A railway vehicle derailment intelligent monitoring system based on wheel-rail contact trajectory

Abstract

The application discloses a railway vehicle derailment intelligent monitoring system based on wheel-rail contact trajectory, which comprises a contour identification module, a camera posture self-calibration verification module, a contour identification verification module, a wheel set transverse displacement identification verification module and a mechanical simulation module, and is electrically connected between the contour identification module, the camera posture self-calibration verification module, the contour identification verification module, the wheel set transverse displacement identification verification module and the mechanical simulation module. The contour identification module comprises a high-definition camera, the high-definition camera scans the wheel and rail contour, the camera posture self-calibration verification module is used for adjusting the shooting angle of the high-definition camera, the contour identification verification module identifies and verifies the wheel and rail contour shot by the high-definition camera, and the data information obtained by the contour identification module, the camera posture self-calibration verification module, the contour identification verification module and the wheel set transverse displacement identification verification module is displayed through the mechanical simulation module.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of rail transit, and particularly relates to a railway vehicle derailment intelligent monitoring system based on wheel-rail contact trajectory. BACKGROUND

[0002] The lateral stability of a train is an important index for judging the running performance of the train, and the main purpose is to avoid the snake instability of the train. When the train appears to be in snake instability, the wheelset lateral displacement of the train will increase, and even the wheel flange will collide with the track. In addition, the lateral vibration of the wheelset will also induce the large lateral vibration of the bogie and the car body. Once such a phenomenon occurs, the running performance of the train will deteriorate, the ride comfort will decrease, the dynamic load acting on each part of the train will increase, the wheel-rail force will rise, and the vehicle and the track line will be damaged, and even a derailment accident may occur.

[0003] When the train is running, the wheel-rail contact trajectory is constantly changing due to the influence of factors such as vehicle system characteristics, wheel-rail profile, curve line, track irregularity, etc. In order to realize the online detection of the wheel-rail contact trajectory, a specific method is used to detect the wheel-rail profile and the wheelset lateral displacement, and then the wheel-rail contact trajectory is obtained. At present, there are few online detection methods for wheel-rail relative displacement, which are mainly divided into the following three categories: (1) thermal imaging and laser detection method; (2) template matching and edge detection method; (3) virtual point detection method based on deep learning. At the same time, there are very few studies on the safety evaluation using the wheel-rail contact trajectory, and there is no clear standard for judging safety. SUMMARY

[0004] The purpose of the present application is to solve the above problems, and to provide a railway vehicle derailment intelligent monitoring system based on wheel-rail contact trajectory, which has high recognition accuracy and can automatically generate a wheel-rail contact trajectory dynamic video, and dynamically displays the wheel-rail contact point in a three-dimensional coordinate system.

[0005] In order to solve the above technical problems, the technical scheme of the present application is: a railway vehicle derailment intelligent monitoring system based on wheel-rail contact trajectory, comprising a contour recognition module, a camera pose self-calibration verification module, a contour recognition verification module, a wheel set lateral displacement recognition verification module and a mechanical simulation module, the contour recognition module, the camera pose self-calibration verification module, the contour recognition verification module, the wheel set lateral displacement recognition verification module and the mechanical simulation module are electrically connected, the contour recognition module comprises a high-definition camera, the high-definition camera scans the wheel and rail contour, the camera pose self-calibration verification module is used for adjusting the shooting angle of the high-definition camera, the contour recognition verification module identifies and verifies the wheel and rail contour photographed by the high-definition camera, the wheel set lateral displacement recognition verification module identifies and verifies the wheel set lateral displacement in the wheel and rail photographed by the high-definition camera, and the data information obtained by the contour recognition module, the camera pose self-calibration verification module, the contour recognition verification module and the wheel set lateral displacement recognition verification module is displayed through the mechanical simulation module.

[0006] The present application also discloses a railway vehicle derailment safety evaluation method based on wheel-rail contact trajectory, comprising the following steps:

[0007] S1, through dynamic simulation and bench test, the mapping relationship between the safety indexes such as derailment coefficient, wheel load reduction rate and wheel axle lateral force and the wheel-rail contact point position is obtained by using the force wheel set;

[0008] S2, the collected image is preprocessed;

[0009] S3, the difference of contour identification caused by different angles between the laser line and the camera is measured;

[0010] S4, the algorithm is verified by static test;

[0011] S5, dynamic verification is carried out on the rolling vibration test bed.

[0012] Further, in the step S1, the different positions of the wheel-rail contact point position on the tread are divided into safety levels through the mapping relationship, and more accurate safety evaluation criteria are formulated in combination with the motion trend of the contact point.

[0013] Further, the step S2 further comprises the following sub-steps:

[0014] S21, the preprocessing comprises irrelevant area cropping, light supplementing, color space conversion, adaptive brightness correction, gray processing and noise reduction processing;

[0015] S22, after the image preprocessed in the step S21 is processed by the Canny edge detection algorithm, the steel rail rough positioning processing is carried out, then the left and right edges of the steel rail are detected by using the Hough straight line detection algorithm, and the steel rail positioning is realized;

[0016] S23, the laser trajectory on the rail is recognized as a reference line, and the original profile of the wheel and the rail is obtained by constructing a conversion model of pixel distance to actual distance.

[0017] Further, the laser line measured in the step S3 is measured by a 2D laser displacement sensor to measure the rail profile, and the laser auxiliary line of the 2D laser displacement sensor is used as an auxiliary line to identify the rail profile.

[0018] Further, the included angle between the laser auxiliary line of the 2D laser displacement sensor and the camera is greater than 40°.

[0019] Further, in the step S4, the algorithm is verified by a static test, the wheelset lateral displacement curve is obtained by the difference between the distances of each frame of image, and the wheelset lateral displacement curve in the time domain is restored according to the actual sampling frequency of the high-definition camera according to the sampling theorem.

[0020] Further, the dynamic verification is performed in the step S5.

[0021] The beneficial effects of the present application are that the railway vehicle derailment intelligent monitoring system based on the wheel-rail contact trajectory provided by the present application proposes an image recognition method assisted by a grating laser light source, which greatly improves the accuracy of recognition. The camera self-calibration technology and the conversion model are used to obtain the wheel-rail profile and the wheelset lateral displacement, and to provide conditions for wheel-rail contact calculation. The snake instability recognition based on the wheelset lateral displacement and the derailment safety evaluation system, through a large number of test bench tests, establishes the mapping relationship between the safety evaluation system based on the wheel-rail contact trajectory and the safety evaluation system based on the wheel-rail force. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 is a curve passing through the derailment coefficient and the wheel-rail contact point position relationship diagram in the railway vehicle derailment intelligent monitoring system based on the wheel-rail contact trajectory of the present application;

[0023] Figure 2 is a quasi-static wheelset lateral displacement recognition verification diagram of the present application;

[0024] Figure 3 is a wheel and rail profile recognition principle flowchart of the present application;

[0025] Figure 4 is a wheel and rail profile recognition verification diagram of the present application;

[0026] Figure 5 is a wheel lateral displacement verification diagram of the present application;

[0027] Figure 6 is a wheel lateral displacement time domain curve diagram of the present application;

[0028] Figure 7 This is the verification work of the present invention based on a rolling vibration test bench;

[0029] Figure 8 This is a flowchart of the workflow of the present invention;

[0030] Figure 9 This is a schematic diagram of the test bench verification of the present invention;

[0031] Figure 10 This invention compares the measured and identified values ​​of the lateral relative displacement between the wheel and rail. Detailed Implementation

[0032] The present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0033] like Figures 1 to 10 As shown, this invention provides an intelligent monitoring system for railway vehicle derailment based on wheel-rail contact trajectory. It includes a contour recognition module, a camera attitude self-calibration verification module, a contour recognition verification module, a wheelset lateral displacement recognition verification module, and a mechanical simulation module. These modules are electrically connected. The contour recognition module includes a high-definition camera that scans the wheel and rail contours. The camera attitude self-calibration verification module adjusts the shooting angle of the high-definition camera. The contour recognition verification module identifies and verifies the wheel and rail contours captured by the high-definition camera. The wheelset lateral displacement recognition verification module identifies and verifies the wheelset lateral displacement captured by the high-definition camera. The data obtained by these modules is displayed through the mechanical simulation module. All modules are electrically connected according to actual usage needs to facilitate data transmission.

[0034] This invention focuses on four main aspects: wheel and rail contour recognition, camera attitude self-calibration verification, wheel and rail contour recognition verification, and wheelset lateral displacement recognition verification. The result is a railway vehicle derailment safety evaluation method and intelligent monitoring system.

[0035] In the embodiment, the application is based on a kinetic simulation or a bench test to carry out a correlation study between a wheel-rail running contact trajectory and a vehicle dynamic performance, to study a wheel-rail contact trajectory law of a lateral motion of a wheel set when the wheel set is in a hunting instability, to study a position of a wheel-rail contact point on a wheel tread before derailment, to realize a judgment of a risk of hunting instability and derailment of the vehicle through the wheel-rail running contact trajectory, to construct a vehicle running safety evaluation system based on the wheel-rail contact trajectory, to construct an intelligent monitoring system of the railway vehicle based on the wheel-rail contact trajectory, to construct a set of wheel-rail running contact trajectory calculation and analysis software, to study a wheel-rail profile accurate matching method, to realize calculation and analysis of the wheel-rail running contact trajectory of the heavy haul railway wagon, to automatically generate a wheel-rail contact trajectory dynamic video, to dynamically display the wheel-rail contact point in a three-dimensional coordinate system, to automatically analyze and identify the risk of hunting instability and derailment of the vehicle according to the wheel-rail running contact point, to divide the risk degree into three levels of normal, early warning and alarm, and to timely send early warning and alarm information.

[0036] The application further discloses a railway vehicle derailment safety evaluation method based on a wheel-rail contact trajectory, which comprises the following steps.

[0037] S1, through kinetic simulation and bench test, a mapping relationship between safety indexes such as a derailment coefficient, a wheel load reduction rate and a wheel axle lateral force and a wheel-rail contact point position is obtained by using a force measuring wheel set.

[0038] In step S1, different positions of the wheel-rail contact point position on the tread are divided into safety levels through the mapping relationship, and a more accurate safety evaluation criterion is formulated in combination with a motion trend of the contact point.

[0039] Through kinetic simulation and bench test, a mapping relationship between safety indexes such as a derailment coefficient, a wheel load reduction rate and a wheel axle lateral force and a wheel-rail contact point position is obtained by using a force measuring wheel set, as shown in FIG. Figure 1 It can be seen from the figure that there is a relatively obvious linear relationship between the position of the outer rail wheel-rail contact point and the derailment coefficient when the position of the contact point is more than 30 mm. Different positions of the wheel-rail contact point position on the tread can be divided into safety levels through the mapping relationship, and a more accurate safety evaluation criterion is formulated in combination with a motion trend of the contact point.

[0040] S2, the collected images are preprocessed.

[0041] Step S2 further comprises the following sub-steps:

[0042] S21, the preprocessing comprises irrelevant area cropping, light supplementing processing, color space conversion, adaptive brightness correction, gray scale processing and noise reduction processing. The result is shown in FIG. Figure 3 .

[0043] S22, after the image pre-processed in step S21 is processed by Canny edge detection algorithm, rough positioning of the rail is processed, and then Hough straight line detection algorithm is used to detect the left and right edges of the rail to realize positioning of the rail.

[0044] S23, the laser track on the rail is identified and taken as a reference line, and a conversion model of pixel distance to actual distance is constructed to obtain the original profile of the wheel and the rail.

[0045] S3, the difference of profile identification caused by different angles between the laser line and the camera is measured.

[0046] In step S3, the 2D laser displacement sensor is used to measure the rail profile, and the laser line of the 2D laser displacement sensor is used as an auxiliary line to identify the rail profile.

[0047] As shown in Figure 4 , the 2D laser displacement sensor is used to measure the rail profile, and the laser line of the 2D laser displacement sensor is used as an auxiliary line to identify the rail profile, and the difference of profile identification caused by different angles between the laser line and the camera is measured. As can be seen from the results, in order to achieve sufficient identification accuracy, the angle between the laser auxiliary line of the 2D laser displacement sensor and the camera is greater than 40°.

[0048] S4, static test verification of the algorithm.

[0049] In step S4, the static test verification of the algorithm is performed, the wheelset lateral displacement curve is obtained by subtracting the distance between each frame of image, and the wheelset lateral displacement curve in time domain is restored according to the actual sampling frequency of the high-definition camera according to the sampling theorem.

[0050] As shown in Figure 5 and Figure 6 , the test verification of the algorithm is performed, and the wheelset lateral displacement curve is obtained by subtracting the distance between each frame of image. Figure 5 In the figure, the maximum error is 0.2mm, and in the lower Figure 6 , the wheelset lateral displacement curve in time domain is restored according to the actual sampling frequency of the camera according to the sampling theorem.

[0051] S5, dynamic verification on a rolling vibration test bench.

[0052] In step S5, in the dynamic verification, the rollers of the test bench are adjusted by the hydraulic system of the test bench, the lateral displacement of the hydraulic system is y r , the lateral displacement of the wheel relative to the absolute coordinate system can be tested, the lateral displacement of the wheel is y w , and the lateral relative displacement of the wheel and the roller is defined as y wr1, represents the lateral relative displacement of the wheel on the rail, and the calculation formula is y wr1 = y r -y w .

[0053] In the dynamic verification process, the lateral displacement of the roller is 2mm, 4mm, 6mm and 8mm in turn, and a certain time is kept at each lateral displacement, and the lateral relative displacement y wr1 of the wheel and the roller obtained is y wr2 As shown in Figure 10 , from the results, the change laws of the two are similar, and the maximum error is not more than 0.2mm.

[0054] As shown in Figure 7 , after static verification, dynamic verification is also carried out on the rolling vibration test bed, and the verification test of the application is carried out when the vehicle dynamics rolling vibration test is carried out.

[0055] As shown in Figure 8 , the application first identifies the profiles of the wheel and the rail, then self-calibrates the camera posture, then sequentially verifies the profile identification of the wheel and the rail and the lateral displacement identification of the wheelset, and finally evaluates the safety.

[0056] Those skilled in the art will realize that the embodiments described herein are for the purpose of helping the reader to understand the principles of the application and should be understood as not limiting the protection scope of the application to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations according to the technical inspirations disclosed in the application without departing from the essence of the application, and these modifications and combinations are still within the protection scope of the application.

Claims

1. A railway vehicle derailment intelligent monitoring system based on wheel-rail contact trajectory, characterized in that: The system includes a contour recognition module, a camera attitude self-calibration verification module, a contour recognition verification module, a wheelset lateral displacement recognition verification module, and a mechanical simulation module. These modules are electrically connected. The contour recognition module includes a high-definition camera that scans the contours of the wheels and rails. The camera attitude self-calibration verification module is used to adjust the shooting angle of the high-definition camera. The contour recognition verification module identifies and verifies the wheel and rail contours captured by the high-definition camera. The wheelset lateral displacement recognition verification module identifies and verifies the lateral displacement of the wheelset in the wheels and rails captured by the high-definition camera. The data obtained by these modules is displayed through the mechanical simulation module. The system is used to implement a railway vehicle derailment safety assessment method based on wheel-rail contact trajectory, including the following steps: S1. Through dynamic simulation and bench tests, the mapping relationship between the derailment coefficient, wheel load reduction rate and wheel axle lateral force safety index and the wheel-rail contact point position is obtained using force-measuring wheelsets. In step S1, the safety levels of the wheel-rail contact point at different positions on the tread are classified through mapping relationships, and a more accurate safety evaluation criterion is formulated by combining the movement trend of the contact point. Through dynamic simulation and bench testing, the mapping relationship between the derailment coefficient, wheel load reduction rate, and wheel axle lateral force safety index and the wheel-rail contact point position is obtained using force-measuring wheelsets. Through the mapping relationship, the safety levels of the wheel-rail contact point at different positions on the tread can be classified, and a more accurate safety evaluation criterion can be formulated by combining the movement trend of the contact point. S2. Preprocess the acquired images; S3. Measure the difference in profile recognition at different angles between the laser line and the camera; In step S3, the laser line is measured using a 2D laser displacement sensor to measure the rail profile, and the laser steel line of the 2D laser displacement sensor is used as an auxiliary line for rail profile recognition. S4. Perform static experiments to verify the algorithm; S5. Perform dynamic verification on a rolling vibration test bench; Step S2 further includes the following sub-steps: S21. Preprocessing includes: irrelevant area cropping, fill light processing, color space conversion, adaptive brightness correction, grayscale processing, and noise reduction processing. S22. After processing the preprocessed image in step S21 with the Canny edge detection algorithm, perform coarse positioning of the rail, and then use the Hough line detection algorithm to detect the left and right edges of the rail to achieve rail positioning. S23. Identify the laser trajectory on the rail and use it as a reference line. By constructing a conversion model from pixel distance to actual distance, obtain the original contours of the wheel and rail.

2. The intelligent monitoring system for railway vehicle derailment based on wheel-rail contact trajectory according to claim 1, characterized in that, In step S3, the laser line is measured using a 2D laser displacement sensor to measure the rail profile, and the laser steel line of the 2D laser displacement sensor is used as an auxiliary line for rail profile recognition.

3. The intelligent monitoring system for railway vehicle derailment based on wheel-rail contact trajectory according to claim 1, characterized in that, The angle between the laser auxiliary line of the 2D laser displacement sensor and the camera is greater than 40°.

4. The intelligent monitoring system for railway vehicle derailment based on wheel-rail contact trajectory according to claim 1, characterized in that, In step S4, the algorithm is statically tested and verified. The wheel pair lateral movement curve is obtained by subtracting the distance between each frame image. The time-domain wheel pair lateral movement curve is restored based on the actual sampling frequency of the high-definition camera using the sampling theorem.

5. The intelligent monitoring system for railway vehicle derailment based on wheel-rail contact trajectory according to claim 1, characterized in that, Dynamic verification is performed in step S5.

Images