A method and system for automatic wheel flaw detection and wheelset parallelism correction

Abstract

The present application provides a method and system for automated wheel flaw detection and wheelset parallelism correction. Specifically, the method collects dual distance data from the inner side of the rim in real time through a distance measuring sensor, and calculates the wheel deflection angle in combination with the vertical installation spacing. A first error signal is generated based on a closed-loop model comparison with a preset threshold value, and the air film pressure of the leveling mechanism is driven to dynamically compensate for the offset. The air film pressure change and time domain characteristics are synchronously extracted to generate an inclination instruction, and the servo motor is controlled to adjust the inclination of the support surface to form an adaptive detection trajectory. During the movement of the wheelset, a second error signal is generated by the difference between the two sets of distances, and when the limit is exceeded, a multi-parameter collaborative correction is triggered until the trajectory is stabilized and meets the standard. Adaptive dynamic tracking and millimeter-level precision control of the wheelset detection path are realized. The present application realizes micron-level adaptive tracking of the detection trajectory for complex wheelsets through air film dynamic compensation and multi-parameter collaborative control.

Description

Technical Field

[0001] The present application relates to the technical field of wheel flaw detection and parallelism correction, and in particular to a method and system for automated wheel flaw detection and wheelset parallelism correction. Background Art

[0002] Rail vehicle wheels are susceptible to fatigue defects under long-term alternating stress, posing a direct threat to operating safety. This requires efficient online ultrasonic testing. Because testing must be performed without disassembling the wheelset, the inspection system must possess precise wheelset positioning capabilities to ensure strict alignment between the lifting device and the wheelset axis, preventing secondary damage such as wheel rim abrasions caused by misaligned lifting. With the high frequency of rail transit and increasingly complex line conditions, dynamic offset caused by the serpentine motion of wheelsets is becoming increasingly prominent, placing higher demands on the dynamic adaptive adjustment capabilities of the inspection equipment.

[0003] The current mainstream technology utilizes a coordinated operation of a fixed rail bridge and a mobile flaw detection trolley. This dual-track system, consisting of a dedicated overhead rail bridge and an underground flaw detection track, utilizes a jacking and rotating wheel device mounted on the flaw detection trolley to perform wheel lift and rotation testing. The system uses laser ranging and a mechanical compensation mechanism to correct static installation deviations between the rail and the flaw detection track. A multi-degree-of-freedom adjustable probe mount achieves adaptive alignment to the wheel's curved surface. The inspection process relies on preset track position parameters, with a servo control system providing coarse positioning of the trolley and wheelset.

[0004] However, the existing system can only compensate for the static position deviation during the track installation phase and cannot effectively deal with the dynamic offset caused by the serpentine motion of the train's rear wheels due to braking. Summary of the Invention

[0005] The present application provides an automated wheel flaw detection and wheelset parallelism correction method and system to solve the problems of low detection accuracy, poor trajectory adaptability and insufficient correction stability in the prior art.

[0006] In a first aspect, the present application provides an automated wheel flaw detection and wheelset parallelism correction method, wherein a set of distance measuring sensors are arranged in opposite directions of top rotating wheel devices on both sides, and the top rotating wheel device on each side is used to support one wheel of the wheelset. The method includes:

[0007] The first distance data and the second distance data of the inner side of the corresponding wheel rim are collected by each set of distance measuring sensors respectively, and the deflection angle between the top rotating wheel device on each side and the corresponding wheel is calculated based on the installation vertical distance of each set of distance measuring sensors;

[0008] A closed-loop control model is constructed to compare the deflection angles corresponding to the wheels on both sides with a preset angle threshold to generate a first error signal. Based on the direction and amplitude of the first error signal, the leveling mechanisms on both sides are driven to generate an air film pressure distribution through the air bearing, wherein the air film pressure distribution forms a dynamic compensation relationship with the deflection direction of the wheel;

[0009] Synchronously extracting the air film pressure change of the leveling mechanism and, in combination with the time domain characteristics of the first error signal, generating inclination adjustment instructions for the servo motors on both sides, thereby controlling the servo motors on both sides to adjust the spatial inclination of the support surface through the inclination adjustment instructions, so that the wheelset tread and the detection path of the flaw detector form an adaptive following trajectory;

[0010] During the continuous rotation of the wheelset, the distance signal difference of each group of distance measuring sensors is calculated respectively, and the distance signal difference of the two groups is input into the closed-loop control model to generate a second error signal; if the second error signal exceeds the preset tolerance range, the multi-parameter coordinated correction of the leveling mechanisms and servo motors on both sides is triggered until the stability of the adaptive following trajectory meets the preset control accuracy.

[0011] Optionally, it also includes:

[0012] During the rotation of the wheelset, the temperature distribution data of the two wheel rims are collected respectively by the temperature sensors attached to the top rotating wheel devices on both sides;

[0013] Based on the temperature distribution data and the thermal expansion coefficient of the rim material, the first distance data and the second distance data collected by each set of distance measuring sensors are corrected to generate the temperature-compensated deflection angles of the wheels on both sides;

[0014] The deflection angles of the wheels on both sides are input into a closed-loop control model to drive the leveling mechanisms on both sides and the servo motors to synchronously perform temperature-adaptive pressure distribution adjustment and inclination adjustment to offset the effect of thermal deformation on the parallelism correction of the wheelset.

[0015] Optionally, it also includes:

[0016] When the wheelset is stationary, the distance measuring sensor groups of the top rotating wheel devices on both sides scan the circumferential contours of the inner side surfaces of the two wheel rims respectively to extract the asymmetric geometric features of the two wheels;

[0017] generating structural compensation parameters of the two wheels according to the asymmetric geometric features, and embedding the structural compensation parameters into the closed-loop control model respectively;

[0018] When the wheelset rotates, the pressure distribution gradient of the leveling mechanisms on both sides and the inclination adjustment amount of the servo motor are corrected according to the superposition result of the deflection angle of the wheels on both sides and the corresponding structural compensation parameters, so that the wheelset tread and the adaptive following trajectory adapt to the dynamic balance requirements of the asymmetric wheelset.

[0019] Optionally, comparing the deflection angle with a preset angle threshold to generate a first error signal, and driving a leveling mechanism to generate a non-uniform air film pressure distribution through an air bearing according to a direction and an amplitude of the first error signal, includes:

[0020] Comparing the deflection angles of the wheels on both sides with a preset angle threshold to generate a first error signal including direction and amplitude;

[0021] determining, according to the direction of the first error signal, the left side or the right side of the air bearing in the leveling mechanisms on both sides as the pressure compensation side;

[0022] Based on the amplitude of the first error signal, proportionally adjust the air film pressure on the pressure compensation side of the leveling mechanisms on both sides so that the air film pressure of the leveling mechanisms on both sides forms a non-uniform pressure difference opposite to the deflection direction of the corresponding wheel;

[0023] The non-uniform pressure difference drives the wheels to dynamically adjust in a direction that offsets the deflection.

[0024] Optionally, the synchronous extraction of the air film pressure variation of the leveling mechanism, combined with the time domain characteristics of the first error signal, generates an inclination adjustment instruction for the servo motor, and controls the servo motor to adjust the spatial inclination of the support surface through the inclination adjustment instruction, so that the wheelset tread and the detection path of the flaw detector form an adaptive following trajectory, including:

[0025] By extracting the air film pressure changes of the leveling mechanisms on both sides, the air film pressure difference of the leveling mechanisms on both sides is determined, and the change trend of the first error signals on both sides over time is extracted;

[0026] Generate the dynamic gain values ​​of the servo motors on both sides according to the pressure difference on both sides and the rate of change of the corresponding side;

[0027] Determining the inclination adjustment amount of the servo motors on both sides according to the correlation relationship between the dynamic gain value and the first error signal on the corresponding side;

[0028] The servo motor drives the supporting surfaces on both sides to independently rotate the inclination adjustment amount around the transverse axis, so that the inclination direction of the wheelset tread is dynamically matched with the detection path of the flaw detector.

[0029] Optionally, during the continuous rotation of the wheelset, a distance signal difference of the distance measuring sensor group is calculated, and the distance signal difference is input into the closed-loop control model to generate a second error signal; if the second error signal exceeds a preset tolerance range, a multi-parameter coordinated correction of the leveling mechanism and the servo motor is triggered until the stability of the adaptive following trajectory meets a preset control accuracy, including:

[0030] When the wheelset rotates continuously, the distance signal difference of the distance measuring sensor group of the top rotating wheel device on both sides is calculated respectively, and the difference between the two distance signal differences is input into the closed-loop control model to generate a second error signal;

[0031] According to the duration of the second error signal, the pressure compensation ratio of the leveling mechanisms on both sides and the inclination adjustment sensitivity of the servo motor are adjusted, and by cyclically adjusting the pressure compensation ratio and the inclination adjustment sensitivity, the difference between the two distance signals falls back to the preset tolerance range, and the adaptive following stability of the wheelset and the detection path is maintained.

[0032] Optionally, collecting the first distance data and the second distance data of the inner side of the corresponding wheel rim through each set of distance measuring sensors, and calculating the deflection angle between the top rotating wheel device on each side and the corresponding wheel in combination with the installation vertical distance of each set of distance measuring sensors, includes:

[0033] A first distance measuring sensor and a second distance measuring sensor are fixed to the top rotating wheel device on each side, and the detection directions of the first distance measuring sensor and the second distance measuring sensor are set to point vertically to symmetrical positions on the inner side of the rim;

[0034] The first distance value and the second distance value of the inner side of the wheel rim to the sensor are respectively collected in real time by the first distance measuring sensor and the second distance measuring sensor;

[0035] Obtaining the installed vertical distance between the first distance measuring sensor and the second distance measuring sensor in the top rotating wheel device on each side;

[0036] The deflection angle α between the top turning wheel device on each side and the corresponding wheel is calculated according to the formula: sinα=(L1-L2) / L, where L1 represents the first distance value, L2 represents the second distance value, and L represents the installation vertical distance; when L1>L2, α is a positive deflection angle, and when L1<L2, α is a negative deflection angle.

[0037] In a second aspect, the present application provides an automated wheel flaw detection and wheelset parallelism correction system, comprising:

[0038] a calculation module, configured to collect first distance data and second distance data of the inner side surface of the rim of the corresponding wheel through each set of distance measuring sensors, and calculate the deflection angle between the top rotating wheel device on each side and the corresponding wheel in combination with the vertical distance between the installation of each set of distance measuring sensors;

[0039] a generation module, configured to construct a closed-loop control model, compare the deflection angles corresponding to the wheels on both sides with a preset angle threshold to generate a first error signal, and drive the leveling mechanisms on both sides to generate an air film pressure distribution through the air bearing based on the direction and amplitude of the first error signal, wherein the air film pressure distribution forms a dynamic compensation relationship with the deflection direction of the wheel;

[0040] an adjustment module, configured to synchronously extract the air film pressure variation of the leveling mechanism and, in combination with the time domain characteristics of the first error signal, generate an inclination adjustment instruction for the servo motors on both sides, so as to control the servo motors on both sides to adjust the spatial inclination of the support surface through the inclination adjustment instruction, so that the wheelset tread and the detection path of the flaw detector form an adaptive following trajectory;

[0041] The correction module is used to calculate the distance signal difference of each group of distance measuring sensors during the continuous rotation of the wheelset, and input the distance signal difference of the two groups into the closed-loop control model to generate a second error signal; if the second error signal exceeds the preset tolerance range, it triggers the multi-parameter coordinated correction of the leveling mechanisms and servo motors on both sides until the stability of the adaptive following trajectory meets the preset control accuracy.

[0042] In a third aspect, an embodiment of the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an automated wheel flaw detection and wheelset parallelism correction method as described in the first aspect above.

[0043] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements an automated wheel flaw detection and wheelset parallelism correction method as described in the first aspect.

[0044] In this application's technical solution, dual-sided distance sensors collect rim data in real time and calculate deflection angles. This is combined with a closed-loop control model to dynamically generate error compensation instructions, driving the air film pressure distribution and wheel deflection to form inverse dynamic compensation. Servo motor inclination adjustment allows precise tracking of the wheelset tread and the inspection path, while maintaining trajectory stability in real time during continuous rotation based on a multi-parameter collaborative correction mechanism. This effectively addresses the low efficiency and poor accuracy of traditional manual correction, significantly improving the reliability of wheelset flaw detection and the adaptive capabilities of parallelism adjustment, ensuring dynamic matching of the inspection process with the wheel position and posture.

[0045] Furthermore, by integrating temperature compensation and structural compensation mechanisms, the accuracy and environmental adaptability of wheelset flaw detection and parallelism correction are further improved. Based on the temperature sensor, the rim temperature distribution is collected in real time and the distance measurement data is corrected. In combination with the thermal expansion characteristics, the pressure and inclination are dynamically adjusted to effectively eliminate the measurement error and correction deviation caused by thermal deformation. At the same time, the asymmetric geometric features of the rim are extracted through static contour scanning, and the structural compensation parameters are generated and embedded in the control model. The compensation correction adjustment instructions are superimposed during dynamic rotation, enabling the system to actively adapt to the asymmetric shape of the wheel and balance the force and posture of the wheelset during dynamic operation. The two compensation mechanisms work together to overcome the influence of ambient temperature fluctuations on measurement accuracy and solve the trajectory mismatch problem caused by wheel geometric irregularities. This significantly improves the robustness and correction consistency of the system under complex working conditions, ensuring that the adaptive following trajectory always maintains the dynamic optimal match with the actual shape of the wheelset.

[0046] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0048] Figure 1 A schematic structural diagram of an automated wheel flaw detection and wheelset parallelism correction system provided in an embodiment of the present application is shown;

[0049] Figure 2 A flow chart of an automated wheel flaw detection and wheelset parallelism correction method provided by an embodiment of the present application is shown;

[0050] Figure 3 A scene diagram showing an automated wheel flaw detection and wheelset parallelism correction method provided by an embodiment of the present application is shown;

[0051] Figure 4 An architectural diagram of an automated wheel flaw detection and wheelset parallelism correction system is provided for an embodiment of the present application.

[0052] Figure 5 A schematic diagram of the structure of a computing device that can be implemented in an automated wheel flaw detection and wheelset parallelism correction system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0054] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0055] Research has found that while current wheel flaw detection technology can compensate for static installation deviations through track correction, it still has significant limitations when dealing with the dynamic offset of wheelsets caused by the serpentine motion of trains. Studies have shown that when the offset between the centerline of the wheelset and the axis of the flaw detection trolley exceeds 3mm, the non-centering contact between the rigid jacking device and the wheelset will cause the local contact stress of the wheel rim to surge to 15-20MPa. At the same time, an angle deviation of 0.5°-1.5° is formed between the jacking axis and the actual axis of the wheelset. This deviation will cause periodic scratches on the wheel rim during wheel rotation. Test data shows that at a speed of 200r / min, a damage groove with a depth of more than 0.2mm can be formed in just 5 minutes, posing a serious threat to the service safety of the wheelset. Existing technologies lack the ability to perceive dynamic offsets in real time and adaptively compensate for them, making it difficult to effectively suppress this type of secondary damage.

[0056] To address these shortcomings, this application proposes an automated wheel flaw detection and wheelset parallelism correction method based on dynamic deflection closed-loop control. By installing distance sensor arrays on both sides of the top-rotating wheel assembly, real-time rim inboard distance data is collected and the deflection angle is calculated. An error-driven closed-loop control model is constructed, combining a coordinated control mechanism of air bearing film pressure distribution and servo motor tilt adjustment to achieve real-time compensation for dynamic wheel offset. Specifically, the system uses a first error signal to drive a leveling mechanism to generate an air film pressure field that dynamically matches the deflection direction. Simultaneously, the system adjusts the spatial posture of the support surface using servo motor tilt adjustment commands, ensuring that the wheelset tread and the detection path form an adaptive tracking trajectory. Furthermore, multi-parameter coordinated corrections are triggered based on the distance signal difference during continuous wheel rotation to ensure trajectory stability within a ±0.1mm control accuracy. Through a three-stage control system consisting of dynamic sensing, pressure compensation, and tilt tracking, this method reduces wheel rim contact stress to below 5MPa, eliminating the risk of scratches caused by non-coaxial rotation and fundamentally addressing the existing problem of wheelset damage caused by dynamic offset.

[0057] Before describing the automated wheel flaw detection and wheelset parallelism correction method provided by this application, a brief description of the overall structural arrangement of the solution is given:

[0058] like Figure 1 As shown, the present application arranges a group of ranging sensors in the relative directions of the top rotating wheel devices A and B on both sides respectively. The top rotating wheel device on each side is used to support one wheel in the wheelset, wherein the ranging sensor group of the top rotating wheel device A includes a first ranging sensor A1 and a second ranging sensor A2. The A1 and A2 use at least one of laser ranging sensors, ultrasonic ranging sensors or infrared ranging sensors, which are fixed on both sides of the transverse support frame of the top rotating wheel device A, and the detection direction is perpendicular to the axially symmetrical point on the inner side of the corresponding wheel rim.

[0059] The distance measuring sensor group of the top rotating wheel device B includes a first distance measuring sensor B1 and a second distance measuring sensor B2. B1 and B2 are arranged in the same structure as A1 and A2, and are symmetrically embedded in the axial end face of the top rotating wheel device B. The detection direction is at a preset inclination angle with the wheel rotation plane to adapt to the rim curvature.

[0060] The lengths of the installation baselines between A1 and A2 and between B1 and B2 are equal and parallel to the transverse axis of the top rotating wheel device, forming a bilaterally symmetrical rim deformation detection array.

[0061] This application mainly realizes automated wheel flaw detection and parallelism correction of wheelsets by separately collecting relevant data of the top rotating wheel device on each side. The technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making any creative work shall fall within the scope of protection of this application.

[0062] Figure 2 A flowchart of an automated wheel flaw detection and wheelset parallelism correction method is provided for the embodiment of the present application. Figure 2 As shown, the method includes:

[0063] 201. Collect first distance data and second distance data of the inner side of the corresponding wheel rim through each set of distance measuring sensors, and calculate the deflection angle between the top rotating wheel device on each side and the corresponding wheel in combination with the installation vertical distance of each set of distance measuring sensors;

[0064] As a possible implementation solution, step 201 may specifically include:

[0065] 2011. Fix a first distance measuring sensor and a second distance measuring sensor to the top rotating wheel device on each side, and set the detection directions of the first distance measuring sensor and the second distance measuring sensor to point vertically to symmetrical positions on the inner side of the rim;

[0066] 2012. Using the first distance measuring sensor and the second distance measuring sensor, respectively, collect in real time a first distance value and a second distance value from the inner side of the corresponding wheel rim to the sensor;

[0067] 2013. Obtain the installation vertical distance between the first distance measuring sensor and the second distance measuring sensor in the top rotating wheel device on each side;

[0068] 2014. The deflection angle α between the top turning wheel device on each side and the corresponding wheel is calculated according to the formula: sinα=(L1-L2) / L, where L1 represents the first distance value, L2 represents the second distance value, and L represents the installation vertical distance; when L1>L2, α is a positive deflection angle, and when L1<L2, α is a negative deflection angle.

[0069] In the above solution, the distance sensor is a detection device used to measure the distance between an object and the sensor. The inner side of the wheel rim is the annular surface on the inner side of the wheel opposite the sensor. The first distance data is the raw distance measurement value collected by the first distance sensor. The second distance data is the raw distance measurement value collected by the second distance sensor. The installation vertical distance is the fixed vertical spacing between two distance sensors on the same side. The top rotating wheel assembly is a mechanical structure used to support and adjust the position of the wheel. The yaw angle is the tilt angle of the wheel relative to the standard position. The first distance sensor is the primary distance measuring device mounted on the top rotating wheel assembly. The second distance sensor is the auxiliary distance measuring device mounted on the top rotating wheel assembly. The detection direction is the direction in which the distance sensor transmits and receives signals. The symmetrical position is the mirror image distribution point on the inner side of the wheel rim relative to the sensor mounting point. The first distance value is the specific value measured in real time by the first distance sensor. The second distance value is the specific value measured in real time by the second distance sensor. A positive yaw angle is the angle value when the wheel is tilted in a specific direction. A negative yaw angle is the angle value when the wheel is tilted in the opposite direction.

[0070] In this embodiment, first, in step 2011, a first ranging sensor and a second ranging sensor are mounted on the top rotating wheel assembly on either side of the vehicle. Mechanical calibration is performed to ensure that the detection directions of the sensors are strictly perpendicular to the inner side of the wheel rim and symmetrically distributed on either side of the rim centerline. After the sensors are installed, a laser calibration tool is used to verify the perpendicularity of the detection direction to the inner side of the wheel rim to ensure the geometric accuracy of subsequent measurements.

[0071] Subsequently, in step 2012, the first and second distance measuring sensors are activated. Using laser or ultrasonic ranging technology, they collect first and second distance values ​​from the inner side of the rim to the sensors in real time. The sensors transmit detection signals to the inner side of the rim at a fixed frequency (e.g., 100 Hz) and calculate distance data based on the signal reflection time. During the data collection process, a filtering algorithm eliminates transient noise interference caused by wheel vibration or surface contamination to ensure data stability and continuity.

[0072] At the same time, in step 2013, the vertical distance between the first and second distance measuring sensors on each top wheel assembly is retrieved from the device parameter library. This distance is a fixed value measured using a high-precision ruler during sensor installation and stored in the system configuration file. If the sensor assembly installation position is changed, this parameter must be remeasured and updated to ensure the accuracy of subsequent calculations.

[0073] Finally, in step 2014, the first and second distance values ​​obtained in step 2012 are substituted into the geometric relationship formula to calculate the deflection angle between the top-rotating wheel assembly and the wheel. The specific process is as follows: First, the difference between the two distance values ​​is calculated (assuming the difference between the two distance values ​​can be L1-L2), and then divided by the installation vertical distance L in step 2013 to obtain the sine function value; then, the inverse sine operation is performed to calculate the angle α. The system automatically determines the deflection direction based on the size relationship between L1 and L2: if L1 is greater than L2, it is determined to be a positive deflection (e.g., the wheel tilts outward), otherwise it is a negative deflection. The calculation results are transmitted in real time to the vehicle posture control system, which is used to dynamically adjust the position of the top-rotating wheel assembly to maintain operational stability.

[0074] In the overall solution of step 201 above, accurate dynamic detection of the wheelset deflection angle is achieved through the vertically symmetrical layout of the dual distance measuring sensors and the principle of triangulation. A triangular geometric model is constructed using the vertical distance and real-time distance difference between the two sensors, and the direction and amplitude of the deflection angle are directly calculated by trigonometric functions, effectively eliminating the measurement error caused by a single distance measuring point. Through the symmetrical design of the detection direction and the positive and negative angle discrimination mechanism, the instantaneous tilt state of the wheel can be reflected in real time, providing a high-precision feedback reference for subsequent control. This method breaks through the mechanical limitations of traditional contact detection, effectively eliminates the interference of wheel surface wear through non-contact dynamic measurement, and significantly improves the sensitivity and reliability of deflection detection. The combination of dual-channel data fusion and mathematical model calculation establishes accurate initial input for the closed-loop leveling system, laying the foundation for the basic parameter support for dynamic correction of air film pressure compensation and servo adjustment.

[0075] 202. Construct a closed-loop control model to compare the deflection angles corresponding to the wheels on both sides with a preset angle threshold to generate a first error signal. Based on the direction and amplitude of the first error signal, drive the leveling mechanisms on both sides to generate an air film pressure distribution through the air bearing, wherein the air film pressure distribution forms a dynamic compensation relationship with the deflection direction of the wheel;

[0076] As a possible implementation solution, step 202 may specifically include:

[0077] 2021. Compare the deflection angles of the wheels on both sides with a preset angle threshold to generate a first error signal including a direction and an amplitude;

[0078] 2022. Determine, based on the direction of the first error signal, the left side or the right side of the air bearing in the leveling mechanisms on both sides as the pressure compensation side;

[0079] 2023. Based on the amplitude of the first error signal, proportionally adjust the air film pressure on the pressure compensation side of the leveling mechanisms on both sides so that the air film pressure of the leveling mechanisms on both sides forms a non-uniform pressure difference opposite to the deflection direction of the corresponding wheel;

[0080] 2024. Dynamically adjust the direction of the wheel counteracting the deflection through the non-uniform pressure difference driving wheel.

[0081] In the above scheme, the closed-loop control model is a control system that automatically adjusts output based on a feedback mechanism. The preset angle threshold is the maximum allowable wheel deflection angle set by the system. The first error signal is the deviation between the actual wheel deflection angle and the preset threshold. The direction, the positive or negative polarity of the error signal, indicates the direction of wheel deflection. The amplitude, the magnitude of the error signal, indicates the severity of the deflection. The leveling mechanism is a mechanical device used to adjust wheel balance. An air bearing is a bearing structure that reduces friction through air film support. The air film pressure distribution is the non-uniform pressure field generated by the air bearing. The dynamic compensation relationship is a corresponding mechanism for adjusting the air film pressure in real time with wheel deflection. The pressure compensation side is the side of the bearing that requires increased air film pressure to offset deflection. The non-uniform pressure difference is the asymmetric distribution of air film pressure on both sides of the leveling mechanism. The wheelset is the vehicle's running mechanism consisting of two wheels. Dynamic adjustment is the process of automatically correcting wheel posture based on the real-time error signal. Proportional adjustment is the operation of changing control parameters proportionally based on the error amplitude. Deflection compensation is the control objective of driving the wheel to restore balance through pressure difference. Feedback comparison is the real-time comparison of measured values ​​against set values. Control actuation converts error signals into mechanical action. Air film adjustment changes the support characteristics by adjusting the air pressure distribution. Automatic balancing is the system's ability to maintain stable wheelset operation.

[0082] In an embodiment of the present application, first, through step 2021, the real-time deflection angle α of the wheels on both sides is obtained, and each angle is compared in real time with a preset safety angle threshold (such as the safety angle threshold can be ±0.5°). The comparison process is implemented by a threshold comparison algorithm: when the actual angle exceeds the threshold (for example, the real-time deflection angle α can be 0.8°), the system generates a first error signal. The direction of the signal is determined by the positive and negative signs of the deflection angle (positive value means outward deflection, negative value means inward deflection), and the amplitude is determined by calculating the absolute difference between the actual angle and the threshold (for example, the difference between 0.8° and 0.5° is 0.3°). This error signal is transmitted to the leveling module through the controller to provide input parameters of direction and amplitude for subsequent steps.

[0083] Next, in step 2022, based on the direction of the error signal generated in step 2021, the system uses a logic judgment module to determine the pressure compensation side of the air bearing in the leveling mechanism. The specific rule is: if the error direction is positive (wheel deflection is outward), the inner area of ​​the air bearing (the side closer to the wheel center) is selected as the compensation side; if it is negative (inward deflection), the outer area (the side away from the wheel center) is selected. This determination is directly related to the physical structure layout of the leveling mechanism, ensuring that the pressure compensation direction is opposite to the direction of wheel deflection.

[0084] Subsequently, through step 2023, according to the error margin of step 2021 (for example, the error margin can be 0.3°), the system proportionally adjusts the air film pressure on the pressure compensation side selected in step 2022 through the electronically controlled proportional valve. The larger the error margin, the higher the air pressure increase (for example, 0.3° corresponds to an air pressure increase of 30kPa), while the air pressure on the other side maintains the baseline value or decreases appropriately. This process adopts the proportional-integral (PI) control algorithm to adjust the air pressure value through real-time feedback, so that a non-uniform pressure difference is formed on both sides of the air bearing (for example, the pressure on the inside increases and the pressure on the outside remains unchanged). This pressure difference acts directly on the top turntable device to provide driving force for wheelset adjustment.

[0085] Finally, through step 2024, the generated non-uniform pressure difference is used to adjust the driving wheel pair in the direction that offsets the deflection. For example, when the wheel deflects outward, the increased air film pressure on the inside of the air bearing will push the top wheel device inward, thereby correcting the wheel posture. During the adjustment process, the system continuously monitors the wheel deflection angle in real time through the distance measuring sensor in step 201, and feeds the latest data back to the error signal generation module in step 2021 to form a closed-loop control. When the deflection angle returns to within the threshold range, the system gradually reduces the air pressure compensation amount, and finally stabilizes the wheelset in a preset safe posture. The entire process achieves dynamic balancing and real-time correction of the wheel posture through the synergy of mechanics and control.

[0086] Specifically, the above steps 2021 to 2024 can be implemented as follows:

[0087] Compare the real-time deflection angles α1 and α2 of the wheels on both sides with the preset angle threshold α0 respectively to generate first error signals Δα1=α1-α0 and Δα2=α2-α0;

[0088] The deflection direction of the wheels on both sides is determined by the positive and negative signs of Δα1 and Δα2, and the deflection amplitude levels on both sides are determined by |Δα1| and |Δα2|. Based on the deflection direction and amplitude levels of the wheels on both sides, the opening ratio of the left and right pressure regulating valves of the air bearings in the leveling mechanisms on both sides is independently controlled so that the left air film pressure Pleft and the right air film pressure Pright of the leveling mechanism on each side meet the relationship:

[0089] When Δα1>0, Pleft / Pright of the left leveling mechanism = 1+k1|Δα1|; when Δα2>0, Pleft / Pright of the right leveling mechanism = 1+k2|Δα2|; a dynamic compensation torque opposite to the deflection direction of the corresponding wheel is generated through the non-uniform air film pressure distribution on both sides.

[0090] The system first uses sensors to obtain the wheel deflection angles α1 and α2 on both sides in real time. These values ​​are then compared with a preset angle threshold α0 (e.g., α0 = 0°, representing an ideal vertical position) to generate first error signals Δα1 = α1 - α0 and Δα2 = α2 - α0. This process utilizes a real-time threshold comparison algorithm. For example, if the left wheel deflects outward by α1 = +0.8°, Δα1 = +0.8°, indicating an outward deflection of 0.8°; if the right wheel deflects by α2 = -0.6°, Δα2 = -0.6°, indicating an inward deflection of 0.6°. The signs of the error signals Δα1 and Δα2 are used to directly determine the direction of wheel deflection (positive values ​​indicate outward, negative values ​​indicate inward), while the absolute values ​​|Δα1| and |Δα2| are used to quantify the severity of the deflection using an amplitude-to-level mapping rule.

[0091] Based on the direction determination results, the system independently controls the left and right pressure regulating valves of the air bearings of the leveling mechanisms on both sides. For example, when Δα1 on the left side is greater than 0 (outward deflection), the opening of the left pressure regulating valve of the left leveling mechanism increases proportionally, and the opening of the right valve decreases, so that the ratio of the left air film pressure Pleft to the right air film pressure Pright satisfies Pleft / Pright=1+k1|Δα1| (k1 is the preset pressure gain coefficient, such as k1 can be = 0.5 / °, indicating that the pressure ratio increases by 0.5 for every 1° error). In the specific implementation, the error amplitude |Δα1| is converted into a valve opening instruction through an electronically controlled proportional valve and a PID control algorithm. For example, when Δα1 = 0.8°, Pleft / Pright on the left side = 1+0.5×0.8=1.4, that is, the opening of the left valve increases by 40%. Similarly, the leveling mechanism on the right side is independently adjusted according to the direction and amplitude of Δα2 to ensure that the pressure distribution on both sides is asymmetric and opposite to the deflection direction.

[0092] Ultimately, the leveling mechanisms on both sides generate a dynamic compensation torque between the wheel and the top turn wheel device through the non-uniform air film pressure distribution generated by the air bearing (such as Pleft>Pright on the left side or Pright>Pleft on the right side). For example, when the left wheel deflects outward, the increase in the left air film pressure will push the top turn wheel device inward, generating an inward corrective torque through the mechanical lever effect. At the same time, the system continuously feeds back the real-time deflection angle through the sensor to form a closed-loop control. When the deflection angle returns to the threshold range, the pressure regulating valve gradually restores the balanced opening to stabilize the wheel in the preset posture. The entire process realizes real-time dynamic correction of wheel deflection through the complete link of "error detection-direction determination-pressure regulation-torque compensation-closed-loop feedback".

[0093] In the overall solution of step 202 above, rapid dynamic compensation of wheelset deflection is achieved through closed-loop linkage control of the error signal and the air film pressure. The pressure compensation side is precisely positioned based on the error direction, and the air film pressure difference is adjusted in combination with the amplitude ratio to form a non-uniform pressure field in the opposite direction of the wheel deflection, and the wheelset posture is accurately corrected by the flexible air film force. Through the real-time response mechanism of direction-pressure mapping, the system can actively offset the instantaneous deflection torque in the wheelset rotation, breaking through the mechanical hysteresis limitations of traditional rigid adjustment. The non-contact air film pressure compensation strategy avoids mechanical wear and maintains high-frequency dynamic adjustment capabilities, ensuring that the wheelset maintains high-precision posture stability during continuous rotation. This closed-loop control method significantly improves the adaptability and anti-interference ability of parallelism correction, and effectively solves the trajectory offset problem caused by adjustment hysteresis or overshoot in traditional methods.

[0094] 203. Synchronously extract the air film pressure change of the leveling mechanism and, in combination with the time domain characteristics of the first error signal, generate an inclination adjustment instruction for the servo motors on both sides. The inclination adjustment instruction is used to control the servo motors on both sides to adjust the spatial inclination of the support surface, so that the wheelset tread and the detection path of the flaw detector form an adaptive following trajectory.

[0095] As a possible implementation solution, step 203 may specifically include:

[0096] 2031. Determine the air film pressure difference of the leveling mechanisms on both sides by extracting the air film pressure change of the leveling mechanisms on both sides, and extract the change trend of the first error signals on both sides over time;

[0097] 2032. Generate dynamic gain values ​​of the servo motors on both sides according to the pressure difference on both sides and the rate of change of the corresponding side;

[0098] 2033. Determine the tilt angle adjustment amount of the servo motors on both sides according to the correlation between the dynamic gain value and the first error signal on the corresponding side;

[0099] 2034. The servo motor drives the support surfaces on both sides to independently rotate the inclination adjustment amount around the transverse axis, so that the inclination direction of the wheelset tread is dynamically matched with the detection path of the flaw detector.

[0100] In the above scheme, the air film pressure variation of the leveling mechanism is the temporal variation of the air film pressure generated by the air bearing. The time-domain characteristics of the first error signal are the temporal regularity of the deviation signal. The servo motor is an electric motor used to precisely control the mechanical position. The inclination adjustment command is the command signal that controls the servo motor's angle adjustment. The spatial inclination of the support surface is the tilt angle of the support structure in three-dimensional space. The wheelset tread is the contact surface between the wheel and the track. The flaw detector's inspection path is the preset scanning trajectory of the inspection equipment. Adaptive tracking is the automatic adjustment of the wheelset to match the motion path of the inspection path. The air film pressure difference between the two leveling mechanisms is the asymmetry of the air film pressures of the two leveling mechanisms. The temporal trend of the first error signal is the temporal regularity of the deviation signal. The dynamic gain value is the amplification factor in the servo motor control system that changes with the error. The inclination adjustment value is the angle the servo motor needs to rotate. The transverse axis is the horizontal axis perpendicular to the vehicle's forward direction. The inclination direction of the wheelset tread is the deflection of the wheel contact surface relative to the horizontal plane. The dynamic matching of the flaw detector's inspection path is a process in which the wheelset is adjusted in real time to keep it consistent with the inspection trajectory.

[0101] In the embodiment of the present application, first, through step 2031, the system collects the change in the air film pressure of the leveling mechanisms on both sides in real time, calculates the air film pressure difference of each leveling mechanism (the left side is the left air film pressure minus the right air film pressure, and the same is true for the right side), and simultaneously extracts the change trend of the first error signal on both sides over time. The calculation of the air film pressure difference is achieved by directly reading the left and right air film pressures through the pressure sensor and subtracting them in real time; the change trend of the first error signal is obtained by recording historical data and analyzing its slope or fluctuation direction. For example, a time series analysis method is used to determine whether the error signal is continuously expanding or gradually converging. These data provide basic input for subsequent dynamic adjustment.

[0102] Then, in step 2032, the system combines the air film pressure difference obtained in step 2031 with the rate of change of the first error signal on the corresponding side to generate dynamic gain values ​​for the servo motors on both sides. Specifically, through a control algorithm (such as fuzzy control or adaptive rules), it is determined that if the air film pressure difference on one side is large and the error signal changes dramatically (for example, the error increases rapidly), the dynamic gain value of the servo motor on that side is increased to enhance the adjustment sensitivity; conversely, if the pressure difference is small and the error trend is gentle, the dynamic gain value is reduced to avoid overshoot. This process ensures that the response strength of the servo motor matches the actual demand.

[0103] Next, in step 2033, based on the correlation between the dynamic gain value from step 2032 and the corresponding first error signal, the system uses a weighted allocation algorithm to determine the tilt adjustment for each servo motor. For example, if the dynamic gain value on the left side is higher, the tilt adjustment for the left servo motor is proportionally amplified by the amplitude of the first error signal (e.g., the larger the error, the larger the adjustment angle), while the tilt adjustment for the right side is calculated independently. This step converts the abstract gain value and actual error signal into a specific angle correction instruction.

[0104] Finally, based on the inclination adjustment, the system independently controls the servo motors on both sides to rotate the corresponding angle around the transverse axis. For example, if the left side requires inward tilt correction, the servo motor drives the left support surface to rotate counterclockwise by a certain angle, so that the inclination angle of the wheelset tread it supports forms a dynamic tracking trajectory with the flaw detector's inspection path. During the adjustment process, the system monitors the wheelset's posture in real time and continuously optimizes the inclination command through a closed-loop feedback mechanism to ensure that the inspection path and the contact surface of the wheel tread are always aligned.

[0105] Specifically, the above steps 2031 to 2034 can be implemented as follows:

[0106] Obtain in real time the pressure differences ΔP1 = |Pleft1-Pright1| and ΔP2 = |Pleft2-Pright2| of the left air film pressure Pleft and the right air film pressure Pright of the leveling mechanisms on both sides;

[0107] Extract the change rates d(Δα1) / dt and d(Δα2) / dt of the first error signals Δα1 and Δα2 on both sides in the time series;

[0108] The left tilt angle adjustment gain coefficient G1 = ΔP1 × d(Δα1) / dt is generated according to the product of ΔP1 and d(Δα1) / dt;

[0109] The right tilt angle adjustment gain coefficient G2 = ΔP2 × d(Δα2) / dt is generated according to the product of ΔP2 and d(Δα2) / dt;

[0110] Multiply G1 by Δα1, G2 by Δα2 respectively to generate the inclination adjustment instructions θ1 = G1 × Δα1 and θ2 = G2 × Δα2 for the servo motors on both sides;

[0111] The servo motors on both sides independently drive the corresponding support surfaces to rotate around the transverse axis by angles θ1 and θ2, so that the tangent planes of the two wheel treads form a synchronous dynamic following inclination with the detection path.

[0112] The system first uses pressure sensors to collect real-time data on the left and right air film pressures Pleft and Pright of the leveling mechanisms on both sides, and calculates the pressure differences ΔP1 and ΔP2 for each leveling mechanism (ΔP1 on the left is the absolute difference between Pleft1 and Pright1, and ΔP2 on the right is the absolute difference between Pleft2 and Pright2). Simultaneously, the real-time rates of change (i.e., the rate of increase or decrease of the error signals per unit time) of the first error signals Δα1 and Δα2 on both sides are extracted from historical data. Using differential calculations or derivative algorithms, the system calculates d(Δα1) / dt and d(Δα2) / dt, which are used to quantify the dynamic trends of the error signals.

[0113] Based on these parameters, the system multiplies the pressure difference ΔP on each side by the corresponding side's error signal change rate d(Δα) / dt in real time to generate the tilt adjustment gain coefficients G1 and G2 on both sides (G1 = ΔP1 × d(Δα1) / dt on the left side, G2 = ΔP2 × d(Δα2) / dt on the right side). This process is implemented through a signal fusion algorithm, whose logic is: if the pressure difference on one side is large and the error signal deteriorates rapidly (high rate of change), the gain coefficient will increase significantly, and vice versa, thereby dynamically adapting the adjustment sensitivity under different operating conditions.

[0114] Next, the system multiplies the gain coefficients G1 and G2 by the first error signals Δα1 and Δα2 on the corresponding sides, respectively, to generate the inclination adjustment instructions θ1 and θ2 for the servo motors on both sides (θ1 = G1 × Δα1, θ2 = G2 × Δα2). This calculation process is completed by the embedded controller to ensure the real-time and independence of the instructions. Finally, the inclination adjustment instructions θ1 and θ2 drive the support surfaces on both sides to rotate the corresponding angles around the transverse axis through the servo motor. For example, if the left side θ1 is a positive value, the servo motor controls the left support surface to tilt inward, and the same is true for the right side, so that the tangent planes of the wheel treads on both sides follow the detection path of the flaw detector in real time, forming a dynamically adapted contact inclination angle. The entire process achieves high-precision synchronization of the wheelset posture and the detection path through the closed-loop coupling of the pressure difference, error change rate and gain coefficient.

[0115] In the complete solution of step 203 above, multi-dimensional coordinated control of wheelset posture adjustment is achieved by dynamically integrating the air film pressure change and the error time domain characteristics. Dynamic gain is generated based on the pressure difference and the error change rate, so that the servo motor's inclination adjustment has both response speed and stability, avoiding the lag of traditional single parameter adjustment. By independently adjusting the inclination angles of the support surfaces on both sides, the spatial posture of the wheelset tread and the detection path is accurately matched, forming an adaptive dynamic following mechanism. The system deeply couples pressure compensation with inclination adjustment, and uses air film pressure feedback to correct the servo motor movement in real time. It retains the advantages of flexible air film compensation and enhances the precise control of wheelset posture through rigid inclination adjustment. This closed-loop control mechanism with multi-physical quantity coordination significantly improves the synchronization and stability of trajectory following during dynamic rotation of the wheelset, ensuring that the flaw detection path always maintains optimal fit with the wheelset tread, effectively solving the detection blind spot problem caused by mechanical hysteresis or path deviation in traditional methods.

[0116] 204. During the continuous rotation of the wheelset, the distance signal difference of each group of distance measuring sensors is calculated respectively, and the distance signal difference of the two groups is input into the closed-loop control model to generate a second error signal; if the second error signal exceeds the preset tolerance range, the multi-parameter coordinated correction of the leveling mechanisms and servo motors on both sides is triggered until the stability of the adaptive following trajectory meets the preset control accuracy.

[0117] As a possible implementation solution, step 204 may specifically include:

[0118] 2041. When the wheelset rotates continuously, the distance signal difference of the distance measuring sensor groups of the top rotating wheel devices on both sides is calculated respectively, and the difference between the two distance signal differences is input into the closed-loop control model to generate a second error signal;

[0119] 2042. According to the duration of the second error signal, adjust the pressure compensation ratio of the leveling mechanisms on both sides and the inclination adjustment sensitivity of the servo motor, and cyclically adjust the pressure compensation ratio and inclination adjustment sensitivity to make the difference between the two distance signals fall back to the preset tolerance range, and maintain the adaptive following stability of the wheelset and the detection path.

[0120] In the above scheme, the continuous rotation process of the wheelset refers to the state in which the wheels rotate continuously while traveling. The distance signal difference of the distance measuring sensors is the difference between the measured values ​​of two sensors in the same group. The closed-loop control model is a control system that automatically adjusts the output based on feedback signals. The second error signal is the deviation signal generated by comparing the distance signal difference with a preset tolerance. The preset tolerance range is the maximum range of allowable fluctuations in the distance signal difference. Multi-parameter coordinated correction is the process of jointly adjusting multiple control variables of the leveling mechanism and the servo motor. The stability of the adaptive tracking trajectory is the smoothness of the dynamic matching detection path of the wheelset. The preset control accuracy is the required tracking accuracy of the system. The distance measuring sensor groups of the two top runners are distance measuring devices installed on both sides of the wheel. The difference of the distance signal difference is the comparison of the difference between the signals of the two sensor groups. The duration of the second error signal is the duration of the deviation signal exceeding the tolerance range. The pressure compensation ratio is the distribution coefficient of the air film pressure adjustment amplitude of the leveling mechanism. The tilt adjustment sensitivity is the speed parameter of the servo motor's response to the error signal. The cyclic adjustment is the iterative optimization process of the pressure compensation ratio and the tilt adjustment sensitivity. The preset tolerance interval is the range of values ​​within which the distance signal difference is allowed to remain stable. Adaptive following stability is the reliable state in which the wheelset maintains a dynamic match with the detection path.

[0121] In the embodiment of the present application, first, through step 2041, the system obtains the inner side distance data of the rim collected by each group of distance measuring sensors on the top rotating wheel devices on both sides in real time during the continuous rotation of the wheelset, and calculates the distance signal difference of each group of sensors (that is, the difference between the two distance measuring sensor data in the same group). For example, the distance signal difference of the left sensor group is the difference between the readings of the first distance measuring sensor and the second distance measuring sensor, and the same is true for the right side. Subsequently, the distance signal difference between the left and right sides is calculated again (that is, the left difference minus the right difference), and the result is input into the closed-loop control model. The closed-loop control model compares the input with the preset tolerance range to generate a second error signal that characterizes the dynamic offset state of the wheelset. If the second error signal exceeds the tolerance range, it indicates that there is a risk of instability in the following trajectory of the wheelset and the detection path, triggering subsequent coordinated correction actions.

[0122] Then, through step 2042, the system dynamically adjusts the pressure compensation ratio of the leveling mechanisms on both sides and the inclination adjustment sensitivity of the servo motor according to the duration of the second error signal (for example, the error lasts for more than 2 seconds). Specifically, if the second error signal is not eliminated for a long time, the system increases the pressure compensation ratio according to the gradient (such as gradually increasing the left and right pressure difference of the air bearing) and simultaneously increases the response speed of the servo motor (such as increasing the inclination adjustment step). By adjusting these two parameters through cyclic iteration, the difference in the distance signal on both sides is gradually reduced to within the preset tolerance range. In this process, the air film pressure distribution of the leveling mechanism and the inclination adjustment of the servo motor form a synergistic effect to jointly offset the deviation trend of the wheelset, and ultimately ensure that the adaptive following stability of the wheelset tread and the detection path is maintained within the preset control accuracy range.

[0123] Specifically, the above steps 2041 to 2042 can be implemented as follows:

[0124] When the wheelset rotates continuously, the distance difference values ​​of the distance measuring sensor groups of the top rotating wheel devices on both sides are synchronously calculated ΔL1 = |L11-L21| and ΔL2 = |L12-L22|;

[0125] The difference between ΔL1 and ΔL2, ΔL_e = ΔL1 - ΔL2, is input into the closed-loop control model and compared with the preset tolerance range [ΔL_min, ΔL_max].

[0126] If ΔL_e>ΔL_max or ΔL_e<ΔL_min, generate a second error signal ΔL_error=ΔL_e-ΔL_max (or ΔL_min-ΔL_e);

[0127] According to the duration t_e of ΔL_error, the pressure compensation coefficients k1 and k2 of the leveling mechanisms on both sides and the inclination adjustment gain coefficients G1 and G2 of the servo motor are adjusted synchronously to make:

[0128] k1=k10×(1+μ1×t_e), k2=k20×(1+μ2×t_e);

[0129] G1=G10×(1+ν1×t_e), G2=G20×(1+ν2×t_e);

[0130] The adjustment is performed cyclically until ΔL_e falls within the range of [ΔL_min, ΔL_max].

[0131] When the wheelset rotates continuously, the system collects data from the distance sensor groups of the top runner devices on both sides in real time, and calculates the distance difference ΔL1 and ΔL2 of the sensor groups on each side respectively (the left side is the absolute value difference between the data of the first distance sensor and the second distance sensor, and the same is true on the right side). Subsequently, the difference ΔL_e = ΔL1-ΔL2 of the distance difference between the two sides is input into the closed-loop control model and compared with the preset tolerance range [ΔL_min, ΔL_max]. If ΔL_e exceeds the tolerance range (such as ΔL_e>ΔL_max or ΔL_e<ΔL_min), the system generates a second error signal ΔL_error that represents the degree of deviation (when ΔL_e>ΔL_max, ΔL_error=ΔL_e-ΔL_max; when ΔL_e<ΔL_min, ΔL_error=ΔL_min-ΔL_e).

[0132] Based on the duration t_e of the second error signal ΔL_error (i.e., the length of time the error signal exceeds the tolerance range), the system synchronously adjusts the pressure compensation coefficients k1 and k2 of the leveling mechanisms on both sides and the inclination adjustment gain coefficients G1 and G2 of the servo motors. The specific adjustment logic is: based on the initial values ​​k10 and k20, the pressure compensation coefficients k1 and k2 are gradually increased proportionally with the duration t_e (e.g., k1 = initial value × (1 + μ1 × t_e), where μ1 is the time gain factor of the left compensation coefficient). Similarly, k2 and the servo motor gain coefficients G1 and G2 are adjusted. This process is achieved through an adaptive gradient adjustment algorithm, which dynamically enhances the pressure compensation strength and inclination adjustment sensitivity as the error duration increases.

[0133] The adjusted parameters are then used to gradually correct the wheelset's deviation through the pressure distribution of the leveling mechanism's air bearings and the inclination of the servo motors. The system continuously monitors changes in ΔL_e and iterates through the aforementioned parameter adjustment process until ΔL_e falls within the tolerance range. Ultimately, the multi-parameter synergy between the leveling mechanisms on both sides and the servo motors ensures that the adaptive tracking trajectory stability of the wheelset's tread and detection path meets the preset control accuracy requirements.

[0134] In the overall solution of step 204 above, a multi-parameter collaborative correction mechanism is constructed by dynamically monitoring the distance signal difference changes during the rotation of the wheelset, further improving the accuracy and robustness of the adaptive following trajectory. By generating a second error signal in real time through a closed-loop control model, the system can identify trajectory deviations caused by mechanical vibration or environmental interference, and trigger the linkage compensation of the leveling mechanism and the servo motor. Based on the duration of the error signal, the pressure compensation ratio and the tilt adjustment sensitivity are dynamically adjusted to achieve deep coupling of flexible pressure regulation and rigid posture adjustment, effectively eliminating the residual deflection error in the dynamic rotation of the wheelset. By cyclically optimizing the control parameters, the system can actively suppress the influence of external disturbances on the wheelset posture, maintain high-precision dynamic matching between the detection path and the wheelset tread under complex working conditions, solve the problem of trajectory instability caused by error accumulation or response lag in the traditional correction process, and significantly enhance the system's real-time correction capability and control stability for continuously operating wheelsets.

[0135] The following is an overall example of steps 201 to 204:

[0136] When a high-speed train is traveling in a straight line at a speed of 350 kilometers per hour, the wheelset undergoes dynamic deflection due to sudden crosswind disturbances, causing the wheel rim and the flaw detector's detection path to deviate. At this time, the first and second distance measuring sensors installed on the top runners on both sides of the bogie (step 2011) start real-time monitoring: the detection directions of the two sets of sensors point vertically to the symmetrical positions on the inner side of the wheel rim, and respectively collect the first distance value of the inner side of the left wheel from the sensor and the second distance value of the inner side of the right wheel (step 2012). For example, if the left wheel group deviates toward the outside of the track due to the crosswind, its first distance value is significantly smaller than the second distance value on the right side. Based on the fixed vertical installation distance of the two sensors (step 2013), the system calculates the positive deflection angle α of the left wheel group using the deflection angle formula (step 2014).

[0137] The closed-loop control model compares the left wheel's deflection angle with a safety threshold (step 2021) and generates a first error signal indicating outboard deviation. Based on the signal direction (step 2022), the system determines that the right air bearing of the left wheel's leveling mechanism is the pressure compensation side and proportionally increases the right air film pressure based on the error magnitude (step 2023), creating a non-uniform pressure differential with a higher left and lower right pressure (step 2024). This pressure differential propels the left wheel inward. Simultaneously, the system extracts the change in the leveling mechanism's air film pressure (step 2031). The air film pressure differential on the right side of the left wheel continuously increases with the compensation action, and the first error signal exhibits a temporal pattern of first increasing and then decreasing. Combining the pressure differential acceleration and the error decay rate (step 2032), the system generates a dynamic gain value for the left wheel's servo motor, driving the support surface to fine-tune the inclination clockwise around the transverse axis (step 2033), realigning the left wheel's tread tilt with the flaw detector's laser scanning path (step 2034).

[0138] When the train continues to travel at high speed, the wheelset experiences low-frequency oscillation due to the aftermath of crosswind disturbances. During continuous rotation, the system calculates the difference in distance signals from the distance sensor groups on both sides (step 2041). The difference between the first and second distance values ​​of the left wheel group fluctuates beyond the tolerance range, while the difference for the right wheel group remains stable. The difference between the two differences triggers a second error signal. The closed-loop model adjusts the pressure compensation ratio of the left wheel leveling mechanism and the inclination adjustment sensitivity of the servo motor based on the duration of the signal (step 2042). The pressure compensation ratio is gradually reduced to avoid overshoot, while the servo motor switches to high-frequency micro-motion mode, dynamically correcting the support surface inclination with a 0.1-second cycle. For example, when the left wheel distance signal difference approaches the threshold for the third time, the servo motor, with its increased sensitivity, drives the support surface counterclockwise to compensate by 0.5°. This, combined with the leveling mechanism's air film pressure release, allows the difference to quickly converge to a safe range.

[0139] During this process, distance measurement data analysis (step 201), air film pressure control (step 202), servo tilt optimization (step 203) and dynamic collaborative correction (step 204) form a closed-loop link. For example, when the crosswind influence weakens, the system gradually restores the reference pressure distribution of the leveling mechanism based on the stable distance signal difference and the attenuated first error signal, and the servo motor synchronously returns to the standard tilt parameters to ensure the accuracy of the wheelset's fit with the detection path during steady-state driving. Through a multi-level control mechanism, this solution achieves adaptive adjustment of wheelset parallelism and flaw detection follow-up, providing key technical support for all-weather safe operation and maintenance of high-speed trains.

[0140] like Figure 3 The figure shows the technical architecture of the dynamic flaw detection and parallelism correction system for high-speed train wheelsets. Its core structure consists of two symmetrically distributed wheelsets. Each wheelset is equipped with an air bearing marked in purple, which uses high-pressure air film suspension technology to achieve low-friction support, ensuring the free rotation of the wheelset while avoiding mechanical wear. The top rotating wheel devices arranged symmetrically on both sides of the wheel axle are driven by servo motors and form an intelligent control loop with the closed-loop controller. They cooperate with two sets of high-precision distance measurement sensors distributed on the left and right of the wheelset to capture wheelset offset data in real time. When parallelism deviation is detected, the system immediately initiates millimeter-level dynamic calibration to ensure that the distance between the two wheels strictly meets safety standards. The dotted flaw detection path surrounding the center area of ​​the wheelset is equipped with a non-contact sensor network. It performs full-circumferential metal flaw detection scanning while the wheelset is in continuous operation, and can simultaneously detect surface cracks, internal defects and abnormal wear. The entire system simulates the real operating environment through the black base track. The control center composed of servo motors and closed-loop controllers integrates sensor data and mechanical execution to realize a closed-loop operation process from dynamic detection, intelligent analysis to real-time correction. While ensuring the accuracy of wheelset detection, it significantly improves maintenance efficiency and provides fully automated technical support for the safety of high-speed rail operations.

[0141] Optionally, the method also includes: during the rotation of the wheelset, respectively collecting temperature distribution data of the rims of the two wheels through temperature sensors attached to the top rotating wheel devices on both sides; based on the temperature distribution data and the thermal expansion coefficient of the rim material, correcting the first distance data and the second distance data collected by each group of distance measuring sensors to generate the deflection angles of the wheels on both sides after temperature compensation; inputting the deflection angles of the wheels on both sides into a closed-loop control model, driving the leveling mechanisms and servo motors on both sides to synchronously perform temperature-adaptive pressure distribution adjustment and inclination adjustment to offset the influence of thermal deformation on the parallelism correction of the wheelset.

[0142] In this step, the temperature sensor is a detection device used to measure the surface temperature of an object. The two-side top-wheel devices are support and adjustment mechanisms installed on both sides of the wheel. The wheel rim is a ring-shaped metal structure that forms the outer edge of the wheel. Temperature distribution data is a collection of temperature measurements at different locations on the rim surface. The thermal expansion coefficient of the rim material is a physical parameter that describes the degree to which metal expands when heated. Temperature compensation is the process of correcting thermal deformation errors in the measurement data. The first distance data is the raw distance measurement value collected by the first distance sensor. The second distance data is the raw distance measurement value collected by the second distance sensor. The temperature-compensated deflection angle is the wheel tilt angle after eliminating the effects of thermal deformation. The closed-loop control model is an automatic control system based on feedback regulation. The leveling mechanism is a mechanical device used to adjust the wheel balance. The servo motor is an electric motor that performs precise angle control. Temperature-adaptive pressure distribution adjustment is the dynamic adjustment of air film pressure based on temperature changes. Tilt angle adjustment is the precise adjustment of the support surface angle by the servo motor. Thermal deformation is the change in shape of metal materials caused by temperature changes. Wheelset parallelism correction is the process of ensuring that both wheels maintain parallel positions. Pressure distribution adjustment is the non-uniform control of the air film pressure by the leveling mechanism. Synchronous execution is the coordinated linkage operation of multiple control mechanisms.

[0143] In this embodiment, temperature sensors attached to the top rotating wheel devices on both sides first collect real-time temperature distribution data on the rims of both wheels during rotation. These temperature sensors utilize infrared or contact temperature measurement technology, with multiple measurement points arranged along the circumference or axial direction of the rims to cover temperature variations on the rim surface. For example, if a wheelset experiences a localized temperature rise due to friction or environmental factors, the sensors capture the temperature differences between these areas and transmit the data to the control system.

[0144] The system then corrects the first and second distance data collected by each set of distance sensors based on the collected temperature distribution data and the pre-stored thermal expansion coefficient of the rim material. The specific process involves calculating the thermal expansion of the rim based on temperature changes (e.g., changes in rim thickness or diameter due to material expansion in high-temperature areas) and converting this expansion into a measurement error compensation value for the distance sensors. For example, if the rim temperature on one side increases and expands, the system will scale down the raw data from the distance sensor on that side by the thermal expansion coefficient, thereby eliminating temperature interference on the distance measurement and generating temperature-compensated wheel deflection angles for both sides.

[0145] Finally, the corrected wheel deflection angles on both sides are input into the closed-loop control model, driving the leveling mechanisms and servo motors on both sides to synchronously perform temperature-adaptive pressure distribution adjustment and inclination adjustment. Based on the compensated angular deviation, the closed-loop control model dynamically adjusts the air film pressure distribution of the air bearing (for example, increasing pressure on the high-temperature side to offset expansion deformation) and the inclination angle of the servo motor (for example, tilting toward the low-temperature side to balance wheelset parallelism). This process, through real-time feedback and loop optimization, ensures that the wheelset can maintain stable tracking of the detection path under thermal deformation conditions, ultimately offsetting the impact of thermal deformation on wheelset parallelism correction.

[0146] In practical applications, high-speed train wheelsets experience uneven rim temperature distribution due to prolonged braking friction during summer heat. The outer side of the left wheel rim continues to heat up due to the contact area with the brake pads, while the right wheel experiences a temperature gradient due to differential heat dissipation. Temperature sensors attached to the top runners on both sides collect real-time temperature distribution data at different locations on the inner side of the rim, detecting significantly higher temperatures on the outer side of the left wheel than on the inner side. Based on the dynamic thermal expansion coefficient of the rim material, the system corrects the raw distance data from the first and second distance sensors on the left wheel. The first distance value corresponding to the high-temperature area is lowered due to metal expansion, while the second distance value on the low-temperature side is proportionally calibrated to generate the temperature-compensated left wheel deflection angle. The corrected angle shows a 30% reduction in the actual left wheel deflection compared to the initial measurement, effectively eliminating the distance measurement error caused by thermal expansion. The closed-loop control model uses these compensated deflection angles as input, driving the left wheel leveling mechanism to reduce the air film pressure compensation ratio of the right air bearing. Simultaneously, the right wheel servo motor increases the tilt adjustment sensitivity based on the temperature gradient. Through temperature-adaptive pressure distribution adjustment and inclination angle linkage correction, the wheelset maintains a parallel posture under the influence of thermal deformation, ensuring that the flaw detection laser path accurately follows the tread profile.

[0147] In the above steps, the environmental adaptability and long-term stability of wheelset parallelism correction are significantly improved by integrating a temperature compensation mechanism. The system monitors the rim temperature distribution in real time and integrates the material's thermal expansion characteristics to dynamically correct thermal deformation errors in the distance measurement data, ensuring the accuracy of the deflection angle calculation. The closed-loop control model synchronously adjusts the pressure distribution and inclination angle based on the temperature-compensated angle, actively offsetting the deformation deviation of the wheelset caused by temperature changes. This temperature-adaptive dynamic compensation mechanism effectively solves the measurement distortion and adjustment lag problems caused by alternating hot and cold in traditional correction, making the parallelism correction process unaffected by ambient temperature fluctuations. Through a closed-loop control strategy with thermal coupling, the system enhances the ability to precisely control the wheelset posture under complex working conditions, ensuring that the flaw detection path and wheelset tread are always dynamically matched, significantly improving the reliability of the test results and the robustness of the correction system.

[0148] Optionally, the method also includes: when the wheelset is stationary, scanning the circumferential contours of the inner sides of the rims of the two wheels respectively through the distance measuring sensor groups of the top rotating wheel devices on both sides to extract the asymmetric geometric features of the two wheels; generating structural compensation parameters of the two wheels according to the asymmetric geometric features, and embedding the structural compensation parameters into the closed-loop control model respectively; when the wheelset rotates, correcting the pressure distribution gradient of the leveling mechanisms on both sides and the inclination adjustment amount of the servo motor according to the superposition result of the deflection angle of the wheels on both sides and the corresponding structural compensation parameters, so that the wheelset tread and the adaptive following trajectory adapt to the dynamic balance requirements of the asymmetric wheelset.

[0149] In this step, the wheelset's stationary state refers to the fixed state when the wheels have stopped rotating. The distance sensor groups on the top-rotating wheel devices on both sides are sensors installed on both sides of the wheels for distance measurement. The inner side of the wheel rim is the annular surface on the inner side of the wheel opposite the sensors. The circumferential profile is the geometric feature along the circumference of the rim. Asymmetric geometric features are differences in shape or size between the two wheels. The structural compensation parameter is a correction factor for the wheel's asymmetric features. The closed-loop control model is an automatic control system based on feedback regulation. The deflection angle is the tilt angle of the wheel relative to the standard position. The superposition result is a comprehensive calculation of the deflection angle and the compensation parameter. The pressure distribution gradient is the spatial rate of change of the air film pressure of the leveling mechanism. The inclination angle adjustment is the angle value required by the servo motor to adjust. The wheelset tread is the working surface where the wheel contacts the track. The adaptive following trajectory is the detection path automatically adjusted by the system. An asymmetric wheelset refers to a wheelset with geometric differences between the two wheels. Dynamic balancing requirements are the control conditions required to maintain stability during wheelset motion. The correction process involves adjusting control parameters based on the compensation parameters. The servo motor is the drive unit that performs precise angular control. The leveling mechanism is the mechanical device used to maintain wheel balance. Geometric feature scanning is the process of digitally measuring the rim shape. Compensation parameter embedding is the process of incorporating correction coefficients into the control model.

[0150] In the embodiment of the present application, first, with the wheelset stationary, the circumferential profile of the inner side of the rims of both wheels is scanned using a set of distance-measuring sensors on the top runners on both sides. The distance-measuring sensors utilize high-precision laser or ultrasonic scanning technology, with multiple measurement points evenly distributed along the circumference of the rims, to collect real-time geometric data of the inner side of the rims. Using profile analysis algorithms (such as discrete point cloud fitting or curvature feature extraction), the system identifies asymmetric geometric features of the rims (e.g., localized depressions, uneven thickness, or eccentric deformation) and quantifies these features as geometric deviation parameters.

[0151] Subsequently, based on the extracted asymmetric geometric features, the system generates corresponding structural compensation parameters through a structural compensation algorithm (such as a machine learning model or empirical mapping rules). These parameters are used to describe the degree of influence of the inherent deformation of the rim due to manufacturing errors or wear on the parallelism of the wheelset. The compensation parameters are written into the underlying logic of the closed-loop control model through the parameter embedding module, so that the model can dynamically adjust the control strategy based on the actual geometric state of the rim during operation. For example, if the left rim has a feature of a smaller inner diameter, the compensation parameters will correct the reference distance value of the ranging sensor on that side, thereby eliminating the interference of geometric asymmetry on the deflection angle calculation.

[0152] Finally, when the wheelset rotates, the system superimposes the deflection angles of the wheels on both sides measured in real time with the structural compensation parameters of the corresponding side (for example, the angle compensation value is superimposed on the original deflection angle) to generate a corrected control input signal. The corrected signal drives the leveling mechanisms on both sides to adjust the air film pressure distribution according to a non-uniform gradient (for example, the pressure gradient on the left side is enhanced to compensate for the rim depression), and at the same time controls the servo motor to rotate the support surface according to the corrected inclination adjustment amount. Through this dynamic compensation mechanism, the contact state between the wheelset tread and the adaptive follow-up trajectory can adapt to the geometric characteristics of the asymmetric rim, ensuring that the wheelset maintains dynamic balance during rotation and avoiding correction errors caused by inherent rim deformation. The entire process achieves precise parallelism control of asymmetric wheelsets through the closed-loop logic of "static scanning-feature compensation-dynamic correction".

[0153] In practice, after a train wheelset is replaced and overhauled, manufacturing tolerances lead to asymmetric geometric features on the inner circumferential profiles of the left and right wheel rims. During the stationary commissioning phase, the ranging sensors on both top runners activate a high-density scanning mode. A circumferential concave band is detected in the middle of the inner circumference of the left wheel rim, while a localized bulge forms at the root of the right wheel rim due to manufacturing errors. The system uses point cloud analysis to extract the asymmetric geometric features of both wheelsets and generates corresponding structural compensation parameters. For example, the concave area on the left wheel is marked as a negative compensation zone for dynamic tracking, while the convex area on the right wheel is encoded as a positive compensation coefficient. These parameters are preloaded into the closed-loop control model. When the train starts and enters the test run phase, the deflection angles collected in real time during wheelset rotation are automatically superimposed with the embedded structural compensation parameters. The additional deflection signal generated by the concave band on the left wheel during the rotation cycle is offset by parameter correction, while the angular fluctuation caused by the convex area on the right wheel is dynamically smoothed by the compensation coefficient. The leveling mechanism enhances the air film pressure gradient on the left wheel's outer side based on the corrected combined deviation. Simultaneously, a servo motor compensates for the inclination adjustment of the right wheel's support surface, maintaining dynamic balance between the asymmetric wheelset's tread and the inspection path during high-speed rotation. This process ensures that even with inherent geometric discrepancies, the wheelset can still achieve optimized parallelism and inspection accuracy through pre-embedded structural compensation.

[0154] In the above steps, the correction and adaptation problems of asymmetric wheelsets are effectively solved by superimposing static contour scanning and dynamic compensation. In a stationary state, the asymmetric geometric features of the rim are extracted and structural compensation parameters are generated. The inherent morphological deviation of the wheel is pre-identified and embedded in the control model as a dynamic adjustment benchmark. When the wheelset rotates, the real-time deflection angle is integrated with the structural compensation parameters to accurately correct the pressure gradient of the leveling mechanism and the inclination angle of the servo motor, so that the adaptive following trajectory actively adapts to the asymmetric shape of the wheelset. This dual compensation mechanism combining dynamic and static methods not only overcomes the trajectory deviation problem caused by ignoring the geometric shape of the wheel in traditional methods, but also enhances the system's compatibility with complex wheelset structures through parametric compensation, ensuring dynamic fit between the flaw detection path and the tread shape. The system significantly improves the accuracy and operational stability of parallelism correction of asymmetric wheelsets through pre-learning of structural features and real-time correction of closed-loop control.

[0155] Figure 4 A schematic diagram of the structure of an automated wheel flaw detection and wheelset parallelism correction system is provided for an embodiment of the present application. The system comprises a set of distance measuring sensors disposed in opposite directions on two top rotating wheel devices, each of which supports one wheel in the wheelset.

[0156] like Figure 4 As shown, the system includes:

[0157] a calculation module 41 for collecting first distance data and second distance data of the inner side of the wheel rim of the corresponding wheel through each set of distance measuring sensors, and calculating the deflection angle between the top rotating wheel device on each side and the corresponding wheel based on the vertical installation distance of each set of distance measuring sensors;

[0158] a generation module 42 for constructing a closed-loop control model, comparing the deflection angles corresponding to the wheels on both sides with a preset angle threshold to generate a first error signal, and, based on the direction and amplitude of the first error signal, driving the leveling mechanisms on both sides to generate an air film pressure distribution through the air bearing, wherein the air film pressure distribution forms a dynamic compensation relationship with the deflection direction of the wheel;

[0159] an adjustment module 43 for synchronously extracting the air film pressure variation of the leveling mechanism and, in combination with the time domain characteristics of the first error signal, generating an inclination adjustment instruction for the servo motors on both sides, thereby controlling the servo motors on both sides to adjust the spatial inclination of the support surface through the inclination adjustment instruction, so that the wheelset tread and the detection path of the flaw detector form an adaptive following trajectory;

[0160] The correction module 44 is used to calculate the distance signal difference of each group of distance measuring sensors during the continuous rotation of the wheelset, and input the distance signal difference of the two groups into the closed-loop control model to generate a second error signal; if the second error signal exceeds the preset tolerance range, it triggers the multi-parameter coordinated correction of the leveling mechanisms and servo motors on both sides until the stability of the adaptive following trajectory meets the preset control accuracy.

[0161] Figure 4 The automated wheel flaw detection and wheelset parallelism correction system can perform Figure 2 The implementation principles and technical effects of the automated wheel flaw detection and wheelset parallelism correction method described in the illustrated embodiment will not be elaborated upon. The specific manner in which the various modules and units of the automated wheel flaw detection and wheelset parallelism correction system in the aforementioned embodiment perform their operations has been described in detail in the relevant embodiments of the method and will not be further elaborated upon here.

[0162] In one possible design, Figure 5 The automated wheel flaw detection and wheelset parallelism correction system of the embodiment shown can be implemented as a computing device, such as Figure 5 As shown, the computing device may include a storage component 51 and a processing component 52;

[0163] The storage component 51 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 52 .

[0164] The processing component 52 is used for the above Figure 1The embodiment of the invention provides an automated wheel flaw detection and wheelset parallelism correction method.

[0165] The processing component 52 may include one or more processors to execute computer instructions to perform all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0166] The storage component 51 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0167] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0168] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0169] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0170] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0171] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 2 The embodiment shown is an automated wheel flaw detection and wheelset parallelism correction method.

[0172] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0173] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0174] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0175] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An automated wheel flaw detection and wheelset parallelism correction method, characterized in that: A set of distance measuring sensors is arranged in opposite directions of the top rotating wheel devices on both sides, and the top rotating wheel device on each side is used to support one wheel of the wheelset. The method includes: The first distance data and the second distance data of the inner side of the corresponding wheel rim are collected by each set of distance measuring sensors respectively, and the deflection angle between the top rotating wheel device on each side and the corresponding wheel is calculated based on the installation vertical distance of each set of distance measuring sensors; A closed-loop control model is constructed to compare the deflection angles corresponding to the wheels on both sides with a preset angle threshold to generate a first error signal. Based on the direction and amplitude of the first error signal, the leveling mechanisms on both sides are driven to generate an air film pressure distribution through the air bearing, wherein the air film pressure distribution forms a dynamic compensation relationship with the deflection direction of the wheel; Synchronously extracting the air film pressure change of the leveling mechanism and, in combination with the time domain characteristics of the first error signal, generating inclination adjustment instructions for the servo motors on both sides, thereby controlling the servo motors on both sides to adjust the spatial inclination of the support surface through the inclination adjustment instructions, so that the wheelset tread and the detection path of the flaw detector form an adaptive following trajectory; During the continuous rotation of the wheelset, the distance signal difference of each set of distance measuring sensors is calculated and input into the closed-loop control model to generate a second error signal. If the second error signal exceeds a preset tolerance range, the multi-parameter coordinated correction of the leveling mechanisms and servo motors on both sides is triggered until the stability of the adaptive tracking trajectory meets the preset control accuracy. Also includes: During the rotation of the wheelset, the temperature distribution data of the two wheel rims are collected respectively by the temperature sensors attached to the top rotating wheel devices on both sides; Based on the temperature distribution data and the thermal expansion coefficient of the rim material, the first distance data and the second distance data collected by each set of distance measuring sensors are corrected to generate the temperature-compensated deflection angles of the wheels on both sides; Inputting the deflection angles of the wheels on both sides into a closed-loop control model drives the leveling mechanisms and servo motors on both sides to synchronously perform temperature-adaptive pressure distribution adjustment and inclination adjustment to offset the effect of thermal deformation on wheel parallelism correction; Also includes: When the wheelset is stationary, the distance measuring sensor groups of the top rotating wheel devices on both sides scan the circumferential contours of the inner side surfaces of the two wheel rims respectively to extract the asymmetric geometric features of the two wheels; generating structural compensation parameters of the two wheels according to the asymmetric geometric features, and embedding the structural compensation parameters into the closed-loop control model respectively; When the wheelset rotates, the pressure distribution gradient of the leveling mechanisms on both sides and the inclination adjustment amount of the servo motor are corrected according to the superposition result of the deflection angle of the wheels on both sides and the corresponding structural compensation parameters, so that the wheelset tread and the adaptive following trajectory adapt to the dynamic balance requirements of the asymmetric wheelset.

2. The method according to claim 1, characterized in that The step of comparing the deflection angle with a preset angle threshold to generate a first error signal, and driving a leveling mechanism to generate a non-uniform air film pressure distribution through an air bearing according to a direction and an amplitude of the first error signal, comprises: Comparing the deflection angles of the wheels on both sides with a preset angle threshold to generate a first error signal including direction and amplitude; determining, according to the direction of the first error signal, the left side or the right side of the air bearing in the leveling mechanisms on both sides as the pressure compensation side; Based on the amplitude of the first error signal, proportionally adjust the air film pressure on the pressure compensation side of the leveling mechanisms on both sides so that the air film pressure of the leveling mechanisms on both sides forms a non-uniform pressure difference opposite to the deflection direction of the corresponding wheel; The non-uniform pressure difference drives the wheels to dynamically adjust in a direction that offsets the deflection.

3. The method according to claim 1, characterized in that The synchronous extraction of the air film pressure change of the leveling mechanism and the generation of an inclination adjustment instruction for the servo motor in combination with the time domain characteristics of the first error signal are performed, so as to control the servo motor to adjust the spatial inclination of the support surface through the inclination adjustment instruction, so that the wheelset tread and the detection path of the flaw detector form an adaptive following trajectory, including: By extracting the air film pressure changes of the leveling mechanisms on both sides, the air film pressure difference of the leveling mechanisms on both sides is determined, and the change trend of the first error signals on both sides over time is extracted; Generate the dynamic gain values ​​of the servo motors on both sides according to the pressure difference on both sides and the rate of change of the corresponding side; Determining the inclination adjustment amount of the servo motors on both sides according to the correlation relationship between the dynamic gain value and the first error signal on the corresponding side; The servo motor drives the supporting surfaces on both sides to independently rotate the inclination adjustment amount around the transverse axis, so that the inclination direction of the wheelset tread is dynamically matched with the detection path of the flaw detector.

4. The method according to claim 1, wherein calculating the distance signal difference of the distance measuring sensor group during the continuous rotation of the wheelset, and inputting the distance signal difference into the closed-loop control model to generate a second error signal; If the second error signal exceeds a preset tolerance range, triggering the multi-parameter coordinated correction of the leveling mechanism and the servo motor until the stability of the adaptive following trajectory meets the preset control accuracy, including: When the wheelset rotates continuously, the distance signal difference of the distance measuring sensor group of the top rotating wheel device on both sides is calculated respectively, and the difference between the two distance signal differences is input into the closed-loop control model to generate a second error signal; According to the duration of the second error signal, the pressure compensation ratio of the leveling mechanisms on both sides and the inclination adjustment sensitivity of the servo motor are adjusted, and by cyclically adjusting the pressure compensation ratio and the inclination adjustment sensitivity, the difference between the two distance signals falls back to the preset tolerance range, and the adaptive following stability of the wheelset and the detection path is maintained.

5. The method according to claim 1, wherein The method comprises collecting first distance data and second distance data of the inner side of the wheel rim of the corresponding wheel through each set of distance measuring sensors, and calculating the deflection angle between the top rotating wheel device on each side and the corresponding wheel in combination with the installation vertical distance of each set of distance measuring sensors, including: A first distance measuring sensor and a second distance measuring sensor are fixed to the top rotating wheel device on each side, and the detection directions of the first distance measuring sensor and the second distance measuring sensor are set to point vertically to symmetrical positions on the inner side of the rim; The first distance value and the second distance value of the inner side of the wheel rim to the sensor are respectively collected in real time by the first distance measuring sensor and the second distance measuring sensor; Obtaining the installed vertical distance between the first distance measuring sensor and the second distance measuring sensor in the top rotating wheel device on each side; The deflection angle α between the top turning wheel device on each side and the corresponding wheel is calculated according to the formula: sinα=(L1-L2) / L, where L1 represents the first distance value, L2 represents the second distance value, and L represents the installation vertical distance; when L1>L2, α is a positive deflection angle, and when L1<L2, α is a negative deflection angle.

6. An automated wheel flaw detection and wheelset parallelism correction system, used to execute the automated wheel flaw detection and wheelset parallelism correction method according to any one of claims 1 to 5, characterized in that: The system is provided with a set of distance measuring sensors in opposite directions of top rotating wheel devices on both sides, each top rotating wheel device is used to support one wheel in the wheelset, and the system includes: a calculation module, configured to collect first distance data and second distance data of the inner side surface of the rim of the corresponding wheel through each set of distance measuring sensors, and calculate the deflection angle between the top rotating wheel device on each side and the corresponding wheel in combination with the vertical distance between the installation of each set of distance measuring sensors; a generation module, configured to construct a closed-loop control model, compare the deflection angles corresponding to the wheels on both sides with a preset angle threshold to generate a first error signal, and drive the leveling mechanisms on both sides to generate an air film pressure distribution through the air bearing based on the direction and amplitude of the first error signal, wherein the air film pressure distribution forms a dynamic compensation relationship with the deflection direction of the wheel; an adjustment module, configured to synchronously extract the air film pressure variation of the leveling mechanism and, in combination with the time domain characteristics of the first error signal, generate an inclination adjustment instruction for the servo motors on both sides, so as to control the servo motors on both sides to adjust the spatial inclination of the support surface through the inclination adjustment instruction, so that the wheelset tread and the detection path of the flaw detector form an adaptive following trajectory; The correction module is used to calculate the distance signal difference of each group of distance measuring sensors during the continuous rotation of the wheelset, and input the distance signal difference of the two groups into the closed-loop control model to generate a second error signal; if the second error signal exceeds the preset tolerance range, it triggers the multi-parameter coordinated correction of the leveling mechanisms and servo motors on both sides until the stability of the adaptive following trajectory meets the preset control accuracy.

7. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the automated wheel flaw detection and wheelset parallelism correction method as described in any one of claims 1 to 5.

8. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, the automated wheel flaw detection and wheelset parallelism correction method according to any one of claims 1 to 5 is implemented.

Images