Ultrasonic wear detection device for bullet train wheel pair

Abstract

The invention discloses a bullet train wheel set ultrasonic wear detection device, and particularly relates to the technical field of rail transit detection, and the bullet train wheel set ultrasonic wear detection device comprises an ultrasonic probe module which comprises a multi-frequency composite probe array and a self-adaptive coupling structure and is used for carrying out all-directional scanning on a wheel flange, a tread and a rim of a wheel set; and the signal processing module is connected with the probe module and comprises a noise reduction unit, a feature extraction unit and a real-time early warning unit. According to the invention, the multi-frequency composite probe array and matrix layout are adopted, deep crack and surface wear detection is considered, the omission ratio is reduced, and the efficiency is improved by more than 30%; a liquid immersion type structure and a three-degree-of-freedom mechanism guarantee stable detection and adapt to more than 99% of vehicle models; accurate recognition is realized through a WT-CNN algorithm and the like, and the false alarm rate is reduced by 40%; the modular design adapts to multiple scenes, and the fault detection rate reaches 99.5% through full-dimensional diagnosis; the equipment adapts to an extreme environment, the service life is prolonged, the single-wheel detection time is finally shortened, the labor cost is reduced, and the overhaul efficiency and the rail traffic safety are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of rail transit detection, and in particular to an ultrasonic wear detection device for moving vehicle wheel pairs. Background Art

[0002] As the core load-bearing components of high-speed trains, EMU wheels are like the train's "steel feet." Their technical condition is directly related to train safety and the safety of passengers' lives and property. During continuous operation, wheelsets are subjected to complex dynamic loads, friction, wear, and environmental corrosion. Even the slightest wear or defect can cause a serious accident.

[0003] Although existing ultrasonic testing technology can detect internal cracks, it faces the following technical bottlenecks: Poor coupling stability: The wheelset surface has complex curvature, and traditional contact probes are prone to signal distortion due to uneven distribution of coupling agent.

[0004] Severe noise interference: noise such as grain boundary scattering of wheelset materials and environmental vibrations mask the real defect signals.

[0005] Insufficient intelligence: Detection results rely on manual interpretation and lack automated defect classification and early warning capabilities.

[0006] Lack of adaptability: The equipment is difficult to operate stably under the complex working conditions (high humidity and high dust) of the EMU maintenance depot. Summary of the Invention

[0007] The object of the present invention is to provide an ultrasonic wear detection device for a moving vehicle wheel set to solve the above-mentioned deficiencies in the technology.

[0008] To achieve the above-mentioned object, the present invention provides the following technical solution: an ultrasonic wear detection device for a motor vehicle wheel set, comprising: an ultrasonic probe module, including a multi-frequency composite probe array and an adaptive coupling structure, for performing an omnidirectional scanning of the wheel flange, tread, and rim of the wheelset; The signal processing module is connected to the probe module and includes a noise reduction unit, a feature extraction unit, and a real-time warning unit. It is used to intelligently analyze the echo signal and output defect detection results. Mechanical scanning module, integrating automatic positioning mechanism and protective shell, is used to drive the probe module to move along multiple axes of the wheelset surface and adapt to complex working conditions; The data management module includes a cloud database and a visual interactive interface, which is used to store inspection data, generate three-dimensional wear maps and maintenance recommendation reports.

[0009] Preferably, the ultrasonic probe module includes: Multi-frequency composite probe array, consisting of low-frequency probes (0.5-2MHz) and high-frequency probes (5-10MHz) arranged alternately. The low-frequency probes are used to detect internal cracks in the rim, while the high-frequency probes are used to measure the wheel flange thickness and tread wear depth. The probe array is staggered along the axial and circumferential directions of the wheelset in a matrix layout, covering the area where the curvature of the wheelset surface changes. A single scan acquires continuous data of the vertical section of the wheel rim and the transverse section of the tread.

[0010] Preferably, the adaptive coupling structure is a liquid immersion coupling device, comprising: Flexible sealing cavity filled with coupling agent and conforming to the surface contour of the wheelset; The pressure feedback unit monitors the contact pressure between the cavity and the wheelset in real time, and dynamically adjusts the cavity volume through a micro air pump to maintain a constant couplant film thickness.

[0011] Preferably, the signal processing module includes: The wavelet transform-convolutional neural network (WT-CNN) hybrid denoising unit uses wavelet transform to remove environmental noise, and then uses a pre-trained CNN model to identify and filter out material grain boundary scattering interference; The dynamic threshold warning unit automatically adjusts the defect judgment threshold according to the wheelset material, mileage and historical data, and triggers a graded alarm when it detects a wheel rim thickness deviation of >1mm, tread wear depth of >2mm or crack length of >5mm.

[0012] Preferably, the mechanical scanning module includes: A modular quick-release bracket allows the device to switch between portable mode and online mode. In portable mode, it uses a magnetic fixing structure, while in online mode, it is integrated into the robot end effector of the wheelset maintenance line. The three-degree-of-freedom scanning mechanism consists of an electric rotating platform, a linear slide and a surface tracking sensor, which enables the probe to move in axial translation, circumferential rotation and radial adaptive fitting along the wheelset.

[0013] Preferably, the data management module includes: The wheelset digital twin model maps the inspection data to a 3D wheelset model, uses color gradients to represent the degree of wear, and annotates the crack location and propagation direction; The remaining life prediction algorithm, based on the wheelset material fatigue curve and wear rate model, calculates the remaining mileage before the rim thickness drops to the safety threshold.

[0014] Preferably, the device further comprises a multi-physical field composite detection unit, integrating an eddy current sensor and an infrared thermal imager, which are respectively used to synchronously detect abnormalities in the conductive layer on the surface of the wheelset and internal stress concentration areas.

[0015] Preferably, the protective shell of the device meets the IP67 protection level, is provided with a shock-absorbing honeycomb structure inside, and has an operating temperature range of -25°C to 60°C, which is suitable for high-humidity and dusty EMU maintenance environments.

[0016] In the above technical solution, the technical effects and advantages provided by the present invention are: 1. Utilizing multi-frequency composite probe array technology, the system cleverly combines the advantages of low-frequency (0.5-2MHz) and high-frequency (5-10MHz) probes. The low-frequency probe, with its powerful penetration, accurately detects deep cracks in the rim; the high-frequency probe, with its high resolution, keenly captures surface wear on the rim and tread, eliminating the blind spots typically associated with traditional single-frequency probes. The matrix-like staggered layout provides full axial and circumferential coverage, enabling continuous data capture of vertical and transverse cross-sections of the wheelset in a single scan. This significantly reduces missed detection rates and improves detection efficiency by over 30%.

[0017] 2. The liquid-immersion adaptive coupling structure utilizes a flexible sealing cavity and a pressure feedback adjustment mechanism to dynamically maintain a constant 0.1mm thickness of the coupling agent film, perfectly resolving the problem of uneven coupling agent distribution caused by the curved surface of the wheelset and improving signal stability by 50%. Combined with a three-degree-of-freedom mechanical scanning mechanism with axial translation, circumferential rotation, and radial bonding capabilities, and a surface tracking sensor, the probe maintains perpendicular contact with the wheelset surface regardless of how the wheelset diameter varies between Φ840mm and Φ920mm or what type of wear it exhibits, ensuring detection adaptability covering over 99% of EMU models.

[0018] 3. The WT-CNN hybrid noise reduction algorithm innovatively integrates wavelet transform and convolutional neural network to effectively filter out environmental noise and suppress interference from material grain boundary scattering, improving the signal-to-noise ratio to over 20dB and achieving a defect recognition accuracy of over 98%. The dynamic threshold warning model intelligently adjusts the alarm threshold based on parameters such as wheelset material hardness and mileage. For example, the wheel rim thickness deviation threshold can be flexibly adjusted between 0.8-1.2mm, reducing the false alarm rate by 40%. Digital twin and life prediction technology intuitively displays wear distribution through 3D visualization. Combined with a remaining mileage calculation model with an error of only ±5%, it predicts the wheelset replacement cycle in advance, reducing the risk of sudden failure by 90%.

[0019] 4. The modular quick-release design enables rapid switching between portable and online modes. The magnetically fixed portable mode supports single-person operation, while the online mode can be integrated into a maintenance robot, reducing inspection preparation time from 30 minutes to 5 minutes. It is suitable for various scenarios such as temporary spot checks and automated assembly line inspections. The multi-physics field composite detection unit integrates eddy current and infrared technologies to simultaneously detect surface conductive layer anomalies and internal stress concentration, enabling full-dimensional diagnosis of the wheelset's "surface, interior, and material properties," increasing the comprehensive fault detection rate to 99.5%.

[0020] 5. The IP67 protective casing and shock-absorbing honeycomb structure enable the equipment to operate stably in high humidity (RH95%), dusty environments, and strong vibration (≤5g), extending its service life to over 8 years. The wide operating temperature range (-25°C to 60°C) adapts to extreme climates in northern and southern China, eliminating the need for additional temperature control equipment and reducing operation and maintenance costs by 60%. The inspection time for a single wheelset has been reduced from 20 minutes to 5 minutes, reducing labor costs by 70% and increasing annual maintenance efficiency by 300%. Preventive maintenance reduces unplanned wheelset replacements, reducing the average annual maintenance cost of a single EMU by 500,000 yuan. The risk of derailment due to wheelset defects has also been significantly reduced, significantly improving rail transit operational safety and generating significant social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0022] Figure 1 It is a schematic diagram of the overall framework structure of the present invention; Figure 2 This is a schematic diagram of the probe array layout structure of the present invention; Figure 3 is a cross-sectional schematic diagram of the adaptive coupling structure of the present invention; Figure 4 Schematic diagram of the signal processing steps of the present invention; Figure 5 It is a structural schematic diagram of the mechanical scanning mechanism of the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0024] The present invention provides Figures 1 to 5 The ultrasonic wear detection device for a moving vehicle wheel set shown in the figure comprises: Ultrasonic probe module, including a multi-frequency composite probe array and an adaptive coupling structure, is used to perform all-round scanning of the wheel flange, tread, and rim of the wheelset; The signal processing module is connected to the probe module and includes a noise reduction unit, a feature extraction unit, and a real-time warning unit. It is used to intelligently analyze the echo signal and output defect detection results. Mechanical scanning module, integrating automatic positioning mechanism and protective shell, is used to drive the probe module to move along multiple axes of the wheelset surface and adapt to complex working conditions; Data management module, including a cloud database and visual interactive interface, used to store inspection data, generate 3D wear maps and maintenance recommendation reports; The ultrasonic probe module includes: Multi-frequency composite probe array, consisting of low-frequency probes (0.5-2MHz) and high-frequency probes (5-10MHz) arranged alternately. The low-frequency probes are used to detect internal cracks in the rim, while the high-frequency probes are used to measure the wheel flange thickness and tread wear depth. The probe array is staggered along the axial and circumferential directions of the wheelset in a matrix layout, covering the area where the curvature of the wheelset surface changes. A single scan acquires continuous data of the vertical section of the wheel rim and the transverse section of the tread.

[0025] The adaptive coupling structure is a liquid immersion coupling device, comprising: Flexible sealing cavity filled with coupling agent and conforming to the surface contour of the wheelset; The pressure feedback unit monitors the contact pressure between the cavity and the wheelset in real time, and dynamically adjusts the cavity volume through a micro air pump to maintain a constant couplant film thickness.

[0026] The signal processing module includes: The wavelet transform-convolutional neural network (WT-CNN) hybrid denoising unit uses wavelet transform to remove environmental noise, and then uses a pre-trained CNN model to identify and filter out material grain boundary scattering interference; The dynamic threshold warning unit automatically adjusts the defect judgment threshold according to the wheelset material, mileage and historical data, and triggers a graded alarm when it detects a wheel rim thickness deviation of >1mm, tread wear depth of >2mm or crack length of >5mm.

[0027] The mechanical scanning module includes: A modular quick-release bracket allows the device to switch between portable mode and online mode. In portable mode, it uses a magnetic fixing structure, while in online mode, it is integrated into the robot end effector of the wheelset maintenance line. The three-degree-of-freedom scanning mechanism consists of an electric rotating platform, a linear slide and a surface tracking sensor, which enables the probe to move in axial translation, circumferential rotation and radial adaptive fitting along the wheelset.

[0028] The data management module includes: The wheelset digital twin model maps the inspection data to a 3D wheelset model, uses color gradients to represent the degree of wear, and annotates the crack location and propagation direction; The remaining life prediction algorithm, based on the wheelset material fatigue curve and wear rate model, calculates the remaining mileage before the rim thickness drops to the safety threshold.

[0029] The device also includes a multi-physical field composite detection unit, an integrated eddy current sensor and an infrared thermal imager, which are used to synchronously detect abnormalities in the conductive layer on the surface of the wheelset and internal stress concentration areas.

[0030] The protective shell of the device meets the IP67 protection level, is equipped with a shock-absorbing honeycomb structure inside, and has an operating temperature range of -25°C to 60°C. It is suitable for high-humidity and dusty EMU maintenance environments.

[0031] Example 1: Ultrasonic probe module (1) Multi-frequency composite probe array The low-frequency probe (1.5MHz) and the high-frequency probe (7.5MHz) are arranged alternately in a ratio of 1:3 to cover the area where the wheelset curvature changes.

[0032] During scanning, the low-frequency probe transmits a pulse width of Tp=2μs, and the high-frequency probe transmits Tp=0.5μs, ensuring a balance between penetration depth and resolution.

[0033] A single scan simultaneously acquires the rim vertical section (low-frequency data) and the tread transverse section (high-frequency data), generating a complete wheelset profile through data fusion. (2) Adaptive coupling structure The flexible sealing cavity is made of silicone, which conforms to the curvature of the wheelset surface. It is filled with a water-based coupling agent and water-based gel. A micro pump circulates the fluid to prevent air bubbles from being trapped. The cavity pressure P is adjusted by a micro air pump to meet the following requirements:

[0034] Where k is the elastic coefficient, Δh is the gap between the wheelset surface and the cavity, and P0 is the initial pressure; Example 2: Signal processing module (1) WT-CNN hybrid denoising algorithm Wavelet transform: multi-scale decomposition of the echo signal s(t):

[0035] Among them, a is the scale factor, b is the translation factor, and ψ(t) is the Morlet wavelet basis function; CNN denoising: Construct a 5-layer convolutional network with the wavelet coefficient matrix W(a,b) as input, the output layer activation function as Sigmoid, and the loss function as:.

[0036] .

[0037] in, is the prediction coefficient, is the pure signal coefficient.

[0038] (2) Dynamic threshold warning model The threshold T is dynamically adjusted according to the wheelset mileage L and material hardness H:

[0039] Among them, T0 is the initial threshold, α, β are the attenuation coefficients, Lmax, Hmax are the maximum mileage and hardness calibration values.

[0040] Example 3: Data Management Module (1) Wheelset digital twin model 3D reconstruction: Generates a 3D mesh model of the wheelset based on the inspection data, and vertex shading maps the wear depth (red: wear > 2mm, green: normal).

[0041] Crack propagation prediction: Finite element analysis (FEA) is used to simulate the stress distribution of cracks within the rim and predict the propagation direction and rate.

[0042] (2) Remaining life prediction algorithm Based on the wheel rim thickness wear rate v and the safety threshold Smin, the remaining mileage R is calculated as:

[0043] Among them, Scurrent is the current thickness, and v is obtained through regression analysis of historical data.

[0044] The results are displayed on a visual interface, prompting "Recommend wheelset replacement" or "Continue monitoring."

[0045] Example 4: Mechanical Scanning Module Three-degree-of-freedom motion control: The probe's axial translation velocity vx = 10 mm / s, circumferential rotation angular velocity ω = 5∘ / s, and radial fitting displacement d is controlled by feedback from the surface tracking sensor:

[0046] Where, e(t) is the distance error between the probe and the wheelset surface, and Kp and Ki are PID control parameters.

[0047] Modular quick-release design Portable mode: The magnetic bracket (adsorption force ≥ 200N) is fixed to the wheelset surface and is suitable for temporary spot checks.

[0048] Online mode: Integrate into the end of the maintenance robot through a standard interface (such as ISO9409) to achieve automated inspection of the assembly line.

[0049] Industrial applicability verification 1. Laboratory testing Detection accuracy verification: Wheel rim thickness testing: A high-frequency probe (7.5MHz) was used to perform 100 repeated measurements on a standard test block (thickness 25-35mm), with an error range of ±0.08mm (95% confidence level), which is better than traditional caliper measurement (±0.2mm).

[0050] Tread wear depth: In simulated wear gradient specimens (depth 0.5-5mm), the system resolution reaches 0.05mm, capable of identifying tiny peeling defects.

[0051] Crack detection capability: Detection of artificial prefabricated cracks (depth 2-10mm) shows a crack detection rate of 98.5% (depth ≥ 2mm) and a false alarm rate of <1.5%.

[0052] Environmental adaptability test: Temperature and humidity limits: Continuously operate for 24 hours in a -25°C low-temperature chamber and a 60°C high-temperature chamber, with detection error fluctuation less than 0.1mm and no freezing or evaporation of the coupling agent.

[0053] Vibration resistance: Simulating the 5g vibration environment of a train maintenance depot, the device maintained 99% signal stability and showed no structural damage to the protective casing.

[0054] Efficiency verification: Online mode: Full parameter detection of a single wheelset takes 4 minutes and 30 seconds (traditional manual detection takes 20 minutes), and the production line can detect 12 wheelsets per hour.

[0055] Algorithm processing speed: The WT-CNN hybrid noise reduction algorithm single-frame signal processing time is ≤50ms, meeting the needs of real-time detection.

[0056] 2. Field Application CRH380B EMU wheelset maintenance line integration: Continuous operation test: 500 hours of cumulative operation, 1,200 wheelset inspections completed, 0 failure shutdowns, and a false alarm rate of <0.8%.

[0057] IP67 protection verification: Dust test: at dust concentration 200mg / m 3 The internal circuits can run for 72 hours in a safe environment without any dust accumulation.

[0058] Waterproof test: Immersed in water at a depth of 1 meter for 30 minutes, the shell has no leakage and functions normally.

[0059] Verification of compatibility with multiple vehicle models: It has been successfully adapted to models such as the CR400AF "Fuxing" and CRH3G, with wheel set diameters ranging from Φ840-Φ920mm, without the need for hardware adjustment.

[0060] Economic benefit analysis: Cost savings: The annual maintenance cost of a single EMU train is reduced by approximately RMB 520,000 (reducing unplanned wheelset replacement and labor costs).

[0061] Improved efficiency: The average daily inspection volume in the maintenance workshop increased from 40 groups to 150 groups, and the manpower requirement was reduced from 6 to 2 people.

[0062] in conclusion Through the collaborative innovation of multi-frequency ultrasonic composite detection, WT-CNN intelligent noise reduction algorithm and modular three-degree-of-freedom scanning mechanism, this invention overcomes the technical bottlenecks of traditional wheelset detection, such as unstable coupling, large noise interference and low degree of automation, and achieves three major breakthroughs: Full-dimensional detection capabilities: By integrating low-frequency penetration testing (rim cracks) with high-frequency surface analysis (rim / tread wear), and synchronously integrating eddy current (surface conductive layer) and infrared (internal stress) testing, we build an integrated "depth-surface-material performance" diagnostic system.

[0063] Intelligent decision support: The three-dimensional wear map and remaining life prediction model based on digital twins transforms post-maintenance into preventive maintenance, with the prediction accuracy of wheelset replacement cycle reaching 93%, and the efficiency of equipment life cycle management increased by 40%.

[0064] Industrial-grade reliability design: The IP67 protection and wide temperature range adaptability design break through the harsh working conditions of the EMU maintenance depot. The average annual failure rate of the equipment is less than 0.5 times, and the comprehensive operation and maintenance costs are reduced by 60%.

[0065] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

Claims

1. An ultrasonic wear detection device for a moving vehicle wheel pair, characterized in that: include: Ultrasonic probe module, including a multi-frequency composite probe array and an adaptive coupling structure, is used to perform all-round scanning of the wheel flange, tread, and rim of the wheelset; The signal processing module is connected to the probe module and includes a noise reduction unit, a feature extraction unit, and a real-time warning unit. It is used to intelligently analyze the echo signal and output defect detection results. Mechanical scanning module, integrating automatic positioning mechanism and protective shell, is used to drive the probe module to move along multiple axes of the wheelset surface and adapt to complex working conditions; The data management module includes a cloud database and a visual interactive interface, which is used to store inspection data, generate three-dimensional wear maps and maintenance recommendation reports.

2. The ultrasonic wear detection device for a moving vehicle wheel set according to claim 1, characterized in that: The ultrasonic probe module includes: Multi-frequency composite probe array, consisting of low-frequency probes (0.5-2MHz) and high-frequency probes (5-10MHz) arranged alternately. The low-frequency probes are used to detect internal cracks in the rim, while the high-frequency probes are used to measure the wheel flange thickness and tread wear depth. The probe array is staggered along the axial and circumferential directions of the wheelset in a matrix layout, covering the area where the curvature of the wheelset surface changes. A single scan acquires continuous data of the vertical section of the wheel rim and the transverse section of the tread.

3. The ultrasonic wear detection device for a moving vehicle wheel set according to claim 1, characterized in that: The adaptive coupling structure is a liquid immersion coupling device, comprising: Flexible sealing cavity filled with coupling agent and conforming to the surface contour of the wheelset; The pressure feedback unit monitors the contact pressure between the cavity and the wheelset in real time, and dynamically adjusts the cavity volume through a micro air pump to maintain a constant couplant film thickness.

4. The ultrasonic wear detection device for a moving vehicle wheel set according to claim 1, characterized in that: The signal processing module includes: The wavelet transform-convolutional neural network (WT-CNN) hybrid denoising unit uses wavelet transform to remove environmental noise, and then uses a pre-trained CNN model to identify and filter out material grain boundary scattering interference; The dynamic threshold warning unit automatically adjusts the defect judgment threshold according to the wheelset material, mileage and historical data, and triggers a graded alarm when it detects a wheel rim thickness deviation of >1mm, tread wear depth of >2mm or crack length of >5mm.

5. The ultrasonic wear detection device for a moving vehicle wheel set according to claim 1, characterized in that: The mechanical scanning module includes: A modular quick-release bracket allows the device to switch between portable mode and online mode. In portable mode, it uses a magnetic fixing structure, while in online mode, it is integrated into the robot end effector of the wheelset maintenance line. The three-degree-of-freedom scanning mechanism consists of an electric rotating platform, a linear slide and a surface tracking sensor, which enables the probe to move in axial translation, circumferential rotation and radial adaptive fitting along the wheelset.

6. The ultrasonic wear detection device for a motor vehicle wheel set according to claim 1, characterized in that: The data management module includes: The wheelset digital twin model maps the inspection data to a 3D wheelset model, uses color gradients to represent the degree of wear, and annotates the crack location and propagation direction; The remaining life prediction algorithm, based on the wheelset material fatigue curve and wear rate model, calculates the remaining mileage before the rim thickness drops to the safety threshold.

7. The ultrasonic wear detection device for a motor vehicle wheel set according to claim 1, characterized in that: The device also includes a multi-physical field composite detection unit, an integrated eddy current sensor and an infrared thermal imager, which are used to synchronously detect abnormalities in the conductive layer on the surface of the wheelset and internal stress concentration areas.

8. The ultrasonic wear detection device for a motor vehicle wheel set according to any one of claims 1 to 7, characterized in that: The protective shell of the device meets the IP67 protection level, is equipped with a shock-absorbing honeycomb structure inside, and has an operating temperature range of -25°C to 60°C. It is suitable for high-humidity and dusty EMU maintenance environments.

Images