Method for detecting defects in small-diameter pipeline

Abstract

The invention relates to the technical field of pipeline safety detection, in particular to a method for detecting defects in a small-diameter pipeline, which comprises the following steps of: planning a detection track of a magnetic robot based on a circumferential welding seam of the pipeline; the stress condition of the robot in the track path is analyzed, and the magnetic force of the robot is determined; the robot collects a pipeline image and identifies defects in the pipeline through the deep learning network; a control system of the robot is built, a mobile computer is used for remotely controlling the Raspberry Pi 4b, and then the Raspberry Pi 4b is used for controlling a robot motor to achieve autonomous operation. According to the detection method, the defect detection efficiency of the small-pipe-diameter pipeline is improved, the automation degree is high, the designed robot can freely run in the iron pipeline, and the good prospect is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline safety detection, and in particular to a method for detecting defects in small-diameter pipelines. Background Art

[0002] In the national energy system, pipelines are a vital link and play an irreplaceable role in the processing, allocation, operation and configuration of energy. Mobile robots are widely used in the field of pipeline maintenance and play an extremely critical role in the safety of pipeline systems, especially for the detection of circumferential welds at pipeline joints. Due to their own structural design and detection methods, the existing mobile robots still have problems with low automation and low detection efficiency in the detection of defects in pipelines. Summary of the invention

[0003] In order to solve the above technical problems existing in the prior art, the present invention proposes a method for detecting defects in small-diameter pipelines, and its specific technical solution is as follows: A method for detecting defects in small-diameter pipelines, which plans the detection trajectory of a magnetic suction robot based on the pipeline circumferential weld; Analyze the force applied to the robot in the trajectory path to determine the robot's magnetic force; The robot collects pipeline images and identifies defects inside the pipeline through a deep learning network; Build the robot's control system, use a mobile computer to remotely control the Raspberry Pi 4b, and then use the Raspberry Pi 4b to control the robot's motor to achieve autonomous operation.

[0004] Furthermore, the magnetic robot is a two-wheeled mobile robot in the shape of a bicycle.

[0005] Furthermore, the detection trajectory of the magnetic suction robot is to first walk along the pipeline busbar to a distance in front of the weld, and then The angle is turned to the circumference of the weld, and then moves along the weld for a circle, and finally The angle turns to the direction of the pipeline busbar and continues to move forward. The angle is [-90°, 90°].

[0006] Furthermore, the expression for analyzing the force condition of the robot in the trajectory path is as follows: , , Where G is the robot's own weight, l =0.03m, G≈40N, ≈0.14m, The angle is [-90°, 90°], which can be calculated as: .

[0007] Furthermore, the robot control system includes a host computer and a slave computer. The host computer is a mobile computer, namely a PC, and the slave computer includes: Raspberry Pi 4b. Raspberry Pi 4b is equipped with Velodyne laser radar, camera, and mobile light source. The slave computer collects sensor data from the radar and camera, and transmits it back to the host computer after preprocessing. After processing, the host computer controls the next action of the robot and outputs the pipeline defect detection results.

[0008] Furthermore, the mobile computer and Raspberry Pi 4b build a local area network through WIFI, and use ROS nodes to publish information to transmit radar data to the PC for processing. Similarly, the PC also uses ROS nodes to publish the processed results and instructions back to the Raspberry Pi 4b. After calculating the next control signal, the PC publishes the control signal to the drive board of the front and rear wheels and servo of the robot through the USB-CAN connection line to control the robot to move.

[0009] Furthermore, the robot identifies defects in the pipeline by preprocessing the images captured by the robot camera, applying Gaussian filtering and Canny edge detection to extract edge coordinates, then using dilation operation and contour detection to estimate the number of cluster centers, and then using these edge coordinates and the estimated number of cluster centers as input parameters for K-means clustering, thereby achieving adaptive segmentation of the image.

[0010] The detection method of the invention improves the defect detection efficiency and automation of small-diameter pipelines. The designed robot can move freely in the iron pipeline and has good prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 is a schematic diagram of the detection trajectory of the robot of this embodiment for the circumferential weld in the pipeline; Figure 2 and Figure 3 They are schematic diagrams of mechanical analysis of the upper and lower sides of the robot of this embodiment in the pipeline; Figure 4 Schematic diagram of the control system structure of the robot of this embodiment; Figure 5 It is a flow chart of the defect detection method of this embodiment. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical scheme and technical effect of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0013] This embodiment discloses a method for detecting defects in pipelines with a diameter of 600mm to 1000mm. A two-wheeled mobile robot is used. The robot is equipped with a camera to enter the pipeline to collect data, and the defects in the pipeline are identified through a deep learning network. In particular, during the detection process, the detection trajectory of the robot is designed according to the characteristics of the circumferential weld of the pipeline, the force conditions of the robot in different postures in the detection trajectory are analyzed, and the adsorption force required for the robot to run in the trajectory is calculated, and the final structure of the robot is determined by simulation. Finally, the control system of the detection robot is designed, and a mobile computer is used to remotely control the Raspberry Pi 4b, and then the robot motor is controlled by the Raspberry Pi 4b to achieve autonomous operation.

[0014] Specifically, the two-wheeled mobile robot is a bicycle-shaped magnetic adsorption robot. For the circumferential weld in the pipeline, its detection path is planned as first walking along the pipeline busbar to a distance in front of the weld, then turning to the weld circumference at a certain angle, then moving along the weld for one circle and taking images, and finally turning to the pipeline busbar direction at a certain angle and continuing to move forward. Figure 1 In this process, the robot will travel on the busbars at different angles, such as the bottom, side, and top of the pipeline. Therefore, it is necessary to analyze the magnetic force of the robot's magnetic wheel to ensure that the robot can operate normally. Figure 2 and Figure 3 As shown, the mechanical analysis expression of the robot is as follows: , , Where G is the robot's own weight, l =0.03m, G≈40N, ≈0.14m, The angle is [-90°, 90°], which can be calculated as: .

[0015] like Figure 4 As shown, the control system of the robot includes an upper computer and a lower computer. The upper computer is a mobile computer, i.e., a PC. The lower computer includes: a Raspberry Pi 4b, equipped with a Velodyne laser radar, a camera, and a mobile light source. The lower computer is responsible for collecting sensor data such as radar and camera, and transmitting it back to the upper computer after preprocessing. After processing, the upper computer controls the next action of the robot and determines the pipeline defect detection result. The mobile computer and the Raspberry Pi build a local area network through WIFI, and use the ROS node to publish information to transmit the radar data to the PC for processing. Similarly, the PC also uses the ROS node to publish the processed results and instructions to the Raspberry Pi. After calculating the next control signal, the PC publishes the control signal to the front and rear wheels of the robot and the drive board of the servo through the USB-CAN connection line to control the robot to move.

[0016] The robot identifies defects in the pipeline, such as Figure 5 As shown in FIG, specifically, after preprocessing the image captured by the robot camera, Gaussian filtering and Canny edge detection are applied to extract edge coordinates. Then, the number of cluster centers is estimated using dilation operation and contour detection. Subsequently, these edge coordinates and the estimated number of cluster centers are used as input parameters for K-means clustering, thereby achieving adaptive image segmentation.

[0017] The above is only a preferred implementation case of the present invention and does not limit the present invention in any form. Although the implementation process of the present invention is described in detail above, for those familiar with the art, they can still modify the technical solutions recorded in the above examples, or replace some of the technical features therein with equivalents. All modifications, equivalent replacements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting defects in small-diameter pipelines, characterized in that: Planning the detection trajectory of the magnetic suction robot based on the circumferential weld of the pipeline; Analyze the force applied to the robot in the trajectory path to determine the robot's magnetic force; The robot collects pipeline images and identifies defects inside the pipeline through a deep learning network; Build the robot's control system, use a mobile computer to remotely control the Raspberry Pi 4b, and then use the Raspberry Pi 4b to control the robot's motor to achieve autonomous operation.

2. The method for detecting defects in a pipeline according to claim 1, characterized in that: The magnetic robot is a two-wheeled mobile robot in the shape of a bicycle.

3. The method for detecting defects in a pipeline according to claim 1, characterized in that: The detection trajectory of the magnetic robot is to first walk along the pipeline busbar to a distance in front of the weld, and then The angle is turned to the circumference of the weld, and then moves along the weld for a circle, and finally The angle turns to the direction of the pipeline busbar and continues to move forward. The angle is [-90°, 90°].

4. The method for detecting defects in a pipeline according to claim 1, characterized in that: The expression for analyzing the force of the robot in the trajectory path is as follows: ， ， Where G is the robot's own weight, l =0.03m, G≈40N, ≈0.14m, The angle is [-90°, 90°], which can be calculated as: 。 5. The method for detecting defects in a pipeline according to claim 1, characterized in that: The robot control system includes a host computer and a slave computer. The host computer is a mobile computer, namely a PC, and the slave computer includes: Raspberry Pi 4b. Raspberry Pi 4b is equipped with Velodyne laser radar, camera, and mobile light source. The slave computer collects sensor data from the radar and camera, and transmits it back to the host computer after preprocessing. After processing, the host computer controls the next action of the robot and outputs the pipeline defect detection results.

6. The method for detecting defects in a pipeline according to claim 5, characterized in that: The mobile computer and Raspberry Pi 4b build a local area network through WIFI, and use ROS nodes to publish information to transmit radar data to the PC for processing. Similarly, the PC also uses ROS nodes to publish the processed results and instructions back to the Raspberry Pi 4b. After calculating the next control signal, the PC publishes the control signal to the drive board of the front and rear wheels and servo of the robot through the USB-CAN connection line to control the robot to move.

7. The method for detecting defects in a pipeline according to claim 1, characterized in that: The robot identifies defects in the pipeline by preprocessing the images captured by the robot camera, applying Gaussian filtering and Canny edge detection to extract edge coordinates, then using dilation operation and contour detection to estimate the number of cluster centers, and then using these edge coordinates and the estimated number of cluster centers as input parameters of K-means clustering, thereby achieving adaptive segmentation of the image.

Images