Intelligent laser detection equipment for factory-welded steel rail base metal

Abstract

The invention belongs to the technical field of steel rail detection, and provides intelligent laser detection equipment for factory welding steel rail base metal. The detection mechanism is arranged on the outer side of a steel rail, the detection mechanism comprises a main frame, a movable frame is movably installed on the inner side of the main frame, a laser frame is connected to one side of the movable frame, four laser sensors are arranged on the inner side of the laser frame, and the four laser sensors are independent high-density multi-point laser sensors; laser rays of the four laser sensors cover the whole section of the steel rail, and the distances between the four laser devices and the steel rail are calibrated through calibration rails; the laser ranging technology is adopted, the geometric dimension within 2.5 m of a rail head is statically measured, the rail head is compared with a standard rail, meanwhile, the image recognition technology is adopted, damage judgment and software automatic switching are carried out in the steel rail running process, abnormal data are intelligently screened, abnormal alarm is found, manual detection and human eye recognition are replaced, the accuracy rate is increased, and the labor cost is reduced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of rail detection, and particularly relates to a laser intelligent detection equipment for mill-welded rail parent material. BACKGROUND

[0002] High-speed rails adopt long rail seamless laying technology to ensure smooth running and low noise of vehicles. The seamless long rail is formed by welding and is divided into two stages of mill welding and field welding. First, 100-meter rails are rolled in a steel mill, transported to a rail welding plant, welded into 500-meter rails, i.e. mill welding, and then transported to the field for field welding. Although the rails are strictly inspected when they are rolled out of the mill, they may be deformed or bruised during multiple hoisting from the steel mill to the rail welding plant. Deformation of the rail head may cause inaccurate welding alignment, and damage caused by bruising may produce cracks after the rail is laid, which poses a safety hazard to train operation. In order to ensure the safety of train operation, the rail welding plant must strictly detect the rail parent material again before welding.

[0003] However, the rail cross-section curve is extremely complex, which is formed by multiple arcs with different diameters and multiple inclined lines with different angles. The rail welding plant implements the "People's Republic of China Railway Industry Standard TBT2344.1-2020" to manually detect the rail parent material. Multiple workers hold special measuring tools such as clamps, rulers, calipers, and twist gauges to determine whether the geometric dimensions of the two ends of the rail exceed the standard. It is very difficult to accurately find the detection points on the arcs and inclined lines using standard measuring tools. Due to the constantly changing curvature of the arcs and inclined lines, the standard measuring tools cannot perfectly fit them, resulting in large measurement errors. For example, when measuring the rail crown fullness, the standard measuring tool may not accurately capture the highest point of the rail crown curve, resulting in a deviation between the measurement result and the actual value. Moreover, this measurement method relies on the human eye to observe the gap to determine whether the size meets the standard, which is inevitably affected by the operator's experience. Different experience levels of operators may have different judgments of the gap size. Even the same operator's eyes are prone to fatigue after a long time of work, further affecting the accuracy of the judgment. SUMMARY

[0004] The purpose of the present application is to provide a laser intelligent detection equipment for mill-welded rail parent material to solve the problems of large measurement error, dependence on operator experience, and low detection efficiency in the existing manual detection method of rail parent material.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0006] A laser intelligent detection equipment for mill-welded rail parent material, comprising:

[0007] a rail;

[0008] Detection mechanism, detection mechanism is arranged in the outside of the steel rail, the detection mechanism includes the main frame, the inside of the main frame is movably installed with the moving frame, the moving frame one side is connected with the laser frame, the laser frame inside is provided with No. 1 laser sensor, No. 2 laser sensor, No. 3 laser sensor, No. 4 laser sensor, four laser sensors are independent high-density multi-point laser sensor, the laser line of four laser sensors covers the full section of the steel rail, the distance of four laser sensors from the steel rail is calibrated by the calibration rail, the main frame top is provided with a rack, the moving frame one side is provided with a gear, the gear is engaged with the rack, the moving frame one side is provided with a servo motor, the servo motor output end is connected with the shaft of the gear through the shaft coupling, the main frame top is provided with left-right symmetrical double sets of gear and rack transmission structure, the servo motor is commanded by programmable logic controller and synchronously walks smoothly, and the track movement coordinate value is output in real time, the steel rail is scanned in 3D, all data are collected, four laser sensors are equipped with software program, the software is used for laser ranging and image recognition, laser ranging technology is used for static detection of geometric size within 2.5 meters at both ends of the rail head, and image recognition technology is used for dynamic detection of damage in full rail operation, and the software is used for realizing automatic switching of two modes of geometric size detection and damage detection.

[0009] Preferably, the main frame top two sides are symmetrically installed with slide rails, the slide rails are slidably installed with sliding blocks, and the moving frame top is fixedly installed with a top plate close to one side of the sliding block.

[0010] Preferably, the main frame two sides are symmetrically installed with racks, the moving frame two sides are symmetrically installed with side plates, the side plates are movably installed with gears through bearings close to one side of the rack, the gears are engaged with the racks, and the side plates are installed with servo motors through fixing bases away from one side of the gear.

[0011] Preferably, the laser frame is fixedly installed with two upper installation plates close to one side of the No. 1 laser sensor, and the No. 1 laser sensor is fixedly installed between the two upper installation plates.

[0012] Preferably, the laser frame is fixedly installed with a left installation plate close to one side of the No. 2 laser sensor, and the No. 2 laser sensor is fixedly installed on one side of the left installation plate.

[0013] Preferably, the laser frame is fixedly installed with a right installation plate close to one side of the No. 3 laser sensor, and the No. 3 laser sensor is fixedly installed on one side of the right installation plate.

[0014] Preferably, the laser frame is fixedly installed with an inclined installation plate close to one side of the No. 4 laser sensor, and the No. 4 laser sensor is fixedly installed on one side of the inclined installation plate.

[0015] Compared with the prior art, the beneficial effects of the present invention are:

[0016] (1) This invention uses laser ranging technology to statically measure the geometric dimensions of the rail head within 2.5 meters and compare them with the standard rail. At the same time, it uses image recognition technology to judge the damage during rail operation. The software automatically switches between two modes, intelligently filters abnormal data, and alarms when abnormalities are detected. This replaces manual inspection with special measuring tools and human eye recognition, reduces labor costs, and improves the accuracy and precision of the inspection. The rail profile data is generated by a high-precision linear laser sensor (0.05mm) and undergoes multiple lateral data comparisons. Compared with the error of manual measurement, it effectively increases the measurement accuracy and has a complete data storage and archiving function.

[0017] (2) The present invention effectively enhances the reliability of the detection. The laser detection equipment does not rely on human eye observation. It directly acquires the geometric dimension data of the rail surface through the laser sensor and automatically compares and analyzes it with the standard contour. During the detection process, the software automatically processes the collected data, makes judgments and alarms according to the preset algorithm, eliminates the influence of the operator's experience on the detection results, avoids the risk of judgment errors caused by operator fatigue, and ensures the reliability of the detection results. Moreover, the laser detection equipment can work continuously without fatigue and can always maintain stable detection performance.

[0018] (3) This invention uses only 4 laser sensors to simultaneously realize both laser ranging and image recognition functions. It uses laser ranging technology to statically detect geometric dimensions within 2.5 meters at both ends of the rail head, and uses image recognition technology to dynamically detect damage during the entire rail operation. This equipment is low in cost, small in size, simple to operate, easy to maintain, and effective.

[0019] (4) The present invention also takes into account manual operation. The height and width of the crossbeam are designed to facilitate manual operation. Manual operation is required in three situations: during the maintenance period of intelligent equipment, traditional methods can be used for inspection; when intelligent equipment detects surface defects and alarms, it is handed over to manual review and processing; and it is convenient to push in the calibration rail and damaged rail for periodic calibration.

[0020] (5) The software program equipped with this invention has the function of measuring the geometric dimensions of the rail. The point cloud coordinates collected by four lasers are stitched together to form a complete profile. Multiple profiles are collected and compared laterally during movement. The key parameters related to the rail profile, such as rail head width, rail web width, rail height, twist, and straightness, are automatically calculated. This function can replace most of the manual measurement links in the pre-welding inspection. At the same time, it can perform damage detection. By comparing with the standard rail profile, abnormal depressions and protrusions on the surface of the rail being tested are found. Damage is automatically marked according to the judgment standard such as depth and area. An alarm is triggered when an abnormality is detected. Manual processing is required, thereby replacing manual inspection with special measuring tools and human eye recognition, improving the accuracy of inspection and reducing labor costs. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall structure of the present invention;

[0022] Figure 2 This is a bottom-view structural diagram of the present invention;

[0023] Figure 3 This is an enlarged structural diagram of point A in the present invention;

[0024] Figure 4 This is a schematic diagram of the main frame structure of the present invention;

[0025] Figure 5 This is an enlarged structural diagram of point B in the present invention;

[0026] Figure 6 This is a schematic diagram of the mobile frame structure of the present invention;

[0027] Figure 7 This is a schematic diagram of the laser frame structure of the present invention;

[0028] Figure 8 This is a front view schematic diagram of the laser frame structure of the present invention.

[0029] In the diagram: 1. Rail; 2. Detection mechanism; 21. Main frame; 22. Moving frame; 23. Laser frame; 24. Laser sensor No. 1; 25. Laser sensor No. 2; 26. Laser sensor No. 3; 27. Laser sensor No. 4; 28. Slide rail; 29. ​​Slider; 210. Top plate; 211. Rack; 212. Side plate; 213. Gear; 214. Servo motor; 215. Upper mounting plate; 216. Left mounting plate; 217. Right mounting plate; 218. Angled mounting plate. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0031] Example 1:

[0032] Please see Figure 1 - Figure 8 As shown, a laser intelligent inspection equipment for welded steel rail base materials includes:

[0033] Rail 1;

[0034] The detection mechanism 2 is located outside the rail 1. The detection mechanism 2 includes a main frame 21, with a movable frame 22 movably mounted inside the main frame 21. A laser frame 23 is connected to one side of the movable frame 22. Inside the laser frame 23 are four independent high-density multi-point laser sensors: a first laser sensor 24, a second laser sensor 25, a third laser sensor 26, and a fourth laser sensor 27. The laser lines of the four sensors cover the entire cross-section of the rail 1. The distance between each of the four lasers and the rail 1 is calibrated using a calibration rail. A rack 211 is mounted at the top of the main frame 21, and a gear 213 is mounted on one side of the movable frame 22. The gear 213 meshes with the rack 211. A servo motor 214 is installed on one side. The output end of the servo motor 214 is connected to the shaft of the gear 213 via a coupling. The top of the main frame 21 is equipped with a symmetrical double set of gears 213 and racks 211 transmission structure. The servo motor 214 is commanded by a programmable logic controller and moves synchronously and smoothly, and outputs the track direction motion coordinate value in real time to perform 3D scanning of the rail 1 and collect all data. Four laser sensors are equipped with software programs. The software is used for laser ranging and image recognition. Laser ranging technology is used to statically detect geometric dimensions within 2.5 meters at both ends of the rail head, and image recognition technology is used to dynamically detect damage during the operation of the entire rail. The software is used to realize the automatic switching between the two modes of geometric dimension detection and damage detection.

[0035] Please see Figure 2 - Figure 6 As shown, specifically, slide rails 28 are symmetrically installed on both sides of the top of the main frame 21, and sliders 29 are slidably installed on the outer side of the slide rails 28. A top plate 210 is fixedly installed on the top of the movable frame 22 near the slider 29, and the top plate 210 is fixedly connected to the slider 29. Racks 211 are symmetrically installed on both sides of the main frame 21, and side plates 212 are symmetrically installed on both sides of the movable frame 22. A gear 213 is movably installed on the side plate 212 near the rack 211 through a bearing. The gear 213 meshes with the rack 211. A servo motor 214 is installed on the side plate 212 away from the gear 213 through a fixed seat. The output end of the servo motor 214 is connected to the shaft of the gear 213 through a coupling.

[0036] As can be seen from the above, the main frame 21 plays a supporting and stabilizing role. The main frame 21 uses I-beams as crossbeams and columns, with two I-beam crossbeams at the top and bottom, and the crossbeams are fixed with diagonal braces. The slide rail 28 and rack 211 are set on both sides of the I-beams. The minimum height of the crossbeam is 1.8 meters, and the width of the column is 3.5 meters. This facilitates manual operation during equipment maintenance, manual verification of equipment alarms, and calibration of calibrated and damaged rails. The moving frame 22 can move back and forth inside the main frame 21, and synchronously drive the laser frame 23 to move, thereby driving the four laser sensors to move at a constant speed outside the rail 1 for detection. The moving frame 22 achieves stable movement inside the main frame 21 through the cooperation of the top plate 210, the slider 29, and the slide rail 28. Since the gear 213 meshes with the rack 211, when the gear 213 rotates, it will drive the entire moving frame 22 to move. The servo motor 214 provides power output for the rotation of the gear 213.

[0037] Please see Figure 7 - Figure 8 As shown, specifically, two upper mounting plates 215 are fixedly installed on the side of the laser frame 23 near the first laser sensor 24, and the first laser sensor 24 is fixedly installed between the two upper mounting plates 215. A left mounting plate 216 is fixedly installed on the side of the laser frame 23 near the second laser sensor 25, and the second laser sensor 25 is fixedly installed on one side of the left mounting plate 216. A right mounting plate 217 is fixedly installed on the side of the laser frame 23 near the third laser sensor 26, and the third laser sensor 26 is fixedly installed on one side of the right mounting plate 217. An oblique mounting plate 218 is fixedly installed on the side of the laser frame 23 near the fourth laser sensor 27, and the fourth laser sensor 27 is fixedly installed on one side of the oblique mounting plate 218.

[0038] As shown above, laser sensor 24 is located above rail 1 and is responsible for collecting and displaying the rail top curve. Laser sensor 25 is located to the left of rail 1 and is used to collect and display the left side curve of rail 1. Laser sensor 26 is located to the right of rail 1 and can collect and display the right side curve of rail 1. Laser sensor 27 is located diagonally below rail 1 and is specifically used to collect and display the rail bottom curve. Laser sensors 24, 25, 26, and 27 all use domestic Qingbo laser sensors. Among them, laser sensors 24 and 27 are model LH-2340, with an installation distance of 200-220mm and a measurement width of 110-198mm. Laser sensors 25 and 26 are model LH2350, with an installation distance of 357-414mm and a measurement width of 183-357mm.

[0039] Using a multi-point laser displacement sensor, the laser is evenly divided into hundreds or thousands of points. The laser beam is emitted from the emission port towards the object being measured, forming a laser line on the surface of the object. The multi-point laser sensor automatically converts the reflection angle of each laser point into height distance, and the distance between the laser points is calculated as width. Thus, each laser point has two coordinate values. The multi-point laser sensor is an advanced, high-speed, and high-precision measurement tool in the laser field. The longitudinal coordinate value is obtained by the track movement of the laser frame, thus obtaining the 3D data of the rail.

[0040] Four laser sensors are used, placed in the top, bottom, left and right directions of rail 1. The top, left and right lasers face rail 1 directly, while the bottom laser is placed at an angle to avoid affecting the passage of rail 1. The laser lines cover the entire cross-section of rail 1 and collect all data on the surface of rail 1.

[0041] The laser sensor is equipped with software for measuring the geometric dimensions of both ends of the rail head in a static state, and image recognition technology for judging damage in motion. It automatically stitches together point cloud data from multiple laser sensors to generate a full cross-sectional profile. Multiple laser sensors are mutually anti-interference, operate in rotation at high speed, automatically calibrate sensor positions to reduce the frequency of manual maintenance, and have anti-shake structure and software processing for automatic shake correction. The acquired profile can be compared with a standard profile, and 60N and 75N high-density data acquisition modes are available. The minimum acquisition interval for the cross-sectional profile is 1mm to ensure that the missed detection rate is minimized. It has high-efficiency processing capabilities for large batches of data, intelligently filters unnecessary information, that is, data with normal profiles, and only saves abnormal data, reducing hardware computing / storage load and improving the accuracy of manual review. For the acquired point cloud data, it first performs preliminary registration to confirm its basic shape, then converts it into a 2D image and performs grayscale compensation. Machine vision is used to accurately identify surface damage of rail 1. The point cloud data acquired by the four lasers are processed separately and simultaneously to automatically generate key parameters related to the profile of rail 1, such as torsion and straightness, and automatically save and upload the data.

[0042] Before testing, a calibration rail is made and sent to the metrology institute for testing to obtain accurate geometric dimensions. The positions of the four independent lasers are then calibrated one by one, and the distance between the lasers and the rail 1 is adjusted. In daily work, the lasers need to be calibrated regularly using the calibration rail, and the software is used for fine-tuning (see patent: A calibration rail for a laser intelligent inspection equipment for factory-welded rail base materials).

[0043] Before inspection, 18 surface defects were artificially created on the rail head, rail web, and rail bottom to create a damaged rail. The damaged rail was placed at the inspection location, and the laser inspection equipment was activated to allow the laser sensor to scan the damaged rail. Through deep learning algorithms, the laser sensor memorized the shape and depth of the defects. At the same time, image recognition technology was used to store the defect image features in the memory bank so that similar defects could be compared and identified in real time during subsequent inspections.

[0044] During the inspection process, the software automatically performs image stitching, anti-interference processing, automatic calibration, and anti-shake correction. After comparing the collected profile data with the standard profile, it generates an inspection report, which includes inspection data such as rail height, rail head width, rail crown fullness, cross-sectional asymmetry, rail web thickness, rail base width, rail base edge thickness, and rail base indentation, as well as information such as the location and type of damage. At the same time, the software performs efficient calculations on the data, intelligently filters abnormal data, reduces the hardware computing / storage load, and improves the accuracy of manual review.

[0045] Measurement steps:

[0046] S1, Measure the geometric dimensions of the end of rail 1 from 0 to 2.5 meters: Rail 1 is advanced to the designated position and stops. The laser frame 23 moves at a constant speed along the rail direction to perform a full-section scan of the surface of rail 1. The laser sensor emits a laser beam and projects it onto the surface of rail 1 to obtain the coordinate information of each point on the surface of rail 1. The data comparison program is started, and the corresponding built-in rail standard profile database is called. The collected data is compared with the standard profile in real time. If the data exceeds the range, an alarm is triggered, prompting the staff that there is a problem with the geometric dimensions of the end of rail 1.

[0047] S2, Measuring damage across the entire rail: Laser frame 23 returns to the starting point and remains stationary. Rail 1 begins to pass rapidly at a speed of not less than 0.4 meters per second. During the passage of rail 1, the laser sensor collects data in real time, converts the point cloud data into a grayscale image, and starts the image comparison program. The laser collects data in real time, converts the point cloud data into a grayscale image, and uses image recognition technology to call the deep learning memory library to identify surface damage on rail 1. If damage is detected, an alarm is triggered in time, and information such as the location and type of damage is recorded.

[0048] S3, Measure the geometric dimensions of the tail end of rail 1 from 0 to 2.5 meters: Same as the data processing method in the first step, the tail end of rail 1 reaches the designated position and stops. The laser frame 23 moves at a constant speed along the rail direction to perform a full-section scan of the surface of rail 1. Start the data comparison program, call the corresponding built-in rail standard profile database, and compare the collected data with the standard profile in real time. If the data exceeds the range, an alarm will be triggered.

[0049] S4, End of Inspection: Laser frame 23 returns to the starting point, ending the inspection of one rail 1.

[0050] Compared with traditional manual inspection methods, this equipment greatly improves inspection accuracy, can accurately measure the geometric dimensions of rail 1, accurately identify minor damage on the surface of rail 1, enhance the reliability of inspection, eliminate the influence of operator experience on inspection results, avoid judgment errors caused by operator fatigue, and significantly improve inspection efficiency, shorten inspection time, and improve the production efficiency of rail welding plants.

[0051] The accompanying drawings of the embodiments disclosed in this invention only involve structures related to the embodiments disclosed in this invention. Other structures can refer to general designs. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.

[0052] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A laser intelligent inspection equipment for welded steel rail base materials, characterized in that, include: Rail (1); The detection mechanism (2) is located outside the rail (1). The detection mechanism (2) includes a main frame (21). A movable frame (22) is movably installed inside the main frame (21). A laser frame (23) is connected to one side of the movable frame (22). A first laser sensor (24), a second laser sensor (25), a third laser sensor (26), and a fourth laser sensor (27) are installed inside the laser frame (23). The four laser sensors are independent high-density multi-point laser sensors. The laser lines of the four laser sensors cover the entire cross section of the rail (1). The distance between the four lasers and the rail (1) is calibrated using a calibration rail. A rack (211) is installed at the top of the main frame (21). A gear (213) is installed on one side of the movable frame (22). The gear (211) 3) Engages with rack (211). A servo motor (214) is provided on one side of the moving frame (22). The output end of the servo motor (214) is connected to the shaft of the gear (213) through a coupling. A double set of gears (213) and rack (211) transmission structure with left and right symmetry is provided at the top of the main frame (21). The servo motor (214) is directed by a programmable logic controller and moves synchronously and smoothly, and outputs the track direction motion coordinate value in real time to perform 3D scanning on the rail (1) and collect all data. The four laser sensors are equipped with software programs. The software is used for laser ranging and image recognition. Laser ranging technology is used to statically detect geometric dimensions within 2.5 meters at both ends of the rail head. Image recognition technology is used to dynamically detect damage during the operation of the entire rail. The software is used to realize the automatic switching between the two modes of geometric dimension detection and damage detection.

2. The laser intelligent inspection equipment for welded steel rail base materials according to claim 1, characterized in that, The main frame (21) has slide rails (28) symmetrically installed on both sides of the top end. A slider (29) is slidably installed on the outside of the slide rails (28). A top plate (210) is fixedly installed on the top end of the movable frame (22) near the slider (29). The top plate (210) is fixedly connected to the slider (29).

3. The laser intelligent inspection equipment for welded steel rail base materials according to claim 2, characterized in that, The main frame (21) is symmetrically equipped with racks (211) on both sides, and the movable frame (22) is symmetrically equipped with side plates (212) on both sides. A gear (213) is movably mounted on the side plate (212) near the rack (211) via a bearing. The gear (213) meshes with the rack (211). A servo motor (214) is mounted on the side plate (212) away from the gear (213) via a fixed seat. The output end of the servo motor (214) is connected to the shaft of the gear (213) via a coupling.

4. The laser intelligent inspection equipment for welded steel rail base materials according to claim 1, characterized in that, Two upper mounting plates (215) are fixedly installed on the side of the laser frame (23) near the first laser sensor (24), and the first laser sensor (24) is fixedly installed between the two upper mounting plates (215).

5. The laser intelligent inspection equipment for welded rail base materials according to claim 1, characterized in that, The laser frame (23) has a left mounting plate (216) fixedly installed on the side near the second laser sensor (25), and the second laser sensor (25) is fixedly installed on the side of the left mounting plate (216).

6. The laser intelligent inspection equipment for welded steel rail base materials according to claim 1, characterized in that, The laser frame (23) has a right mounting plate (217) fixedly installed on the side near the third laser sensor (26), and the third laser sensor (26) is fixedly installed on the side of the right mounting plate (217).

7. The laser intelligent inspection equipment for welded rail base materials according to claim 1, characterized in that, The laser frame (23) has a slanted mounting plate (218) fixedly installed on the side near the fourth laser sensor (27), and the fourth laser sensor (27) is fixedly installed on the side of the slanted mounting plate (218).

Images