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for trajectory planning in nondestructive testing of model-less components, characterized in that, To automate the ultrasonic testing of model-less components using a robotic arm to hold an ultrasonic probe, the following steps are included: S1. Obtain the point cloud file of the surface of the component under test, and perform spatial correction on the point cloud file to ensure that the centroid of the component under test is in a horizontal position in space. S2. Pre-plan the ultrasonic testing path, extract the point cloud along the pre-planned path, fit it with a high-order curve, and fit the component surface testing path corresponding to each scanning path in turn according to the pre-planned testing parameters to obtain the surface testing path of the component under test. The point cloud file is pre-planned for detection. Scanning parameters including scanning range, scanning interval, step distance, and scanning direction are set, and the point cloud data passed through the preset path is extracted again. The surface point cloud on each scanning path is extracted in sequence, and high-order curve fitting is performed in space. Some points with excessive errors due to scanning accuracy or selection accuracy are removed. The fitted high-order curve is the scanning path line on the surface of the component being measured. S3. All sampling points are extended normally on the fitted curve to obtain the trajectory coordinates of the robotic arm flange and the pose matrix of the robotic arm during the detection. All spatial motion points of the flange are connected to the pose matrix of the robotic arm to guide the detection path planning of the multi-degree-of-freedom robotic arm. According to the preset scanning parameters, all sampling points on the surface path are discretized, and the sampling points are extended normally on the fitted curve in turn to obtain the position of the ultrasonic probe in the actual detection. After obtaining the ultrasonic probe positions corresponding to all sampling points in the preset scanning path, these positions are used as the flange space coordinates of the multi-degree-of-freedom robotic arm during the inspection process. The corresponding flange motion trajectory and the robotic arm pose matrix are obtained to guide the robotic arm movement and realize automated ultrasonic non-destructive testing trajectory planning for modelless components.

2. The method for non-destructive testing trajectory planning of model-less components according to claim 1, characterized in that: The point cloud information of the surface of the component under test is obtained by using a scanning device, and the point cloud file is spatially corrected to ensure that its centroid is in a horizontal position in space.

Images