Railway axle counting equipment management system

Abstract

The invention discloses a railway axle counting equipment management system which comprises a hyper-fusion sensing terminal, an edge intelligent diagnosis unit, a digital twin management platform, a block chain evidence storage subsystem and an intelligent interaction terminal. The super-fusion sensing terminal is integrated with an orthogonal magnetic head array, a vibration sensor, a temperature and humidity sensor, a tilt angle sensor, a laser ranging module, a current sensor and a dust concentration sensor; the edge intelligent diagnosis unit is internally provided with a lightweight neural network and a fault feature knowledge graph; the digital twinborn management platform can construct a three-dimensional digital twinborn body of axle counting equipment, and integrates a multi-physics field coupling simulation module; the block chain evidence storage subsystem can perform uplink evidence storage on the fault original signal and the maintenance work order record; the intelligent interaction terminal comprises a three-dimensional visual interface and a mobile operation and maintenance App. The intelligent level of railway axle counting equipment management can be effectively improved, railway transportation safety and operation efficiency are enhanced, and the method is suitable for various scenes such as high-speed railways and urban rail transit.

Description

Technical Field

[0001] The present invention relates to the technical field of railway transportation, and particularly to a railway axle counter equipment management system. Background Art

[0002] In the field of railway transportation, axle counter equipment, as the core signal equipment for judging whether a track section is occupied by a train, its operating state is directly related to railway operation safety and transportation efficiency. With the development of railway transportation towards high-speed and intelligent directions, higher requirements are put forward for the accuracy, reliability and intelligent level of the axle counter equipment management system.

[0003] Currently, traditional railway axle counter equipment management systems mainly rely on a single magnetic head sensor to collect train axle number signals, and monitor equipment status and diagnose faults through simple threshold judgment and rule engines.

[0004] However, there are many drawbacks in this management mode:

[0005] In terms of axle counting accuracy and fault diagnosis, traditional systems mostly use a single magnetic head sensor to collect axle number signals, which are extremely vulnerable to factors such as track electromagnetic interference and metal corrosion, resulting in frequent problems of missed counting and mis-counting, and the axle counting accuracy is generally lower than 99.5%. Especially when a train runs at a low speed or passes through with a short formation, there are obvious recognition blind spots. At the same time, the fault diagnosis method based on threshold alarm can only identify sudden obvious faults, and lacks effective prediction ability for early potential faults such as sensor sensitivity decline and circuit board component aging. Fault location relies on manual section-by-section inspection, resulting in an average fault repair time exceeding 2 hours, seriously affecting the timeliness and safety of railway transportation.

[0006] At the level of equipment maintenance management, traditional systems adopt a preventive maintenance strategy with a fixed cycle. This "one-size-fits-all" method is prone to "over-maintenance", causing a large number of unnecessary maintenance operations and increasing operation and maintenance costs. It may also lead to "insufficient maintenance" due to the inability to detect potential equipment faults in time. The replacement of key components lacks a scientific basis and mostly relies on the experience judgment of operation and maintenance personnel, making it difficult to achieve precise maintenance. In addition, traditional systems only record axle counting and simple alarm information, and key information such as equipment operating environment data and mechanical stress data are not effectively collected and utilized. There is a serious data island phenomenon between the railway signal system and the track maintenance monitoring system, and multi-dimensional correlation analysis cannot be carried out, resulting in a recurrence rate of up to 18% for the same type of faults within half a year, greatly reducing the equipment management efficiency.

[0007] In terms of data management and operation and maintenance interaction, the traditional axle counting equipment management system lacks a data trustworthy evidence mechanism, and the integrity and traceability of fault data are difficult to guarantee, which is not conducive to fault responsibility definition and experience summary. At the same time, operation and maintenance management mostly relies on manual operations and paper records, lacks intelligent interactive means, and cannot achieve real-time visual monitoring of equipment status and remote operation and maintenance guidance. The standardization of on-site operations and data traceability are poor, which makes it difficult to meet the development requirements of intelligent railway operation and maintenance.

[0008] Therefore, how to provide a railway axle counting equipment management system is a problem that those skilled in the art need to solve urgently. Summary of the invention

[0009] One purpose of the present invention is to propose a railway axle counting equipment management system. The present invention can improve the axle counting accuracy and realize real-time fault reasoning through multimodal sensor fusion and intelligent algorithms, improve the early fault recognition rate, and use digital twins and prediction models to predict the life of key components in advance. By using blockchain evidence and intelligent interaction technology, it can ensure that fault data is credible and traceable, while realizing real-time visualization of equipment status and full-process digital operation and maintenance, significantly improving the level of intelligent railway management.

[0010] A railway axle counting equipment management system according to an embodiment of the present invention includes a hyper-converged sensing terminal, an edge intelligent diagnosis unit, a digital twin management platform, a blockchain evidence storage subsystem, and an intelligent interactive terminal;

[0011] The hyper-converged sensing terminal integrates an orthogonal magnetic head array, a vibration sensor, a temperature and humidity sensor, an inclination sensor, a laser ranging module, a current sensor and a dust concentration sensor, and is used to collect train axle number signals, equipment vibration data, environmental parameters, installation posture data, power consumption data and dust concentration data;

[0012] The edge intelligent diagnosis unit has a built-in lightweight neural network and fault feature knowledge graph, performs real-time fusion processing on multi-modal sensor data, realizes shaft pulse signal noise reduction, real-time diagnosis of equipment status and local fault caching, and integrates a vibration energy harvester and a LoRaWAN wireless module, using a dynamic sleep mechanism;

[0013] The digital twin management platform can construct a three-dimensional digital twin of the axle counting equipment, integrate a multi-physics field coupling simulation module, and achieve real-time mapping with the physical equipment through the OPC UA protocol to realize fault preview, remaining life prediction and dynamic optimization of maintenance strategy;

[0014] The blockchain evidence storage subsystem can store the original fault signal and maintenance work order records on the chain, use the improved PoS consensus algorithm to build a consortium chain, and realize fault responsibility tracing and cross-domain data sharing through smart contracts;

[0015] The intelligent interaction terminal includes a 3D visualization interface and a mobile operation and maintenance App, supporting real-time monitoring of device status, intelligent dispatching of work orders, and virtual pre-rehearsal of maintenance plans.

[0016] Further, the orthogonal magnetic head array is arranged in a three-group X / Y / Z-axis three-dimensional layout, with a distance of 4-6 cm between adjacent magnetic head groups. Combined with a differential amplification circuit and a Kalman filtering algorithm, the anti-electromagnetic interference ability is improved by 40%-50%, and the low-speed direction judgment accuracy rate is ≥99.5%. The laser ranging module monitors the vertical distance between the magnetic head and the wheel rim in real time, with an accuracy of ±0.1 mm. When the offset > 2 mm, the installation angle compensation algorithm is automatically triggered, and the deviation is calibrated by adjusting the inclination angle of the magnetic head bracket.

[0017] Further, the vibration energy collector adopts a piezoelectric ceramic and electromagnetic induction composite structure, generating ≥50 mW of continuous power supply when the train passes, and realizing a battery life of ≥15 days in the power-off section in cooperation with a super capacitor. The edge intelligent diagnosis unit is built with a self-calibration module, triggering a standard wheel set simulation device during the skylight time every day to automatically complete the calibration of the magnetic head sensitivity, and the calibration error ≤0.5%.

[0018] Further, the lightweight neural network is a Transformer model with a parameter scale ≤10 MB, including twelve encoder layers and a hidden layer dimension of 128, realizing real-time fault inference at the 5-10 ms level. The fault feature knowledge graph contains ≥200 fault mode association rules, and the dust concentration sensor is used to trigger the automatic cleaning instruction of the magnetic head surface.

[0019] Further, the digital twin obtains a device geometric model with an accuracy of ±0.05 mm through point cloud scanning, and integrates multi-physical field simulation modules of mechanics, thermotics, and electromagnetics, supporting device performance simulation in the temperature range of -50°C to +60°C and the humidity environment of 0-100% RH. The digital twin management platform is data-interconnected with the railway signal system and the track status monitoring system of the railway engineering department, and analyzes the coupling relationship between device failures and external factors such as track settlement and catenary fluctuations through a spatio-temporal correlation algorithm.

[0020] Further, the digital twin management platform is built with an LSTM-Attention hybrid model. Based on parameters such as the insulation resistance of the magnetic head coil and the capacitance of the circuit board, the prediction accuracy rate of the remaining life of key components is ≥85%, and a replacement warning is generated 30-60 days in advance.

[0021] Further, the blockchain evidence storage subsystem adopts a consortium chain architecture, and the nodes include the railway administration dispatching center, equipment manufacturers, and operation and maintenance units. The 100 ms fault original waveform data and work order operation timestamps are stored in an immutable manner through the SHA-256 hash algorithm, and the block generation interval ≤30 seconds.

[0022] Furthermore, the three-dimensional visualization interface of the intelligent interactive terminal supports real-time mapping of equipment status, three-dimensional positioning of fault sections and virtual preview of maintenance processes, and the mobile operation and maintenance App supports offline work order downloads, on-site data entry and real-time positioning check-in.

[0023] Furthermore, the current sensor is used to monitor abnormal power consumption of the axle counting equipment, and data interaction is achieved between the hyper-converged sensing terminal, edge intelligent diagnosis unit, digital twin management platform, blockchain evidence subsystem and intelligent interactive terminal through a railway-specific secure data network.

[0024] The beneficial effects of the present invention are:

[0025] 1. The present invention adopts an orthogonal magnetic head array combined with multi-sensor fusion technology, and cooperates with a differential amplifier circuit and a Kalman filter algorithm to significantly improve the ability to resist electromagnetic interference and greatly improve the accuracy of axle counting. The edge intelligent diagnosis unit has a built-in lightweight Transformer model to achieve 5-10ms real-time fault reasoning, greatly improve the early fault recognition rate, and greatly shorten the fault location and repair time, thereby ensuring railway driving safety and transportation efficiency.

[0026] 2. In the present invention, the digital twin management platform uses the LSTM-Attention hybrid model to predict the remaining life of key components, combines multi-physical field coupling simulation to rehearse maintenance operations, generates optimal maintenance plans, reduces unnecessary maintenance work, reduces operation and maintenance costs, and avoids operational losses caused by sudden equipment failures.

[0027] 3. In the present invention, the blockchain evidence storage subsystem is used to ensure that fault data cannot be tampered with, responsibility traceability and cross-domain sharing are achieved through smart contracts, and the three-dimensional visualization interface and mobile operation and maintenance App realize real-time mapping of equipment status and full-process digital management, support offline operations and precise positioning sign-in, and improve the convenience, standardization and data traceability of operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0029] In the attached picture:

[0030] Figure 1 This is a schematic diagram of the framework structure of the railway axle counting equipment management system proposed by the present invention;

[0031] Figure 2 This is a diagram of the railway axle counting equipment data collection architecture of the railway axle counting equipment management system proposed by the present invention;

[0032] Figure 3This is a processing flow chart of the edge intelligent diagnosis algorithm of the railway axle counting equipment management system proposed by the present invention;

[0033] Figure 4 This is a working principle diagram of the digital twin management platform of the railway axle counting equipment management system proposed by the present invention;

[0034] Figure 5 This is a schematic diagram of the full-process interaction of the railway axle counting equipment management system proposed in the present invention. DETAILED DESCRIPTION

[0035] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0036] Railway axle counting equipment management system, including hyper-converged sensing terminal, edge intelligent diagnosis unit, digital twin management platform, blockchain evidence storage subsystem and intelligent interactive terminal;

[0037] The hyper-converged sensing terminal adopts a modular integrated design, is encapsulated in a metal chassis with IP67 protection level, and is installed on a special bracket next to the track.

[0038] The hyper-converged sensing terminal integrates an orthogonal magnetic head array, a vibration sensor, an environmental monitoring module, a laser ranging module, and a current sensor;

[0039] The orthogonal magnetic head array consists of three groups of high-precision magnetoelectric sensors in a three-dimensional layout of the X / Y / Z axes. The spacing between adjacent magnetic head groups is set to 5 cm. Each group of magnetic heads has a built-in differential amplifier circuit, which can amplify the induced voltage signal by 1000 times and filter out high-frequency noise through a low-pass filter. The surface of the magnetic head is covered with a nano-level anti-corrosion coating, and the salt spray resistance test time exceeds 1000 hours.

[0040] The vibration sensor uses a three-axis MEMS vibration sensor with a sampling frequency of 10kHz and a resolution of 0.001g. The sensor is fixed to the center of the metal substrate inside the axle counter box with epoxy resin glue to ensure complete collection of vibration signals.

[0041] The environmental monitoring module includes a temperature and humidity sensor and a dust concentration sensor. 2 C communication protocol, temperature measurement range -40℃~+125℃, accuracy ±0.1℃, humidity measurement range 0~100%RH, accuracy ±1.5%RH;

[0042] The dust concentration sensor is based on the principle of infrared scattering and can detect particles larger than 0.8μm with a detection range of 0 to 1000μg / m 3 When the concentration exceeds the set threshold, such as 500 μg / m 3Trigger the automatic cleaning instruction when

[0043] The laser ranging module uses a TOF laser ranging sensor with a measurement range of 20mm to 200mm, an accuracy of ±0.1mm. The module is fixed directly above the magnetic head through an adjustable bracket to continuously monitor the vertical distance between the magnetic head and the wheel rim.

[0044] The current sensor selects a closed-loop Hall current sensor with a measurement range of 0 to 5A and an accuracy of ±0.5%. It is connected in series to the power supply circuit of the axle counting device to continuously monitor the power consumption of the device.

[0045] Specifically, when the train enters the axle counting section, the vibration sensor first detects the vibration signal, triggering the start of the hyper-converged perception terminal. Each sensor synchronously collects data:

[0046] The orthogonal magnetic head array outputs three-axis magnetoelectric signals with a voltage range of 0 to 5V. After being converted into digital signals by a 24-bit ADC, the sampling frequency is 10kHz.

[0047] The laser ranging module outputs the distance data between the magnetic head and the wheel rim at a frequency of 100Hz.

[0048] The environmental monitoring module collects temperature, humidity, and dust concentration data every 10 seconds.

[0049] The current sensor collects the device current data at a frequency of 1kHz.

[0050] And all data is aggregated to the edge intelligent diagnosis unit through the internal bus.

[0051] Embodiment 2

[0052] The edge intelligent diagnosis unit uses a low-power ARM Cortex-A72 processor with a main frequency of 2.0GHz, integrated with 1GB DDR4 memory and 8GB eMMC storage. Its core modules include a vibration energy collector, a LoRaWAN wireless module, and a self-calibration module:

[0053] Among them, the vibration energy collector consists of a piezoelectric ceramic sheet and an electromagnetic induction coil to form a composite structure. When the train passes by, the piezoelectric ceramic generates alternating current under the vibration of 5 - 20Hz, and after rectification and filtering, it outputs 5V direct current; the electromagnetic induction coil generates an induced electromotive force under the change of the train's magnetic field. The two are superimposed to supply power to the system. Combined with a 200mF super capacitor, it can achieve a 15-day battery life in a power-off section.

[0054] Secondly, the LoRaWAN wireless module selects the RAK4631 chip, with a working frequency band of 470 - 510MHz, a transmission power of 20dBm, supporting a hybrid networking mode of star and Mesh. The single-hop transmission distance reaches 5km, and the communication error rate is lower than 10 -6 ;

[0055] In addition, the self-calibration module is built with a standard wheel set simulation device, which includes simulation wheel sets with different wheel diameters of 1000mm, 1050mm, and 1100mm and wheel flange thicknesses of 23mm, 25mm, and 27mm. Driven by a stepper motor, it can simulate the passing scenario of a train in the speed range of 0 - 350 km / h.

[0056] Specifically, first, the wavelet denoising algorithm is used to decompose the magnetoelectric signal into 5 layers. After reconstruction, the signal-to-noise ratio is increased by 15 dB. Then, the dynamic time warping algorithm is used to align data with different sampling rates such as vibration and current with the head signal as the reference to achieve data preprocessing. The lightweight Transformer model contains 12 encoder layers with a hidden layer dimension of 128. Through an input of 32-dimensional feature vectors, such as the slope of the rising edge of the shaft pulse and the energy value in the 10 - 50 kHz frequency band of the vibration signal, etc., the model is trained at the edge through transfer learning, and the inference time is controlled within 5 - 10 ms. The fault feature knowledge graph pre-stores more than 200 association rules. For example: "The vertical distance deviation of the head > 2 mm and the vibration acceleration > 0.5 g → Head installation offset fault". When the fault probability output by the model exceeds 0.8, an alarm is triggered and 100 ms of original waveform data is cached to the local Flash to complete fault diagnosis. Finally, from 0:00 to 4:00 during the early morning skylight time every day, the self-calibration module is started, the standard wheel set passes through the head array at different speeds, the edge terminal collects the standard shaft pulse signal, compares it with the historical reference data, and automatically adjusts the counting threshold through the PID algorithm, with the calibration error controlled within 0.5% to achieve self-calibration.

[0057] Embodiment 3

[0058] The digital twin management platform can perform point cloud scanning on the axle counter equipment using a handheld laser scanner to obtain surface geometric data, construct a high-precision 3D model in 3dsMax software with the number of faces controlled within 100,000, and import it into the Unity3D engine for material texture mapping to achieve 3D modeling. Subsequently, by integrating the ANSYS Twin Builder simulation module, a coupled model of mechanics, thermotics, and electromagnetics is established;

[0059] Among them, for the mechanics field: simulate the vibration stress distribution in the frequency range of 0 - 500 Hz when the train passes;

[0060] For the thermotics field: set the boundary conditions of the environmental temperature from -50°C to +60°C and the device power consumption from 5 - 20 W to simulate the temperature rise process of the circuit board components;

[0061] For the electromagnetics field: calculate the magnetic field intensity distribution around the head within 0 - 10 mT when the wheel set passes based on the Maxwell software;

[0062] Specifically, through the OPC UA protocol, 12 key parameters of physical devices, such as head current, vibration acceleration, temperature and humidity, etc., are synchronized to the digital twin at intervals of 100 ms. When the state deviation between the digital twin and the physical device exceeds 5%, a deep diagnosis process is triggered. Using the LSTM-Attention hybrid model, historical data such as the insulation resistance of the head coil and the capacitance of the circuit board are analyzed to predict the remaining life of key components. For example, when the predicted remaining life of the head coil is less than 30 days, the system automatically generates a replacement warning and pushes it to the operation and maintenance personnel through the intelligent interaction terminal. It realizes data interconnection with the railway signal system, CTC / interlocking, and the track status monitoring system of the engineering department through the dedicated railway safety data network. Using the spatio-temporal correlation algorithm, when the axle counter equipment in a certain section has counting anomalies continuously for 3 times, the synchronous track settlement data and the catenary voltage fluctuation records are automatically retrieved to identify potential correlation factors. For example, when the track settlement > 3 mm and the catenary voltage fluctuation > 10%, the probability of axle counter failure increases by 60%.

[0063] Embodiment 4

[0064] The blockchain evidence storage subsystem adopts the consortium chain mode and includes 7 consensus nodes: 3 railway administration dispatching centers, 2 equipment manufacturers, and 2 operation and maintenance units. The node hardware configuration is an Intel Xeon Gold 6248R processor, 64 GB of memory, and 1 TB of SSD storage. The improved PoS consensus algorithm is adopted, and the node weights are dynamically adjusted according to the historical maintenance quality scores;

[0065] Specifically, when the edge intelligent diagnosis unit detects a fault, information such as 100 ms of original waveform data, the fault occurrence timestamp, and the device ID is used to generate a hash value through the SHA-256 algorithm;

[0066] At the same time, the operation records of the maintenance work orders, including the personnel ID, operation steps, and completion time, are also hashed;

[0067] The above hash values and metadata are packaged into blocks. The block generation interval is 30 seconds, and they are synchronized among nodes through the P2P network and stored on the chain after being verified by the majority of nodes;

[0068] It should be noted that a Solidity smart contract can be written to realize the traceability of fault responsibility: when the same type of fault of the same device occurs continuously for 3 times, the work order creation permission of the corresponding operation and maintenance unit is automatically frozen until a detailed fault analysis report is submitted;

[0069] By designing a data sharing contract, cross-roadway fault case desensitized sharing is realized. The shared data automatically masks sensitive information such as device numbers and geographical locations, and only key contents such as fault phenomena and treatment methods are open.

[0070] Embodiment 5

[0071] The intelligent interaction terminal includes a three-dimensional visualization interface and a mobile operation and maintenance App;

[0072] Among them, the three-dimensional visualization interface is developed based on WebGL technology, deployed on the railway dispatching center server, and the interface displays the status of axle counting equipment in real time, supporting the following functions:

[0073] Three-dimensional positioning: When the equipment fails, the faulty section is automatically highlighted in the three-dimensional model of the railway line, and the specific coordinates are marked;

[0074] Virtual rehearsal: Simulate maintenance operations in the digital twin, generate an optimal maintenance plan including a tool list and safety procedures, and guide on-site operations;

[0075] Secondly, the mobile operation and maintenance App supports Android and iOS systems and has the following functions:

[0076] Offline work order: In areas without network coverage, work order tasks within 72 hours can be downloaded in advance, including the three-dimensional model of the equipment, illustrated instructions for maintenance steps, and safety video tutorials;

[0077] On-site data entry: By scanning the RFID tag of the equipment, the equipment ledger is automatically associated, supporting taking photos and uploading photos of the fault site and filling in maintenance records. The data is temporarily stored in the local database and synchronized to the blockchain evidence storage subsystem after the network is restored;

[0078] Real-time positioning and check-in: Integrated with Beidou / GPS dual-mode positioning module, with a positioning accuracy of ±5m. When the operation and maintenance personnel arrive at the operation section, they are automatically checked in, and the arrival time and stay duration are recorded to ensure the traceability of the operation process.

[0079] Working principle: First, the hyper-converged perception terminal adopts a vibration trigger mechanism. When the train enters the axle counter section, the vibration sensor activates other sensors to work together. The orthogonal magnetic head array outputs magnetoelectric signals, which are converted into digital signals by a 24-bit ADC. The differential amplification circuit and low-pass filter improve the signal quality. Sensors such as vibration, current, and laser ranging respectively collect data on equipment vibration, power consumption, and magnetic head offset. Temperature, humidity, and dust concentration sensors monitor environmental parameters. After the collected data is transmitted to the edge intelligent diagnosis unit, the signal-to-noise ratio is improved by the wavelet denoising algorithm, and the dynamic time warping algorithm is used to achieve the time series alignment of multi-source data. Subsequently, through the lightweight Transformer model built in the edge intelligent diagnosis unit, 32-dimensional feature vectors are extracted from the preprocessed data, and the analysis is carried out in combination with the fault feature knowledge graph with more than 200 pre-stored rules. When the fault probability exceeds 0.8, an alarm is triggered and the original waveform data 100 ms before the fault is cached to the local Flash. At the same time, if it is judged as a magnetic head sensitivity fault, the standard wheel set simulation device is started during the skylight time, and the counting threshold is automatically calibrated by the PID algorithm, with the error controlled within 0.Within 5%, the processed characteristic data is transmitted to the digital twin management platform through the LoRaWAN wireless module to achieve two-way communication. Then, the digital twin management platform constructs a three-dimensional model of the device through point cloud scanning, uses ANSYS Twin Builder to realize multi-physical field coupling simulation of mechanics, thermotics, and electromagnetics. With the help of the OPC UA protocol, 12 key parameters of the physical device are synchronized to the digital twin at intervals of 100 ms. When the deviation exceeds 5%, in-depth diagnosis is triggered. The LSTM-Attention hybrid model is used to analyze the historical data of the device, predict the remaining life of key components, combine with the digital twin to pre-enact maintenance operations, generate the optimal maintenance plan, and at the same time communicate with the data of the railway signal and track maintenance monitoring system, use the spatio-temporal correlation algorithm to optimize the maintenance strategy. The blockchain evidence storage subsystem adopts a consortium chain architecture, and 7 consensus nodes reach consensus based on the improved PoS algorithm. After the edge intelligent diagnosis unit detects a fault, it hashes information such as the original waveform data and timestamp, and packages it with the maintenance work order record into a block. A block is generated every 30 seconds and stored on the chain after being verified by the nodes. The fault liability traceability is realized through the Solidity smart contract. When the same type of fault of the same device occurs continuously 3 times, the work order permission of the corresponding operation and maintenance unit is frozen. At the same time, the fault cases across railway administrations are desensitized and shared to improve the industry operation and maintenance level. When in use, the three-dimensional visualization interface is based on WebGL technology, which displays the device status in real time, automatically locates the fault section during a fault, and guides on-site operations through the digital twin pre-enactment of the maintenance plan. The mobile operation and maintenance App supports Android and iOS systems, and has functions such as offline work order download, RFID code scanning to associate with the account, on-site data collection and synchronization, and Beidou / GPS positioning check-in, ensuring the normal development of offline operations, the integrity and traceability of data, and real-time monitoring of the operation progress. Each module collaborates through the railway dedicated security data network. The data collected by the hyper-converged perception terminal is processed by the edge intelligent diagnosis unit and then transmitted to the digital twin management platform. The platform generates diagnosis results and maintenance strategies, pushes them to the intelligent interaction terminal and synchronizes the data to the blockchain evidence storage subsystem. The interaction terminal receives the instructions and feedbacks them to the platform to drive the system to adjust. The blockchain evidence storage subsystem ensures the credibility of the data and realizes the intelligent management of the entire device process.

[0080] The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and all should be covered within the protection scope of the present invention.

Claims

1. Railway axle counter equipment management system, characterized in that, Including hyper-converged sensing terminals, edge intelligent diagnosis units, digital twin management platforms, blockchain evidence storage subsystems and intelligent interactive terminals; The hyper-converged sensing terminal integrates an orthogonal magnetic head array, a vibration sensor, a temperature and humidity sensor, an inclination sensor, a laser ranging module, a current sensor and a dust concentration sensor, and is used to collect train axle number signals, equipment vibration data, environmental parameters, installation posture data, power consumption data and dust concentration data; The edge intelligent diagnosis unit has a built-in lightweight neural network and fault feature knowledge graph, performs real-time fusion processing on multi-modal sensor data, realizes shaft pulse signal noise reduction, real-time diagnosis of equipment status and local fault caching, and integrates a vibration energy harvester and a LoRaWAN wireless module, using a dynamic sleep mechanism; The digital twin management platform can construct a three-dimensional digital twin of the axle counting equipment, integrate a multi-physics field coupling simulation module, and achieve real-time mapping with the physical equipment through the OPC UA protocol to realize fault preview, remaining life prediction and dynamic optimization of maintenance strategy; The blockchain evidence storage subsystem can store the original fault signal and maintenance work order records on the chain, use the improved PoS consensus algorithm to build a consortium chain, and realize fault responsibility tracing and cross-domain data sharing through smart contracts; The intelligent interactive terminal includes a three-dimensional visualization interface and a mobile operation and maintenance App, which supports real-time monitoring of equipment status, intelligent scheduling of work orders, and virtual preview of maintenance plans.

2. The railway axle counter equipment management system according to claim 1, wherein, The orthogonal magnetic head array is a three-dimensional layout of three groups of X / Y / Z axes, with a spacing of 4-6cm between adjacent magnetic head groups. With the differential amplifier circuit and Kalman filter algorithm, the anti-electromagnetic interference capability is improved by 40%-50%, and the low-speed direction determination accuracy is ≥99.5%. The laser ranging module monitors the vertical distance between the magnetic head and the wheel rim in real time with an accuracy of ±0.1mm. When the offset is >2mm, the installation angle compensation algorithm is automatically triggered, and the deviation calibration is achieved by adjusting the inclination angle of the magnetic head bracket.

3. The railway axle counter equipment management system according to claim 1, characterized in that, The vibration energy harvester adopts a composite structure of piezoelectric ceramics and electromagnetic induction, which generates ≥50mW of continuous power supply when a train passes, and cooperates with supercapacitors to achieve a battery life of ≥15 days in the power-free section. The edge intelligent diagnostic unit has a built-in self-calibration module, which uses the skylight time every day to trigger the standard wheelset simulation device to automatically complete the head sensitivity calibration, and the calibration error is ≤0.5%.

4. The railway axle counter equipment management system according to claim 1, characterized in that The lightweight neural network is a Transformer model with a parameter scale of ≤10MB, including a twelve-layer encoder and a hidden layer dimension of 128, achieving 5-10ms level real-time fault reasoning. The fault feature knowledge graph contains ≥200 fault mode association rules, and the dust concentration sensor is used to trigger an automatic cleaning instruction for the head surface.

5. The railway axle counter equipment management system according to claim 1, characterized in that The digital twin obtains a device geometric model with an accuracy of ±0.05 mm through point cloud scanning, and integrates multi-physical field simulation modules for mechanics, thermotics, and electromagnetics, supporting device performance simulation in the temperature range of -50°C to +60°C and humidity environment of 0 - 100% RH. The digital twin management platform is data-interconnected with the railway signal system and the track status monitoring system for the maintenance of tracks, and analyzes the coupling relationship between equipment failures and external factors such as track settlement and catenary fluctuations through spatio-temporal correlation algorithms.

6. The railway axle counter equipment management system according to claim 1, wherein The digital twin management platform has an LSTM-Attention hybrid model built in. Based on parameters such as the insulation resistance of the magnetic head coil and the capacitance of the circuit board, the prediction accuracy of the remaining life of key components is ≥85%, and a replacement warning is generated 30 - 60 days in advance.

7. The railway axle counter equipment management system according to claim 1, characterized in that The blockchain evidence storage subsystem adopts a consortium chain architecture. The nodes include the railway bureau dispatching center, equipment manufacturers, and operation and maintenance units. The SHA-256 hash algorithm is used to perform immutably store the 100ms fault original waveform data and work order operation timestamps, and the block generation interval is ≤30 seconds.

8. The railway axle counter equipment management system according to claim 1, characterized in that, The 3D visualization interface of the intelligent interaction terminal supports real-time mapping of device status, 3D positioning of fault sections, and virtual pre-enactment of the maintenance process. The mobile operation and maintenance App supports offline work order downloading, on-site data entry, and real-time positioning check-in.

9. The railway axle counter equipment management system according to claim 1, characterized in that The current sensor is used to monitor abnormal power consumption of the axle counter equipment. Data interaction is achieved between the hyper-converged perception terminal, the edge intelligent diagnosis unit, the digital twin management platform, the blockchain evidence storage subsystem, and the intelligent interaction terminal through the railway dedicated secure data network.

Images