Wheel center positioning method and system for rail vehicle

Abstract

The present application provides a method and system for locating the wheel center of a rail vehicle. Specifically, a first and a second detection unit are installed on both sides of the wheel, and the wheel chord length is measured by the first unit. When the second unit triggers the detection, it is controlled to move in the same direction by half the chord length to determine the preliminary positioning of the wheel center. By integrating the dynamic contact spot deformation and wheel rim degradation data, a collaborative analysis model is constructed, and deep learning is used to extract the nonlinear relationship between the wear gradient and the contact spot offset to generate a dynamic correction value. Based on the dynamic correction value, millimeter-level dynamic compensation is performed on the positioning center to achieve stability control of the wheel-rail contact center in the lateral, vertical and rotational dimensions, and suppress stress concentration caused by wheel rim wear. The technical solution provided by the present application improves the positioning accuracy and stability of the wheel-rail contact center.

Description

Technical Field

[0001] The present application relates to the technical field of wheel center positioning, and in particular to a wheel center positioning method and system for a rail vehicle. Background Art

[0002] In railway transportation systems, accurate, real-time positioning of the wheel-rail contact center is crucial for train stability, wheel flange wear control, and track lifespan improvement. Due to the long-term dynamic interaction between wheels and rails, geometric deformation and contact stress variations occur. Existing technologies must provide real-time adaptive positioning capabilities under dynamic conditions while also avoiding contact stress concentration and abnormal wear caused by positioning deviations.

[0003] One of the currently adopted solutions is a dynamic monitoring system for wheel rim geometric parameters based on multi-sensor fusion. The system collects wheel position data in real time through laser ranging and inertial sensor units, calculates the wheel-rail contact center based on a preset geometric model, and performs position compensation through a servo motor-driven positioning mechanism.

[0004] However, the positioning calculation relies on a preset geometric model, but does not deeply analyze the impact of wheel-rail material deformation and stress distribution on contact center offset during dynamic contact. As a result, the correction amount is insufficiently adapted to the mechanical characteristics of the actual working conditions. During long-term operation, the positioning accuracy is easily disturbed by factors such as material creep, resulting in low wheel center positioning accuracy. Summary of the Invention

[0005] The present application provides a method and system for locating the wheel center of a rail vehicle, so as to solve the problem of low wheel center positioning accuracy in the prior art.

[0006] In a first aspect, the present application provides a method for locating the wheel center of a rail vehicle, comprising:

[0007] Pre-installing a first detection unit and a second detection unit on both sides of the wheel;

[0008] determining a chord length of the wheel according to a detection process of the wheel by the first detection unit;

[0009] When the second detection unit detects the wheel, calculating half of the chord length, and controlling the second detection unit to move half of the chord length in the same direction;

[0010] When the second detection unit is controlled to move in the same direction by half the chord length, a current position of the second detection unit is determined as a preliminary positioning position of the wheel center;

[0011] Based on the collaborative learning framework of dynamic contact spot deformation characteristics and rim geometry degradation, an offset correction value of the initial positioning position of the wheel center is generated, and the actuator is driven to dynamically correct the initial positioning position according to the offset correction value to maintain the positioning stability of the wheel-rail contact center.

[0012] Optionally, the collaborative learning framework based on the dynamic contact patch deformation characteristics and the wheel rim geometry degradation generates the offset correction value of the preliminary positioning position of the wheel center, including:

[0013] The wheel surface geometry at different degrees of wear is acquired through 3D laser scanning, and a static parameter set including the wheel flange thickness and tread wear depth is generated. Based on the wheel flange thickness and tread wear depth, a spatial geometric mapping relationship between the wheel flange centerline and the track contact surface is established;

[0014] A millimeter-wave radar network is distributed along the longitudinal direction of the track to capture the spatial coordinates of the dynamic contact patch in the wheel-rail contact area in real time. The deformation characteristics of the dynamic contact patch are analyzed by combining the spatial geometric mapping relationship between the centerline of the wheel flange and the track contact surface.

[0015] Inputting the wheel rim geometry degradation characteristics and the dynamic contact patch deformation characteristics in the static parameter set into a collaborative learning framework, and generating an offset correction value for the preliminary positioning position of the wheel center through an interactive method of local model parameter encryption iteration and global parameter aggregation;

[0016] Driving the actuator to dynamically correct the preliminary positioning position according to the offset correction amount to maintain the positioning stability of the wheel-rail contact center includes:

[0017] Based on the offset correction amount and in combination with the spatial distribution law of the deformation characteristics of the dynamic contact spot, the actuator is driven to adjust the preliminary positioning position, and the positioning stability of the wheel-rail contact center during high-speed operation is maintained through real-time matching of the wheel rim geometric characteristics with the contact spot coordinates.

[0018] Optionally, inputting the rim geometry degradation feature and the dynamic contact patch deformation feature in the static parameter set into a collaborative learning framework, and generating an offset correction value for the preliminary positioning position of the wheel center through an interactive manner of local model parameter encryption iteration and global parameter aggregation, includes:

[0019] Decomposing a geometric deformation gradient tensor of the wheel rim geometric degradation feature, extracting a degradation feature vector related to the wheel-rail contact surface curvature in the geometric deformation gradient tensor, and generating a feature decomposition sequence including a curvature degradation gradient;

[0020] executing a noise masking mechanism on the eigendecomposition sequence at the local computing node, superimposing random mask noise through a dynamic attenuation window, and generating an interference-resistant degenerate gradient encryption chain;

[0021] A joint parameter space is constructed by combining the degradation gradient encryption chain with the stress distribution time series signal in the dynamic contact patch deformation characteristics, and a global parameter topology chain coupled with degradation and deformation is generated through cross-node parameter interpolation operation in the collaborative learning framework;

[0022] Dynamically iteratively encrypting the degenerate gradient in the global parameter topology chain at the local computing node, fusing the stress distribution of the wheel-rail contact area to generate an encrypted weight gradient chain and distributing it to the global aggregation node;

[0023] The encrypted gradient chains of multiple nodes are aggregated in a global aggregation node, the random mask noise is eliminated, and the degradation and deformation coupling phase offsets across the nodes are extracted to generate an offset correction for the preliminary positioning position of the wheel center.

[0024] Optionally, the step of constructing a joint parameter space by combining the degraded gradient encryption chain with the stress distribution time series signal in the dynamic contact patch deformation feature, and generating a global parameter topology chain coupled with degradation and deformation through cross-node parameter interpolation in the collaborative learning framework includes:

[0025] Aligning the sampling timestamps of the spatiotemporal distribution and stress distribution time series signals in the degradation gradient encryption chain, matching the geometric deformation phase and dynamic stress fluctuation period of the wheel-rail contact area, and generating a spatiotemporally aligned degradation and stress parameter matrix;

[0026] Performing parameter interpolation operations between multiple nodes on the parameter matrix to compensate for the difference in deformation gradients of dynamic contact patches between adjacent track sections, and generating an interpolation compensation parameter chain that is continuously distributed across the nodes;

[0027] Fusing the degraded gradient and stress distribution amplitude in the interpolation compensation parameter chain, adjusting the fusion weight of the gradient and stress based on the instantaneous deformation rate, and generating a dynamic weight fusion parameter set;

[0028] Performing spatiotemporal convolution superposition of degradation gradient and stress distribution on the dynamic weight fusion parameter set, extracting parameter aggregation features of the wheel-rail contact center area, and generating a characteristic topological network of degradation and deformation coupling;

[0029] The cooperative oscillation mode of the degradation gradient and stress distribution in the characteristic topological network is analyzed to generate a global parameter topological chain of degradation and deformation coupling.

[0030] Optionally, the millimeter-wave radar network is distributed longitudinally along the track to capture the spatial coordinates of the dynamic contact patch in the wheel-rail contact area in real time, and analyze the deformation characteristics of the dynamic contact patch in combination with the spatial geometric mapping relationship between the center line of the wheel flange and the track contact surface, including:

[0031] A millimeter-wave radar array is deployed longitudinally along the track to synchronously collect reflected signal pulse sequences from the wheel-rail contact area. This captures the multipath echo delay and amplitude fluctuations associated with track contact surface deformation in the reflected signals, generating a raw signal grid of the spatial coordinates of the dynamic contact patch.

[0032] Performing multipath interference suppression processing on the original signal grid to compensate for attenuation distortion of the reflected signal during high-speed operation and generate an interference-resistant dynamic contact patch spatial coordinate grid;

[0033] Extracting the impulse response amplitude and time-series phase fluctuation of the dynamic contact spot spatial coordinate grid to construct an impulse response topology of the dynamic contact spot deformation characteristics;

[0034] Performing spatiotemporal correlation matching between the impulse response topology and the spatial geometric mapping relationship of the rim centerline to generate a coupling parameter set of the dynamic contact patch deformation characteristics and the geometric mapping;

[0035] The pulse response amplitude fluctuations in the coupled parameter set and the spatial offset of the geometric mapping are analyzed to generate the deformation characteristics of the dynamic contact patch.

[0036] Optionally, performing spatiotemporal correlation matching on the impulse response topology and the spatial geometric mapping relationship of the rim centerline to generate a coupling parameter set of the dynamic contact patch deformation characteristics and the geometric mapping includes:

[0037] Calibrate the spatiotemporal coordinate origin of the impulse response topology and the geometric mapping space reference point of the wheel rim centerline to eliminate the coordinate system offset of the propagation path of the wheel-rail contact area deformation and the geometric degradation parameters, and generate a coordinate set of deformation and geometric association synchronized with the spatiotemporal reference;

[0038] Tracking the diffusion rate of the deformation propagation path in the coordinate set, mapping the attenuation trend of the spatial geometric degradation gradient of the rim centerline as the track contact surface wears, and generating a dynamic correlation map between deformation diffusion and geometric degradation;

[0039] quantifying the phase offset angle between the deformation propagation path and the spatial geometric degradation gradient in the dynamic correlation map, integrating the instantaneous load intensity of the wheel-rail contact area to calculate the spatial coupling weight, and generating a coupling weight chain of deformation and geometry;

[0040] Reconstructing the dynamic correlation map in the coupling weight chain, superimposing the multi-dimensional influencing factors of the instantaneous load distribution on the wheel-rail contact surface, and generating a multi-dimensional coupling parameter field of deformation propagation and geometric degradation;

[0041] The oscillation frequency of the deformation propagation path and the attenuation amplitude of the geometric degradation gradient in the multi-dimensional coupling parameter field are analyzed to generate a coupling parameter set of the dynamic contact spot deformation characteristics and geometric mapping.

[0042] Optionally, the step of driving an actuator to adjust the preliminary positioning position based on the offset correction amount and in combination with the spatial distribution pattern of the deformation characteristics of the dynamic contact spot, and maintaining the positioning stability of the wheel-rail contact center during high-speed operation by real-time matching of the wheel rim geometric characteristics with the contact spot coordinates, includes:

[0043] Generate dynamic baseline parameters corresponding to the offset correction, decompose the high-frequency oscillation component and low-frequency attenuation trend in the spatial distribution of the dynamic contact patch deformation characteristics, construct a baseline offset field for wheel-rail contact center positioning and send it to the actuator drive nodes of each track section;

[0044] Mapping the phase difference between the wheel rim geometric features and the contact spot coordinates in the baseline offset field at the actuator drive node, compensating for the time delay effect in the spatial distribution law, and generating a wheel center coordinate offset adjustment drive signal;

[0045] Analyzing the instantaneous amplitude fluctuation of the high-frequency oscillation component in the wheel center coordinate offset adjustment drive signal, and combining the propagation path of the contact patch deformation characteristics, assigning dynamic weights to the actuator drive nodes to generate multi-node dynamic weight drive parameters;

[0046] Converting the multi-node dynamic weighted drive parameters into pulse width modulation signals for the actuators, matching the phase synchronization of the wheel rim geometric features with the contact spot coordinates, and generating a dynamic control instruction set for the wheel-rail contact center;

[0047] The dynamic control instruction set is executed, and the wheel center position is adjusted and the positioning stability of the wheel-rail contact center is maintained through real-time closed-loop feedback of the pressure gradient and contact spot coordinates of the wheel rim contact area by the actuator.

[0048] Optionally, determining the chord length of the wheel according to a detection process of the wheel by the first detection unit includes:

[0049] When the first detection unit detects that the wheel enters the detection area, the encoder is activated to record a first displacement signal of the wheel movement;

[0050] When the first detection unit fails to detect the wheel, stopping the encoder from recording and reading the second displacement signal recorded by the encoder;

[0051] determining, based on the first displacement signal recorded by the encoder, an initial displacement value when the wheel edge begins to enter the detection area;

[0052] determining, based on the second displacement signal recorded by the encoder, a final displacement value when the wheel edge completely passes through the first detection unit;

[0053] A chord length of the wheel is determined according to the final displacement value and the initial displacement value.

[0054] Optionally, when the second detection unit detects the wheel, calculating half of the chord length and controlling the second detection unit to move half of the chord length in the same direction includes:

[0055] When the second detection unit detects the wheel, a microprocessor is used to calculate half of the chord length, and a servo motor is used to drive the second detection unit to move half of the chord length in the same direction.

[0056] In a second aspect, the present application provides a wheel center positioning system for a rail vehicle, comprising:

[0057] An installation module is used to pre-install the first detection unit and the second detection unit on both sides of the wheel;

[0058] a measuring module, configured to determine the chord length of the wheel according to a detection process of the wheel by the first detection unit;

[0059] a calculation module, configured to calculate half the chord length when the second detection unit detects the wheel, and control the second detection unit to move a distance half the chord length in the same direction;

[0060] a positioning module, configured to determine a current position of the second detection unit as a preliminary positioning position of the wheel center when controlling the second detection unit to move in the same direction by half the chord length;

[0061] A correction module is used to generate an offset correction value for the initial positioning position of the wheel center based on a collaborative learning framework of dynamic contact spot deformation characteristics and rim geometry degradation, and drive an actuator to dynamically correct the initial positioning position according to the offset correction value to maintain the positioning stability of the wheel-rail contact center.

[0062] In an embodiment of the present application, a first detection unit and a second detection unit are pre-installed on both sides of the wheel; the chord length of the wheel is determined based on the detection process of the wheel by the first detection unit; when the second detection unit detects the wheel, half of the chord length is calculated, and the second detection unit is controlled to move half of the chord length in the same direction; when the second detection unit is controlled to move half of the chord length in the same direction, the current position of the second detection unit is determined to be the preliminary positioning position of the wheel center; based on a collaborative learning framework of dynamic contact spot deformation characteristics and rim geometric degradation, an offset correction amount of the preliminary positioning position of the wheel center is generated, and the actuator is driven to dynamically correct the preliminary positioning position according to the offset correction amount to maintain the positioning stability of the wheel-rail contact center.

[0063] The technical solution of this application has the following beneficial effects:

[0064] This application significantly improves the positioning accuracy and stability of the wheel-rail contact center through the collaborative positioning and dynamic correction mechanism of dual detection units. Based on chord length measurement and half-distance movement, the wheel center is quickly and initially positioned. Combined with the collaborative analysis of dynamic contact spot deformation characteristics and flange degradation data, offset correction values are generated in real time to drive the actuator for multi-dimensional dynamic compensation. This effectively suppresses the accumulation of positioning deviations caused by geometric wear and uneven stress distribution in dynamic wheel-rail contact, reduces the risk of abnormal wear caused by contact stress concentration, and enhances the system's adaptability to wheel-rail material deformation and dynamic load changes, extending the service life of wheels and tracks and ensuring the smoothness and safety of train operation.

[0065] Furthermore, by integrating static wheel flange geometry degradation characteristics with dynamic contact patch deformation data, a full-dimensional dynamic compensation mechanism for the wheel-rail contact center is constructed, significantly improving positioning stability under high-speed operating conditions. Based on 3D laser scanning, a precise geometric mapping of wheel flange thickness and tread wear is generated. Combined with a millimeter-wave radar network, the dynamic contact patch deformation pattern is captured in real time. A collaborative learning framework analyzes the correlation between wheel flange degradation and contact stress distribution, driving a multi-axis actuator to achieve dynamic correction compensation in the lateral, vertical, and rotational dimensions. This effectively suppresses contact center offset caused by wheel-rail geometry deformation and dynamic load fluctuations, reduces the risk of abnormal wear, enhances the adaptive matching capability of the wheel-rail contact interface, extends the service life of the track system, and provides technical support for the smoothness and safety of trains under high-speed operation.

[0066] 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

[0067] 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.

[0068] Figure 1 A flow chart showing a method for locating the wheel center of a rail vehicle provided by the present application is shown;

[0069] Figure 2 A schematic structural diagram of a wheel center positioning system for a rail vehicle provided in the present application is shown. DETAILED DESCRIPTION

[0070] 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.

[0071] 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.

[0072] This application aims to overcome the defects of traditional wheelset positioning technology that relies on static geometric calibration and cannot adapt to the dynamic wear of the wheel flange and the interference of contact spot deformation, and to solve the hidden risks of wheel-rail eccentric wear and derailment caused by low efficiency of manual calibration and positioning reference drift. This application constructs a wheel center dynamic correction and contact mechanics collaborative control system by integrating multi-source detection data and dynamic characteristics of wheel-rail coupling, aiming to achieve autonomous perception, real-time correction and stability maintenance of the wheel contact center under complex working conditions, and improve the monitoring accuracy of the wheel-rail system throughout its life cycle and the safety redundancy of train operation.

[0073] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0074] Figure 1 A flow chart of a method for locating the wheel center of a rail vehicle is provided in an embodiment of the present application. Figure 1 As shown, the method includes:

[0075] 101. Pre-install a first detection unit and a second detection unit on both sides of the wheel;

[0076] In this step, the first detection unit refers to a laser ranging sensor installed on the outside of the wheel for capturing the geometric features of the wheel rim.

[0077] The second detection unit is a capacitive displacement sensor arranged under the wheel tread for assisting positioning.

[0078] In this embodiment, the detection units are deployed through a combination of laser ranging sensors and capacitive displacement sensors. The first detection unit is installed on the outside of the wheel rim (with a spacing of 15 mm), and the second detection unit is located below the wheel tread (at an elevation angle of 30°). The sampling timing of the two units is synchronized using the CAN bus (at a frequency of 1 kHz). Based on the geometric symmetry of the wheelset, adaptive threshold segmentation (Otsu algorithm) is used to extract wheel profile point cloud data. Combined with the Hough transform, the highest point of the wheel rim and the lowest point of the tread are detected. The installation angle compensation parameters of the two sensors are dynamically calibrated (with a compensation accuracy of ±0.02 mm) to ensure that the dual detection units are in the wheel turning workshop of the Baotou Depot. The wheel turning worker first starts the Hegenscheidt CNC wheel turning machine and installs a standard sample wheelset (860 mm diameter, 1353 mm inside distance) into the processing position. The sample wheels are measured three times using a laser wheel diameter ruler to ensure that the wheel diameter error is ≤0.05 mm and the inside distance fluctuation is controlled within a range of ±0.2 mm. During the calibration process, the operator simultaneously calibrates the machine tool's axial positioning accuracy, checks the guideway straightness with a dial gauge (error <0.01mm / m), and verifies the cutting tool holder's feed resolution (0.001mm / pulse). Upon completion of the calibration, the system generates an "Equipment Status Confirmation Sheet," providing a reference coordinate system for subsequent maintenance.

[0079] 102. Determine the chord length of the wheel according to the detection process of the wheel by the first detection unit;

[0080] In this step, the chord length refers to the maximum straight-line distance of the local arc segment of the wheel tread measured by the first detection unit.

[0081] The detection process refers to the process of fitting the geometric features of the rim based on the sliding window method and the least squares method.

[0082] In this embodiment, based on the wheel rim contact point sequence collected by the first detection unit (sampling interval 0.1mm), a sliding window method (window width 50mm) is used to extract the continuous curvature characteristics of the wheel tread. The initial chord length value is generated by fitting local arc segments using the least squares method (fitting residual <0.05mm). The effective range of the chord length is constrained by the prior parameters of the wheel diameter (standard value 840-920mm). A Kalman filter is used to eliminate lateral vibration noise of the wheelset (vibration amplitude ±3mm). The final dynamic chord length measurement value (update frequency 200Hz) is output. Its accuracy is optimized to ±0.1mm through cubic polynomial interpolation, providing a geometric reference for subsequent half-chord length movement.

[0083] After receiving a CRH5A EMU wheelset (with a cumulative mileage of 128,000 kilometers), a wheel turner scanned the tread using a portable wheel flange profiler and measured a flange thickness of 28.5mm (close to the scrap threshold of 28mm) and an equivalent taper of 0.18mm (0.05mm above the standard). Comparing the wheel diameter attenuation curve recorded by the TCMS system revealed that the diameter of wheelset No. 3 had worn from 860mm to 854.3mm. The operator entered "progressive cutting" mode into the CNC system, which automatically generated turning parameters: a target diameter of 853mm (with a 0.3mm safety margin), a single-side cutting depth of 0.35mm, and a Q value of 0.12. The system simultaneously calculated tool path compensation to eliminate the 0.07mm radial error caused by wheelset installation eccentricity.

[0084] 103. When the second detection unit detects the wheel, calculate half of the chord length, and control the second detection unit to move half of the chord length in the same direction;

[0085] In this step, half the chord length refers to the displacement amount for adjusting the position of the second detection unit calculated based on the chord length measurement value.

[0086] The same direction refers to the longitudinal movement path of the rail that is consistent with the direction of wheel movement.

[0087] In this embodiment of the present application, when the second detection unit triggers a wheel rim contact signal, a timestamp alignment mechanism is used to synchronize the two unit data streams. The target displacement of the half-chord length is calculated by dividing the instantaneous chord length by two (resolution 0.01mm). A linear motor (with a repeatability of ±5μm) is driven to move the second detection unit longitudinally along the track. PID closed-loop control (with a proportional coefficient Kp = 0.8) is used to compensate for track deviations caused by track unevenness in real time. Simultaneously, inertial measurement unit (IMU) data is integrated to eliminate mechanical vibration interference (acceleration threshold 0.3g), ensuring that the position error during the half-chord length movement is stable within ±0.15mm, completing the target position adjustment of the detection unit.

[0088] After the turning process is initiated, the carbide tool cuts into the wheel rim at a speed of 650 r / min, with a feed rate of 0.02 mm per revolution. The cutting depth is fine-tuned in real time by a hydraulic servo system. The operator monitors the cutting zone temperature (peaking at 320°C) via an infrared monitor. When the thickness of the blue chips exceeds 0.1 mm, a high-pressure air blower (at 0.6 MPa) is immediately triggered, breaking the continuous ribbon of chips into fragments less than 5 cm in length. Cut-resistant gloves and goggles are worn throughout the operation, and the chip chute is cleaned every 15 minutes to prevent the accumulation of hot chips, which could trigger an overheating alarm. After 2 hours and 36 minutes of machining, the wheelset diameter has returned to 853.1 mm, with a surface roughness of Ra ≤ 1.6 μm.

[0089] 104. When the second detection unit is controlled to move in the same direction by half the chord length, a current position of the second detection unit is determined as a preliminary positioning position of the wheel center;

[0090] In this step, the current position refers to the absolute spatial coordinates after the second detection unit completes the half-chord length movement.

[0091] The preliminary positioning position refers to the spatial reference point of the wheel center generated based on the data fusion of the dual detection units.

[0092] In the embodiment of the present application, after the second detection unit completes the half-chord length displacement, its absolute position coordinates are read by a high-precision grating ruler (resolution 0.001mm), and the spatial coordinate transformation is performed in combination with the initial installation matrix of the first detection unit (4×4 homogeneous transformation matrix). The weighted average method is used to fuse the position feedback data of the two units (weight ratio 6:4) to eliminate the nonlinear error of the sensor, and at the same time, the wheel set thermal expansion compensation parameters (temperature coefficient × / ℃) to correct the influence of ambient temperature drift, and finally generate the preliminary positioning coordinates of the wheel center (three-dimensional error ellipse radius ≤ 0.25mm), and establish the initial spatial reference system of the wheel-rail contact center.

[0093] After the repair was completed, the wheelwright used a three-point contact wheel gauge to remeasure the wheelset diameter (the average of three measurements was 853.09mm, with a standard deviation of 0.03mm). He also used a rim template ruler to check the contour fit, ensuring that the clearance with the standard R100 arc surface was less than 0.05mm. Wheel-rail contact simulation software verified that the equivalent taper was reduced to 0.03mm, and the lateral force fluctuation value was optimized from 14.6kN before the correction to 8.2kN. All data was entered into the "Wheelset Repair Record Sheet" and cross-checked with the values stored in the on-board TCMS system, confirming that the wheel diameter parameter error rate was less than 0.05%. After discovering that the inner distance deviation of wheelset No. 4 was +0.3mm, the secondary calibration procedure was immediately initiated.

[0094] 105. Based on the collaborative learning framework of dynamic contact spot deformation characteristics and rim geometry degradation, an offset correction value for the initial positioning position of the wheel center is generated, and the actuator is driven to dynamically correct the initial positioning position according to the offset correction value to maintain the positioning stability of the wheel-rail contact center.

[0095] In this step, the dynamic contact patch deformation characteristics refer to the geometric changes in the wheel-rail contact area caused by forces during operation.

[0096] Flanged geometry degradation refers to the reduction in the geometric dimensions of the wheel rim due to wear during long-term operation.

[0097] The offset correction refers to the compensation value calculated by the collaborative learning framework for adjusting the wheel center positioning.

[0098] The actuator refers to the linear servo motor used to achieve dynamic adjustment of the wheel center position.

[0099] Positioning stability refers to the ability of the wheel-rail contact center to maintain position error within the allowable range during dynamic correction.

[0100] In this embodiment, a time-series correlation model is constructed based on an LSTM neural network to correlate contact patch deformation (sampling rate 500Hz) with wheel flange wear (3D scanning accuracy 0.02mm). Dynamic contact force (six-dimensional force sensor data) and wheel diameter degradation parameters (monthly mean change) are input. An attention mechanism is then used to extract synergistic features between contact patch area fluctuation (standard deviation threshold 0.8mm²) and wheel flange thickness loss. A random forest regression algorithm is used to generate position corrections (output resolution 0.01mm), driving a linear servo motor (response time 10ms) to perform submicron position compensation. The Lyapunov stability criterion is used to constrain the correction amplitude (maximum correction ±1.5mm), ensuring a stable wheel-rail contact center positioning error within ±0.3mm, ensuring the dynamic balance of the train's running posture.

[0101] The corrected data was uploaded to the vehicle health management system. Combined with the wheelset material fatigue model (based on the Archard wear equation), the remaining service life of the wheelset was predicted to increase from 70,000 kilometers to 110,000 kilometers. The system automatically generated a "Rotation Effect Evaluation Report," demonstrating an 18% reduction in wheel-rail contact stress and a 23dB reduction in vibration energy. The system also optimized the next rotation cycle to 90,000 kilometers and recommended adjusting the train's route to reduce the frequency of curves with a radius of less than 600 meters. Finally, the wheelset was equipped with RFID tags to enable full lifecycle data traceability.

[0102] In summary, steps 101 to 105 achieve high-precision dynamic self-compensating positioning of the wheel-rail contact center. Through collaborative measurement of dual detection units and a chord-length adaptive motion mechanism, combined with real-time fusion analysis of contact patch deformation characteristics and flange degradation parameters, a fully closed-loop correction model for wheel center position is constructed. Based on dynamic contact mechanical feedback and a geometric feature decoupling algorithm, the system automatically eliminates positioning deviations caused by flange wear, tread peeling, and lateral wheel-rail vibration, establishing a stable tracking capability for the wheel-rail contact center that is resistant to mechanical deformation interference. This improves the robustness and positioning repeatability of dynamic detection of rail vehicle wheelsets, providing a real-time dynamic benchmark for optimizing the wheel-rail relationship.

[0103] In order to overcome the defects of traditional wheel-rail contact center positioning technology that relies on offline calibration and cannot adapt to dynamic wear of the wheel flange and random deformation of the contact patch, and to solve the positioning drift and derailment risks caused by static parameter update lag and inaccurate dynamic deformation modeling, this technology integrates multi-dimensional wear characteristics and real-time contact mechanics data to construct a wheel-rail geometry degradation compensation system based on collaborative learning. It aims to achieve adaptive perception, dynamic correction and stability closed-loop control of the wheel center position, providing rail transit systems with high-precision wheel-rail relationship optimization support under all working conditions and throughout the entire cycle.

[0104] In some embodiments, the step 105 of generating the offset correction value of the preliminary positioning position of the wheel center based on the collaborative learning framework of the dynamic contact patch deformation characteristics and the rim geometry degradation includes:

[0105] 201. Obtaining the surface geometry of wheels with different degrees of wear through three-dimensional laser scanning, generating a static parameter set including the thickness of the wheel rim and the wear depth of the tread, and establishing a spatial geometric mapping relationship between the wheel rim centerline and the track contact surface based on the wheel rim thickness and the wear depth of the tread;

[0106] In step 201 , three-dimensional laser scanning refers to a technology that uses a laser beam to collect high-precision three-dimensional point clouds on the wheel surface.

[0107] The thickness of the rim refers to the vertical distance from the highest point to the base surface of the wheel rim.

[0108] Tread wear depth refers to the amount of surface material lost from the wheel tread due to wear.

[0109] The static parameter set refers to a fixed data set of wheel geometric characteristics obtained through measurement.

[0110] The spatial geometric mapping relationship refers to the spatial position and angle association model between the wheel rim centerline and the track contact surface.

[0111] In this embodiment, a high-precision 3D laser scanner (resolution 0.02mm) is used to collect a full-circumferential point cloud of the wheel surface. A multi-view registration algorithm (such as ICP Iterative Closest Point) is used to fuse wheel geometry data at different wear stages to generate a static parameter set for rim thickness (measurement error ±0.1mm) and tread wear depth (quantization accuracy ±0.05mm). A 3D wheel model is constructed using non-uniform rational B-spline (NURBS) surface reconstruction technology. Combined with the track contact surface inclination (range 5°-15°) and standard wheel-rail clearance parameters, a spatial geometric projection algorithm is used to map the rim centerline to the track contact surface coordinate system. This generates a spatial mapping relationship matrix that includes rim geometric degradation and track compatibility, providing a static benchmark for dynamic contact analysis.

[0112] 202. A millimeter-wave radar network is distributed longitudinally along the track to capture the spatial coordinates of the dynamic contact patch in the wheel-rail contact area in real time, and the deformation characteristics of the dynamic contact patch are analyzed by combining the spatial geometric mapping relationship between the center line of the wheel flange and the track contact surface;

[0113] In step 202 , the millimeter wave radar network refers to an array of high-frequency radar devices arranged along the track for capturing dynamic contact patches.

[0114] The dynamic contact spot spatial coordinates refer to the spatial position data of the wheel-rail contact area that changes with time during operation.

[0115] The deformation characteristics refer to the geometric shape change pattern of the contact spot after being subjected to force.

[0116] In this embodiment, millimeter-wave radar nodes (operating at 77 GHz) are deployed every 10 meters along the track. MIMO array technology is used to capture the three-dimensional coordinates of the wheel-rail contact patch in real time (with a refresh rate of 100 Hz). A multi-target tracking algorithm (such as JPDA joint probabilistic data association) is used to separate the wheel-rail contact patch from other interfering signals. Combined with the wheel flange centerline mapping relationship in step 201, a curvature matching algorithm is used to calculate the contact patch shape variable (such as ellipticity change of ±0.3 mm). Time series analysis (ARIMA model) is introduced to extract contact patch area fluctuations (standard deviation threshold 1.2 mm²) and position drift trends (speed 0.5 mm / s). A dynamic contact patch deformation characteristic spectrum is constructed to achieve real-time state perception of the wheel-rail contact mechanics.

[0117] 203. Inputting the wheel rim geometric degradation characteristics and the dynamic contact patch deformation characteristics in the static parameter set into a collaborative learning framework, and generating an offset correction value for the preliminary positioning position of the wheel center through an interactive method of local model parameter encryption iteration and global parameter aggregation;

[0118] In step 203, the collaborative learning framework refers to a distributed learning architecture that implements model training through multi-node data interaction.

[0119] Local model parameter encryption iteration refers to the process of encrypting, updating, and optimizing model parameters at the local node.

[0120] Global parameter aggregation refers to the operation of fusing and unifying the model parameters of multiple local nodes.

[0121] The offset correction refers to the compensation value used to adjust the wheel center position.

[0122] In this embodiment, based on a federated learning framework, local nodes use a lightweight convolutional network (MobileNet-V3) to extract features from the wheel rim thickness degradation rate (monthly average 0.05-0.2mm) and contact patch deformation rate (0.1-0.8mm / s). Homomorphic encryption is then used to securely transmit the model gradient parameters (dimension 256). A global server fuses multi-node features using an adaptive weighted aggregation algorithm (weights assigned based on the confidence level of the node data). A gated recurrent unit (GRU) is used to construct a time series prediction model, outputting wheel center lateral offset corrections (resolution 0.01mm) and longitudinal compensations (step size 0.05mm), forming a distributed collaborative correction strategy that resists data siloing.

[0123] 204. Driving the actuator to dynamically correct the preliminary positioning position according to the offset correction amount to maintain the positioning stability of the wheel-rail contact center includes:

[0124] In step 204 , the actuator refers to a mechanical device for dynamically adjusting the wheel center position.

[0125] Positioning stability refers to the ability of the wheel-rail contact center to maintain position error within the allowable range during dynamic correction.

[0126] In this embodiment, the actuator uses a linear voice coil motor (thrust 120N, response time 5ms) to receive offset correction commands. The Lyapunov stability criterion constrains the correction amplitude (maximum ±2mm). Combined with a heat map of the spatial distribution of the contact patch deformation (Gaussian kernel radius 50mm), the PID control parameters (proportional coefficient 0.6-1.2) are dynamically adjusted. The wheel-rail contact force feedback loop integrates six-dimensional force sensor data (sampling rate 1kHz) with wheel rim geometry scans. Kalman filtering is used to achieve real-time matching of the actuator displacement (accuracy ±0.03mm) with the contact patch coordinates, ensuring synchronization of the correction process with the wheel-rail dynamics.

[0127] 205. Based on the offset correction amount and in combination with the spatial distribution law of the deformation characteristics of the dynamic contact spot, the actuator is driven to adjust the initial positioning position, and the positioning stability of the wheel-rail contact center during high-speed operation is maintained through real-time matching of the wheel rim geometric characteristics and the contact spot coordinates.

[0128] In step 205 , the spatial distribution law refers to the spatial variation trend and pattern of the dynamic contact spot deformation characteristics.

[0129] Rim geometry refers to the geometric shape and size parameters of the wheel rim.

[0130] Real-time matching of contact spot coordinates refers to the process of dynamically aligning the spatial position of the contact spot with the geometric features of the rim.

[0131] In this embodiment, principal component analysis (cumulative variance contribution >85%) of the spatial distribution of contact patch deformation is used to extract deformation directional feature vectors (e.g., axial contribution of 60%), driving the actuator to perform dynamic compensation along a composite transverse-vertical track motion trajectory (interpolation period 0.1ms). By real-time registration (matching error <0.1mm) of wheel flange geometry (e.g., a 35°-42° wheel flange angle) with the contact patch coordinates, and adaptively adjusting the correction frequency based on the wheelset's meandering frequency (1-3Hz), the wheel-rail contact center offset is controlled within ±0.25mm at a speed of 380km / h, achieving closed-loop control of positioning stability under high-speed operating conditions.

[0132] Here's a specific example:

[0133] In the wheel-rail contact center location scenario for heavy-haul freight trains, the system uses a high-frame-rate 3D laser scanner (resolution 0.05mm) to perform full circumferential scans of long-service wheels. A phase unwrapping algorithm is used to reconstruct the wheel flange thickness (measurement range 30-40mm) and tread wear gradient (accuracy ±0.08mm). The wheel-rail spatial geometry mapping matrix is then constructed based on the rail contact surface curvature radius (standard value 300-350mm). 77GHz millimeter-wave radar nodes are deployed every 15 meters along the freight line. Doppler shift compensation technology is used to mitigate the impact of low-speed train oscillation (amplitude ±5mm) on contact patch detection. A Bayesian filtering algorithm is used to fuse the time series of contact patch area (baseline value 200-300mm²) with position offset (detection error ±0.3mm). A blockchain-based federated learning framework trained 20 freight trains on wheel flange degradation rates (average daily rate of 0.02-0.05mm) and contact patch fluctuation characteristics (frequency of 0.5-2Hz). A gated attention mechanism was used to screen key features and generate offset correction parameters (compensation step size of 0.03-0.1mm). This system then actuated a six-degree-of-freedom parallel mechanism (with a repeatability of ±0.02mm) to perform dynamic compensation. By combining real-time matching of contact patch heat maps (grid density 5mm×5mm) with wheel flange angles (35°-45°), the system achieved a stable wheel-rail contact center offset within a ±0.4mm threshold under heavy load conditions at 80km / h, effectively addressing the snaking instability caused by eccentric wheel flange wear on freight trains.

[0134] In summary, steps 201 to 205 achieve dynamic self-healing positioning of the wheel-rail contact center throughout its lifecycle. By integrating static wear parameters from 3D laser scanning with dynamic contact patch deformation characteristics captured by millimeter-wave radar, a multimodal data-based rim geometry degradation compensation model is constructed. Based on a federated learning framework, the system implements encrypted training of local wear data and distributed aggregation of global dynamic parameters, generating a positioning correction strategy that is resistant to mechanical deformation interference. By combining the spatial distribution of the contact patch with a real-time matching mechanism for rim geometry, this system overcomes the hysteresis limitations of traditional static detection and achieves submillimeter dynamic compensation of wheel center position under high-speed conditions. This effectively suppresses wheel-rail eccentric wear and snaking instability, improving train operation safety and wheelset maintenance efficiency.

[0135] In order to overcome the difficulty in decoupling geometric degradation characteristics from dynamic stress and deformation in wheel-rail contact center positioning technology, and the correction lag problem caused by multi-node data noise interference, and to solve the positioning drift risk caused by traditional methods relying on single-dimensional features and lack of cross-node data security, this technology integrates gradient tensor decomposition and anti-noise encryption mechanism to construct a global topological analysis model of degradation-deformation coupling, aiming to achieve highly robust dynamic correction and cross-node collaborative optimization of the wheel center position, and improve the collaborative control accuracy and system stability of wheel-rail contact mechanics under complex scenarios such as heavy load and high speed.

[0136] In some embodiments, as described in step 203, the rim geometry degradation characteristics and the dynamic contact patch deformation characteristics in the static parameter set are input into a collaborative learning framework, and an offset correction value for the preliminary positioning position of the wheel center is generated through an interactive method of local model parameter encryption iteration and global parameter aggregation, including:

[0137] 301. Decompose the geometric deformation gradient tensor of the wheel rim geometric degradation feature, extract the degradation feature vector related to the wheel-rail contact surface curvature in the geometric deformation gradient tensor, and generate a feature decomposition sequence including the curvature degradation gradient;

[0138] In step 301 , the geometric deformation gradient tensor refers to a mathematical structure that describes the rate and direction of change of the rim geometric degradation characteristics in space.

[0139] The curvature degradation gradient refers to the rate of change of geometric deformation related to the curvature of the wheel-rail interface.

[0140] The eigendecomposition sequence refers to the set of eigenvectors containing curvature degenerate gradients obtained by decomposing the geometric deformation gradient tensor.

[0141] In this embodiment, a tensor decomposition algorithm is used to reduce the dimensionality of wheel flange geometric degradation features. Principal component analysis is used to extract eigenvectors related to the curvature of the wheel-rail contact surface, generating an eigendecomposition sequence containing the curvature degradation gradient. Combining the wheel flange thickness degradation rate (daily average of 0.02-0.05mm) with the tread wear depth (quantization accuracy of ±0.08mm), singular value decomposition techniques are used to isolate the curvature-sensitive components of the geometric deformation gradient tensor. A correlation matrix is constructed between the degradation features and the rail contact surface curvature (standard value of 300-350mm), providing core feature data for subsequent noise masking and cross-node fusion.

[0142] 302. Execute a noise masking mechanism on the eigendecomposition sequence at the local computing node, superimpose random mask noise through a dynamic attenuation window, and generate an interference-resistant degenerate gradient encryption chain;

[0143] In step 302, the noise mask mechanism refers to a technology that enhances data security and anti-interference capability by superimposing random noise.

[0144] The dynamic attenuation window refers to the data processing window that adjusts the noise amplitude over time.

[0145] The degraded gradient encryption chain refers to the interference-resistant degraded gradient data sequence after noise mask processing.

[0146] In this embodiment, a dynamic attenuation window (50ms) is used at the local computing node to perform noise masking on the eigendecomposition sequence. A Gaussian random function is used to generate zero-mean masking noise (standard deviation 0.1) and superimposed on the degraded gradient sequence. An adaptive threshold segmentation algorithm (OTSU) is used to screen for valid eigenvalues. The noise-masked gradient sequence is encrypted using homomorphic encryption technology to generate an interference-resistant degraded gradient encryption chain, ensuring security and noise resistance during data transmission.

[0147] 303. Construct a joint parameter space using the degraded gradient encryption chain and the stress distribution time series signal in the dynamic contact patch deformation feature, and generate a global parameter topology chain coupled with degradation and deformation through cross-node parameter interpolation operation in the collaborative learning framework;

[0148] In step 303 , the joint parameter space refers to a multi-dimensional data space constructed by fusing the degraded gradient encryption chain with the dynamic contact spot stress distribution signal.

[0149] Cross-node parameter interpolation refers to the operation of data interpolation and fusion between different computing nodes.

[0150] The global parameter topology chain refers to the global data chain of degenerate and deformation coupling generated by cross-node fusion.

[0151] In this embodiment, a kernel function mapping technique is used to construct a joint parameter space based on a degraded gradient encryption chain and a dynamic contact patch stress distribution time series signal (sampling rate 1kHz). Through cross-node parameter interpolation (e.g., bicubic spline interpolation) within a collaborative learning framework, local degradation characteristics are integrated with global stress distribution patterns to generate a global parameter topology chain that couples degradation and deformation. Combined with wheel-rail contact force feedback data (from a six-dimensional force sensor), the phase alignment accuracy between the degradation gradient and stress distribution in the topology chain is optimized, providing a foundation for subsequent dynamic iterative encryption.

[0152] 304. Dynamically iteratively encrypt the degenerate gradient in the global parameter topology chain at the local computing node, fuse the stress distribution of the wheel-rail contact area to generate an encrypted weight gradient chain, and distribute it to the global aggregation node;

[0153] In step 304 , dynamic iterative encryption refers to a process of performing multiple encryption processes on the degraded gradient data to enhance security.

[0154] The encrypted weight gradient chain refers to the gradient data sequence with encrypted weights generated after fusing stress distribution data.

[0155] In this embodiment, a dynamic iterative encryption algorithm is used at local computing nodes to encrypt the degradation gradients in the global parameter topology chain. This encrypted weight gradient chain is generated using stress distribution data (quantization accuracy ±0.05 MPa) in the wheel-rail contact area. This encrypted gradient chain is distributed to global aggregation nodes using distributed hash table (DHT) technology. The encrypted weight distribution strategy is optimized based on the wheel flange geometry degradation rate (monthly average 0.2-0.5 mm) and the contact patch deformation frequency (0.5-2 Hz), ensuring the accuracy and efficiency of cross-node data fusion.

[0156] 305. Aggregate the encrypted gradient chains of multiple nodes in a global aggregation node, eliminate the random mask noise, extract the degradation and deformation coupling phase offset across nodes, and generate an offset correction for the preliminary positioning position of the wheel center.

[0157] In step 305 , the cross-node degradation and deformation coupling phase offset refers to the phase difference between the degradation and deformation features extracted by aggregating multi-node data.

[0158] The offset correction refers to the compensation value used to adjust the wheel center position.

[0159] In this embodiment, a weighted averaging algorithm is used to aggregate encrypted gradient chains across multiple nodes at a global aggregation node. Fourier transforms are then used to extract the cross-node phase offset of degradation and deformation coupling (resolution 0.01mm). A low-pass filter (cutoff frequency 10Hz) is used to eliminate random mask noise, and the offset correction is generated based on the wheel-rail contact center positioning error (threshold ±0.3mm). The correction amplitude is constrained (maximum ±1.5mm) based on the Lyapunov stability criterion, achieving high-precision dynamic compensation and stability control of the wheel center position.

[0160] Here's a specific example:

[0161] In the high-speed rail wheel center positioning scenario, the wheel rim of a certain train set showed non-uniform wear after long-term operation. First, the wheel rim cross-section geometric data was obtained by a 3D laser scanner, and the deformation gradient tensor matrix was constructed based on the curvature differential algorithm to extract the contact surface curvature change rate exceeding The feature vector of the wheel / rail contact patch is used to generate a degraded gradient sequence containing different wear phases (0°, 120°, and 240°). Local edge computing nodes use an adaptive Hamming window to dynamically suppress noise in the gradient sequence and superimpose quantum random noise (amplitude controlled to ±5μm) that meets NIST standards to form an encryption chain. During the cross-node collaboration phase, the encryption chain is spatiotemporally aligned with the finite element stress cloud map (sampling rate 1kHz) of the wheel / rail dynamic contact patch. A joint feature space containing 12-dimensional deformation parameters is constructed through bicubic spline interpolation. After three rounds of federated learning iterations, the global aggregation node uses an improved K-SVD algorithm to eliminate multi-node noise interference. Based on the phase offset analysis of the contact patch stress extreme point (maximum stress 380MPa), the final output is the wheel center alignment correction. In actual tests on a CR400BF vehicle, this method reduced the lateral positioning error of the wheelset from 2.3mm to 0.7mm, effectively eliminating the harmonic component of the snaking motion caused by wheel rim wear (amplitude reduced by 63%), and meeting the sub-millimeter positioning accuracy requirements at a speed of 350km / h.

[0162] In summary, steps 301 to 305 achieve multi-dimensional, noise-resistant collaborative positioning compensation for the wheel-rail contact center. By jointly modeling the geometric deformation gradient tensor decomposition and dynamic stress distribution signals, a coupled analysis system for degradation characteristics and contact patch deformation is constructed. The system generates an interference-resistant encrypted gradient chain based on a noise masking mechanism. It utilizes cross-node parameter interpolation to fuse local degradation characteristics with global deformation patterns, overcoming the characteristic deviation limitations of traditional single-node learning. This allows for phase-synchronized analysis of wheel flange degradation gradients and contact stresses, resulting in a high-precision correction value generation capability that is resistant to data noise and node heterogeneity. This ensures the dynamic stability and interference resistance of wheel center positioning under complex operating conditions, providing a holographic compensation benchmark for dynamic wheel-rail adaptation.

[0163] In order to overcome the problems of compensation lag and deviation accumulation caused by spatiotemporal data asynchrony and isolated local parameters in wheel-rail contact center compensation technology, and to solve the problem of positioning inaccuracy caused by the inability of traditional methods to coordinate degradation gradients and dynamic stress fluctuations and cross-regional data faults, this technology integrates spatiotemporal alignment and multi-node collaborative optimization mechanism to construct a degradation-deformation global topological analysis framework, aiming to achieve real-time phase matching and dynamic weight adaptive compensation of the wheel-rail contact center, providing a highly robust wheel-rail collaborative control solution for extreme working conditions such as high speed and heavy load.

[0164] In some embodiments, as described in step 303, a joint parameter space is constructed by combining the degradation gradient encryption chain and the stress distribution time series signal in the dynamic contact patch deformation feature, and a global parameter topology chain coupled with degradation and deformation is generated through cross-node parameter interpolation operation in the collaborative learning framework, including:

[0165] 401. Align the sampling timestamps of the spatiotemporal distribution and stress distribution time series signals in the degradation gradient encryption chain, match the geometric deformation phase and dynamic stress fluctuation period of the wheel-rail contact area, and generate a spatiotemporally aligned degradation and stress parameter matrix;

[0166] In step 401, the degraded gradient encryption chain refers to the interference-resistant degraded gradient data sequence after noise mask processing.

[0167] The stress distribution time series signal refers to the stress fluctuation data in the wheel-rail contact area that changes with time.

[0168] The geometric deformation phase refers to the periodic variation characteristics of the geometric deformation in the wheel-rail contact area over time.

[0169] The dynamic stress fluctuation period refers to the periodic change pattern of the stress distribution signal in time.

[0170] The parameter matrix refers to the data matrix generated by aligning the degradation gradient with the stress distribution in time and space.

[0171] In this embodiment, a time series alignment algorithm (such as dynamic time warping) is used to match the sampling timing of the degradation gradient encryption chain with the stress distribution signal. A phase-locked loop (PLL) technique is then used to synchronize the geometric deformation phase (accuracy ±0.5°) of the wheel-rail contact area with the dynamic stress fluctuation period (frequency 0.1-5Hz). Based on the wheel flange degradation rate (average daily rate of 0.02mm) and the stress amplitude change rate (gradient 0.3MPa / s), a spatiotemporal synchronization compensation matrix (dimension 128×128) is constructed. Null values caused by timestamp deviations are filled using cubic spline interpolation (interpolation error <0.1%). Ultimately, a parameter matrix is generated that contains the spatiotemporal correlations between the degradation gradient and the stress distribution, providing a reference data source for cross-node interpolation.

[0172] 402. Perform parameter interpolation operations between multiple nodes on the parameter matrix to compensate for the difference in deformation gradients of dynamic contact patches between adjacent track sections, and generate an interpolation compensation parameter chain that is continuously distributed across nodes.

[0173] In step 402 , the parameter interpolation operation among multiple nodes refers to the operation of performing data interpolation and fusion among different computing nodes.

[0174] The dynamic contact patch deformation gradient difference refers to the degree of difference in the deformation characteristics of the contact patches between adjacent track sections.

[0175] The interpolation compensation parameter chain refers to a continuous sequence of distribution parameters generated by interpolation across nodes.

[0176] In this embodiment, a spatial kriging interpolation algorithm is used between multiple nodes to compensate for parameter discontinuities between adjacent nodes based on the track segment length (standard: 25m) and the difference in dynamic contact patch deformation gradients (threshold: ±0.5mm / m). The spatial continuity of the contact patch deformation gradient (range: 1-8mm / m) is adjusted using an adaptive interpolation step size (0.5-2m). The interpolation parameter distribution is optimized using node confidence weights (0.6-0.9). This generates a continuous interpolation compensation chain (with a resolution of 0.1mm) covering the entire track, eliminating compensation lags caused by local data silos and forming a global spatial correlation model between degradation gradients and deformation characteristics.

[0177] 403. Fusing the degradation gradient and stress distribution amplitude in the interpolation compensation parameter chain, adjusting the fusion weight of the gradient and stress based on the instantaneous deformation rate, and generating a dynamic weight fusion parameter set;

[0178] In step 403 , the degradation gradient refers to the rate of change of the rim geometry degradation characteristics.

[0179] The stress distribution amplitude refers to the intensity of stress fluctuation in the contact spot.

[0180] The instantaneous deformation rate refers to the instantaneous change rate of the geometric deformation of the wheel-rail contact area.

[0181] The dynamic weight fusion parameter set refers to the fusion parameter set generated after adjusting the weight according to the instantaneous deformation rate.

[0182] In this embodiment, the fusion weights of degradation gradient and stress amplitude are dynamically adjusted based on the instantaneous deformation rate (quantization accuracy ±0.05 mm / s). An exponential decay function (time constant 50 ms) is used to assign gradient weights (0.3-0.7) and stress weights (0.7-0.3). A sliding window (width 100 ms) is used to extract the contact patch stress peak (threshold 2 MPa) and degradation gradient extreme value (threshold 0.1 mm / m). A fuzzy logic rule base (32 rules) is used to optimize the weight assignment strategy and generate a dynamic weight fusion parameter set (update frequency 50 Hz), achieving real-time adaptive fusion of degradation and deformation characteristics.

[0183] 404. Perform spatiotemporal convolution of degradation gradient and stress distribution on the dynamic weight fusion parameter set to extract parameter aggregation features of the wheel-rail contact center area and generate a feature topology network of degradation and deformation coupling;

[0184] In step 404, the spatiotemporal convolution superposition refers to the process of performing convolution operations on the degradation gradient and the stress distribution in the time and space dimensions.

[0185] Parameter aggregation features refer to the key features of the wheel-rail contact center area extracted by convolutional superposition.

[0186] The feature topology network refers to the relationship network between features of degenerate and deformation coupling.

[0187] In this embodiment, a three-dimensional spatiotemporal convolution kernel (size 3×3×5) is used to extract features from the dynamic weight fusion parameter set. Dilated convolution (with a dilation rate of 2) is used to enhance the long-range dependencies between degradation gradients and stress distributions. This is combined with a max pooling layer (stride length 2) to compress redundant feature dimensions. Residual connections are used to preserve the original parameter distribution characteristics, ultimately generating a feature topology network (512 nodes) that couples degradation and deformation. This topology utilizes a graph attention mechanism (8 heads) to enhance parameter aggregation in the contact center region, forming a global feature representation of wheel-rail dynamic adaptation.

[0188] 405. Analyze the cooperative oscillation mode of the degradation gradient and the stress distribution in the characteristic topological network to generate a global parameter topological chain of degradation and deformation coupling.

[0189] In step 405 , the cooperative oscillation mode refers to the law of synchronous changes between the degradation gradient and the stress distribution.

[0190] The global parameter topological chain refers to the global parameter association chain generated by analyzing the cooperative oscillation mode.

[0191] In this embodiment, a Fourier transform is used to extract the fundamental oscillation mode (main frequency 0.5-2 Hz) of the degradation gradient and stress distribution in the characteristic topological network. Coherence analysis (threshold 0.85) is then used to screen the cooperative oscillation components. A spectral clustering algorithm (number of clusters 5) is used to divide the phase synchronization region. The Lyapunov exponent (threshold 0.2) is combined to evaluate the oscillation stability. Ultimately, a global parameter topological chain (dimension 256) is generated for the coupling of degradation and deformation. The topological relationships are optimized using node betweenness centrality (threshold 0.4), providing a full-dimensional parameter correlation benchmark for dynamic compensation of the wheel-rail contact center.

[0192] In summary, steps 401 to 405 achieve global spatiotemporal dynamic compensation of the wheel-rail contact center. By aligning the spatiotemporal phases of degradation gradients and stress distributions and performing multi-node interpolation compensation, a cross-regional parameter fusion system with continuous distribution is constructed. Based on a dynamic weight adjustment mechanism and a spatiotemporal convolution superposition algorithm, the system analyzes the co-oscillation laws of degradation and deformation, breaking through the spatiotemporal limitations of traditional local compensation. This model forms a global topological correlation between degradation gradients and stress fluctuations, achieving high-precision anti-disturbance correction of the contact center position, effectively suppressing phase mismatch and stress concentration issues in dynamic wheel-rail contact, and improving the dynamic adaptability and service life of the wheel-rail system under complex operating conditions.

[0193] Here's a specific example:

[0194] In the scenario of dynamic gap compensation for urban maglev trains, the system uses a submillimeter-wave 3D profilometer (wavelength 0.3mm) to scan the surface topography of the electromagnetic pole piece in real time. A dynamic time warping algorithm aligns the levitation force fluctuation signal (frequency range 50-500Hz) with the pole piece wear gradient data (daily variation 0.005-0.015mm) to construct a spatiotemporal correlation matrix (256×256 dimensions). 60GHz phased array radar nodes are deployed along the levitation track (8m spacing). An adaptive kriging interpolation algorithm is used to compensate for differences in levitation gap gradients between adjacent sections (threshold ±0.08mm / m), generating a continuous gap compensation chain across nodes. Based on the levitation electromagnetic field intensity (0.8-1.5T) and current ripple characteristics (harmonic distortion <5%), a fuzzy logic controller dynamically adjusts the fusion weights (0.4-0.6) of the wear gradient and electromagnetic stress. A 3D spatiotemporal convolution kernel (size 5×5×7) is used to extract the magnetic-mechanical coupling characteristics of the levitation center region, constructing a feature topology network consisting of 512 nodes. Wavelet coherence analysis is used to analyze the phase synchronization pattern of the suspension force pulsation (main frequency 120Hz) and the pole shoe wear gradient (weekly change 0.1mm). Spectral clustering is used to divide the dynamic compensation domain, and a multi-stage voice coil motor array (response time 2ms) is driven to implement nanometer-level suspension gap compensation (accuracy ±0.02mm). Under the operating condition of 430km / h, the suspension center offset is controlled within ±0.15mm, effectively suppressing the suspension force oscillation instability problem caused by eccentric pole shoe wear.

[0195] In some embodiments, as described in step 202, the millimeter-wave radar network distributed longitudinally along the track captures the spatial coordinates of the dynamic contact patch in the wheel-rail contact area in real time, and analyzes the deformation characteristics of the dynamic contact patch based on the spatial geometric mapping relationship between the center line of the wheel flange and the track contact surface, including:

[0196] 501. Deploy a millimeter-wave radar array longitudinally along the track to synchronously collect a pulse sequence of reflected signals from the wheel-rail contact area, capture multipath echo delays and amplitude fluctuations related to track contact surface deformation in the reflected signals, and generate an original signal grid of the spatial coordinates of the dynamic contact patch;

[0197] In step 501, the millimeter wave radar array refers to a combination of high-frequency radar devices arranged longitudinally along the track for capturing reflected signals from the wheel-rail contact area.

[0198] The reflected signal pulse sequence refers to the reflected signal data stream from the wheel-rail contact area collected by the millimeter-wave radar.

[0199] Multipath echo delay refers to the time delay caused by the reflected signal due to different paths.

[0200] Amplitude fluctuation refers to the change in the intensity of the reflected signal during propagation.

[0201] The original signal grid refers to the initial data matrix of the spatial coordinates of the dynamic contact spot generated by the reflected signal.

[0202] In this embodiment, a 77GHz millimeter-wave radar array (2m spacing) is deployed longitudinally along the track. MIMO technology is used to synchronously collect reflected pulse signals (4GHz bandwidth) from the wheel-rail contact area. A pulse compression algorithm is used to extract multipath echo delays (with a resolution of 0.1ns). Adaptive threshold segmentation (Otsu algorithm) is then used to isolate valid echoes related to contact surface deformation. Cubic spline interpolation is then used to construct a spatial coordinate primitive grid (grid density 5mm×5mm). By jointly analyzing echo amplitude fluctuations (dynamic range 60dB) and delay differences (maximum ±3ns), a primitive signal grid containing the three-dimensional coordinates of the contact patch, reflection intensity, and deformation correlation is generated, providing a data foundation for multipath mitigation.

[0203] 502. Perform multipath interference suppression processing on the original signal grid to compensate for attenuation distortion of the reflected signal during high-speed operation, and generate an interference-resistant dynamic contact patch spatial coordinate grid;

[0204] In step 502, multipath interference suppression processing refers to a process of eliminating signal interference caused by multipath effects through an algorithm.

[0205] Attenuation distortion refers to the signal distortion caused by attenuation of the reflected signal during propagation.

[0206] The anti-interference dynamic contact patch space coordinate grid refers to the space coordinate data matrix after multipath suppression processing.

[0207] In this embodiment, an adaptive filtering algorithm (LMS filter order 32) is applied to the original signal grid, and the multipath interference suppression coefficient is calculated in real time based on the train speed (200-350 km / h). A Kalman filter is used to compensate for signal attenuation distortion (attenuation coefficient 0.05-0.2 dB / m). A backpropagation neural network (hidden layer nodes 64) is used to predict the true spatial distribution of the dynamic contact patch, generating an interference-resistant coordinate grid (positioning error ±0.3 mm). The pulse arrival angle (accuracy ±0.5°) and Doppler frequency shift (±5 kHz) data are integrated to optimize the grid node confidence weights, ultimately outputting a dynamic contact patch spatial grid with a time synchronization error of <1 ms.

[0208] 503. Extracting the impulse response amplitude and time-series phase fluctuation of the dynamic contact spot spatial coordinate grid, and constructing an impulse response topological structure of the dynamic contact spot deformation characteristics;

[0209] In step 503, the impulse response amplitude refers to the intensity value of the reflected signal pulse.

[0210] Timing phase fluctuation refers to the phase change of the reflected signal pulse in time.

[0211] The impulse response topology refers to the deformation feature relationship network constructed based on the impulse response amplitude and phase fluctuations.

[0212] In this embodiment, the impulse response envelope is extracted based on the Hilbert transform. A sliding window (50ms width) is used to calculate the amplitude fluctuation range (threshold 0.5-2V). A phase unwrapping algorithm is then used to eliminate temporal phase jumps (jump threshold π / 4). A graph convolutional network (convolution kernel size 3×3) is used to construct the impulse response topology. Nodes represent the contact patch deformation intensity (0-100%), and edge weights correlate the phase synchronization of adjacent regions (coherence > 0.8). The pulse repetition frequency (10kHz) and contact patch movement speed (0.1-5m / s) are integrated to generate a dynamically updated topological network, enabling spatial correlation modeling of deformation characteristics.

[0213] 504. Performing spatiotemporal correlation matching on the impulse response topology and the spatial geometric mapping relationship of the rim centerline to generate a coupling parameter set of the dynamic contact spot deformation characteristics and the geometric mapping;

[0214] In step 504 , the spatial geometric mapping relationship of the rim centerline refers to a spatial position and angle association model between the rim centerline and the rail contact surface.

[0215] Spatiotemporal correlation matching refers to the process of aligning the impulse response topology with the geometric mapping relationship in time and space.

[0216] The coupled parameter set refers to a data set containing degradation and deformation characteristics generated by spatiotemporal correlation matching.

[0217] In this embodiment, a spatiotemporal registration algorithm (ICP Iterative Closest Point) is used to map the impulse response topological node coordinates to the wheel rim centerline coordinate system (with a registration error of ±0.2mm). The geometric mapping weight (0.3-0.7) is calculated using wheel-rail contact force data (sampling rate of 2kHz using a six-dimensional force sensor). The dynamic time warping (DTW) algorithm is used to align the topological node deformation rate (0.05-0.5mm / s) with the wheel rim geometric degradation gradient (daily average of 0.01-0.03mm). A parameter set is generated, including the spatiotemporal correlation strength, phase offset (±0.8mm), and coupling confidence (0.6-0.95). This allows for the establishment of a quantitative correlation model between contact patch deformation and wheel rim degradation.

[0218] 505. Analyze the pulse response amplitude fluctuation in the coupling parameter set and the spatial offset of the geometric mapping to generate deformation characteristics of the dynamic contact patch.

[0219] In step 505 , the impulse response amplitude fluctuation refers to the variation pattern of the reflected signal pulse intensity.

[0220] The spatial offset of the geometric mapping refers to the spatial position deviation between the center line of the wheel flange and the rail contact surface.

[0221] The deformation characteristics of the dynamic contact patch refer to the deformation laws and characteristics of the contact patch generated by the analytical coupling parameter set.

[0222] In the embodiment of this application, the key fluctuation modes in the coupling parameter set are extracted based on principal component analysis (cumulative variance contribution rate>85%), and the nonlinear relationship between amplitude fluctuation (variance threshold 0.5V²) and spatial offset (0.1-1.2mm) is analyzed using a random forest regression model. Stable oscillation modes are screened by Lyapunov exponent (threshold 0.25), and the curvature of the contact patch movement trajectory ( ) generates a deformation feature spectrum, and finally outputs a contact patch deformation feature set including deformation intensity classification (grades I-V), offset direction (8-direction partitions) and dynamic risk index, providing a quantitative decision-making basis for wheel-rail maintenance.

[0223] Here's a specific example:

[0224] In a high-speed railway tunnel, the system deploys a 24GHz millimeter-wave radar array (10m apart) longitudinally along the tunnel wall. Using beamforming technology, it synchronously collects reflected pulse trains (2GHz bandwidth) from the wheel-rail contact area. Time-frequency analysis is used to extract multipath echo delays (0.2ns resolution) and amplitude fluctuations (50dB dynamic range). This generates a raw signal grid (10mm×10mm density) containing the three-dimensional coordinates of the contact patch. An adaptive filtering algorithm (LMS filter order 64) is used to suppress tunnel wall reflection interference. Signal attenuation distortion (attenuation coefficient 0.1dB / m) is compensated in real time based on train speeds (250-350km / h). This generates an interference-resistant dynamic contact patch spatial coordinate grid (positioning error ±0.5mm). The impulse response envelope is extracted using a Hilbert transform, and a graph convolutional network (5×5 kernel size) is used to construct the impulse response topology. Nodes represent the deformation intensity of the contact patch (0-100%). Using a spatiotemporal registration algorithm (ICP iterative closest point), topological nodes were mapped to the wheel flange centerline coordinate system (with a registration error of ±0.3mm). This was combined with wheel-rail contact force data (sampling rate 1kHz) to generate a coupling parameter set. Principal component analysis was used to extract key fluctuation modes, and random forest regression was used to analyze amplitude fluctuations (with a variance threshold of 0.8V²) and spatial offsets (0.2-1.5mm). The final output was a contact patch deformation feature set, including deformation intensity classification (grades I-V) and a dynamic risk index, providing a quantitative basis for decision-making in wheel-rail maintenance in tunnels.

[0225] In summary, steps 501 to 505 implement millimeter-wave multimodal dynamic sensing and compensation control of wheel-rail contact patch deformation characteristics. Through multipath echo analysis and impulse response topology modeling using the millimeter-wave radar array, the system overcomes the spatiotemporal resolution limitations of traditional single-point detection. Based on multipath interference suppression and pulse sequence spatiotemporal correlation algorithms, the system couples the deformation characteristics of the dynamic contact patch with wheel rim geometry mapping, forming a deformation compensation model that is resistant to attenuation distortion. This enables submillimeter-level real-time analysis of the contact patch's spatial offset, overcoming the challenges of signal distortion and positioning drift caused by multipath effects, significantly improving the accuracy and stability of dynamic monitoring of high-speed wheel-rail contact conditions.

[0226] In some embodiments, as described in step 504, performing spatiotemporal correlation matching on the impulse response topology and the spatial geometric mapping relationship of the rim centerline to generate a coupling parameter set of the dynamic contact patch deformation characteristics and the geometric mapping includes:

[0227] 601. Calibrate the spatiotemporal coordinate origin of the impulse response topology and the geometric mapping space reference point of the wheel rim centerline to eliminate the coordinate system offset of the propagation path of the wheel-rail contact area deformation and the geometric degradation parameters, and generate a coordinate set of deformation and geometric association synchronized with the spatiotemporal reference;

[0228] In step 601, the impulse response topology structure refers to a deformation feature relationship network constructed based on impulse response amplitude and phase fluctuations.

[0229] The origin of the time and space coordinates refers to the time and space reference point of the impulse response topology.

[0230] The geometric mapping spatial reference point of the wheel rim centerline refers to the spatial position reference point between the wheel rim centerline and the rail contact surface.

[0231] The deformation propagation path refers to the diffusion trajectory of the deformation in the wheel-rail contact area in time and space.

[0232] The coordinate set of deformation and geometry association of spatiotemporal datum synchronization refers to the data set generated after aligning the deformation propagation path with the geometric degradation parameters.

[0233] In the embodiment of this application, the iterative closest point (ICP) algorithm is used to calibrate the spatiotemporal coordinate origin of the impulse response topology, and the geometric mapping reference point of the wheel rim centerline is calibrated by a laser tracker (accuracy ±0.02mm). The coordinate system offset of the deformation propagation path is compensated by combining the wheel-rail contact force distribution data (sampling rate 1kHz), and the spatiotemporal reference of the contact spot deformation and the wheel rim degradation parameter is aligned using the spatiotemporal registration technology (registration error <0.1mm). Based on the thermal expansion coefficient of the wheel-rail material ( × / ℃) to compensate for the influence of environmental temperature drift, and finally generate a spatiotemporal synchronized coordinate set including the deformation propagation rate (0.05-0.5mm / s) and geometric degradation gradient (daily average 0.01-0.03mm), providing a benchmark data source for dynamic correlation analysis.

[0234] 602. Track the diffusion rate of the deformation propagation path in the coordinate set, map the attenuation trend of the spatial geometric degradation gradient of the rim centerline as the track contact surface wears, and generate a dynamic correlation map of deformation diffusion and geometric degradation;

[0235] In step 602 , the deformation diffusion rate refers to the diffusion speed of the deformation in the wheel-rail contact area in time and space.

[0236] The spatial geometric degradation gradient refers to the rate of change of the geometric deformation of the rim centerline in space.

[0237] The dynamic correlation graph refers to the dynamic relationship network between deformation diffusion and geometric degradation.

[0238] In this embodiment, the optical flow method is used to track the diffusion rate of deformation propagation paths within a coordinate set (with a quantization accuracy of ±0.1 mm / s). Combined with time-series data of track contact surface wear depth (average monthly 0.1-0.4 mm), a degradation gradient attenuation curve is fitted using polynomial regression (with a fitting error of <2%). A graph convolutional network (GCN) is used to construct a dynamic correlation map between deformation diffusion and geometric degradation. Nodes represent the deformation intensity of contact patches (0-100%), and edge weights correlate the synchronization of degradation rates in adjacent regions (correlation coefficient >0.85). By integrating train speed (200-350 km / h) and wheel-rail clearance data (standard value 10-14 mm), a dynamic correlation map covering the entire track is generated, revealing the quantitative correlation between deformation propagation and geometric degradation.

[0239] 603. Quantify the phase offset angle between the deformation propagation path and the spatial geometric degradation gradient in the dynamic correlation map, integrate the instantaneous load intensity of the wheel-rail contact area to calculate the spatial coupling weight, and generate a coupling weight chain of deformation and geometry;

[0240] In step 603 , the phase offset angle refers to the phase difference between the deformation propagation path and the geometric degradation gradient.

[0241] The instantaneous load intensity refers to the load magnitude in the wheel-rail contact area at a certain moment.

[0242] The spatial coupling weight refers to the fusion weight of deformation and geometric features calculated based on the instantaneous load intensity.

[0243] The coupling weight chain refers to the sequence of deformation and geometric feature data containing spatial coupling weights.

[0244] In the embodiment of this application, a wavelet transform is used to extract the phase offset angle (resolution ±0.5°) between the deformation propagation path and the geometric degradation gradient in the dynamic correlation map, and the spatial coupling weight (0.3-0.7) is calculated based on the instantaneous load intensity (peak value 20-50kN). The weight allocation strategy is dynamically adjusted based on fuzzy logic control rules (48 rules), and combined with real-time data of the contact patch stress distribution (grid density 5mm×5mm), a coupling weight chain of deformation and geometry is generated (update frequency 50Hz). Kalman filtering is used to eliminate sensor noise (signal-to-noise ratio >30dB), ensuring the temporal continuity and spatial consistency of the weight chain, providing input conditions for multidimensional parameter field reconstruction.

[0245] 604. Reconstruct the dynamic correlation map in the coupling weight chain, superimpose the multidimensional influencing factors of the instantaneous load distribution on the wheel-rail contact surface, and generate a multidimensional coupling parameter field of deformation propagation and geometric degradation;

[0246] In step 604 , the multi-dimensional influencing factors refer to the multi-dimensional influencing factors of the instantaneous load distribution at the wheel-rail contact surface.

[0247] The multidimensional coupled parameter field refers to a data field that contains multidimensional correlation characteristics of deformation propagation and geometric degradation.

[0248] In this embodiment, principal component analysis (cumulative variance contribution >90%) is used to reconstruct the dynamic correlation map within the coupling weight chain. Multidimensional factors, such as the temperature gradient (0-50°C), humidity effects (20-95% RH), and vibration spectrum (5-200Hz), are superimposed on the instantaneous load distribution at the wheel-rail interface. Tensor decomposition (Tucker decomposition rank 3) is employed to compress parameter dimensions. A spatiotemporal interpolation algorithm (cubic splines) is then used to generate a multidimensional coupled parameter field (with a grid resolution of 1mm×1mm×0.1s) for deformation propagation and geometric degradation. Lyapunov exponents (threshold 0.25) are then used to screen for stable parameter regions, forming a global coupled field model that is resistant to environmental interference.

[0249] 605. Analyze the oscillation frequency of the deformation propagation path and the attenuation amplitude of the geometric degradation gradient in the multi-dimensional coupling parameter field to generate a coupling parameter set of the dynamic contact spot deformation characteristics and the geometric mapping.

[0250] In step 605 , the oscillation frequency refers to the periodic frequency of the deformation propagation path over time.

[0251] The attenuation amplitude refers to the spatial variation of the geometric degradation gradient.

[0252] The coupling parameter set refers to the deformation and geometric mapping feature data set generated by analyzing the multidimensional coupling parameter field.

[0253] In this embodiment of the present application, a Fourier transform is used to analyze the oscillation frequency (main frequency 0.5-3Hz) of the deformation propagation path in the multidimensional coupling parameter field, and a random forest regression model is used to quantify the attenuation amplitude of the geometric degradation gradient (range 0.1-1.2mm). Coherence analysis (threshold 0.8) is used to extract the phase synchronization region between deformation and degradation. Combined with real-time matching data of the wheel rim angle (35°-45°) and the contact patch curvature radius (200-300mm), a coupling parameter set is generated, including deformation intensity classification (grades I-V), offset direction (8-way partitioning), and risk index. Finally, closed-loop compensation of dynamic contact patch deformation characteristics and geometric mapping is achieved through adaptive PID control (response time 10ms), with the error stabilized within a threshold of ±0.2mm.

[0254] Here's a specific example:

[0255] In the wheel-rail dynamic monitoring scenario at the throat area of urban rail transit stations, the system deploys an 80GHz millimeter-wave radar array (5m apart) longitudinally along the turnout section. Combined with the transient 3D structure reconstruction technology of Hefei Zhongke Junda Vision, high-speed multi-view laser scanning generates a submillimeter 3D point cloud (0.05mm resolution) of the wheel-rail contact area. Simultaneously, a reflected pulse sequence (5GHz bandwidth) is collected to capture the micro-deformation of the turnout point-rail contact patch. An iterative closest point (ICP) algorithm is used to calibrate the impulse response topology to the spatial reference of the wheel flange centerline (with a registration error of ±0.1mm). This is then combined with 360° high-definition image data from a trackside panoramic intelligent inspection system to generate a time- and space-synchronized coordinate set containing the deformation propagation path (0.1-0.8mm / s) and wheel flange wear gradient (daily average of 0.02-0.05mm) in the turnout area. The deformation diffusion trajectory was tracked using the optical flow method, and lateral force data (peak value 15-30 kN) from trains passing through turnouts were superimposed. A dynamic deformation-wear correlation map was constructed (updated at 50 Hz), and wavelet packet decomposition was used to quantify the phase offset angle (accuracy ±0.3°). Multidimensional factors, such as the contact patch temperature gradient (ΔT 5-15°C) and the moisture permeability coefficient (0.1-0.5), were integrated using tensor decomposition to reconstruct the coupled parameter field of deformation propagation and geometric degradation (grid density 2 mm × 2 mm × 0.2 s). Finally, coherent spectrum analysis was used to extract the main oscillation frequency band (0.8-2.5 Hz) and the wear attenuation threshold (0.3-1.0 mm). This coupled parameter set, including the turnout contact patch safety level (AE level) and maintenance priority, was generated, providing a decision-making basis for optimizing the dynamic wheel-rail adaptability in the throat area of the station.

[0256] In summary, steps 601 to 605 achieve full-dimensional dynamic sensing and coordinated compensation of wheel-rail contact deformation and geometric degradation. Through synchronous calibration of spatiotemporal references and deformation propagation path tracking, a dynamic correlation model of deformation diffusion and geometric degradation is constructed. Based on phase offset quantization and multidimensional coupled parameter field reconstruction, the system overcomes the accuracy limitations of traditional single-dimensional compensation, enabling the coordinated analysis of the oscillation frequency and attenuation amplitude of the deformation propagation path and wheel flange degradation gradient. This enables high-precision dynamic matching of contact patch deformation characteristics with geometric mapping, effectively suppressing stress concentration and geometric degradation mismatch in the wheel-rail contact area, and improving wheel-rail dynamic adaptability and service reliability.

[0257] In some embodiments, in step 105, based on the offset correction amount and in combination with the spatial distribution of the deformation characteristics of the dynamic contact spot, driving the actuator to adjust the preliminary positioning position, and maintaining the positioning stability of the wheel-rail contact center during high-speed operation by real-time matching of the wheel rim geometric characteristics with the contact spot coordinates, includes:

[0258] 701. Generate dynamic baseline parameters corresponding to the offset correction, decompose the high-frequency oscillation component and the low-frequency attenuation trend in the spatial distribution of the dynamic contact patch deformation characteristics, construct a baseline offset field for wheel-rail contact center positioning, and transmit it to the actuator drive nodes of each track section;

[0259] In step 701 , the dynamic baseline parameter refers to a reference parameter for dynamic compensation generated according to the offset correction amount.

[0260] The high-frequency oscillation component refers to the rapidly changing periodic component in the deformation characteristics of the dynamic contact patch.

[0261] The low-frequency attenuation trend refers to the slowly changing trend component in the dynamic contact patch deformation characteristics.

[0262] The baseline offset field refers to the wheel-rail contact center positioning benchmark data field covering the entire track.

[0263] In this embodiment, a wavelet packet decomposition algorithm is used to separate the high-frequency oscillation component (frequency range 50-200Hz) and the low-frequency attenuation trend (frequency range 0-5Hz) in the dynamic contact patch deformation characteristics. A spatial interpolation algorithm is then used to construct a baseline offset field (grid accuracy 1mm×1mm) for wheel-rail contact center positioning. Based on train speed (200-350km / h) and dynamic wheel-rail clearance data (range 8-14mm), a baseline parameter field covering the entire track is generated. This parameter field is then distributed to the linear motor drive nodes in each track section using a timestamp synchronization protocol (such as IEEE 1588) (communication latency <1ms), providing a reference spatial distribution model for subsequent dynamic compensation.

[0264] 702. Mapping the phase difference between the wheel rim geometric features and the contact spot coordinates in the baseline offset field at the actuator drive node, compensating for the time delay effect in the spatial distribution law, and generating a wheel center coordinate offset adjustment drive signal;

[0265] In step 702 , the rim geometric features refer to the geometric shape and size parameters of the wheel rim.

[0266] The contact spot coordinates refer to the spatial position data of the wheel-rail contact area.

[0267] Phase difference refers to the phase deviation between the rim geometric features and the contact spot coordinates.

[0268] The time delay effect refers to the time delay phenomenon in the deformation propagation process of the wheel-rail contact area.

[0269] The wheel center coordinate offset adjustment drive signal refers to a compensation control signal used to adjust the wheel center position.

[0270] In this embodiment, a phase detection module (with an accuracy of ±0.5°) is deployed at the actuator drive node. A cross-correlation algorithm is used to calculate the phase difference between the wheel rim geometry and the contact patch coordinates (range: 0-30°). A Kalman filter is then used to predict the time delay compensation (with an accuracy of ±0.02ms) based on the contact patch deformation propagation velocity (0.1-0.8mm / s). Six-dimensional force sensor data (sampling rate: 2kHz) is integrated to generate a drive signal (with a resolution of 0.01mm) to adjust the wheel center coordinate offset. A PID control algorithm (with a proportional coefficient of Kp = 0.6) is then used to correct signal amplitude overshoot in real time, ensuring a strict match between the drive signal and the wheel-rail dynamic state.

[0271] 703. Analyze the instantaneous amplitude fluctuation of the high-frequency oscillation component in the wheel center coordinate offset adjustment drive signal, and assign dynamic weights to actuator drive nodes in combination with the propagation path of the contact patch deformation characteristics to generate multi-node dynamic weight drive parameters;

[0272] In step 703, the instantaneous amplitude fluctuation refers to the intensity change of the high-frequency oscillation component at a certain moment.

[0273] The propagation path refers to the diffusion trajectory of the contact patch deformation in time and space.

[0274] The dynamic weight refers to the actuator compensation weight assigned according to the deformation characteristics of the contact patch.

[0275] Multi-node dynamic weight driving parameters refer to a data set containing weight compensation parameters of multiple nodes.

[0276] In the embodiment of the present application, the Hilbert-Huang transform is used to analyze the instantaneous amplitude of the high-frequency oscillation component in the driving signal (fluctuation range 0.5-3V), combined with the curvature change rate of the contact spot deformation propagation path ( ), dynamically assigning execution node weights (0.3-0.7) via a fuzzy logic controller (64-rule base). The weight allocation strategy was optimized based on a contact patch stress distribution heat map (mesh density 5mm×5mm), generating a multi-node dynamic weight-driven parameter set that includes node priority (levels 1-5) and response speed (10-50ms), achieving adaptive optimization of compensation resource configuration.

[0277] 704. Convert the multi-node dynamic weight drive parameters into pulse width modulation signals of the actuator, match the phase synchronization of the wheel rim geometric features and the contact spot coordinates, and generate a dynamic control instruction set for the wheel-rail contact center;

[0278] In step 704 , the pulse width modulation signal refers to a signal for controlling an actuator by adjusting the pulse width.

[0279] Phase synchronization refers to the phase consistency between the rim geometric characteristics and the contact spot coordinates.

[0280] The dynamic control instruction set refers to the control instruction set used to dynamically adjust the wheel-rail contact center.

[0281] In this embodiment, a pulse width modulator (20kHz carrier frequency) converts the dynamic weight parameters into a PWM signal with an adjustable duty cycle (±0.5%). Phase-locked loop technology is used to synchronize the wheel rim geometry with the contact patch coordinates (synchronization error <0.1°). The Lyapunov stability criterion is used to constrain the PWM signal amplitude (threshold ±5V), generating a dynamic control instruction set containing lateral compensation (±2mm) and vertical compensation (±1mm). This instruction set is then distributed to the actuator via the EtherCAT bus (125μs cycle), ensuring strict synchronization of the instruction timing with the wheel-rail contact state.

[0282] 705. Execute the dynamic control instruction set, adjust the wheel center position and maintain the positioning stability of the wheel-rail contact center through the real-time closed-loop feedback of the pressure gradient and contact spot coordinates of the wheel rim contact area by the actuator.

[0283] In step 705 , the pressure gradient refers to the spatial rate of change of pressure in the rim contact area.

[0284] Real-time closed-loop feedback refers to the control process of monitoring and adjusting the wheel center position in real time through sensors.

[0285] Positioning stability refers to the ability of the wheel-rail contact center to maintain position accuracy during dynamic adjustment.

[0286] In this embodiment, the actuator uses a voice coil motor array (thrust 120N, repeatability ±0.01mm) to receive dynamic control commands. A pressure sensor array (range 0-50MPa) provides real-time feedback on the pressure gradient in the wheel rim contact area (gradient resolution ±0.05MPa / mm). This feedback is combined with laser tracking data (sampling rate 1kHz) of the contact spot coordinates to form a closed-loop control circuit. Model predictive control (prediction step size 10ms) is used to dynamically adjust the wheel center position, stabilizing the contact center offset within a ±0.15mm range. Ultimately, a stress equalization algorithm (equalization error <5%) is used to maintain the dynamic stability of the wheel-rail contact.

[0287] Here's a specific example:

[0288] In the context of dynamic wheel-rail adaptive control on curved sections of high-speed railways, the system uses a contact spot laser Doppler vibrometer (LDV) to collect submicron-level vibration spectra (frequency range 0.5-800 Hz) of the wheel-rail contact surface in real time. This is combined with a wheelset 3D topography scanner (accuracy ±5 μm) to construct a wheel flange geometric feature database. An improved HHT transformation is used to separate the high-frequency vibration component (main frequency 50-200 Hz) of the contact spot deformation from the low-frequency creep trend (period 0.2-2 seconds). The Kriging spatial interpolation algorithm is then used to integrate dynamic inspection vehicle data (sampling interval 10 cm) to generate a dynamic baseline parameter field, including a lateral offset baseline (±0.3-1.2 mm) and a longitudinal creep gradient (0.05-0.15 mm / m) for the curved section. A millimeter-wave radar phase interferometer array (resolution 0.1°) is deployed at the actuator nodes. An improved cross-correlation algorithm is used to match the spatiotemporal differences between the wheel flange profile phase and the contact patch coordinates in real time (delay compensation ±0.8ms). This is combined with six-dimensional wheel-rail contact force sensor data (range ±50kN) to generate an adaptive PID compensation signal (adjustment cycle 5ms). Based on the propagation vector field of the contact patch stress cloud (density 1mm×1mm), a fuzzy neural network is used to dynamically assign compensation weights (0.4-0.9) to 16 actuator groups. Carrier reconstruction technology is used to convert these weight parameters into pulse-width modulation commands (carrier frequency 25kHz). Finally, contact patch stress equalization control is implemented using a distributed piezoelectric actuator array (response time 0.2ms). Combined with real-time pressure gradient feedback from a fiber Bragg grating sensor (accuracy ±0.01MPa / mm), the wheel-rail contact center offset in curved sections is stabilized within ±0.1mm, effectively suppressing wheel flange wear and snaking instability.

[0289] In summary, steps 701 to 705 achieve global dynamic closed-loop control and adaptive compensation of the wheel-rail contact center. By constructing a dynamic baseline parameter field and separating high-frequency oscillation components, the system overcomes the phase lag bottleneck of traditional static compensation. Based on phase difference mapping between wheel rim geometry and contact spot coordinates, combined with a multi-node dynamic weight allocation mechanism, the system develops collaborative control capabilities that decouple oscillation suppression and attenuation trends, enabling millisecond-level dynamic adjustment of the wheel center position. By synchronously matching pressure gradient feedback with the phase of the pulse modulation signal, the system eliminates uneven stress distribution and dynamic offset accumulation in the wheel-rail contact area, ultimately achieving cross-scale collaborative optimization of contact center positioning accuracy and stability.

[0290] To overcome the difficulty of accurately detecting the wheel geometric parameters of mobile vehicles such as trains and cars in real time under dynamic conditions, and to address the low efficiency of traditional manual measurement and the susceptibility of contact detection to environmental interference, this technology integrates encoder displacement signals with edge-triggered logic to construct a closed-loop detection system for wheel motion trajectory and geometric dimensions. This system aims to achieve non-contact dynamic measurement of wheel chord length and adaptive compensation for abnormal conditions, providing reliable data support for wheelset rotation and safety warnings, reducing operation and maintenance costs and improving vehicle operation safety.

[0291] In some embodiments, in step 102, determining the chord length of the wheel according to the detection process of the wheel by the first detection unit includes:

[0292] 801. When the first detection unit detects that the wheel enters the detection area, start the encoder to record a first displacement signal of the wheel movement;

[0293] In step 801 , the first detection unit refers to an infrared or photoelectric sensor for detecting a wheel entering a detection area.

[0294] An encoder is a measuring device that converts the linear displacement of a wheel into a pulse signal.

[0295] The first displacement signal refers to the displacement pulse signal recorded by the encoder when the wheel enters the detection area.

[0296] In an embodiment of the present application, a photoelectric sensor or infrared array detects when a wheel enters the detection area, triggering a rotary encoder to start pulse counting. The encoder uses an incremental design, converting the wheel's linear displacement into an angular displacement pulse signal via a gear rack or magnetic roller. The first detection unit uses an infrared beam sensor (such as a combination of SE303A and PH302) to monitor the wheel rim obstruction status in real time. When the wheel edge first blocks the light path, a trigger signal is sent to the encoder. The encoder's built-in high-speed counter (such as the STM32 TIM module) begins counting pulses with a microsecond response. The initial pulse value is captured via the input capture function of a PLC or microcontroller (such as an S7-200 SMART). Finally, through photoelectric signal synchronization and de-jitter algorithms, the initial displacement value at the moment of wheel entry is determined with an accuracy of ±0.1mm.

[0297] 802. When the first detection unit fails to detect the wheel, stop recording by the encoder and read a second displacement signal recorded by the encoder;

[0298] In step 802 , the second displacement signal refers to a displacement pulse signal recorded by the encoder when the wheel completely passes through the detection area.

[0299] In this embodiment of the present application, when the wheel completely passes through the detection area, the infrared sensor optical path is restored, triggering the encoder to stop counting. The encoder then transmits the accumulated pulse count to a controller (such as an SST89E564RD microcontroller) via the SPI or EtherCAT communication protocol. The system uses anti-interference techniques (such as digital filtering) to eliminate false triggers caused by mechanical vibration and ensures displacement signal continuity through timestamp calibration (accuracy of ±0.5ms). The final displacement value is calculated using a pulse-to-displacement conversion coefficient (e.g., 0.02mm per pulse), combined with a wheel diameter dynamic compensation algorithm (based on laser ranging data) to correct for slip errors and generate a second displacement signal dataset.

[0300] 803. Determine, based on the first displacement signal recorded by the encoder, an initial displacement value when the wheel edge begins to enter the detection area;

[0301] In step 803 , the initial displacement value refers to a displacement reference value when the wheel edge begins to enter the detection area.

[0302] In the embodiments of this application, the initial displacement value is calculated based on the relationship between the encoder pulse count and the mechanical transmission ratio. For example, if a rack-and-pinion encoder generates 10mm of linear displacement per revolution, the initial displacement is calculated by dividing the initial pulse count by the pulse count per revolution (e.g., 2000 PPR). The system removes signal noise through wavelet transforms and employs edge detection algorithms (e.g., the Canny operator) to accurately calibrate the moment of wheel entry. Combined with 3D wheelset scan data (e.g., a tread profile library), the system dynamically corrects for the effects of wheel rim geometry deviation on the initial position, ultimately outputting an initial displacement value with submillimeter accuracy.

[0303] 804. Determine, based on the second displacement signal recorded by the encoder, a final displacement value when the wheel edge completely passes through the first detection unit;

[0304] In step 804 , the final displacement value refers to the final displacement value when the wheel edge completely passes through the detection area.

[0305] In this embodiment, the determination of the final displacement value requires compensation for encoder return error and temperature drift. The system uses dual-channel calibration technology (A / B phase pulse phase difference analysis) to detect the encoder's rotation direction and prevent reverse counting. Kalman filtering fuses multi-sensor data (such as ultrasonic ranging) to calibrate the pulse accumulation value in real time. For example, when the wheel has completely passed, the pulse count recorded by the encoder is converted into displacement using a linear interpolation algorithm, and the geometric correction value of the laser sensor (such as the L2+70mm reference point) is superimposed to generate a final displacement value with an accuracy of ±0.05mm.

[0306] 805. Determine a chord length of the wheel according to the final displacement value and the initial displacement value.

[0307] In step 805 , the chord length refers to the straight-line distance of the wheel tread calculated based on the final displacement value and the initial displacement value.

[0308] In this embodiment, chord length calculation dynamically compensates for wheel-rail contact deformation by using displacement differences. The system subtracts the final displacement from the initial displacement value and, combined with the geometric parameters of the wheel's rolling circle reference point (e.g., at L2+70mm on the tread), uses the least-squares method to fit the true chord length. For curved conditions, a curvature radius compensation factor (e.g., 0.2% for R > 300m) is introduced. Parallel computation is implemented using an FPGA to ensure real-time performance at a 1kHz update rate. The final output is validated using a stress equalization algorithm, stabilizing the chord length measurement error to within ±0.3mm, meeting the wheelset rotation accuracy requirements.

[0309] Here's a specific example:

[0310] In the scenario of measuring the geometric parameters of rail vehicle wheels, a subway depot uses a chord length measurement system based on an incremental encoder to monitor wheelset wear. In specific implementation, when a wheel (840mm in diameter) enters the detection area formed by two laser beam sensors, the 2000-line incremental encoder is triggered to start recording the displacement pulse signal. The system collects the initial displacement value of the wheel edge contacting the first laser beam in real time with a detection accuracy of 0.4mm. =125.6mm. When the wheel completely passes the second laser sensor with a spacing of 500mm, the encoder automatically stops recording and reads the final displacement value. =681.2mm. The displacement difference ΔX=555.6mm is obtained through differential operation. Combined with the fixed spacing L=500mm of the dual laser sensors, the actual chord length is calculated to be 832.7mm using the geometric relationship formula chord length C=√(4ΔX²-L²). During the actual measurement at the Shenzhen Metro Line 6 vehicle base, the system found that the chord length of a certain train wheel was shortened by 3.5mm (exceeding the 2mm maintenance threshold) due to long-term operation by comparing the benchmark data of the 3D laser scanner. This triggered an automatic alarm and generated a wheel rim repair work order. This solution uses an anti-slip frame with a counterweight of 100kg to ensure constant contact pressure between the encoder wheel and the rail surface. Combined with weekly chord length benchmark calibration (using a standard diameter 845mm calibration wheel), the measurement error is controlled within the design index of ±0.8mm, effectively solving the problem of cumulative error caused by wheel slippage in traditional wheel diameter measurement.

[0311] In summary, steps 801 to 805 achieve full automation and dynamic error compensation for wheel chord length measurement. Through the timing control and edge-triggered mechanism of the encoder displacement signal, the system overcomes the technical bottlenecks of positioning ambiguity and response lag in traditional manual measurement. The system calculates chord length based on the displacement difference between the wheel entering and exiting the detection area. Combined with the encoder signal's self-calibration algorithm, it eliminates accumulated errors caused by mechanical vibration and speed fluctuations, enabling submillimeter-level, real-time dynamic measurement of wheel geometry parameters. Precise calibration and compensation of the displacement signal's start and end points simultaneously address measurement deviations caused by wheel tilt or deflection, providing a stable data source for wheel diameter calculation and tread wear analysis.

[0312] To overcome the problem of positioning misalignment caused by wheel geometry deviation and dynamic stress fluctuations in wheel-rail contact center detection, and to address the technical bottleneck of traditional methods that rely on manual calibration and cannot adapt to complex working conditions, this technology integrates half-chord length calculation and servo drive control to construct an adaptive dynamic positioning system for the wheel-rail contact center. This system aims to achieve real-time and accurate detection and dynamic compensation of the wheel center position, providing an efficient and reliable solution for wheel-rail maintenance and safe operation.

[0313] In some embodiments, in step 103, when the second detection unit detects the wheel, calculating half of the chord length and controlling the second detection unit to move half of the chord length in the same direction includes:

[0314] 901. When the second detection unit detects the wheel, a microprocessor is used to calculate half of the chord length, and a servo motor is used to drive the second detection unit to move half of the chord length in the same direction.

[0315] In step 901 , the second detection unit refers to an infrared or photoelectric sensor for detecting whether a wheel enters a target area.

[0316] A microprocessor refers to an embedded computing chip used to perform half-chord length calculations.

[0317] Half the chord length is half the straight-line distance of the wheel tread.

[0318] The servo motor refers to a closed-loop control motor used to drive the second detection unit to move precisely.

[0319] In an embodiment of the present application, a photoelectric sensor or laser array is used to detect in real time when the wheel enters the second detection area, triggering a microprocessor (such as an ARM Cortex-M4) to initiate the calculation of the half-chord length. The microprocessor loads pre-stored chord length data (such as 600mm), uses a floating-point unit (FPU) to quickly calculate the half-value (300mm), and transmits the calculation result to a servo motor controller (such as the Delta ASDA-A2 series) via the SPI or CAN bus. The servo motor uses closed-loop feedback control (encoder resolution 10000PPR) combined with a PID adjustment algorithm (proportional coefficient Kp=0.8) to dynamically adjust the movement speed and acceleration to ensure that the second detection unit moves accurately to the target position (error ±0.1mm). Through real-time position monitoring and dynamic compensation mechanisms, the system eliminates displacement deviations caused by mechanical vibration and track unevenness, ultimately achieving precise positioning of the wheel center position.

[0320] Here's a specific example:

[0321] In a dynamic inspection scenario for heavy-duty truck wheelsets in mining, the system deploys a laser array along the truck's track as a secondary detection unit. When a wheel enters the inspection area, the STM32F407 microprocessor triggers the calculation of the half-chord length. The microprocessor loads pre-stored chord length data (e.g., 850mm) and rapidly calculates the half-chord length (425mm) using the floating-point unit. The result is transmitted to a Yaskawa servo motor controller (SGD7S series) via the CAN bus. The servo motor utilizes a 20-bit high-resolution encoder (accuracy ±0.01mm) and a PID closed-loop control algorithm (proportional coefficient Kp = 0.7) to precisely move the secondary detection unit along the track the half-chord length (425mm). The system monitors its position in real time using a laser ranging sensor (accuracy ±0.05mm) and uses a Kalman filter to eliminate displacement errors caused by track vibration and mechanical backlash (compensation ±0.2mm). Ultimately, the secondary detection unit locates the wheel center within ±0.1mm, providing high-precision benchmark data for wheelset repair and safety monitoring in mining trucks, significantly improving dynamic wheel-rail compatibility and operational stability.

[0322] In summary, step 901 implements automated dynamic positioning and compensation for wheel-rail contact center detection. By using a microprocessor to calculate the half-chord length in real time and control the precise movement of the servo motor, this system overcomes the response lag and positioning deviation limitations of traditional static detection. The system dynamically adjusts the position of the second detection unit based on the half-chord length. Combined with closed-loop feedback control of the servo motor, this eliminates uneven stress distribution and geometric offset in the wheel-rail contact area, achieving submillimeter precision positioning of the wheel center. The coordinated control of half-chord length compensation and the servo drive significantly improves the stability and reliability of wheel-rail contact center detection, providing high-precision data support for dynamic wheel-rail adaptation.

[0323] Figure 2 The present invention provides a structural diagram of a wheel center positioning system for a rail vehicle, as shown in FIG. Figure 2 As shown, the system includes:

[0324] An installation module 21 is used to pre-install the first detection unit and the second detection unit on both sides of the wheel;

[0325] a measuring module 22, configured to determine the chord length of the wheel according to the detection process of the wheel by the first detection unit;

[0326] a calculation module 23, configured to calculate half the chord length when the second detection unit detects the wheel, and control the second detection unit to move a distance half the chord length in the same direction;

[0327] a positioning module 24 for determining a current position of the second detection unit as a preliminary positioning position of the wheel center when controlling the second detection unit to move in the same direction by half the chord length;

[0328] The correction module 25 is used to generate an offset correction value for the initial positioning position of the wheel center based on the collaborative learning framework of the dynamic contact spot deformation characteristics and the rim geometry degradation, and drive the actuator to dynamically correct the initial positioning position according to the offset correction value to maintain the positioning stability of the wheel-rail contact center.

[0329] Figure 2 The wheel center positioning system of a rail vehicle can be implemented Figure 1 The implementation principle and technical effects of the wheel center positioning method for a rail vehicle described in the illustrated embodiment will not be elaborated on here. The specific manner in which the various modules and units of the wheel center positioning system for a rail vehicle in the above embodiment perform their operations has been described in detail in the embodiments of the method and will not be elaborated on here.

[0330] 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. A method for locating the wheel center of a rail vehicle, characterized in that: include: Pre-installing a first detection unit and a second detection unit on both sides of the wheel; determining a chord length of the wheel according to a detection process of the wheel by the first detection unit; When the second detection unit detects the wheel, calculating half of the chord length, and controlling the second detection unit to move half of the chord length in the same direction; When the second detection unit is controlled to move in the same direction by half the chord length, a current position of the second detection unit is determined as a preliminary positioning position of the wheel center; Based on a collaborative learning framework of dynamic contact patch deformation characteristics and wheel rim geometry degradation, an offset correction value for the initial positioning position of the wheel center is generated, and an actuator is driven to dynamically correct the initial positioning position based on the offset correction value to maintain the positioning stability of the wheel-rail contact center; The collaborative learning framework based on the dynamic contact patch deformation characteristics and the wheel rim geometry degradation generates the offset correction value of the preliminary positioning position of the wheel center, including: The wheel surface geometry at different degrees of wear is acquired through 3D laser scanning, and a static parameter set including the wheel flange thickness and tread wear depth is generated. Based on the wheel flange thickness and tread wear depth, a spatial geometric mapping relationship between the wheel flange centerline and the track contact surface is established; A millimeter-wave radar network is distributed along the longitudinal direction of the track to capture the spatial coordinates of the dynamic contact patch in the wheel-rail contact area in real time. The deformation characteristics of the dynamic contact patch are analyzed by combining the spatial geometric mapping relationship between the centerline of the wheel flange and the track contact surface. Inputting the wheel rim geometry degradation characteristics and the dynamic contact patch deformation characteristics in the static parameter set into a collaborative learning framework, and generating an offset correction value for the preliminary positioning position of the wheel center through an interactive method of local model parameter encryption iteration and global parameter aggregation; Driving the actuator to dynamically correct the preliminary positioning position according to the offset correction amount to maintain the positioning stability of the wheel-rail contact center includes: Based on the offset correction amount and in combination with the spatial distribution law of the deformation characteristics of the dynamic contact spot, the actuator is driven to adjust the preliminary positioning position, and the positioning stability of the wheel-rail contact center during high-speed operation is maintained through real-time matching of the wheel rim geometric characteristics with the contact spot coordinates.

2. The method according to claim 1, characterized in that Inputting the rim geometry degradation features and the dynamic contact patch deformation features in the static parameter set into a collaborative learning framework, and generating an offset correction value for the preliminary positioning position of the wheel center through an interactive method of local model parameter encryption iteration and global parameter aggregation, includes: Decomposing a geometric deformation gradient tensor of the wheel rim geometric degradation feature, extracting a degradation feature vector related to the wheel-rail contact surface curvature in the geometric deformation gradient tensor, and generating a feature decomposition sequence including a curvature degradation gradient; Performing a noise masking mechanism on the eigendecomposition sequence at a local computing node, superimposing random mask noise through a dynamic attenuation window, and generating an interference-resistant degenerate gradient encryption chain; A joint parameter space is constructed by combining the degradation gradient encryption chain with the stress distribution time series signal in the dynamic contact patch deformation characteristics, and a global parameter topology chain coupled with degradation and deformation is generated through cross-node parameter interpolation operation in the collaborative learning framework; Dynamically iteratively encrypting the degenerate gradient in the global parameter topology chain at the local computing node, fusing the stress distribution of the wheel-rail contact area to generate an encrypted weight gradient chain and distributing it to the global aggregation node; The encrypted gradient chains of multiple nodes are aggregated in a global aggregation node, the random mask noise is eliminated, and the degradation and deformation coupling phase offsets across the nodes are extracted to generate an offset correction for the preliminary positioning position of the wheel center.

3. The method according to claim 2, characterized in that The method constructs a joint parameter space by combining the degradation gradient encryption chain with the stress distribution time series signal in the dynamic contact patch deformation feature, and generates a global parameter topology chain coupled with degradation and deformation through cross-node parameter interpolation operation in the collaborative learning framework, including: Aligning the sampling timestamps of the time-space distribution of the degradation gradient encryption chain with the stress distribution time series signal in the dynamic contact spot deformation feature, matching the geometric deformation phase of the wheel-rail contact area with the dynamic stress fluctuation period, and generating a time-space aligned degradation and stress parameter matrix; Performing parameter interpolation operations between multiple nodes on the parameter matrix to compensate for the difference in deformation gradients of dynamic contact patches between adjacent track sections, and generating an interpolation compensation parameter chain that is continuously distributed across the nodes; Fusing the degraded gradient and stress distribution amplitude in the interpolation compensation parameter chain, adjusting the fusion weight of the gradient and stress based on the instantaneous deformation rate, and generating a dynamic weight fusion parameter set; Performing spatiotemporal convolution superposition of degradation gradient and stress distribution on the dynamic weight fusion parameter set, extracting parameter aggregation features of the wheel-rail contact center area, and generating a characteristic topological network of degradation and deformation coupling; The cooperative oscillation mode of the degradation gradient and stress distribution in the characteristic topological network is analyzed to generate a global parameter topological chain of degradation and deformation coupling.

4. The method according to claim 1, wherein The millimeter-wave radar network distributed along the longitudinal direction of the track captures the spatial coordinates of the dynamic contact patch in the wheel-rail contact area in real time, and analyzes the deformation characteristics of the dynamic contact patch based on the spatial geometric mapping relationship between the center line of the wheel flange and the track contact surface, including: A millimeter-wave radar array is deployed longitudinally along the track to synchronously collect reflected signal pulse sequences from the wheel-rail contact area. This captures the multipath echo delay and amplitude fluctuations associated with track contact surface deformation in the reflected signals, generating a raw signal grid of the spatial coordinates of the dynamic contact patch. Performing multipath interference suppression processing on the original signal grid to compensate for attenuation distortion of the reflected signal during high-speed operation and generate an interference-resistant dynamic contact patch spatial coordinate grid; Extracting the impulse response amplitude and time-series phase fluctuation of the dynamic contact spot spatial coordinate grid to construct an impulse response topology of the dynamic contact spot deformation characteristics; Performing spatiotemporal correlation matching between the impulse response topology and the spatial geometric mapping relationship of the rim centerline to generate a coupling parameter set of the dynamic contact patch deformation characteristics and the geometric mapping; The pulse response amplitude fluctuations in the coupled parameter set and the spatial offset of the geometric mapping are analyzed to generate the deformation characteristics of the dynamic contact patch.

5. The method according to claim 4, characterized in that The step of performing spatiotemporal correlation matching between the impulse response topology and the spatial geometric mapping relationship of the rim centerline to generate a coupling parameter set of the dynamic contact spot deformation characteristics and the geometric mapping includes: Calibrate the spatiotemporal coordinate origin of the impulse response topology and the geometric mapping space reference point of the wheel rim centerline to eliminate the coordinate system offset of the propagation path of the wheel-rail contact area deformation and the geometric degradation parameters, and generate a coordinate set of deformation and geometric association synchronized with the spatiotemporal reference; Tracking the diffusion rate of the deformation propagation path in the coordinate set, mapping the attenuation trend of the spatial geometric degradation gradient of the rim centerline as the track contact surface wears, and generating a dynamic correlation map between deformation diffusion and geometric degradation; quantifying the phase offset angle between the deformation propagation path and the spatial geometric degradation gradient in the dynamic correlation map, integrating the instantaneous load intensity of the wheel-rail contact area to calculate the spatial coupling weight, and generating a coupling weight chain of deformation and geometry; Reconstructing the dynamic correlation map in the coupling weight chain, superimposing the multi-dimensional influencing factors of the instantaneous load distribution on the wheel-rail contact surface, and generating a multi-dimensional coupling parameter field of deformation propagation and geometric degradation; The oscillation frequency of the deformation propagation path and the attenuation amplitude of the geometric degradation gradient in the multi-dimensional coupling parameter field are analyzed to generate a coupling parameter set of the dynamic contact spot deformation characteristics and geometric mapping.

6. The method according to claim 1, wherein The method of driving the actuator to adjust the preliminary positioning position based on the offset correction amount and in combination with the spatial distribution law of the deformation characteristics of the dynamic contact spot, and maintaining the positioning stability of the wheel-rail contact center during high-speed operation by real-time matching of the wheel rim geometric characteristics with the contact spot coordinates, includes: Generate dynamic baseline parameters corresponding to the offset correction, decompose the high-frequency oscillation component and low-frequency attenuation trend in the spatial distribution of the dynamic contact patch deformation characteristics, construct a baseline offset field for wheel-rail contact center positioning and send it to the actuator drive nodes of each track section; Mapping the phase difference between the wheel rim geometric features and the contact spot coordinates in the baseline offset field at the actuator drive node, compensating for the time delay effect in the spatial distribution law, and generating a wheel center coordinate offset adjustment drive signal; Analyzing the instantaneous amplitude fluctuation of the high-frequency oscillation component in the wheel center coordinate offset adjustment drive signal, and combining the propagation path of the contact patch deformation characteristics, assigning dynamic weights to the actuator drive nodes to generate multi-node dynamic weight drive parameters; Converting the multi-node dynamic weighted drive parameters into pulse width modulation signals for the actuators, matching the phase synchronization of the wheel rim geometric features with the contact spot coordinates, and generating a dynamic control instruction set for the wheel-rail contact center; The dynamic control instruction set is executed, and the wheel center position is adjusted and the positioning stability of the wheel-rail contact center is maintained through real-time closed-loop feedback of the pressure gradient and contact spot coordinates of the wheel rim contact area by the actuator.

7. The method according to claim 1, characterized in that The determining the chord length of the wheel according to the detection process of the wheel by the first detection unit includes: When the first detection unit detects that the wheel enters the detection area, the encoder is activated to record a first displacement signal of the wheel movement; When the first detection unit fails to detect the wheel, stopping the encoder from recording and reading the second displacement signal recorded by the encoder; determining, based on the first displacement signal recorded by the encoder, an initial displacement value when the wheel edge begins to enter the detection area; determining, based on the second displacement signal recorded by the encoder, a final displacement value when the wheel edge completely passes through the first detection unit; A chord length of the wheel is determined according to the final displacement value and the initial displacement value.

8. The method according to claim 1, characterized in that When the second detection unit detects the wheel, calculating half of the chord length and controlling the second detection unit to move half of the chord length in the same direction comprises: When the second detection unit detects the wheel, a microprocessor is used to calculate half of the chord length, and a servo motor is used to drive the second detection unit to move half of the chord length in the same direction.

9. A wheel center positioning system for a rail vehicle, characterized in that: include: An installation module is used to pre-install the first detection unit and the second detection unit on both sides of the wheel; a measuring module, configured to determine the chord length of the wheel according to a detection process of the wheel by the first detection unit; a calculation module, configured to calculate half the chord length when the second detection unit detects the wheel, and control the second detection unit to move a distance half the chord length in the same direction; a positioning module, configured to determine a current position of the second detection unit as a preliminary positioning position of the wheel center when controlling the second detection unit to move in the same direction by half the chord length; a correction module for generating an offset correction for the initial positioning position of the wheel center based on a collaborative learning framework of dynamic contact patch deformation characteristics and wheel rim geometry degradation, and for driving an actuator to dynamically correct the initial positioning position according to the offset correction to maintain the positioning stability of the wheel-rail contact center; The collaborative learning framework based on the dynamic contact patch deformation characteristics and the wheel rim geometry degradation generates the offset correction value of the preliminary positioning position of the wheel center, including: The wheel surface geometry at different degrees of wear is acquired through 3D laser scanning, and a static parameter set including the wheel flange thickness and tread wear depth is generated. Based on the wheel flange thickness and tread wear depth, a spatial geometric mapping relationship between the wheel flange centerline and the track contact surface is established; A millimeter-wave radar network is distributed along the longitudinal direction of the track to capture the spatial coordinates of the dynamic contact patch in the wheel-rail contact area in real time. The deformation characteristics of the dynamic contact patch are analyzed by combining the spatial geometric mapping relationship between the centerline of the wheel flange and the track contact surface. Inputting the wheel rim geometry degradation characteristics and the dynamic contact patch deformation characteristics in the static parameter set into a collaborative learning framework, and generating an offset correction value for the preliminary positioning position of the wheel center through an interactive method of local model parameter encryption iteration and global parameter aggregation; Driving the actuator to dynamically correct the preliminary positioning position according to the offset correction amount to maintain the positioning stability of the wheel-rail contact center includes: Based on the offset correction amount and in combination with the spatial distribution law of the deformation characteristics of the dynamic contact spot, the actuator is driven to adjust the preliminary positioning position, and the positioning stability of the wheel-rail contact center during high-speed operation is maintained through real-time matching of the wheel rim geometric characteristics with the contact spot coordinates.

Images