Urban rail vehicle driving analysis system based on digital twinning

Abstract

The invention relates to the technical field of rail traffic monitoring, and discloses an urban rail vehicle driving analysis system based on digital twinning, which is characterized in that vehicle vibration and traction motor current signals are collected in real time through a vehicle-mounted sensing unit, and an edge feature extraction unit calculates energy ratio features of a predefined fault frequency band; the data transmission unit transmits the characteristic data and the working condition information to a digital twin platform, a digital fault tree model is arranged in the platform, and a cross-physical domain fault conduction path is analyzed and recognized through multi-source data coupling. Through combination of frequency band energy ratio characteristics and fault tree logical reasoning, electromechanical coupling fault risks which cannot be found by traditional isolated monitoring are effectively captured, meanwhile, a diagnosis threshold value is dynamically adjusted according to the wheel-rail adhesion state, and the diagnosis reliability under the complex working condition is improved.

Description

Technical Field

[0001] The present invention relates to a digital twin-based urban rail vehicle travel analysis system, belonging to the technical field of rail transit monitoring. Background Art

[0002] Currently, urban rail vehicle status monitoring mainly uses a subsystem independent alarm mechanism: physical quantities such as mechanical vibration and traction current are set with fixed thresholds within their respective monitoring domains. When a single signal exceeds the limit, an isolated alarm is triggered. This method exposes fundamental limitations in typical scenarios such as electrical phase-splitting sections. The instantaneous current fluctuations in the traction system are easily misjudged as electrical faults, and the mechanical resonance of the transmission chain stimulated by this is missed because it does not reach the vibration threshold. The root cause is that existing technologies cannot capture the energy transfer path across physical domains.

[0003] As the urban rail network evolves towards high frequency and multiple operating conditions, the system faces deeper contradictions: 1. Sudden changes in wheel-rail adhesion, such as rain and snow, distort the transmission efficiency of vibration signals, rendering the diagnostic benchmark based on fixed thresholds invalid; 2. Resonant frequency shifts caused by track foundation settlement cause the preset frequency band monitoring to gradually become inaccurate over time; 3. The industry has tried to integrate multi-source data to improve accuracy, but this has been difficult to implement due to insufficient original signal transmission bandwidth and limited edge computing power.

[0004] Specifically, existing technologies suffer from the following shortcomings: The fragmented processing of mechanical, electrical, and track data prevents the establishment of a logical chain linking fault risk across physical domains; static diagnostic thresholds severely distort monitoring data under dynamic conditions such as sudden wheel-rail adhesion; and characteristic frequency shifts caused by slow infrastructure deformation gradually cause the preset frequency band analysis to deviate from the actual fault-sensitive zone. Therefore, the technical challenges addressed by this invention are to establish cross-physical domain energy transfer path perception capabilities and enable autonomous adaptation of diagnostic benchmarks to complex environments and evolving infrastructure. Summary of the Invention

[0005] The present invention provides a digital twin-based urban rail vehicle driving analysis system, the main purpose of which is to solve the problems that the existing domain monitoring mechanism cannot capture the fault transmission path across physical domains, and the diagnostic benchmark lacks dynamic adaptability to complex environments and infrastructure evolution.

[0006] To achieve the above objectives, the present invention provides a digital twin-based urban rail vehicle driving analysis system, comprising:

[0007] an on-board sensing unit configured to collect multi-source data including at least a vehicle vibration signal and a traction motor current signal in real time;

[0008] an edge feature extraction unit configured to process raw data collected by the on-board sensing unit in real time, and calculate frequency band energy ratios of vehicle vibration signals and frequency band energy ratios of traction motor current signals based on a plurality of target frequency bands associated with urban rail vehicle failure modes predefined through engineering practice;

[0009] The data transmission unit is configured to transmit the frequency band energy ratio of the vehicle vibration signal, the frequency band energy ratio of the traction motor current signal, and the real-time operating condition data of the vehicle to the digital twin platform; the digital twin platform is configured to: have a built-in digital fault tree model, which contains multiple logic gates for defining the cross-physical domain coupling logical relationship between the frequency band energy ratio and the vehicle operating condition data; receive the frequency band energy ratio and the real-time operating condition data of the vehicle, and use the digital fault tree model for logical reasoning to identify potential fault risk transmission paths caused by the coupling of multiple physical domain factors; generate diagnostic conclusions and maintenance recommendations including cross-physical domain causal chain analysis.

[0010] Preferably, in the edge feature extraction unit, the lightweight algorithm for calculating the frequency band energy ratio is configured to be selected from the Goertzel algorithm or a combination of digital bandpass filtering and root mean square operation. The lightweight algorithm achieves extreme simplification and efficient characterization of the complexity of the original signal by calculating the energy aggregation characteristics of a specific frequency band, and can run on a microcontroller.

[0011] Preferably, in the digital twin platform, the digital fault tree model is configured to include a logic branch for determining the risk of coupled resonance between the traction motor current fluctuation and the mechanical vibration, which identifies the risk of coupled resonance by comparing the frequency band energy ratio of the traction motor current signal with a first preset threshold, and comparing the frequency band energy ratio of the vehicle vibration signal with a second preset threshold.

[0012] Preferably, the digital twin platform is also configured to have an adaptive adjustment function. When the real-time operating condition data of the vehicle indicates that the vehicle is in rainy or snowy weather or passing through an electrical phase-separated section, the digital twin platform automatically calls the corresponding sub-logic tree in the digital fault tree model or dynamically adjusts the decision threshold used for comparison with the frequency band energy ratio.

[0013] Preferably, the edge feature extraction unit is also configured to calculate the high-frequency ripple density of the traction motor current signal, and use the high-frequency ripple density as a proxy variable for the wheel-rail adhesion state; the digital twin platform is also configured to adjust the judgment threshold for comparing the frequency band energy ratio of the vehicle vibration signal in real time according to the current ripple density.

[0014] Preferably, in the digital twin platform, the adjustment module for determining the threshold value of the frequency band energy ratio of the vehicle vibration signal is configured to be Exceeds the sticking state threshold calibrated by engineering practice When the vehicle vibration signal frequency band energy ratio is dynamically lowered, the judgment threshold is , and its adjustment relationship satisfies: ,in, is the adjusted decision threshold, is the adjustment coefficient between the adhesion state proxy variable and the judgment threshold, is the current ripple density, is the sticky state threshold, is the judgment threshold before adjustment.

[0015] Preferably, the on-board sensing unit is further configured to control the vehicle vibration sensor to operate in a preset low-frequency sampling mode to collect the track residual vibration signal when the real-time operating condition data of the vehicle indicates that the vehicle is in a parked state; the edge feature extraction unit is further configured to calculate the time domain permutation entropy that characterizes the signal complexity based on the track residual vibration signal; the digital twin platform is further configured to dynamically adjust the frequency range of the target frequency band based on which the frequency band energy ratio of the vehicle vibration signal is calculated according to the long-term change trend of the time domain permutation entropy.

[0016] Preferably, the data transmission unit is configured to transmit the frequency band energy ratio and the real-time vehicle operating condition data via a low-power wide area network, and the low-power wide area network is selected from LTE-M or NB-IoT.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] 1. By simultaneously extracting the frequency band energy ratio of vibration and current signals at the edge and combining it with multi-logic gate coupling analysis of the platform-side digital fault tree, the system dynamically correlates mechanical vibration anomalies with the energy accumulation characteristics of electrical fluctuations in specific frequency bands for the first time. When the traction motor current BER threshold exceeds the threshold and forms a logical AND gate relationship with the vibration BER anomaly in a specific frequency band, the system automatically identifies the electromechanical resonance conduction path that cannot be captured by traditional isolated monitoring. This avoids missed and false alarms caused by cross-domain coupling effects without increasing hardware costs. Furthermore, the platform dynamically adjusts the vibration BER determination threshold based on the traction motor current ripple density as a proxy variable for wheel-rail adhesion. When rainy or snowy weather causes wheel-rail transmission efficiency to decrease, the system actively compensates for signal attenuation by lowering the vibration alarm threshold, ensuring stable diagnostic logic under complex operating conditions. This mechanism enables the current signal to not only perform its primary monitoring function but also simultaneously generate implicit environmental parameters for calibrating the mechanical diagnostic benchmark, achieving a synergistic multiplication of the efficiency of multi-source data.

[0019] 2. By reusing vibration sensors during stop periods to collect residual track vibration and quantifying changes in the track foundation state based on time-domain permutation entropy, the system converts ultra-low-frequency vibration signals into long-term trend indicators representing track stiffness. When the PE value in a specific section continues to rise, the platform automatically expands the target frequency band range of the vibration BER analysis, allowing the core diagnostic parameters to dynamically adjust to the resonant frequency shift caused by track settlement, thereby avoiding the time bottleneck of traditional fixed-frequency band analysis that decays with operating time. While maintaining the constraint of zero increase in on-board hardware, the system automatically switches the sensor to low-frequency sampling mode according to the vehicle's docking status, converting idle periods into track status monitoring windows. Utilizing the extremely low computing power characteristics of the permutation entropy algorithm, real-time calculation of the track foundation health is achieved on the microcontroller, allowing a single on-board terminal to simultaneously perform the dual functions of dynamic fault diagnosis and infrastructure status perception, forming a closed loop of health management for mobile nodes and fixed facilities.

[0020] 3. The digital fault tree directly outputs a treatment plan for the source of resonant energy based on the identified risk transmission path. For example, when it is determined that the wheel-rail resonance in a specific section is induced by rail corrugation, the system automatically generates a rail grinding recommendation accurate to the section number, upgrading the operation and maintenance decision from component maintenance to blocking the energy transmission path, effectively preventing the occurrence of cascading failures in multiple physical domains. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a schematic diagram of the overall architecture of a digital twin-based urban rail vehicle driving analysis system of the present invention;

[0022] Figure 2 This is a diagram showing the change in traction current and vibration energy ratio and fault logic output of the present invention;

[0023] Figure 3 This is the workflow sequence diagram of the urban rail vehicle driving analysis system based on digital twins of the present invention.

[0024] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0025] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0026] The embodiment of the present application provides a city rail vehicle driving analysis system based on digital twins, whose overall architecture is based on the coordinated operation of the vehicle and the cloud. On the vehicle terminal side, an on-board sensing unit, an edge feature extraction unit and a data transmission unit are deployed, while a core digital twin platform is constructed on the cloud. The workflow of the system starts with the synchronous acquisition of key physical signals of the vehicle by the on-board sensing unit, undergoes lightweight feature extraction on the edge side, and then uploads highly condensed feature information and operating condition data through the data transmission unit. Finally, the digital fault tree model built into the digital twin platform performs deep fusion analysis and logical reasoning, and then outputs diagnostic conclusions and maintenance recommendations. In actual operation scenarios, massive amounts of original vibration and current waveform data pose a severe challenge to the computing power and transmission bandwidth of the vehicle terminal. To meet this challenge, the solution of the present invention is configured to perform efficient feature characterization at the source of data generation, which is effective. Specifically, the on-board sensing unit continuously collects multi-source data including at least vehicle vibration signals and traction motor current signals in real time. Then, the edge feature extraction unit does not directly transmit these original signals, but calls a lightweight algorithm. The algorithm is configured to be selected from the Goertzel algorithm or a combination of digital bandpass filtering and root mean square operation, and can run stably on resource-constrained on-board microcontrollers. By targeting multiple target frequency bands that are predefined through a large number of engineering practices and closely related to specific fault modes, the unit calculates the frequency band energy ratio of the vehicle vibration signal and the frequency band energy ratio of the traction motor current signal respectively; in this way, by performing energy aggregation feature calculations for specific frequency bands on the edge side, effective dimensionality reduction and information representation of high-dimensional time series data are achieved, thereby solving the dual constraints of on-board terminal computing power and data transmission bandwidth without sacrificing key fault information.

[0027] Furthermore, the fundamental limitation of traditional domain monitoring is that it cannot provide insight into the potential fault transmission relationship between different physical domains such as mechanical and electrical. In view of this, the solution of the present invention constructs a core mechanism for cross-domain coupling analysis. The data transmission unit is responsible for efficiently and reliably transmitting the frequency band energy ratio of the vehicle vibration signal and the traction motor current signal generated by the aforementioned edge computing, together with the real-time vehicle operating condition data obtained from the vehicle bus network, to the remote digital twin platform through a low-power wide area network (such as LTE-M or NB-IoT). The core of the platform has a built-in digital fault tree model, which contains multiple logic gates to rigorously define the cross-physical domain coupling between the frequency band energy ratio and the vehicle operating condition data. Logical relationship, for example, one of the key logical branches is specially configured to determine the coupling resonance risk between the traction motor current fluctuation and the mechanical vibration. The determination rule is that when the frequency band energy ratio of the traction motor current signal received by the platform exceeds a first preset threshold, and at the same time or within a very short time window, the frequency band energy ratio of the vehicle vibration signal corresponding to the target frequency band also exceeds a second preset threshold, a logical AND gate is triggered; through this logical reasoning based on multi-source data coupling and fault tree model, the system can accurately identify the fault risk transmission path caused by the coupling of multiple physical domain factors that cannot be detected by a single monitoring method, and generate a diagnostic conclusion that includes cross-physical domain causal chain analysis.

[0028] Furthermore, the operating environment of urban rail vehicles is complex and changeable, especially when encountering rainy and snowy weather, the adhesion state between the wheel and rail will change significantly, resulting in distortion of the transmission efficiency of the vibration signal. If a fixed static threshold is still used for diagnosis, the reliability of the monitoring results will be greatly reduced. In order to establish an environmentally adaptive diagnostic benchmark, the present invention introduces a proxy variable for the wheel-rail adhesion state. When processing the traction motor current signal, the edge feature extraction unit is also configured to calculate the high-frequency ripple density thereof. The high-frequency ripple density is transmitted to the digital twin platform as an effective proxy variable for the wheel-rail adhesion state. After receiving the ripple density data, the platform activates a set of adaptive adjustment functions to adjust the judgment threshold used for comparison with the frequency band energy ratio of the vehicle vibration signal in real time according to the current ripple density. The specific adjustment procedure is determined as a precise mathematical relationship: when the current ripple density Exceeds a sticking state threshold calibrated through a large number of circuit tests and engineering practices When the platform dynamically lowers the judgment threshold of the frequency band energy ratio of the vehicle vibration signal , the adjusted threshold Satisfy the formula ,in, It is a pre-calibrated adjustment coefficient that represents the sensitivity between the adhesion state proxy variable and the judgment threshold. With this mechanism, when external working conditions cause the vibration signal to attenuate, the system can actively compensate and lower the alarm threshold to ensure the stability and reliability of the diagnostic logic under complex working conditions.

[0029] In addition, to avoid monitoring attenuation problems caused by long-term slow deformation of infrastructure, such as resonant frequency shift caused by track foundation settlement, the present invention has designed a resource-reusing track state perception and model update mechanism. When the on-board perception unit detects that the vehicle's real-time operating data indicates that the vehicle is in a parked state, it will automatically control the vehicle vibration sensor to switch to a preset low-frequency sampling mode to specifically collect the track residual vibration signal. The edge feature extraction unit then calculates the time domain permutation entropy of this low-frequency signal. As a quantitative indicator of time series complexity with extremely low computational complexity, the time domain permutation entropy can run without burden on the on-board microcontroller. , and can effectively characterize the state changes of the track foundation's stiffness and constraints. After the digital twin platform collects the time-domain permutation entropy data corresponding to a specific track section uploaded by each vehicle over a long period of time, it can reversely infer the evolution of the track foundation state based on its long-term change trend. When it monitors that the time-domain permutation entropy of a specific section shows continuous unidirectional deviation or abnormal fluctuation, the platform will automatically and dynamically adjust the frequency range of the target frequency band used for calculating the energy ratio of the vehicle vibration signal frequency band in that section. By converting the idle period of the vehicle into a track state monitoring window, the system constructs a dynamic feedback closed loop to ensure the long-term effectiveness of the core diagnostic model.

[0030] Ultimately, the value of the system of the present invention lies not only in identifying risks, but also in providing closed-loop decisions that can directly guide operation and maintenance practices. When the digital fault tree model finally identifies a clear fault risk transmission path based on the above analysis, for example, when it is determined that the continuous wheel-rail resonance in a certain section is induced by rail corrugation of a specific wavelength, the platform is also configured to generate and output maintenance recommendations for rail grinding for specific sections based on this analysis result. The recommendations can be accurate to the track section number. In this way, operation and maintenance decisions can evolve from the traditional component maintenance mode to the active intervention mode of blocking specific resonant energy transmission paths, thereby effectively curbing the occurrence and deterioration of multi-physical domain cascading failures and realizing integrated management of the health status of vehicles and lines.

[0031] The target frequency band associated with a specific failure mode in this scenario, such as the heart rate and bandwidth It is set by performing power spectrum density analysis on bench test data containing known faults or historical operation and maintenance diagnosis data. It is a statistical method that distributes signal power in the frequency domain to identify the characteristic frequency points where energy is concentrated and their fluctuation range. It is used to adaptively adjust the adjustment coefficient of the diagnostic threshold. The value is the current ripple density exceeding the threshold value when the lubricating medium is changed to systematically reduce the adhesion coefficient on the wheel-rail simulation test bench. The slope of the fitted line is obtained by performing linear regression analysis on the relative attenuation rate of the vibration signal energy, a statistical method used to establish a linear relationship model between variables; and the adjustment coefficient used to dynamically adjust the target frequency band In the finite element simulation environment, by gradually reducing the track foundation support stiffness parameters, the resulting wheel-rail system resonance frequency offset and the synchronously calculated track residual vibration time domain permutation entropy change rate slope are calculated. Determined by regression analysis; the embedding dimension used in the time domain permutation entropy calculation With time delay These two core calculation parameters are fixed to and This setting is based on the analysis of the dynamic characteristics of vehicle vibration signals. The dimension is sufficient to expand the main dynamic modes of the rail system while requiring moderate computational effort, and the delay can preserve the signal details in the high-frequency band. At the same time, in view of the manufacturing tolerances and wear differences of different individual vehicles, the system performs a baseline calibration procedure during the initial deployment phase. This procedure requires a specified vehicle to run completely on a benchmark line that includes typical operating conditions such as traction, braking, and coasting. The system automatically collects full-time domain data during this process and constructs a background noise statistical database unique to the vehicle. Subsequently, all initial judgment thresholds in the digital fault tree model will be ultimately personalized based on the statistical distribution of this personalized database, such as the mean plus three times the standard deviation, thus completing the deployment of the diagnostic model from general to specialized.

[0032] Example 1: In a continuously operating urban rail line, a long-standing challenge lies in a special section with both large slopes and small radius curves. Especially after a sudden rainstorm, the surface of the rails in this section will be mixed with rainwater and accumulated oil, and the wheel-rail adhesion will deteriorate sharply. When the train passes through this section in traction mode, the wheels are prone to violent dynamic slip. This condition poses a dilemma for traditional monitoring systems: the current of the traction motor increases sharply to resist slip, which can easily be misjudged as an electrical fault by the isolated electrical monitoring system; and the intermittent high-frequency mechanical shock generated by slip excitation may not have enough energy to trigger the statically set higher vibration alarm threshold, thereby resulting in the omission of potential damage to the transmission system; when a vehicle deployed with the solution of the present invention enters the above-mentioned slippery curved slope section, its on-board sensing unit is synchronously collecting the traction motor current signal and the vehicle vibration signal. At this time, the vehicle's edge feature extraction unit simultaneously calculates its high-frequency ripple density when processing the traction motor current signal. Due to the severe wheel slip, the The value instantly soared and significantly exceeded the adhesion state threshold calibrated by engineering practice After the digital twin platform receives this data, its built-in digital fault tree model is based on far beyond This fact determines that the wheel-rail adhesion is seriously bad as the primary boundary condition of the current system. This judgment immediately triggers the adaptive adjustment function of the system, and the platform automatically calls Deterministic procedures, based on real-time incoming value, dynamically lowering the judgment threshold for vehicle vibration signals In this way, the high-frequency ripple density of the traction motor current signal in this scenario is no longer regarded as an isolated electrical disturbance, but is used as an environmental state parameter to calibrate the mechanical vibration diagnostic benchmark, thereby avoiding the lack of adaptability of the static threshold under complex working conditions.

[0033] In other words, the vibration judgment threshold After being adaptively lowered, the frequency-band energy ratio of the transmission chain mechanical vibration induced by wheel-rail slip, which might have been ignored due to its small energy amplitude, is now highlighted and exceeds the new adjusted threshold. At the same time, the frequency-band energy ratio of the traction motor current fluctuations caused by the slip also exceeds the first preset threshold. The digital fault tree model in the digital twin platform captures the synchronization anomalies of the two in the time and frequency domains. Its logical branch for determining the risk of coupled resonance is satisfied, and the system therefore generates a diagnostic conclusion including a cross-physical domain causal chain analysis, which clearly points out that the risk stems from the electromechanical coupling resonance caused by poor wheel-rail adhesion, rather than an isolated electrical or mechanical failure.

[0034] Example 2: This example aims to conduct a controlled bench test to replicate the vibration signal distortion and risk of missed faults faced by traditional fixed-threshold diagnostic methods when the wheel-rail adhesion state of an urban rail vehicle deteriorates. This test serves as a benchmark to verify the ability of the present invention's solution to maintain diagnostic reliability by dynamically adjusting the diagnostic threshold. The test is conducted on a full-scale wheel-rail simulation test platform. This platform can replicate the vehicle's wheelset, traction drive, and suspension system at a 1:1 scale. It is equipped with a high-precision torque and speed controller and can accurately simulate wheel-rail contact conditions from dry to varying degrees of wetness through a controllable liquid spray system on the rail wheel surface. The platform is equipped with a vibration sensor of the same model as the onboard sensing unit at the axle box, and a current sensor is configured at the output of the traction motor to collect real-time vehicle vibration signals and traction motor current signals during the test. First, to simulate a persistent, incipient mechanical fault, an external exciter is used to apply a weak, continuous vibration signal of a specific frequency to the axle box. The vibration energy is precisely set to the static diagnostic threshold under normal dry conditions. About 80% of the The value of is set by using the standard deviation method after statistical analysis of the background noise data of the system running for a long time under dry conditions. The technical trade-off is to ensure the sensitivity to weak faults while avoiding false alarms caused by normal operation fluctuations. Secondly, the current ripple density, which is a proxy variable for the wheel-rail adhesion state, , its functional relationship with the actual adhesion coefficient and key parameters and , is determined by a series of calibration tests before the test officially begins. The calibration test gradually reduces the adhesion coefficient and records the corresponding Stable value, thus establishing a mapping model between the two.

[0035] After the test process was started, the wheel-rail simulation test platform first ran stably under the dry, high adhesion coefficient baseline conditions. At this time, the energy of the weak fault vibration signal at the axle box was stable, but below the static threshold. , the system diagnosis status is normal, then the liquid spraying system starts to inject water-oil mixture into the rail wheel surface to simulate the continuous and gradual decrease of the adhesion coefficient. During this process, the system records the wheel-rail adhesion coefficient and current ripple density in real time. , vibration signal energy, and the diagnostic results based on the static threshold and the adaptive threshold of the present invention, respectively. A key phenomenon was observed in the experiment: with the decrease of the adhesion coefficient, due to the deterioration of the transmission efficiency between the wheel and rail, the vibration signal energy generated by the same fault source collected by the sensor showed a significant attenuation trend. At the same time, the current ripple density that characterizes the intensification of wheel-rail slip It rises significantly. Table 1 is a set of typical data records in this process.

[0036] Table 1: Comparison of diagnostic results under different adhesion states.

[0037]

[0038] As shown in Table 1, when the wheel-rail adhesion coefficient drops to 0.2 or below, the actual collected vibration signal energy has decayed to below 6.8, which is lower than the static threshold of 10 units. , resulting in underreporting in traditional diagnostic methods. However, the digital twin platform in the system of the present invention, based on the current ripple density that has risen to more than 8.5 received synchronously, , calculates the dynamic threshold in real time through its built-in adaptive adjustment mechanism , when the adhesion coefficient is 0.1, It has been automatically lowered to 5.2. At this moment, the attenuated vibration signal energy of 5.9 can still effectively break through the dynamic threshold, thereby triggering an accurate alarm. The internal mechanism of this result is that the scheme of the present invention does not regard the vibration signal as an isolated observation quantity, but uses the characteristics of the current signal to make real-time corrections to the diagnostic model of the vibration signal, thereby compensating for the information attenuation caused by changes in the physical environment.

[0039] Example 3: This example combines Figures 1 to 3 , an urban rail vehicle driving analysis system based on digital twin is described. Figure 1 As shown in the figure, the system is divided into two parts: the vehicle-mounted end and the cloud-end. The vehicle-mounted end includes an on-board sensing unit, an edge feature extraction unit, and a data transmission unit. The on-board sensing unit is configured to collect original signals in real time. These original signals include vehicle vibration signals collected by vibration sensors and traction motor current signals collected by current sensors. The edge feature extraction unit processes the received original signals in real time and calculates the frequency band energy ratio, current ripple density, and time domain permutation entropy. The data transmission unit is responsible for transmitting the feature data generated by these edge feature extraction units, including frequency band energy ratio, current ripple density, and time domain permutation entropy, as well as real-time vehicle operating condition data, to the cloud-end via a low-power wide area network (LTE-M / NB-IoT). The core component of the cloud-end is the digital twin platform, which has a built-in digital fault tree model and uses this model to perform cross-physical domain coupling logic analysis. The digital twin platform also has an adaptive threshold adjustment function and optimizes the system through a feedback adjustment mechanism. Ultimately, the system can output diagnostic conclusions and maintenance recommendations.

[0040] like Figure 2 As shown in the figure, the horizontal axis represents time (seconds), and the vertical axis represents the frequency band energy ratio / logic output. The figure contains three curves: the solid black box represents the traction motor current frequency band energy ratio, the dotted box represents the vehicle vibration frequency band energy ratio, and the dotted dot box represents the fault detection logic output. The figure shows the changing trends of the traction motor current frequency band energy ratio and the vehicle vibration frequency band energy ratio within a specific time period, as well as the output results of the fault detection logic after coupling analysis based on these energy ratios.

[0041] like Figure 3As shown in the figure, the process first involves the on-board sensing unit collecting vibration signals and current signals in real time, and transmitting the original signal data to the edge feature extraction unit. The edge feature extraction unit then calculates the vibration frequency band energy ratio and the current frequency band energy ratio, and sends the feature data to the data transmission unit. The data transmission unit obtains the vehicle operating condition data and transmits the data, including the feature data and vehicle operating condition data, to the digital twin platform through a low-power network. The digital twin platform calls the fault diagnosis logic, and the fault model performs cross-domain coupling analysis to identify the risk transmission path. The digital twin platform then returns the diagnosis results, generates a diagnosis report, and finally outputs maintenance recommendations.

[0042] Example 4: During the long-term operation of the line, the track foundation of a specific section will experience slow non-uniform settlement. This settlement will cause the track stiffness to change, and then cause the resonant frequency of the wheel-rail system to drift. For a monitoring system with a preset fixed fault characteristic frequency band, this drift will cause the preset monitoring frequency band to gradually deviate from the actual fault energy concentration area, thereby resulting in a reduction in the system's detection sensitivity. In order to address the problem of reduced monitoring timeliness, the solution of the present invention has a built-in model adaptive adjustment closed loop based on infrastructure status perception. The starting point of the closed loop is the track residual vibration signal analysis performed by the on-board terminal during each vehicle stop. Specifically, when the vehicle is in a parked state, the on-board sensing unit switches to a preset low-frequency sampling mode to collect the track residual vibration signal, and the edge feature extraction unit then performs a time domain permutation entropy calculation on the on-board microcontroller. The calculation procedure is: first, the collected length is The residual vibration time series is reconstructed in phase space to generate a series of dimensional delay vector; secondly, sort the size of the elements in each delay vector to determine its corresponding One of the possible arrangement patterns; secondly, the frequency of occurrence of each arrangement pattern is counted to obtain its probability distribution; finally, the entropy value of the probability distribution is calculated according to the definition of information entropy, which is the quantized value of the time domain arrangement entropy, which represents the constraint state of the orbital system.

[0043] On the digital twin platform side, its core digital fault tree model undergoes an offline construction and parameter calibration stage before deployment. The logical structure of the model is established based on the analysis of the systematic failure modes and effects of urban rail vehicles. The specific component-level failures are regarded as the underlying basic events, and the system-level hazards are regarded as the top-level events. The fault conduction topology is constructed through the logical relationship of AND gates or gates. The judgment thresholds based on each logic gate in the model are calibrated as follows: a large number of current data of healthy vehicles under various typical working conditions are collected, and the frequency band energy ratio of each target frequency band is calculated to form a benchmark database. By performing statistical analysis on the database, the background noise distribution of the energy ratio of each frequency band under each working condition is determined. Finally, the judgment threshold of the frequency band is set to the background noise mean under the corresponding working condition plus the average value. times the standard deviation, is an integer greater than or equal to 3, so as to balance the detection rate and false alarm rate; for the resonance frequency shift problem caused by track foundation settlement, the digital twin platform executes a long-term adaptive adjustment program. The platform collects and stores the time domain permutation entropy data with precise location tags uploaded by each vehicle for a long time, and for each independent track section, performs linear regression analysis on its time domain permutation entropy sequence in a rolling long time window to calculate the slope of its changing trend. When the platform determines that the time domain permutation entropy slope of a certain section continues to exceed a preset basic state degradation rate threshold in multiple consecutive calculation windows, it triggers the adjustment of the monitoring parameters of the section. The adjustment procedure is: the system will adjust the target frequency band center frequency originally set for the section for vehicle vibration signal analysis according to the degree of slope exceeding the limit. Fine-tune the new center frequency satisfy ,in, is the difference between the measured slope and the degradation rate threshold, and It is an adjustment coefficient that characterizes the relationship between track stiffness changes and resonant frequency shift, pre-calibrated through simulation or bench testing. Through this data-driven slow variable feedback adjustment mechanism, the solution of the present invention ensures that the diagnostic capability can be autonomously adjusted as the infrastructure evolves, thereby maintaining its effectiveness throughout its life cycle.

[0044] Example 5: To further ensure the universality and high reliability of the solution of the present invention in different individual vehicles and special line environments, a baseline characteristic calibration must be performed for each vehicle on which the system is to be deployed. The procedure requires that the vehicle be fully operated on a test track with representative working conditions, in accordance with a preset test procedure, covering traction, braking and coasting states at different speed levels. During this period, the on-board perception unit and the edge feature extraction unit operate in learning mode, continuously collecting and calculating the full-band energy ratio of the vehicle's vibration and traction motor current on the baseline line. The collected data is transmitted to the digital twin platform, and the platform generates a personalized background noise database for the specific vehicle based on this. Subsequently, the platform will use this database as a basis to individually fine-tune the various judgment thresholds in the general digital fault tree model, thereby avoiding individual differences caused by vehicle manufacturing tolerances or different component wear conditions, and ensuring the accuracy of the diagnostic logic.

[0045] In addition, in order to cope with special operating conditions such as passing through the electrical phase-splitting section, several operating condition logic suppression branches are preset in the digital fault tree model. The real-time vehicle operating condition data received by the system through the data transmission unit includes a status flag obtained from the train network control system to indicate whether the vehicle is currently in the electrical phase-splitting section. When the digital twin platform detects that the flag is true, it will automatically activate the corresponding suppression logic branch. The function of this branch is to adjust the trigger threshold of the specific fault judgment branch related to the sudden change of traction motor current according to the preset rules during the period when the flag is valid. When the vehicle exits the electrical phase-splitting section and the status flag returns to false, all judgment thresholds will immediately and automatically return to their normal or adaptively adjusted levels. Through this deterministic operating condition identification and logic suppression mechanism, diagnostic false alarms caused by large fluctuations in electrical signals caused by normal operating operations can be effectively avoided. These are all extended implementation methods known to ordinary technicians in this field.

[0046] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A digital twin-based urban rail vehicle driving analysis system, characterized in that: include: an on-board sensing unit configured to collect multi-source data including at least a vehicle vibration signal and a traction motor current signal in real time; an edge feature extraction unit configured to process raw data collected by the on-board sensing unit in real time, and calculate frequency band energy ratios of vehicle vibration signals and frequency band energy ratios of traction motor current signals based on a plurality of target frequency bands associated with urban rail vehicle failure modes predefined through engineering practice; a data transmission unit configured to transmit the frequency band energy ratio of the vehicle vibration signal, the frequency band energy ratio of the traction motor current signal, and the real-time operating condition data of the vehicle to the digital twin platform; The digital twin platform is configured to: have a built-in digital fault tree model containing multiple logic gates that define the cross-physical domain coupling logical relationship between frequency band energy ratios and vehicle operating condition data; receive frequency band energy ratios and real-time vehicle operating condition data, and use the digital fault tree model for logical reasoning to identify potential fault risk transmission paths caused by the coupling of multiple physical domain factors; Generate diagnostic conclusions and maintenance recommendations that include causal chain analysis across physical domains.

2. The urban rail vehicle driving analysis system based on digital twin according to claim 1 is characterized in that: In the edge feature extraction unit, a lightweight algorithm for calculating the frequency band energy ratio is configured to be selected from the Goertzel algorithm or a combination of digital bandpass filtering and root mean square operation. The lightweight algorithm is calculated by energy aggregation features for preset frequency bands.

3. The urban rail vehicle driving analysis system based on digital twin according to claim 1 is characterized in that: In the digital twin platform, the digital fault tree model is configured to include a logic branch for determining the risk of coupled resonance between traction motor current fluctuations and mechanical vibrations. It identifies the risk of coupled resonance by comparing the frequency band energy ratio of the traction motor current signal with a first preset threshold, and by comparing the frequency band energy ratio of the vehicle vibration signal with a second preset threshold.

4. The urban rail vehicle driving analysis system based on digital twin according to claim 1 is characterized in that: The digital twin platform is also configured with adaptive adjustment capabilities. When the vehicle's real-time operating data indicates that the vehicle is in rainy or snowy weather or passing through an electrical phase-separated section, the digital twin platform automatically calls the corresponding sub-logic tree in the digital fault tree model or dynamically adjusts the decision threshold used for comparison with the frequency band energy ratio.

5. The urban rail vehicle driving analysis system based on digital twin according to claim 1 is characterized in that: The edge feature extraction unit is also configured to calculate the high-frequency ripple density of the traction motor current signal and use the high-frequency ripple density as a proxy variable for the wheel-rail adhesion state; the digital twin platform is also configured to adjust the judgment threshold for comparing the frequency band energy ratio of the vehicle vibration signal in real time according to the current ripple density.

6. The urban rail vehicle driving analysis system based on digital twin according to claim 5 is characterized in that: In the digital twin platform, the module for adjusting the threshold value of the frequency band energy ratio of the vehicle vibration signal is configured to be Exceeds the sticking state threshold calibrated by engineering practice When the vehicle vibration signal frequency band energy ratio is dynamically lowered, the judgment threshold is , and its adjustment relationship satisfies: ,in, is the adjusted decision threshold, is the adjustment coefficient between the adhesion state proxy variable and the judgment threshold, is the current ripple density, is the sticky state threshold, is the judgment threshold before adjustment.

7. The urban rail vehicle driving analysis system based on digital twin according to claim 1 is characterized in that: The on-board sensing unit is also configured to control the vehicle vibration sensor to operate in a preset low-frequency sampling mode to collect the track residual vibration signal when the vehicle's real-time operating condition data indicates that the vehicle is in a parked state; the edge feature extraction unit is also configured to calculate the time domain permutation entropy that characterizes its signal complexity based on the track residual vibration signal; the digital twin platform is also configured to dynamically adjust the frequency range of the target frequency band based on which the frequency band energy ratio of the vehicle vibration signal is calculated according to the long-term change trend of the time domain permutation entropy.

8. The urban rail vehicle driving analysis system based on digital twin according to claim 1 is characterized in that: The data transmission unit is configured to transmit the frequency band energy ratio and the real-time operating condition data of the vehicle through a low-power wide area network, and the low-power wide area network is selected from LTE-M or NB-IoT.

Images