A method of cutting a truck bed side wall
By combining 3D laser sensors and robotic systems, the problem of accurate identification and automatic cutting of the side walls of railway freight cars has been solved, achieving efficient and precise cutting results and improving cutting quality and enterprise benefits.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHAOXING HANLI IND AUTOMATION TECH CO LTD
- Filing Date
- 2023-12-05
- Publication Date
- 2026-05-29
AI Technical Summary
In existing technologies, it is difficult to achieve accurate identification and automated cutting of the side walls of railway freight cars, resulting in low cutting efficiency, quality that depends on the operator's experience, and problems such as bulging, deformation and tearing of steel plates.
By employing 3D laser sensors and robotic systems, combined with specific cut recognition algorithms and collision avoidance systems, precise scanning and automatic cutting of the side walls of truck cargo compartments can be achieved.
It improved cutting precision and efficiency, reduced labor intensity, and enhanced the economic benefits of enterprises.
Smart Images
Figure CN117444369B_ABST
Abstract
Description
Technical Field
[0001] This invention proposes a method for cutting the side walls of railway freight cars for the replacement and maintenance of side and end walls. Background Technology
[0002] The freight cars used for railway transportation are all-steel welded structures, mainly composed of components such as the underframe, side walls, end walls, and doors. During transportation, the freight cars are subject to wear, oxidation, and corrosion over a long period of time, requiring the steel plates in the car walls to be cut and replaced periodically.
[0003] In the current technology, the replacement and maintenance of the end walls and side walls of truck bodies are generally carried out manually using flame cutting. The results of manual cutting are difficult to guarantee, and the labor intensity is high, the cutting efficiency is low, and the cutting quality depends entirely on the experience and skill of the operator.
[0004] The cutting of the truck bed wall panels uses the steel plate welds as a reference. Although existing technologies employ industrial robots to locate and cut welds, accurately identifying and locating welds of varying shapes on the truck bed ends and side walls remains a technical challenge. Due to structural issues with the side and end walls, the steel plates may exhibit bulging deformation, tearing, or holes, making it difficult for robots to automatically cut along a standardized path. Therefore, it is necessary to create specialized models for different wall panels based on actual laser scanning data to ensure accurate positioning. The robot's posture can then be adjusted in real-time based on the identified welds, thus achieving precise automatic cutting.
[0005] Based on the above reasons, the inventors established corresponding cutting templates for several different cuts on the side wall panels of existing truck bodies, and adopted appropriate scanning and cut recognition methods for different cutting templates to ensure the accuracy of cut recognition. This case was thus born. Summary of the Invention
[0006] This invention discloses a method for cutting the side wall of a truck cargo compartment, specifically including the following:
[0007] Step (1): Secure the truck bed, lay guide rails on the front of the side wall, and install the robot equipped with a cutting gun and 3D laser sensor on the guide rails so that it can move along the guide rails.
[0008] Step (2): Start the 3D laser sensor to perform positioning scans on the edge of the workpiece to be cut on the side wall to determine the position of the workpiece to be cut;
[0009] The workpieces to be cut on the side wall include the left trapezoidal plate, the middle trapezoidal plate, the right trapezoidal plate, the left C-shaped plate, and the right C-shaped plate;
[0010] Step (3): The robot moves to the position of the workpiece to be cut, restarts the 3D laser sensor, scans along the kerf according to the kerf recognition algorithm preset in the control system, collects 3D point cloud data of the kerf position and sends it to the control system.
[0011] The five templates for the workpieces to be cut given in step (2) above correspond to their respective cut identification algorithms. In the cut identification algorithm, the left trapezoidal plate contains four cuts, and the bevel types of the four cuts are corner joint, splice, overlap, and splice, respectively; the middle trapezoidal plate contains four cuts, and the bevel types of the four cuts are corner joint, splice, corner joint, and splice, respectively; the right trapezoidal plate contains four cuts, and the bevel types of the four cuts are overlap, splice, corner joint, and splice, respectively; the left C-shaped plate and the right C-shaped plate each contain three cuts, and the bevel types of these three cuts are splice, overlap, and splice, respectively;
[0012] For splicing bevel type, the middle position of the splice is taken as the cut position; for overlapping bevel type, the position with the largest height difference at the overlapping is taken as the cut position; for corner bevel type, the intersection of the two identified diagonal lines is taken as the cut position.
[0013] Step (4): The control system identifies the location of the kerf from the received 3D point cloud data according to the kerf recognition algorithm. Each kerf is identified to generate a cutting straight line trajectory. The intersection point between two adjacent kerfs is calculated to finally form a cutting trajectory that is connected end to end.
[0014] Step (5): The control system sets the cutting posture of the cutting torch for each cutting kerf according to the identified kerf shape. The control system sends instructions to the robot to move to the cutting position and adjust the cutting torch according to the set cutting posture. The control system controls the robot to complete the cutting of the workpiece to be cut according to the generated cutting trajectory.
[0015] Furthermore, multiple robots are installed on the guide rail on the same side of the truck bed. The robots are equipped with an anti-collision system, which divides the area moving along the same guide rail direction into different collision zones. Each collision zone is assigned a different collision code. Before moving from one collision zone to the next, the robot first checks whether there is a collision code in the next zone. If there is, it means that there is already a robot in that zone, and the robot to be moved must wait until the robot in the collision zone to be entered leaves before it can enter.
[0016] Furthermore, once the robot enters the designated collision area, a collision code is immediately set. Before the robot completes the cutting and leaves the collision area, it notifies the other robots on the guide rail and then releases the collision code.
[0017] Furthermore, the acquired 3D point cloud data of the kerf includes the height information of the kerf. The control system generates the cutting trajectory based on the received 3D point cloud data as follows:
[0018] The 3D laser sensor continuously scans each kerf on the workpiece to be cut. During the scan, each kerf generates a frame of data at fixed length intervals. The position of the kerf point is identified in each frame of scan data, thereby obtaining a set of feature points for a kerf.
[0019] The feature point set of each kerf is processed, and a cutting straight line trajectory is generated by fitting along the horizontal direction. The intersection point between two adjacent kerfs is calculated to form a kerf that is connected end to end. The height direction of the kerf is filtered to save the height information and remove noise, and finally the three-dimensional spatial cutting trajectory of each kerf is generated.
[0020] Furthermore, for the same cutting path, the starting cutting angle, intermediate cutting angle, and ending cutting angle of the cutting torch need to be set.
[0021] Furthermore, the robot is model Fanuc M-20iD / 12L, and the cutting torch is a MAXPRO200 plasma cutter.
[0022] The cutting method designed in this invention fully considers the design characteristics of the steel plate kerf shape of the existing freight car sidewall and has tailored a scanning and kerf recognition method to match it, which can obtain accurate kerf information. This method can realize automatic cutting of railway freight car sidewall panels with high cutting accuracy, good cutting quality, and significantly improved work efficiency, thereby improving the economic benefits of enterprises. Attached Figure Description
[0023] Figure 1 This is a schematic diagram of the cutting trajectories corresponding to the five types of templates for the side wall in the embodiment;
[0024] Figure 2 This is a schematic diagram of a robot automatically cutting a sidewall using a cutting torch and a 3D laser sensor.
[0025] Figure 3 A schematic diagram for determining the cut location for three bevel types.
[0026] Explanation of icon numbers:
[0027] 1. Side wall; 11. Left trapezoidal plate; 12. Middle trapezoidal plate; 13. Right trapezoidal plate; 14. Left C-shaped plate; 15. Right C-shaped plate; 2. Guide rail; 3. Robot. Detailed Implementation
[0028] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
[0029] This embodiment discloses a method for cutting the side wall of a freight car used for railway freight transport. Existing freight cars, such as... Figure 2 As shown, each side of a carriage has 6 side walls 1, and each side wall 1 has the following... Figure 1 As shown, it is made of five steel plates of different shapes welded together. If a steel plate is damaged and needs to be replaced, the corresponding steel plate is cut and replaced with a new steel plate of the same shape, and then welded again. The cutting method in this embodiment involves... Figure 1 The paper shows the scanning, identification, and cutting trajectory generation of the seams of five different weld seam shapes, and finally uses a robot to complete the automatic cutting operation of each seam.
[0030] The automatic cutting device includes a human-machine interface unit, a control system, a controller, a robot 3, a cutting torch, a cutting power supply, and a scanning and recognition unit. The human-machine interface unit is connected to the control system and is used for display, control operation, and parameter setting. The control system controls the movement of the robot 3 via the controller. The robot 3 is equipped with a cutting torch and a scanning and recognition unit. The robot 3 moves along a guide rail 2 laid in front of the side wall 1, thereby moving the cutting torch and the scanning and recognition unit together. The control system controls the cutting power supply, which powers the cutting torch. The scanning and recognition unit is connected to the control system for signal transmission. The scanning and recognition unit uses a 3D laser sensor to collect the original information of the kerf. The control system calculates the cutting path using a recognition algorithm, controls the movement of the robot 3 based on this path information, moves the cutting torch, and starts the cutting power supply to control the cutting torch to perform the cutting operation.
[0031] The automatic cutting device already exists in the prior art, and its hardware structure is not the focus of the technical improvement of this invention. To achieve better cutting results, the robot 3 in this embodiment is a Fanuc M-20iD / 12L model, the cutting torch is a MAXPRO200 plasma cutter, and the controller is an R-30iB Plus.
[0032] Since this embodiment involves cutting the side wall 1 of the truck body, and considering that each side of the truck body is composed of multiple side wall pieces 1, to improve cutting efficiency, a guide rail 2 can be installed on each side of the truck body, and multiple robots 3 can be installed on each guide rail 2 (see attached diagram). Figure 2This illustration only shows one robot 3 mounted on a single guide rail 2 (illustrations of multiple robots 3 are omitted). These robots can simultaneously perform cutting operations on different side walls 1. To prevent collisions when multiple robots 3 are operating on the same guide rail 2, this embodiment also installs an anti-collision system on the robots 3. This anti-collision system can directly use existing products; this invention does not modify its hardware structure, but only achieves simultaneous operation of multiple robots 3 through reasonable program settings. The following will focus on explaining how to implement the anti-collision program settings.
[0033] The area moving along the same guide rail 2 is divided into different collision zones. For example, the area containing each side wall 1 on the same side of the carriage can be designated as a collision zone, and each collision zone can be assigned a different collision code. Figure 2 Taking the collision area of the first sidewall 1 on the left as an example, which is the current collision area of the robot 3 to be moved, before the robot 3 moves from this collision area to the next collision area, the control system needs to check whether there is a collision code in the next collision area. If there is, it means that another robot 3 is already in the collision area, and the robot 3 to be moved must wait until the robot 3 in the required collision area leaves before it can enter. Once the robot 3 enters the designated collision area, the control system immediately sets the collision code corresponding to that area. Before the robot 3 completes cutting and leaves the collision area, it first notifies the other robots 3 on the same guide rail 2, and then releases the collision code and drives away from the area.
[0034] When cutting the side wall 1 of the truck body using the cutting device described above, the following method shall be followed:
[0035] Step (1): First, the truck body is hoisted to the designated position and fixed. The cutting area of the side wall 1 is set in the control system. The above cutting equipment is installed on each side of the truck body. The robot 3 is moved along the guide rail 2 to the cutting area by the controller through the control system.
[0036] Step (2): After the robot 3 moves to the area to be cut, it starts the 3D laser sensor to perform positioning scan on the edge of the workpiece to be cut on the side wall 1 to determine the accurate position of the workpiece to be cut, so as to ensure that the cut can be scanned during the subsequent second scan.
[0037] Because the workpieces to be cut on the side wall 1 of the existing carriage are divided into five different kerf shapes, and there are certain differences in the identification of different kerf shapes, the cutting trajectory and the cutting posture adjustment of the cutting torch are also closely related to the kerf shape. Therefore, if... Figure 1 As shown, the present invention divides the five types of plates on the side wall 1 into left trapezoidal plate 11, middle trapezoidal plate 12, right trapezoidal plate 13, left C-shaped plate 14, and right C-shaped plate 15 according to the shape of their cuts.
[0038] Step (3): Based on the determined position of the workpiece to be cut, plan the scanning path of the kerf, restart the 3D laser sensor to continuously scan along the kerf, and send the 3D point cloud data of the scanned kerf position to the control system in real time.
[0039] During the identification and scanning of the cut seams, the scanning and identification of the cut seams are performed according to the cut seam identification algorithm preset in the control system that matches the five types of plates. The specific explanation is as follows:
[0040] When a laser beam is applied to the weld, the shape and location information of the weld obtained from the scanning data can be used to classify the cut groove types involved in the above five types of plates into the following four categories:
[0041] 1) Splicing: such as Figure 3 As shown in the diagram on the left, when the obtained cut data does not have a significant height difference, it can be identified as a splicing bevel type, and the platform in the middle of the splice is taken as the cut location. Figure 1 The location indicated by the gray dot in the middle left image.
[0042] 2) Overlap: such as Figure 3 As shown in the middle diagram, when the obtained cut data has a significant height difference, it can be identified as an overlapping bevel type. The position with the largest height difference at the overlap is taken as the cut position. Figure 1 The gray dot in the middle image indicates the location.
[0043] 3) Corner joint: such as Figure 3 As shown in the diagram on the right, when the obtained cut data has obvious concave or convex angles, it can be identified as a corner joint bevel type, and the intersection of the two identified oblique lines is taken as the cut location. Figure 1 The location indicated by the gray dot in the middle right image.
[0044] 4) Overlapping and splicing: This bevel type is actually a combination of overlapping and splicing. The cut is generally an overlapping type, but there has been repair in the middle, and some data belongs to the splicing type. It can be processed as an overlapping type.
[0045] After clarifying the above four types of bevels, by Figure 1It can be seen that the left trapezoidal plate 11 contains four slits, and the bevel types of the four slits are corner joint, splice, overlap (or overlap splice), and splice, respectively. The middle trapezoidal plate 12 contains four slits, and the bevel types of the four slits are corner joint, splice, corner joint, and splice, respectively. The right trapezoidal plate 13 contains four slits, and the bevel types of the four slits are overlap (or overlap splice), splice, corner joint, and splice, respectively. The left C-shaped plate 14 and the right C-shaped plate are set opposite each other, and the rectangular frame between them is hollow. Therefore, the left C-shaped plate 14 and the right C-shaped plate 15 each contain three slits, and the bevel types of these three slits are splice, overlap (or overlap splice), and splice, respectively.
[0046] Step (4): The control system identifies the exact location of the kerf from the received 3D point cloud data based on the kerf recognition algorithm. It identifies each kerf and generates a cutting straight line trajectory. It finds the intersection point between two adjacent kerfs and connects the intersection points end to end to form a cutting trajectory that is connected end to end.
[0047] Specifically, since the kerf is a spatial curve, its sampled 3D point cloud data contains height information. The control system can identify the kerf height from the point cloud data and then generate the cutting trajectory. The method by which the control system generates the cutting trajectory based on the received 3D point cloud data is as follows:
[0048] (1) The 3D laser sensor continuously scans each cut. During the scan, each cut generates a frame of data at a fixed length interval (e.g., a length sampling interval of 2 mm). The cut point position is identified for each frame of scan data, thereby obtaining a set of feature points for a cut.
[0049] (2) Process the feature point set of each cut, fit a cutting straight line trajectory along the horizontal direction, find the intersection point between two adjacent cuts to form a cut that is connected end to end; filter the height direction of the cut, save the height information and remove noise, and finally generate the three-dimensional spatial cutting trajectory of each cut.
[0050] (3) Based on the shape of the kerf, set the cutting posture of each kerf when using the cutting torch to achieve the best cutting effect. For the intersection of two kerfs, in order to avoid collision between the cutting torch and the workpiece, the cutting angle of the cutting torch needs to be adjusted when cutting to the intersection point. Therefore, for the same cutting path, the starting cutting angle, the middle cutting angle and the ending cutting angle of the cutting torch need to be set.
[0051] (4) Adjust the coordinates of the trajectory points according to the type of the cutting trajectory points formed by each kerf and the corresponding fitting shape to obtain the final position of the trajectory points.
[0052] (5) Add the obtained trajectory to the overall cutting trajectory and send the complete cutting trajectory to robot 3.
[0053] Step (5): After the control system generates a complete cutting trajectory, it sends the complete cutting trajectory to robot 3. Robot 3 moves according to the cutting trajectory and adaptively adjusts the cutting posture of the cutting torch. After adjusting the cutting posture, it starts the cutting power and uses the cutting torch to cut along the kerf. After completing the cutting of one area, the robot 3 moves to the next cutting area and repeats the above cutting process until the cutting of the entire carriage's replacement panels is completed.
[0054] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A method for cutting the side wall of a truck body, characterized in that, Includes the following: Step (1): Secure the truck bed, lay guide rails on the front of the side wall, and install the robot equipped with a cutting gun and 3D laser sensor on the guide rails so that it can move along the guide rails. Step (2): Start the 3D laser sensor to perform positioning scans on the edge of the workpiece to be cut on the side wall to determine the position of the workpiece to be cut; The workpieces to be cut on the side wall include the left trapezoidal plate, the middle trapezoidal plate, the right trapezoidal plate, the left C-shaped plate, and the right C-shaped plate; Step (3): The robot moves to the position of the workpiece to be cut, restarts the 3D laser sensor, scans along the kerf according to the kerf recognition algorithm preset in the control system, collects 3D point cloud data of the kerf position and sends it to the control system. The five templates for the workpieces to be cut given in step (2) above correspond to their respective cut identification algorithms. In the cut identification algorithm, the left trapezoidal plate contains four cuts, and the bevel types of the four cuts are corner joint, splice, overlap, and splice, respectively; the middle trapezoidal plate contains four cuts, and the bevel types of the four cuts are corner joint, splice, corner joint, and splice, respectively; the right trapezoidal plate contains four cuts, and the bevel types of the four cuts are overlap, splice, corner joint, and splice, respectively; the left C-shaped plate and the right C-shaped plate each contain three cuts, and the bevel types of these three cuts are splice, overlap, and splice, respectively; For splicing bevel type, the middle position of the splice is taken as the cut position; for overlapping bevel type, the position with the largest height difference at the overlapping is taken as the cut position; for corner bevel type, the intersection of the two identified diagonal lines is taken as the cut position. Step (4): The control system identifies the location of the kerf from the received 3D point cloud data according to the kerf recognition algorithm. Each kerf is identified to generate a cutting straight line trajectory. The intersection point between two adjacent kerfs is calculated to finally form a cutting trajectory that is connected end to end. Step (5): The control system sets the cutting posture of the cutting torch for each cutting kerf according to the identified kerf shape. The control system sends instructions to the robot to move to the cutting position and adjust the cutting torch according to the set cutting posture. The control system controls the robot to complete the cutting of the workpiece to be cut according to the generated cutting trajectory.
2. The method for cutting the side wall of a truck body according to claim 1, characterized in that: Multiple robots are installed on the guide rail on the same side of the truck bed. The robots are equipped with an anti-collision system, which divides the area that moves along the same guide rail into different collision zones. Each collision zone is assigned a different collision code. Before moving from one collision zone to the next, the robot first checks whether there is a collision code in the next zone. If there is, it means that there is already a robot in that zone, and the robot to be moved must wait until the robot in the collision zone to be entered leaves before it can enter.
3. The method for cutting the side wall of a truck body according to claim 2, characterized in that: Once the robot enters the designated collision area, a collision code is immediately set. Before the robot completes the cutting and leaves the collision area, it notifies the other robots on the guide rail and then releases the collision code.
4. The method for cutting the side wall of a truck body according to claim 1, characterized in that: The acquired 3D point cloud data of the kerf includes the height information of the kerf. The control system generates the cutting trajectory based on the received 3D point cloud data as follows: The 3D laser sensor continuously scans each kerf on the workpiece to be cut. During the scan, each kerf generates a frame of data at fixed length intervals. The position of the kerf point is identified in each frame of scan data, thereby obtaining a set of feature points for a kerf. The feature point set of each kerf is processed, and a cutting straight line trajectory is generated by fitting along the horizontal direction. The intersection point between two adjacent kerfs is calculated to form a kerf that is connected end to end. The height direction of the kerf is filtered to save the height information and remove noise, and finally the three-dimensional spatial cutting trajectory of each kerf is generated.
5. The method for cutting the side wall of a truck body according to claim 1, characterized in that: For the same cutting path, the starting cutting angle, intermediate cutting angle, and ending cutting angle of the cutting torch need to be set.
6. The method for cutting the side wall of a truck body according to claim 1, characterized in that: The robot is model Fanuc M-20iD / 12L, and the cutting torch is a MAXPRO200 plasma cutter.