Railway detection data real-time processing method

Abstract

The invention relates to the technical field of railway detection, in particular to a railway detection data real-time processing method, which comprises the steps of collecting multi-source heterogeneous data, generating a reference library, calculating a wear rate and an offset, determining a failure rate and a fracture risk rate, and realizing parallel computing and feature extraction through a distributed streaming processing framework. Window calculation, state management and complex event processing mechanisms are adopted, a machine learning algorithm is combined to carry out anomaly detection and trend prediction, and a real-time alarm mechanism is configured. And a dynamic resource scheduling and distributed fault-tolerant mechanism is introduced, so that high-throughput real-time analysis and data continuity are ensured. The real-time performance, the reliability and the intelligent level of the railway detection system can be remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of railway inspection information technology, and more specifically, to a method for real-time processing of railway inspection data. Background Technology

[0002] In existing railway inspection technologies, with the continuous increase in train speed and track density, the real-time monitoring of track and overhead contact line conditions is becoming increasingly important for safe railway operation. Track inspection vehicles, overhead contact line inspection equipment, and onboard sensors can acquire multi-source data such as track geometry parameters, overhead contact line height and pull-out value, vehicle vibration and acceleration, providing a foundation for track health status assessment.

[0003] However, existing systems mostly employ offline batch processing, requiring manual aggregation and offline analysis after data collection to generate results. This leads to a time delay between the detection data and the on-site operational status, making it difficult to promptly detect potential faults or abnormal trends. Furthermore, railway inspection data is characterized by high frequency, massive volume, and heterogeneous origins. Different inspection devices exhibit variations in data format, sampling frequency, and accuracy. Traditional single-node or centralized processing architectures struggle to complete data parsing, feature extraction, and anomaly identification within a short timeframe, easily resulting in data backlog and wasted computing resources. Simultaneously, with the continuous operation of railway inspection tasks, the system places higher demands on real-time processing, status maintenance, and dynamic scheduling capabilities for high-concurrency data streams. Existing platforms still suffer from performance bottlenecks in streaming computing and multi-node collaborative processing. More importantly, the railway inspection environment is complex and volatile; network transmission links and computing nodes may experience temporary interruptions or failures. Existing systems lack effective fault tolerance and recovery mechanisms, leading to inspection task interruptions, data loss, or failure of anomaly alarms, impacting the stability and reliability of the railway inspection system.

[0004] Therefore, there is an urgent need for a new technical solution that can realize real-time acquisition, streaming processing and anomaly response of railway inspection data under high-speed operating conditions, so as to improve the intelligence and safety of railway operation and maintenance. Summary of the Invention

[0005] In view of this, the present invention proposes a real-time processing method for railway inspection data to achieve real-time monitoring and anomaly response of track and overhead contact line conditions under high-speed train operation. This method solves the problems of data latency, low parsing efficiency, and insufficient fault tolerance in existing technologies by acquiring, transmitting, streaming, and feeding back multi-source heterogeneous data.

[0006] This invention proposes a real-time processing method for railway inspection data, comprising: Record historical operating data of several track inspection vehicles, overhead contact line inspection equipment and on-board sensors, and generate corresponding benchmark reference libraries based on the historical operating data; Several detection paths and several detection nodes are set according to the railway line; The system collects real-time geometric parameters, real-time status data, and real-time physical data of the detection path and the detection nodes, while also collecting environmental data of the area where the detection path is located. The wear rate of the detection path is determined based on the real-time geometric parameters of the detection path, and the offset of the overhead detection node before and after strong wind conditions is determined based on the real-time physical data of the detection node. Vibration characteristic data is generated by combining historical operation data and environmental data. The actual failure rate of the detection path is determined based on vibration characteristic data and wear rate. The breakage risk rate of the detection path is determined based on vibration characteristic data and the offset of the overhead detection node before and after strong wind conditions. The predicted wear rate of the detection path and the predicted offset of the overhead detection node are determined based on vibration characteristic data and environmental data, respectively. The predicted failure rate is generated based on the actual failure rate and the predicted wear rate, and the predicted fracture risk rate is generated based on the fracture risk rate and the predicted offset. The predicted failure rate and predicted fracture risk rate are matched with a benchmark database, where: If the match is successful, the corresponding physical detection damage rate will be generated; If the matching fails, the real-time status data is verified. If the verification result is normal, the predicted failure rate and the predicted fracture risk rate are recorded as stable states and stored in the benchmark database. If the verification result is abnormal, the predicted failure rate and the predicted fracture risk rate are recorded as unstable states, the corresponding physical detection damage rate is generated, and stored in the benchmark database.

[0007] Furthermore, when monitoring and detecting path security, this includes: Record the vibration characteristic data of the detection path, and take cruise photos of each detection path and overhead detection node before and after passing through the area where each detection path is located under strong wind conditions to generate path images and node images, which are respectively determined as the real-time geometric parameters of the path image or the real-time physical data of the node image. Collect wind speed, temperature, and humidity data for each detection path area as environmental data; The wear rate of the detection path is determined based on the path image; the offset of the overhead detection node before and after strong wind conditions is determined based on the node image; the vibration intensity data of the detection path is determined based on wind speed data and vibration characteristic data, including slight vibration and strong vibration; and the meteorological characteristic data of the detection path is determined based on temperature data and humidity data, including exposure data, rainfall data, icing data and low temperature data. The actual failure rate of the detection path is determined based on vibration intensity data and wear rate. The breakage risk rate of the detection path is determined based on vibration intensity data and the offset of the overhead detection node before and after strong wind conditions. The predicted wear rate of the detection path and the predicted offset of the overhead detection node are determined based on vibration intensity data and meteorological characteristic data, respectively. Fault alarms are issued based on the actual failure rate, and fracture alarms are issued based on the fracture risk rate. Among them, strong wind conditions are flowing air with wind speed data greater than or equal to the preset wind speed; The process of predicting the fracture risk rate includes: The actual failure rate and the predicted wear rate are coupled to generate the predicted failure rate; The fracture risk rate is generated by coupling the fracture risk rate and the predicted offset.

[0008] Furthermore, when determining the vibration intensity data of the detection path based on wind speed data and vibration characteristic data, this includes: If the vibration characteristic data is greater than or equal to the preset vibration threshold when the wind speed data is greater than or equal to the preset wind speed, then the vibration intensity data of the detection path is determined to be strong vibration. If the vibration characteristic data is less than the preset vibration threshold, the vibration intensity data of the detection path is determined to be slight vibration. Among them, the preset wind speed is positively correlated with the quality of the detection path, while the preset vibration threshold is negatively correlated with the quality of the detection path.

[0009] Furthermore, when determining the water vapor content in the air based on humidity data, this includes: If the humidity data is greater than or equal to the preset humidity, it is determined that the water vapor content in the air is high. If the humidity data is lower than the preset humidity, it is determined that the water vapor content in the air is low; The preset humidity is negatively correlated with air pressure.

[0010] Furthermore, when determining that the water vapor content in the air is high, the following factors are considered: Meteorological characteristic data for determining the detection path are based on temperature data, including: If the temperature data is not lower than the first preset temperature, then the meteorological characteristic data is determined to be rainfall data. If the temperature data is lower than the first preset temperature, the meteorological characteristic data is determined to be icing data; The first preset temperature is negatively correlated with atmospheric pressure.

[0011] Furthermore, when determining that the water vapor content in the air is low, the following factors are considered: Meteorological characteristic data for determining the detection path are based on temperature data, including: If the temperature data is not lower than the second preset temperature, the meteorological characteristic data is determined to be data of intense sunlight. If the temperature data is lower than the second preset temperature, the meteorological characteristic data is determined to be low temperature data; The second preset temperature is greater than the first preset temperature and is positively correlated with the duration of sunshine.

[0012] Furthermore, when determining the wear rate of the detection path based on the path image, the following steps are included: The wear rate of the detection path is determined by combining the width and length of the surface cracks in the detection path image with the diameter and length of the detection path.

[0013] Furthermore, when determining the offset of the overhead detection node before and after strong wind conditions based on the node image, the following steps are included: A rectangular coordinate system is established by selecting the position where the bottom of the overhead detection node contacts the ground as the origin. The offset before and after the strong wind condition is determined based on the top pixel coordinates of the overhead detection nodes in the node images before and after the strong wind condition.

[0014] Furthermore, when determining the fracture risk rate and wear rate, the following factors are considered: Vibration intensity data is determined as vibration parameters, meteorological characteristic data is determined as meteorological parameters, and the relationship between vibration parameters and wear rate is used to determine the actual failure rate. The fracture risk rate is determined based on the vibration parameters and the offset of the overhead detection node before and after strong wind conditions. Based on vibration intensity data and meteorological characteristic data, combined with the diameter and length of the detection path, the predicted wear rate of the detection path is determined. Based on vibration intensity data and meteorological characteristic data, combined with the offset of overhead detection nodes, the predicted offset of the detection path is determined.

[0015] Compared with existing technologies, the advantages of this invention are as follows: By recording historical operating data of track inspection vehicles, overhead contact line inspection equipment, and onboard sensors and establishing a benchmark database, the detection system possesses adaptive historical learning capabilities. This mechanism can automatically generate benchmarks based on different lines, equipment, and environmental characteristics, thereby enabling dynamic comparison and deviation identification of data during subsequent real-time detection, significantly improving the accuracy and relevance of detection results. Secondly, by real-time acquisition of geometric parameters, state data, and node physical data of the detection path, combined with synchronous analysis of environmental data, comprehensive monitoring of the detection object under multi-dimensional conditions is achieved. The system can not only evaluate key operating indicators such as track wear rate and overhead contact line offset in real time, but also capture the dynamic response characteristics of equipment under complex environmental conditions such as strong winds and temperature changes, thereby improving the comprehensiveness and robustness of detection. In addition, the multi-index risk calculation model built based on vibration characteristic data can simultaneously calculate the actual failure rate, fracture risk rate, and predicted risk value, and establish a self-correction mechanism between the predicted and measured results. This approach enables the system to identify potential fault trends in advance, achieving proactive early warning of track wear and overhead contact line offset, and significantly reducing the impact of sudden failures on railway safety. Finally, through a benchmark database matching and self-updating mechanism, this invention enables the detection system to possess self-evolving characteristics. When the predicted risk result fails to match the benchmark database, the system automatically performs real-time status verification and marks the data as stable or unstable based on the result before updating it to the benchmark database. This mechanism not only avoids false alarms or false negatives but also continuously optimizes model parameters as detection data accumulates, improving prediction accuracy and system stability over long-term operation. Attached Figure Description

[0016] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart illustrating a real-time processing method for railway inspection data provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating a real-time processing method for railway inspection data provided in an embodiment of the present invention. Detailed Implementation

[0017] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0018] like Figures 1-2 As shown in some embodiments of this application, this embodiment provides a real-time processing method for railway inspection data, including: Step S100: Record the historical operating data of several track inspection vehicles, catenary inspection equipment and on-board sensors, and generate a corresponding benchmark library based on the historical operating data; set several corresponding inspection paths and several inspection nodes according to the railway line.

[0019] Understandably, by recording historical operational data from track inspection vehicles, overhead contact line inspection equipment, and onboard sensors, a comprehensive understanding of the response patterns and characteristic distributions of different inspection devices during long-term operation can be achieved. This data includes trends in track geometry parameters, fluctuations in overhead contact line operating conditions, vehicle vibration characteristics, and the impact of environmental factors on inspection results. By extracting features and statistically modeling these historical data, a representative benchmark library can be generated. This benchmark library serves as a "health record" for railway inspection, providing a comparative reference for subsequent real-time inspections and identifying deviations in track or overhead contact line conditions from normal ranges. Secondly, based on the structural characteristics and operational features of the railway line, corresponding inspection paths and nodes are defined. Inspection paths define the inspection scope and analysis objects, while inspection nodes correspond to specific monitoring locations or key components, such as sleepers, contact points, or support points. Through hierarchical definition of inspection paths and nodes, precise spatial positioning and segmented monitoring can be achieved, making data acquisition and analysis traceable and targeted. This node-based management approach facilitates the structured organization and distributed management of inspection data, laying a technical foundation for subsequent wear analysis, risk assessment, and fault prediction. In summary, by organically combining "data modeling, path definition, and node location," railway inspection has shifted from experience-based judgment to data-driven analysis. Using historical data as its core, it generates a benchmark database and configures path nodes, enabling basic capabilities for comparative identification, trend analysis, and anomaly detection. This provides crucial support for intelligent processing and accurate early warning of railway inspection data.

[0020] Step S200: Collect real-time geometric parameters, real-time status data and real-time physical data of the detection path and the detection nodes, and collect environmental data of the area where the detection path is located; determine the wear rate of the detection path based on the real-time geometric parameters of the detection path, and determine the offset of the overhead detection nodes before and after strong wind conditions based on the real-time physical data of the detection nodes, and generate vibration characteristic data by combining historical operation data and environmental data.

[0021] Specifically, the vibration characteristic data of the detection path is recorded, and the detection path and overhead detection node are cruised and photographed before and after passing through the area where each detection path is located under strong wind conditions to generate path images and node images, which are respectively determined as the real-time geometric parameters of the path image or the real-time physical data of the node image; wind speed data, temperature data and humidity data of the area where each detection path is located are collected as environmental data.

[0022] Specifically, the wear rate of the detection path is determined based on the path image, the offset of the overhead detection node before and after strong wind conditions is determined based on the node image, the vibration intensity data of the detection path is determined based on wind speed data and vibration characteristic data, including slight vibration and strong vibration, and the meteorological characteristic data of the detection path is determined based on temperature data and humidity data, including exposure data, rainfall data, icing data and low temperature data.

[0023] Specifically, the actual failure rate of the detection path is determined based on vibration characteristic data and wear rate; the breakage risk rate of the detection path is determined based on vibration characteristic data and the offset of the overhead detection node before and after strong wind conditions; and the predicted wear rate of the detection path and the predicted offset of the overhead detection node are determined based on vibration characteristic data and environmental data, respectively.

[0024] Specifically, strong wind conditions are flowing air with wind speed data greater than or equal to the preset wind speed.

[0025] Specifically, when determining the vibration intensity data of the detection path based on wind speed data and vibration characteristic data, the following steps are taken: if the wind speed data is greater than or equal to a preset wind speed, and the vibration characteristic data is greater than or equal to a preset vibration threshold, then the vibration intensity data of the detection path is determined to be strong vibration; if the vibration characteristic data is less than the preset vibration threshold, then the vibration intensity data of the detection path is determined to be slight vibration; wherein, the preset wind speed is positively correlated with the quality of the detection path, and the preset vibration threshold is negatively correlated with the quality of the detection path.

[0026] Specifically, when determining the meteorological characteristic data of the detection path based on temperature and humidity data, the following are included: when determining the water vapor content in the air based on humidity data, the following are included: if the humidity data is greater than or equal to the preset humidity, the water vapor content in the air is determined to be high; if the humidity data is less than the preset humidity, the water vapor content in the air is determined to be low; wherein, the preset humidity is negatively correlated with air pressure.

[0027] Specifically, when determining that the water vapor content in the air is high, the method includes: determining the meteorological characteristic data of the detection path based on temperature data, wherein: if the temperature data is not less than a first preset temperature, the meteorological characteristic data is determined to be precipitation data; if the temperature data is less than the first preset temperature, the meteorological characteristic data is determined to be icing data; wherein, the first preset temperature is negatively correlated with atmospheric pressure.

[0028] Specifically, when determining that the water vapor content in the air is low, the method includes: determining the meteorological characteristic data of the detection path based on temperature data, wherein: if the temperature data is not less than the second preset temperature, the meteorological characteristic data is determined to be sunshine data; if the temperature data is less than the second preset temperature, the meteorological characteristic data is determined to be low temperature data; wherein the second preset temperature is greater than the first preset temperature and is positively correlated with the sunshine duration.

[0029] Specifically, when determining the wear rate of the detection path based on the path image, it includes: the surface crack width and length of the detection path in the detection path image, and determining the wear rate of the detection path by combining the diameter and length of the detection path.

[0030] Specifically, when determining the offset of the overhead detection node before and after strong wind conditions based on the node image, the process includes: selecting the position where the bottom of the overhead detection node contacts the ground as the origin of the coordinate system to establish a rectangular coordinate system; and determining the offset before and after strong wind conditions based on the top pixel coordinates of the overhead detection node in the node images before and after strong wind conditions.

[0031] Understandably, by fusing multi-source sensing data and analyzing environmental correlations, a dynamic monitoring and risk assessment model for railway inspection paths and overhead inspection nodes under complex weather conditions is established. This enables the quantitative calculation and prediction of wear, offset, and vibration characteristics, thus providing technical support for intelligent identification and safety early warning of railway equipment status. First, a multi-dimensional data fusion system is constructed by synchronously collecting real-time geometric parameters, status data, and node physical data of the inspection path, combined with regional environmental data (including wind speed, temperature, and humidity). The system not only utilizes data from track inspection vehicles and sensors for real-time numerical measurements but also acquires path and node images through cruise photography, extracting geometric change information from a visual perspective, achieving a complementary fusion of "numerical detection" and "image recognition." This multi-modal data fusion principle allows the system to comprehensively reflect the true physical characteristics of the inspection path and nodes under different operating conditions. Second, through the analysis of path and node images, geometric feature extraction and coordinate difference principles are applied. In path analysis, the wear rate is calculated by identifying crack width and length in the image and combining it with the diameter and length of the track, thus achieving a quantitative assessment of the wear degree. In node analysis, the node offset is obtained by establishing a coordinate system based on the bottom of the node and calculating the change in the pixel coordinates of the top of the node before and after strong wind conditions. This principle based on image geometry calculation enables the system to accurately acquire structural deformation data under non-contact conditions, making it suitable for monitoring needs in high-speed operation and harsh environments. Furthermore, in vibration and environmental feature analysis, this technology constructs an environmental perception model and threshold determination mechanism through the logical coupling of wind speed, temperature, humidity, and vibration feature data. By setting preset wind speed and vibration thresholds, the vibration intensity of the detection path under strong wind conditions is determined, thereby identifying the changing trend of the structural stress state. Simultaneously, based on temperature and humidity data, the water vapor content in the air and meteorological types (such as sun exposure, rainfall, icing, and low temperature) are inferred, achieving intelligent classification of climate conditions. This principle, based on correlation modeling between physical environmental parameters and structural response, can effectively reveal the influence of the environment on track wear and node offset. Finally, by comprehensively analyzing vibration characteristic data with wear rate, offset, and meteorological characteristics, the system utilizes a data-driven causal relationship model to determine the actual failure rate, fracture risk rate, and predicted trend. This model uses real-time data as input and historical patterns as a reference to achieve dynamic prediction and multi-factor analysis of potential failures, thereby accurately assessing the stability and risk level of railway structures under complex conditions such as strong winds and extreme temperatures.

[0032] Step S300: Generate a predicted failure rate based on the actual failure rate and the predicted wear rate, and generate a predicted fracture risk rate based on the fracture risk rate and the predicted offset.

[0033] Specifically, determining the fracture risk rate and wear rate includes: identifying vibration intensity data as vibration parameters, identifying meteorological characteristic data as meteorological parameters, and determining the actual failure rate by combining the relationship between vibration parameters and wear rate; determining the fracture risk rate based on the vibration parameters and the offset of overhead detection nodes before and after strong wind conditions; determining the predicted wear rate of the detection path based on the vibration intensity data and meteorological characteristic data combined with the diameter and length of the detection path; and determining the predicted offset of the detection path based on the vibration intensity data and meteorological characteristic data combined with the offset of overhead detection nodes.

[0034] Understandably, by establishing a quantitative correlation between track wear, node offset, and failure risk through multi-parameter coupling analysis and dynamic predictive modeling, the actual failure rate and predicted failure rate of railway inspection paths can be linked and extrapolated, thereby achieving a forward-looking risk assessment of railway structural status. First, using vibration intensity data and meteorological characteristic data as core inputs, vibration parameters and meteorological parameters are constructed as key features describing external disturbances and environmental impacts. Vibration parameters reflect the dynamic forces on the track and overhead nodes under operating or strong wind conditions, while meteorological parameters reflect the long-term impact of external conditions such as temperature, humidity, and wind speed on the material structure. By combining these two types of parameters with physical characteristics such as wear rate and node offset of the inspection path, multi-dimensional modeling of the structural operating state can be achieved, thereby revealing the physical coupling law between environment, vibration, and structural response. Second, in determining the actual failure rate and fracture risk rate, modeling is based on the synergistic relationship between vibration parameters and structural characteristics. Specifically, the relationship between vibration parameters and wear rate reflects the degradation rate of the track under stress. Their coupling is used to calculate the actual failure rate, reflecting the health status of the structure under current operating conditions. Conversely, the coupling between vibration parameters and overhead node offset is used to calculate the fracture risk rate, assessing the likelihood of node instability under strong winds or abnormal vibrations. This principle embodies the combination of dynamic mechanical response analysis and empirical feature modeling, enabling the system to directly infer risk levels through the correlation between changes in physical parameters. Furthermore, in the calculation of predicted wear rate and predicted offset, a multi-factor fusion model of structural geometric features (such as track diameter and length) and environmental parameters (meteorological characteristics) is introduced. The system establishes a predictive model based on actual measurements and environmental trends by applying vibration intensity data and meteorological characteristic data to the structural geometric parameters. This method is a typical example of multivariate fitting and trend extrapolation, capable of predicting future structural wear and offset trends based on the current state, achieving early quantitative assessment of potential failures. Finally, based on the above modeling results, the actual failure rate is fused with the predicted wear rate, and the fracture risk rate is fused with the predicted offset to generate the predicted failure rate and predicted fracture risk rate, respectively. This process embodies the principles of time series reasoning and state estimation, namely, by linking historical and real-time states, it enables dynamic prediction and trend evolution analysis of structural health status.

[0035] Step S400: Match the predicted failure rate and the predicted fracture risk rate with the benchmark database. If the match is successful, the corresponding physical detection damage rate is generated. If the match fails, the real-time status data is verified. If the verification result is normal, the predicted failure rate and the predicted fracture risk rate are recorded as a stable state and stored in the benchmark database. If the verification result is abnormal, the predicted failure rate and the predicted fracture risk rate are recorded as an unstable state, the corresponding physical detection damage rate is generated, and stored in the benchmark database.

[0036] Understandably, based on the failure rate and fracture risk rate output by the predictive model, a preliminary judgment of the current state of the equipment is achieved by matching it with historical data in a benchmark database. The principle is that the benchmark database stores samples of physical damage rates that have been monitored and verified over a long period, covering different equipment types, operating conditions, and damage modes. Through the matching mechanism, it is possible to quickly identify whether the predicted result is consistent with existing patterns, thereby determining its corresponding physical damage level. Secondly, when the predicted result cannot find a match in the benchmark database, the system enters a self-verification phase, that is, verifying the real-time status data. The technical principle of this step is to use multi-source sensor data (such as vibration, temperature, current, voltage, etc.) for cross-validation to determine whether the predicted value reflects the actual state. If the verification result indicates that the equipment is operating normally, it means that although the predictive model outputs a high risk, no actual failure has occurred. At this time, the system marks the predicted result as a stable state and writes it into the benchmark database for future model continuous correction and expansion. If the verification result shows an anomaly, it means that the predictive model has successfully identified potential failures or structural fracture risks. At this point, the system marks the prediction result as unstable and generates the actual physical detection damage rate, which is simultaneously stored in the benchmark database. This process embodies a feedback-based self-learning mechanism: finally, while continuously accumulating new samples, the model parameters and the control threshold are dynamically adjusted to gradually optimize the prediction performance.

[0037] To enable those skilled in the art to fully understand and implement this invention, the specific implementation principle of this invention is further explained below in conjunction with a specific application scenario.

[0038] In the actual operation of a railway inspection system, the first step is to generate a benchmark database. The system collects historical operational data through track inspection vehicles, overhead contact line inspection equipment, and onboard sensors, and categorizes this data according to time series. For example, for a specific inspection path, the system records multi-dimensional information such as track geometry parameters, overhead contact line status data, and environmental variables at different time periods. After standardization, this data forms the benchmark database, providing a reference for subsequent real-time data analysis. The generation process of the benchmark database must ensure the completeness and accuracy of the data to adapt to changes in railway line conditions. For example, for a specific inspection path, its benchmark database should include key indicators such as wear rate, vibration characteristics, and meteorological characteristics at different time periods.

[0039] During the real-time data acquisition phase, the system collects real-time geometric parameters and status data of the detection path through multiple sensor nodes distributed along the railway line, while simultaneously acquiring real-time physical data from overhead detection nodes. For example, anemometers are installed near the detection path to obtain wind speed data; temperature and humidity sensors are responsible for collecting temperature and humidity data. These sensor nodes are all equipped with high-precision sensors to ensure data reliability. Furthermore, the system employs a distributed streaming processing framework to perform parallel computation and feature extraction on large-scale detection data, thereby improving data acquisition efficiency. For instance, when the wind speed data in an area along a certain detection path reaches a preset wind speed, the system immediately triggers the strong wind condition judgment logic and initiates the relevant analysis process.

[0040] During the calculation of vibration characteristic data, the system determines the wear rate based on the real-time geometric parameters of the detection path and determines its offset by combining the real-time physical data of the overhead detection nodes. For example, the path image is used to detect the crack width and length on the surface of the detection path, and the wear rate is calculated by combining the diameter and length of the detection path. The node image is used to select the position where the bottom of the overhead detection node contacts the ground as the origin to establish a rectangular coordinate system, and the offset is calculated based on the pixel coordinates of the top of the overhead detection node in the node images before and after strong wind conditions. The calculation of vibration characteristic data also needs to combine historical operating data and environmental data to ensure its accuracy and comprehensiveness. For example, when the detection path is under strong wind conditions, the system will determine whether the vibration intensity data belongs to slight vibration or strong vibration based on wind speed data and vibration characteristic data.

[0041] In the prediction process of failure rate and fracture risk rate, the system determines the actual failure rate of the detection path based on vibration characteristic data and wear rate, and determines the fracture risk rate of the detection path based on vibration characteristic data and the offset of overhead detection nodes before and after strong wind conditions. Furthermore, the system also combines vibration characteristic data and environmental data to predict the predicted wear rate of the detection path and the predicted offset of the overhead detection nodes, respectively. For example, when environmental data shows high water vapor content in the air and the temperature data is lower than a first preset temperature, the system will determine that the meteorological characteristic data of the detection path is icing data and adjust the calculation models for predicted wear rate and predicted offset accordingly. The actual failure rate and predicted wear rate are coupled to generate the predicted failure rate, and the fracture risk rate and predicted offset are coupled to generate the predicted fracture risk rate.

[0042] During the results feedback phase, the system matches the predicted failure rate and predicted fracture risk rate with a benchmark database. If the match is successful, a corresponding physical damage rate is generated; if the match fails, real-time status data is verified. If the verification result is normal, the predicted failure rate and predicted fracture risk rate are recorded as a stable state and stored in the benchmark database; if the verification result is abnormal, the predicted failure rate and predicted fracture risk rate are recorded as an unstable state, and a corresponding physical damage rate is generated and stored in the benchmark database. This process ensures the dynamic updating capability of the benchmark database, enabling it to adapt to changes in the railway line's condition.

[0043] During the monitoring of the detection path safety, the system records the vibration characteristic data of the detection path and, before and after strong winds pass through the areas where each detection path is located, performs cruise photography to generate path images and node images for each detection path and overhead detection node. The path images and node images are used as part of the real-time geometric parameters and real-time physical data, respectively, for further analysis of the detection path's condition. For example, path images can be used to detect the crack width and length on the detection path surface, while node images are used to calculate the offset of the overhead detection nodes. In addition, the system also collects wind speed, temperature, and humidity data for the areas where each detection path is located as environmental data, and uses this data to determine the vibration intensity data and meteorological characteristic data of the detection path.

[0044] During the determination of vibration intensity data, the system performs a comprehensive analysis based on wind speed data and vibration characteristic data. For example, when the wind speed data is greater than or equal to a preset wind speed and the vibration characteristic data is greater than or equal to a preset vibration threshold, the system determines the vibration intensity data of the detection path to be strong vibration; otherwise, it determines it to be slight vibration. The preset wind speed is positively correlated with the quality of the detection path, while the preset vibration threshold is negatively correlated with the quality of the detection path. This determination logic ensures that the system can flexibly adjust the calculation model of vibration intensity data according to the characteristics of different detection paths.

[0045] In the process of determining meteorological characteristic data, the system first determines the water vapor content in the air based on humidity data. For example, when the humidity data is greater than or equal to a preset humidity level, the system determines that the water vapor content in the air is high; conversely, it determines that the water vapor content in the air is low. The preset humidity level is negatively correlated with air pressure. When the system determines that the water vapor content in the air is high, it further determines the meteorological characteristic data for the detection path based on temperature data. For example, when the temperature data is not less than a first preset temperature, the system determines the meteorological characteristic data as precipitation data; when the temperature data is less than the first preset temperature, the system determines the meteorological characteristic data as icing data. The first preset temperature is negatively correlated with atmospheric pressure. When the system determines that the water vapor content in the air is low, it also determines the meteorological characteristic data for the detection path based on temperature data. For example, when the temperature data is not less than a second preset temperature, the system determines the meteorological characteristic data as sunshine data; when the temperature data is less than the second preset temperature, the system determines the meteorological characteristic data as low temperature data. The second preset temperature is greater than the first preset temperature and is positively correlated with sunshine duration.

[0046] In predicting the fracture risk rate, the system couples the actual failure rate and the predicted wear rate to generate the predicted failure rate, and couples the fracture risk rate and the predicted offset to generate the predicted fracture risk rate. This process is achieved through statistical analysis, machine learning, or deep learning algorithms, ensuring that the system can accurately predict abnormal data or potential faults. For example, when the vibration intensity data of the detection path indicates strong vibration and the meteorological characteristic data indicates icing, the system will trigger an alarm and push relevant information to the monitoring center or maintenance personnel.

[0047] Furthermore, the system incorporates dynamic resource scheduling and distributed fault tolerance mechanisms, improving resource utilization through dynamic load balancing and automatic scaling of computing nodes. For example, when a computing node fails, the system utilizes checkpointing and RDD recalculation mechanisms to ensure the continuity and accuracy of data processing. Multi-node collaborative processing enables high-throughput real-time analysis of large-scale inspection data, significantly improving the real-time performance, reliability, and intelligence level of the railway inspection system.

[0048] The above embodiments describe the technical solution of the present invention in detail through specific steps and logical relationships, ensuring that those skilled in the art can implement the technology according to the contents of the specification.

[0049] In the above embodiments, by recording historical operating data of the track inspection vehicle, overhead contact line inspection equipment, and onboard sensors and establishing a benchmark database, the inspection system possesses adaptive historical learning capabilities. This mechanism can automatically generate benchmarks based on different lines, equipment, and environmental characteristics, thereby enabling dynamic data comparison and deviation identification during subsequent real-time inspections, significantly improving the accuracy and relevance of inspection results. Secondly, by real-time acquisition of geometric parameters, state data, and node physical data of the inspection path, combined with synchronous analysis of environmental data, comprehensive monitoring of the inspection object under multi-dimensional conditions is achieved. The system can not only assess key operating indicators such as track wear rate and overhead contact line offset in real time, but also capture the dynamic response characteristics of equipment under complex environmental conditions such as strong winds and temperature changes, thereby improving the comprehensiveness and robustness of inspections. Furthermore, the multi-index risk calculation model built based on vibration characteristic data can simultaneously calculate the actual failure rate, fracture risk rate, and predicted risk value, and establish a self-correction mechanism between the predicted and measured results. This approach enables the system to identify potential fault trends in advance, achieving proactive early warning of track wear and overhead contact line offset, significantly reducing the impact of sudden failures on railway safety. Finally, through a benchmark database matching and self-updating mechanism, this invention enables the detection system to possess self-evolving characteristics. When the predicted risk result fails to match the benchmark database, the system automatically performs real-time status verification and marks the data as stable or unstable based on the result before updating it to the benchmark database. This mechanism not only avoids false alarms or false negatives but also continuously optimizes model parameters as detection data accumulates, improving prediction accuracy and system stability over long-term operation.

[0050] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0051] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0052] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0053] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for real-time processing of railway inspection data, characterized in that, include: Record historical operating data of several track inspection vehicles, overhead contact line inspection equipment and on-board sensors, and generate corresponding benchmark reference libraries based on the historical operating data; Several detection paths and several detection nodes are set according to the railway line; The system collects real-time geometric parameters, real-time status data, and real-time physical data of the detection path and the detection nodes, while also collecting environmental data of the area where the detection path is located. The wear rate of the detection path is determined based on the real-time geometric parameters of the detection path, and the offset of the overhead detection node before and after strong wind conditions is determined based on the real-time physical data of the detection node. Vibration characteristic data is generated by combining historical operation data and environmental data. The actual failure rate of the detection path is determined based on vibration characteristic data and wear rate. The breakage risk rate of the detection path is determined based on vibration characteristic data and the offset of the overhead detection node before and after strong wind conditions. The predicted wear rate of the detection path and the predicted offset of the overhead detection node are determined based on vibration characteristic data and environmental data, respectively. The predicted failure rate is generated based on the actual failure rate and the predicted wear rate, and the predicted fracture risk rate is generated based on the fracture risk rate and the predicted offset. The predicted failure rate and predicted fracture risk rate are matched with a benchmark database, where: If the match is successful, the corresponding physical detection damage rate will be generated; If the matching fails, the real-time status data is verified. If the verification result is normal, the predicted failure rate and the predicted fracture risk rate are recorded as stable states and stored in the benchmark database. If the verification result is abnormal, the predicted failure rate and the predicted fracture risk rate are recorded as unstable states, the corresponding physical detection damage rate is generated, and stored in the benchmark database.

2. The real-time processing method for railway inspection data as described in claim 1, characterized in that, When monitoring and detecting path security, the following are included: Record the vibration characteristic data of the detection path, and take cruise photos of each detection path and overhead detection node before and after passing through the area where each detection path is located under strong wind conditions to generate path images and node images, which are respectively determined as the real-time geometric parameters of the path image or the real-time physical data of the node image. Collect wind speed, temperature, and humidity data for each detection path area as environmental data; The wear rate of the detection path is determined based on the path image; the offset of the overhead detection node before and after strong wind conditions is determined based on the node image; the vibration intensity data of the detection path is determined based on wind speed data and vibration characteristic data, including slight vibration and strong vibration; and the meteorological characteristic data of the detection path is determined based on temperature data and humidity data, including exposure data, rainfall data, icing data and low temperature data. Among them, strong wind conditions are flowing air with wind speed data greater than or equal to the preset wind speed.

3. The real-time processing method for railway inspection data as described in claim 2, characterized in that, When determining the vibration intensity data for the detection path based on wind speed data and vibration characteristic data, the following should be included: If the vibration characteristic data is greater than or equal to the preset vibration threshold when the wind speed data is greater than or equal to the preset wind speed, then the vibration intensity data of the detection path is determined to be strong vibration. If the vibration characteristic data is less than the preset vibration threshold, the vibration intensity data of the detection path is determined to be slight vibration. Among them, the preset wind speed is positively correlated with the quality of the detection path, while the preset vibration threshold is negatively correlated with the quality of the detection path.

4. The real-time processing method for railway inspection data as described in claim 3, characterized in that, When determining the water vapor content in the air based on humidity data, the following should be included: If the humidity data is greater than or equal to the preset humidity, it is determined that the water vapor content in the air is high. If the humidity data is lower than the preset humidity, it is determined that the water vapor content in the air is low; The preset humidity is negatively correlated with air pressure.

5. The real-time processing method for railway inspection data as described in claim 4, characterized in that, When determining that the water vapor content in the air is high, the following are included: Meteorological characteristic data for determining the detection path are based on temperature data, including: If the temperature data is not lower than the first preset temperature, then the meteorological characteristic data is determined to be rainfall data. If the temperature data is lower than the first preset temperature, the meteorological characteristic data is determined to be icing data; The first preset temperature is negatively correlated with atmospheric pressure.

6. The real-time processing method for railway inspection data as described in claim 4, characterized in that, When determining that the water vapor content in the air is low, the following are included: Meteorological characteristic data for determining the detection path are based on temperature data, including: If the temperature data is not lower than the second preset temperature, the meteorological characteristic data is determined to be data of intense sunlight. If the temperature data is lower than the second preset temperature, the meteorological characteristic data is determined to be low temperature data; The second preset temperature is greater than the first preset temperature and is positively correlated with the duration of sunshine.

7. The real-time processing method for railway inspection data as described in claim 2, characterized in that, When determining the wear rate of the detection path based on the path image, the following is included: The wear rate of the detection path is determined by combining the width and length of the surface cracks in the detection path image with the diameter and length of the detection path.

8. The real-time processing method for railway inspection data as described in claim 2, characterized in that, When determining the offset of overhead detection nodes before and after strong wind conditions based on node images, the following is included: A rectangular coordinate system is established by selecting the position where the bottom of the overhead detection node contacts the ground as the origin. The offset before and after the strong wind condition is determined based on the top pixel coordinates of the overhead detection nodes in the node images before and after the strong wind condition.

9. The real-time processing method for railway inspection data as described in claim 2, characterized in that, Determining the fracture risk rate and wear rate includes: Vibration intensity data is determined as vibration parameters, meteorological characteristic data is determined as meteorological parameters, and the relationship between vibration parameters and wear rate is used to determine the actual failure rate. The fracture risk rate is determined based on the vibration parameters and the offset of the overhead detection node before and after strong wind conditions. Based on vibration intensity data and meteorological characteristic data, combined with the diameter and length of the detection path, the predicted wear rate of the detection path is determined. Based on vibration intensity data and meteorological characteristic data, combined with the offset of overhead detection nodes, the predicted offset of the detection path is determined.

Images