Method and system for detecting welding seam of metro vehicle framework

Abstract

The invention discloses a metro vehicle framework welding seam detection method and system, and the method comprises the steps: S1, generating a digital twin model of a framework according to a design model of a metro vehicle framework or actually collected point cloud data; s2, loading a digital twinborn model and a robot model of the framework in an ROS simulation environment, and building a simulation test environment; s3, marking defect positions in a simulation test environment; s4, in a simulation test environment, performing S41 path planning and S42 path simulation; and S5, according to the defect position and the path plan in the simulation test environment, enabling the robot to execute an actual detection process. According to the method, the digital twinborn model of the framework is established, the framework is subjected to simulated flaw detection, the flaw detection position is freely configured, and the planning path is preset, so that guidance is provided for actual flaw detection, meanwhile, real-time data of the framework can be acquired through the robot, and comparison and analysis with the digital twinborn model are carried out, so that accurate flaw detection of the framework is realized.

Description

Technical Field

[0001] The invention relates to the field of subway vehicle maintenance, and in particular to a method and system for detecting welds of a subway vehicle frame. Background Art

[0002] At present, the common method for inspecting frame welds in subway vehicle maintenance is manual magnetic particle inspection. After the vehicle is dismantled, the frame needs to go through a series of processes including cleaning, paint removal, cleaning, magnetization, inspection, demagnetization, cleaning, and painting. This method has the disadvantages of long cycle, environmental pollution, low efficiency, high cost, and low level of data informationization and intelligence.

[0003] In the prior art, in order to overcome the above-mentioned shortcomings of manual magnetic particle inspection, a robotic arm is often introduced to perform fixed-point teaching eddy current scanning of welds to realize automatic detection of weld data collection. However, there are defects such as a fixed and single operation path, weak adaptability to surface protrusions such as welding slag, and an inflexible scanning range. Summary of the invention

[0004] In view of this, the present invention proposes a robot online detection implementation method, which can solve the above problems to a certain extent.

[0005] Specifically, the technical solution of the present invention is: a method for detecting a weld of a subway vehicle frame, comprising the following steps: S1: Generate a digital twin model of the subway vehicle structure based on the design model of the structure or the point cloud data actually collected; S2: Load the digital twin model and robot model of the architecture in the ROS simulation environment to build a simulation test environment; S3: Marking the defect location under the simulation test environment; S4: Performing S41 path planning and S42 path simulation under the simulation test environment; S5: According to the defect location and path planning in the simulation test environment, the robot executes the actual detection process.

[0006] According to a preferred embodiment, in S1, If a design model exists, a digital twin model of the architecture is directly generated; If a design model exists, a digital twin model of the architecture is directly generated; If the design model does not exist, the appearance of the structure is manually captured using a 3D camera, and the complete structure point cloud is obtained by splicing, and the point cloud modeling is loaded through software to obtain a digital twin model of the structure.

[0007] According to a preferred embodiment, S3 includes: S31: Selecting a registration map at each parking point, wherein the registration map is used to detect the probe end alignment correction amount when the robot performs the actual detection process, and the correction amount is used for path compensation during weld scanning; S32: Mark the defect locations and use different colors to distinguish different types of welds to be inspected.

[0008] According to a preferred embodiment, the defect location and weld type are displayed in a graphical manner in the digital twin model.

[0009] According to a preferred embodiment, in S4: The S41 path planning includes: S411: Selecting the starting and ending points of the weld to be scanned on the frame, and generating a reciprocating vertical scanning path trajectory along the weld direction; S412: Automatically adjust the path to suit the weld shape and position, optimizing the inspection path to improve efficiency and accuracy; The S42 path simulation includes: S421: Check whether the collision volume between the end of the robot detection probe, the arm tube of the robot arm and the frame model exceeds the safety threshold condition; S422: Scan the path trajectory differential, inversely calculate the robot arm posture at each differential point, and simulate the automatic operation of the robot arm in the ROS environment.

[0010] Accordingly, the present invention also proposes a subway vehicle frame weld detection system for implementing the above method, comprising: Digital twin model building module: used to generate the digital twin model of the architecture; Image positioning module: used to select the registration map at each parking point and mark the defect location; Path planning module: used to perform path planning in a simulation test environment and optimize it in combination with the correction amount obtained by image positioning; Path simulation module: used to simulate the automatic operation of the robot arm in the ROS environment and check the collision volume; Weld scanning path compensation module: used to calculate the alignment correction amount of the robot detection probe end and compensate the correction amount to the scanning path; Robot: used to execute the actual inspection process according to the defect location and path planning in the simulation test environment.

[0011] According to a preferred embodiment, the digital twin model construction module generates a digital twin model of the framework based on a design model or point cloud data.

[0012] According to a preferred embodiment, the image positioning module is used to: select a registration map at each parking point, the registration map is used to detect the probe end alignment correction when the robot executes the actual detection process, and the correction is used for path compensation during weld scanning; mark the defect position and distinguish different detection weld categories with different colors.

[0013] According to a preferred embodiment, The path planning module is used to: select the starting and ending points of the weld to be scanned on the frame, and generate a reciprocating vertical scanning path trajectory along the weld direction; automatically adjust the path to adapt to the shape and position of the weld, and optimize the detection path to improve efficiency and accuracy; The path simulation module is used to check whether the collision volume between the end of the robot detection probe, the mechanical arm tube and the frame model exceeds the safety threshold condition; scan the path trajectory difference, inversely calculate the mechanical arm posture of each differential point, and simulate the automatic operation of the mechanical arm in the ROS environment.

[0014] According to a preferred embodiment, the robot includes a 3D camera and a detection probe, the 3D camera is used to collect structure appearance data, and the detection probe is used to scan the weld.

[0015] By establishing a digital twin model of the frame, the present invention can simulate flaw detection of the frame in a virtual environment and freely configure the flaw detection position, preset a planning path, provide guidance and reference for actual flaw detection, and at the same time, collect real-time data of the frame through a robot, and compare and analyze these data with the digital twin model to achieve accurate flaw detection of the frame. The digital twin model can simulate different flaw detection scenarios and conditions, improve the accuracy and reliability of flaw detection, improve the management efficiency and production benefits of related equipment, and can detect frames in different states and different repair processes, that is, the frame on the vehicle before the vehicle is disassembled and the frame after the vehicle is disassembled, with strong adaptability and flexible scanning range. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a step diagram of a method for detecting a weld of a subway vehicle frame of the present invention; Figure 2 It is a composition diagram of a subway vehicle frame weld detection system of the present invention; Figure 3 The present invention is a flow chart for detecting welds of a subway vehicle frame. DETAILED DESCRIPTION

[0017] At present, the common method for inspecting frame welds in subway vehicle maintenance is manual magnetic particle inspection. After the vehicle is dismantled, the frame needs to go through a series of processes such as cleaning, paint removal, cleaning, magnetization, inspection, demagnetization, cleaning, and painting. This process has the disadvantages of long cycle, environmental pollution, low efficiency, high cost, and low level of data informationization and intelligence.

[0018] In the prior art, in order to overcome the above-mentioned shortcomings of manual magnetic particle inspection, a robotic arm is often introduced to perform fixed-point teaching eddy current scanning of welds to realize automatic detection of weld data collection. However, there are defects such as a fixed and single operation path, weak adaptability to surface protrusions such as welding slag, and a too small scanning adaptation range.

[0019] In view of this, the present invention proposes a robot online detection implementation method, which can solve the above problems to a certain extent.

[0020] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0021] Specifically, Figure 1 As shown, the technical solution of this application is: A method for detecting a weld of a subway vehicle frame comprises the following steps: S1: Generate a digital twin model of the subway vehicle structure based on the design model of the structure or the point cloud data actually collected; S2: Load the digital twin model and robot model of the architecture in the ROS simulation environment to build a simulation test environment; S3: Marking the defect location under the simulation test environment; S4: Performing S41 path planning and S42 path simulation under the simulation test environment; S5: According to the defect location and path planning in the simulation test environment, the robot executes the actual detection process.

[0022] Point cloud data is a data structure that represents a set of points in three-dimensional space. Each point contains its coordinates in space (usually X, Y, Z coordinates) and other information, such as color, intensity, normal and other attributes. Optional point cloud processing software, such as CloudCompare or PCL (Point Cloud Library), pre-processes the collected point cloud data, including denoising, filtering, registration and splicing, to generate a complete digital twin model of the architecture. This model will serve as the basis for subsequent simulation tests and environment construction.

[0023] Among them, the ROS (Robot Operating System) simulation environment is a computer simulation technology used for robotics research and development. It allows developers to simulate the robot's behavior and interaction with the environment in a virtual environment without the need for actual hardware.

[0024] After completing the path planning and path simulation in the simulation test environment, the robot will perform the actual inspection process based on the defect location and path planning information. During the actual inspection process, the robot will move to each defect location according to the planned path and scan the weld with the detection probe it carries. The detection probe will collect images or data of the weld and identify and classify the defects of the weld through image processing or machine learning algorithms. Finally, the inspection system will generate a detailed inspection report, including information such as the defect location, type, size, and possible repair suggestions of the weld.

[0025] By establishing a digital twin model of the frame, the present invention can simulate flaw detection of the frame in a virtual environment and freely configure the flaw detection position, preset a planning path, provide guidance and reference for actual flaw detection, and at the same time, collect real-time data of the frame through a robot, and compare and analyze these data with the digital twin model to achieve accurate flaw detection of the frame. The digital twin model can simulate different flaw detection scenarios and conditions, improve the accuracy and reliability of flaw detection, improve the management efficiency and production benefits of related equipment, and detect frames in different states and different repair processes, that is, the frame on the vehicle before the vehicle is disassembled and the frame after the vehicle is disassembled.

[0026] Furthermore, in said S1, If a design model exists, a digital twin model of the architecture is directly generated: the design model is imported into simulation software, such as the Gazebo simulator of ROS (Robot Operating System), and a digital twin model highly similar to the real architecture is generated through conversion and processing. This model will serve as the basis for subsequent simulation tests and environment construction, and is used to simulate the robot's detection process and path planning.

[0027] If there is no design model, the appearance of the structure is collected manually using a 3D camera. The 3D camera will capture a large amount of point cloud data on the surface of the structure, which contains the three-dimensional shape and size information of the structure. The complete point cloud of the structure is obtained by splicing, and the digital twin model of the structure is obtained by loading the point cloud modeling through software.

[0028] As described above, regardless of whether a design model exists, the present invention provides an appropriate method to generate a digital twin model of a subway vehicle frame, providing strong support for subsequent simulation testing, path planning, and robot detection processes.

[0029] Further, the S3 includes: S31: Selecting a registration map at each parking point, the registration map is used to perform alignment correction of the end of the detection probe when the robot performs the actual detection process, and the correction is used for path compensation during weld scanning; specifically, in a simulation test environment, according to the robot's motion path and detection requirements, a registration map is selected at each parking point (i.e., the parking position of the robot when performing the detection operation). The registration map refers to an image or area with obvious features and easy to identify, which is used to perform alignment correction of the end of the robot detection probe during the actual detection process.

[0030] When the robot performs the actual inspection process, it will capture the actual image of the current parking point through its onboard camera or sensor and compare it with the pre-stored registration map. Through the comparison results, the deviation between the position of the end of the detection probe in the actual environment and the theoretical position (i.e., the alignment correction) can be calculated. This correction will be used for path compensation during weld scanning to ensure that the probe can accurately scan along the planned path.

[0031] S32: Mark the defect locations and use different colors to distinguish different types of welds to be inspected.

[0032] In order to more intuitively display the defect location and distinguish different types of inspected welds, different colors are used to mark the defect location and weld categories.

[0033] As a preferred embodiment, high-risk welds can be marked in red, medium-risk welds can be marked in yellow, and low-risk or no-risk welds can be marked in green. At the same time, a unique identifier or number can be assigned to each defect location for subsequent identification and recording.

[0034] The sub-steps S31 and S32 in step S3 together constitute the core content of defect position marking and robot detection probe end alignment correction, which helps to improve the accuracy and efficiency of robot detection.

[0035] According to a preferred embodiment, the defect location and weld category are graphically displayed in the digital twin model. As mentioned above, in addition to color marking and identifier marking, interactive functions can also be added to the digital twin model. For example, when a user clicks on a weld or defect location, detailed information of the location such as defect type, size, risk level, etc. can be displayed.

[0036] According to a preferred embodiment, in S4: The S41 path planning includes: S411: Selecting the starting and ending points of the weld to be scanned on the frame, and generating a reciprocating vertical scanning path trajectory along the weld direction; S412: Automatically identify the shape and position characteristics of the weld. Based on the identification results, the software automatically adjusts the generated scanning path trajectory to adapt to the actual shape and position of the weld. This includes adjusting the curvature, direction and other parameters of the path to ensure that the probe can fit the weld closely, improving the accuracy and efficiency of detection; The S42 path simulation includes: S421: Check whether the collision volume of the robot detection probe end, the robot arm tube and the frame model exceeds the safety threshold condition. If the collision volume exceeds the preset safety threshold condition, a warning is issued and the user is prompted to adjust the path or model parameters to avoid collision during the actual detection process; S422: Differentiate the scanning path trajectory, inversely calculate the robot arm posture at each differential point, and simulate the automatic operation of the robot arm in the ROS environment. Specifically, it includes: differential processing of the scanning path trajectory to obtain a series of tiny path segments or differential points; for each differential point, use the inverse algorithm to calculate the posture of the robot arm at that point; this includes determining parameters such as the joint angle of the robot arm, the position and direction of the end effector; in the ROS environment, simulate the automatic operation of the robot arm according to the calculated posture sequence, which includes simulating the robot arm's start, move, stop and other actions, as well as the scanning process of the end of the detection probe on the weld; verify the accuracy and feasibility of the path planning and robot arm posture calculation through the simulation results. If problems or deficiencies are found, return to the S41 stage for adjustment and optimization.

[0037] Through the above implementation, the S4 stage can generate and optimize the scanning path trajectory that adapts to the shape and position of the weld, and verify the automatic operation capability and safety of the robot arm in the ROS environment through simulation.

[0038] Corresponding to the above method, Figure 2 As shown, the present invention provides a subway vehicle frame weld detection system, comprising: Digital twin model building module: used to generate the digital twin model of the architecture; Image positioning module: used to select the registration map at each parking point and mark the defect location; Path planning module: used to perform path planning in a simulation test environment and optimize it in combination with the correction amount obtained by image positioning; Path simulation module: used to simulate the automatic operation of the robot arm in the ROS environment and check the collision volume; Weld scanning path compensation module: used to calculate the alignment correction amount of the robot detection probe end and compensate the correction amount to the scanning path; Robot: used to execute the actual inspection process according to the defect location and path planning in the simulation test environment.

[0039] Among them, the digital twin model construction module generates a digital twin model of the framework based on the design model or point cloud data.

[0040] Among them, the image positioning module is used to: select a registration map at each parking point, the registration map is used to detect the probe end alignment correction when the robot executes the actual detection process, and the correction is used for path compensation during weld scanning; mark the defect position and distinguish different detection weld categories with different colors.

[0041] The path planning module is used to: select the starting and ending points of the weld to be scanned on the frame, and generate a reciprocating vertical scanning path trajectory along the weld direction; automatically adjust the path to adapt to the weld shape and position, and optimize the detection path to improve efficiency and accuracy; The path simulation module is used to check whether the collision volume between the end of the robot detection probe, the mechanical arm tube and the frame model exceeds the safety threshold condition; scan the path trajectory difference, inversely calculate the mechanical arm posture of each differential point, and simulate the automatic operation of the mechanical arm in the ROS environment.

[0042] The robot includes a 3D camera and a detection probe, wherein the 3D camera is used to collect structure appearance data, and the detection probe is used to scan welds.

[0043] like Figure 3 As shown, a flow chart for detecting welds of a subway vehicle frame is provided, which includes pre-shift calibration and post-shift calibration, that is, calibration of each module in advance and after detection.

[0044] The above is a detailed introduction to a method and system for detecting welds of a subway vehicle frame provided by an embodiment of the present invention. The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same and similar parts between the various embodiments can be referenced to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description. It should be pointed out that for ordinary technicians in this technical field, without departing from the principle of the present invention, several improvements and modifications can be made to the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

[0045] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0046] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

Claims

1. A method for detecting welds of a subway vehicle frame, characterized in that: The following steps are involved: S1: Generate a digital twin model of the subway vehicle structure based on the design model of the structure or the point cloud data actually collected; S2: Load the digital twin model and robot model of the architecture in the ROS simulation environment to build a simulation test environment; S3: Marking the defect location under the simulation test environment; S4: Performing S41 path planning and S42 path simulation under the simulation test environment; S5: According to the defect location and path planning in the simulation test environment, the robot executes the actual detection process.

2. The method for detecting welds of subway vehicle frames according to claim 1, characterized in that: In S1, If a design model exists, a digital twin model of the architecture is directly generated; If the design model does not exist, the appearance of the structure is manually captured using a 3D camera, and the complete structure point cloud is obtained by splicing, and the point cloud modeling is loaded through software to obtain a digital twin model of the structure.

3. The method for detecting welds of subway vehicle frames according to claim 1, characterized in that: The S3 includes: S31: Selecting a registration map at each parking point, wherein the registration map is used to detect the probe end alignment correction amount when the robot performs the actual detection process, and the correction amount is used for path compensation during weld scanning; S32: Mark the defect locations and use different colors to distinguish different types of welds to be inspected.

4. The method for detecting welds of subway vehicle frames according to claim 3, characterized in that: The defect locations and weld categories are graphically displayed in the digital twin model.

5. The method for detecting welds of subway vehicle frames according to claim 1, characterized in that: In S4: The S41 path planning includes: S411: Selecting the starting and ending points of the weld to be scanned on the frame, and generating a reciprocating vertical scanning path trajectory along the weld direction; S412: Automatically adjust the path to suit the weld shape and position, optimizing the inspection path to improve efficiency and accuracy; The S42 path simulation includes: S421: Check whether the collision volume between the end of the robot detection probe, the arm tube of the robot arm and the frame model exceeds the safety threshold condition; S422: Scan the path trajectory differential, inversely calculate the robot arm posture at each differential point, and simulate the automatic operation of the robot arm in the ROS environment.

6. A subway vehicle frame weld detection system, characterized in that: include: Digital twin model building module: used to generate the digital twin model of the architecture; Image positioning module: used to select the registration map at each parking point and mark the defect location; Path planning module: used to perform path planning in a simulation test environment and optimize it in combination with the correction amount obtained by image positioning; Path simulation module: used to simulate the automatic operation of the robot arm in the ROS environment and check the collision volume; Weld scanning path compensation module: used to calculate the alignment correction amount of the robot detection probe end and compensate the correction amount to the scanning path; Robot: used to execute the actual inspection process according to the defect location and path planning in the simulation test environment.

7. The subway vehicle frame weld detection system according to claim 6, characterized in that: The digital twin model construction module generates a digital twin model of the framework based on the design model or point cloud data.

8. The subway vehicle frame weld detection system according to claim 6, characterized in that: The image positioning module is used to select a registration map at each parking point, the registration map is used to detect the probe end alignment correction amount when the robot performs the actual detection process, and the correction amount is used for path compensation during weld scanning; Mark the defect locations and use different colors to distinguish different types of welds to be inspected.

9. The subway vehicle frame weld detection system according to claim 6, characterized in that: The path planning module is used to: select the starting and ending points of the weld to be scanned on the frame, and generate a reciprocating vertical scanning path trajectory along the weld direction; automatically adjust the path to adapt to the shape and position of the weld, and optimize the detection path to improve efficiency and accuracy; The path simulation module is used to check whether the collision volume between the end of the robot detection probe, the mechanical arm tube and the frame model exceeds the safety threshold condition; scan the path trajectory difference, inversely calculate the mechanical arm posture of each differential point, and simulate the automatic operation of the mechanical arm in the ROS environment.

10. The subway vehicle frame weld detection system according to claim 6, characterized in that: The robot comprises a 3D camera and a detection probe, wherein the 3D camera is used to collect the appearance data of the structure, and the detection probe is used to scan the weld.

Images