Intelligent equipment operation and maintenance method and system based on railway operation and maintenance

Abstract

The invention relates to the technical field of railway intelligent operation and maintenance, in particular to an equipment intelligent operation and maintenance method and system based on railway operation and maintenance. Performing data cleaning and self-adaptive multi-stage encryption on the multi-dimensional operation and maintenance sensing data, transmitting the multi-dimensional operation and maintenance sensing data to a master station by adopting a dynamic transmission protocol switching algorithm, and performing restoration and data quality verification on a security data packet received by the master station to obtain cleaned data with spatial association; a satellite positioning technology and a three-dimensional visualization technology are adopted to obtain a dynamically interactive equipment operation and maintenance digital twin environment; and carrying out equipment potential fault identification by adopting a time sequence feature extraction and multi-factor correlation analysis algorithm. Through the big data analysis technology, the operation and maintenance efficiency of the railway equipment is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of railway intelligent operation and maintenance, and in particular to an equipment intelligent operation and maintenance method and system based on railway operation and maintenance. Background Art

[0002] With the rapid expansion of the railway network and the increase in operational complexity, railway monitoring equipment has been increasingly used. Railway monitoring equipment includes temperature sensors, humidity sensors, smoke concentration sensors, and video cameras installed in the machine rooms along the railway. These devices monitor the temperature, humidity, smoke concentration, and environmental video of the machine rooms along the railway online to ensure that the environment of the server and control room in the railway machine room is in a normal state, thereby maintaining the normal dispatch and operation of the railway system. However, the current operation and maintenance of railway monitoring equipment mainly relies on manual regular inspections, which have the following technical problems: First, most railway monitoring equipment is distributed over a wide geographical range and in large numbers, and manual inspections are inefficient and costly. Taking ordinary railways as an example, there are a large number of machine room facilities distributed on hundreds of kilometers of railway lines. Traditional inspection methods require a lot of manpower and material resources, and it is difficult to achieve all-weather monitoring. Second, since railway lines often cross various terrains, the network status of the monitoring equipment in the railway machine rooms along the railway is different. For example, the network environment of the monitoring equipment in the plain area is often good, while the network environment of the monitoring equipment in the mountainous area is often poor and fluctuates greatly. How to optimize the data transmission according to the network environment of the monitoring equipment and the attributes of the monitoring data is also a difficult point that needs to be studied. Third, there may be some potential faults in the operation of railway monitoring equipment. How to use monitoring data to identify potential faults is a problem that needs to be studied.

[0003] In addition, with the development of modern information technologies such as the Internet of Things, big data, and artificial intelligence, railway equipment operation and maintenance technologies are also facing the need for upgrading. Relevant research institutions and enterprises at home and abroad have begun to explore equipment monitoring solutions based on the Internet of Things, fault prediction models based on big data, and monitoring platforms based on three-dimensional visualization, but these technologies have not yet formed a complete solution, and there are problems such as insufficient data security transmission, insufficient correlation analysis of environmental factors, and low accuracy in identifying potential faults.

[0004] In summary, there is an urgent need for an intelligent equipment operation and maintenance method and system based on railway operation and maintenance to solve the above technical problems. Summary of the invention

[0005] (1) Technical issues to be resolved The purpose of the present invention is to provide an intelligent operation and maintenance method and system for equipment based on railway operation and maintenance, so as to realize intelligent operation and maintenance and potential fault identification of railway monitoring equipment.

[0006] (2) Technical solution To achieve the above object, the present invention provides an equipment intelligent operation and maintenance method based on railway operation and maintenance, the method comprising the following steps: S1, collect multi-dimensional operation and maintenance perception data, including equipment environment data and operation status data; the equipment environment data includes temperature data, humidity data, smoke concentration data, and video image data; the operation status data includes equipment operation parameters, alarm information, and historical operation and maintenance records; the equipment operation parameters include equipment CPU usage.

[0007] S2, cleans the multi-dimensional operation and maintenance perception data, performs adaptive multi-level encryption according to the sensitivity level and importance of the data to obtain a secure data packet, and uses a dynamic transmission protocol switching algorithm to transmit the secure data packet to the main station according to the stability and bandwidth conditions of the network transmission link, restores and verifies the data quality of the secure data packet received by the main station, and obtains cleaned data with spatial correlation.

[0008] S3, based on the cleaned data with spatial correlation, uses satellite positioning technology and three-dimensional visualization technology to obtain a dynamic and interactive equipment operation and maintenance digital twin environment.

[0009] S4 uses time series feature extraction and multi-factor correlation analysis algorithms to identify potential equipment failures based on pre-set historical failure modes and cleaned data with spatial correlation, and diagnoses, processes offline, and repairs equipment with potential failures.

[0010] Furthermore, the method of performing data cleaning on the multi-dimensional operation and maintenance perception data, performing adaptive multi-level encryption according to the sensitivity level and importance of the data to obtain a secure data packet, transmitting the secure data packet to the main station using a dynamic transmission protocol switching algorithm according to the stability and bandwidth conditions of the network transmission link, restoring and verifying the data quality of the secure data packet received by the main station, and obtaining the cleaned data with spatial association includes: According to the null values, duplicate values ​​and outliers existing in the multi-dimensional operation and maintenance perception data, a statistical filtering algorithm and an intelligent value filling method are used to obtain the second temperature data, second humidity data, second smoke concentration data, second video image data, second equipment operation parameters, second alarm information, and second historical operation and maintenance records with unified format and complete data; according to the sensitivity level and importance of the data, the second temperature data, second humidity data, and second smoke concentration data are divided into three-level sensitive data, the second video image data and the second historical operation and maintenance records are divided into two-level sensitive data, and the second equipment operation parameters and the second alarm information are divided into first-level sensitive data; an adaptive multi-level encryption method is used, RSA asymmetric encryption is used for the first-level sensitive data to obtain first-level encrypted data, AES symmetric encryption is used for the second-level sensitive data to obtain second-level encrypted data, and lightweight encryption is used for the third-level sensitive data to obtain third-level encrypted data; the first-level encrypted data, the second-level encrypted data, and the third-level encrypted data are combined to obtain a secure data packet.

[0011] According to the stability and bandwidth conditions of the network transmission link, a dynamic transmission protocol switching algorithm is used to transmit the security data packet to the main station; based on the security data packet received by the main station, a decryption algorithm is used to calculate the restored original data information; according to the preset data quality verification algorithm, the integrity rate of the restored original data information is evaluated; data with an integrity rate less than a preset integrity rate threshold is intelligently supplemented to obtain valid restored data; based on the valid restored data and the spatial location information of the computer room, a geographic coordinate mapping algorithm is used to obtain cleaned data with spatial association that includes valid restored data and spatial location.

[0012] Furthermore, the method of transmitting the security data packet to the master station by using a dynamic transmission protocol switching algorithm according to the stability and bandwidth conditions of the network transmission link includes: The signal strength, packet loss rate and delay time of the network transmission link are detected in real time with the preset first period as the sliding time window, and the network environment evaluation algorithm is used to obtain the network environment quality level within the sliding time window; the network environment quality level is one of the first quality level, the second quality level and the third quality level; when the network environment quality level is the first quality level, the first level encrypted data, the second level encrypted data and the third level encrypted data are transmitted simultaneously by the standard data transmission protocol; when the network environment quality level is the second quality level, the first level encrypted data is first transmitted by the high-reliability transmission protocol, and after the transmission is completed, the second level encrypted data and the third level encrypted data are transmitted in sequence; when the network environment quality level is the third quality level, the first level encrypted data is transmitted by the lightweight transmission protocol, and the second level encrypted data and the third level encrypted data are temporarily stored in the local cache, and the network environment quality level is upgraded to the second quality level or the first quality level before being transmitted.

[0013] Furthermore, the method of detecting the signal strength, packet loss rate and delay time of the network transmission link in real time using the preset first period as the sliding time window, and using the network environment evaluation algorithm to obtain the network environment quality level within the sliding time window includes: In the sliding time window, the signal strength, packet loss rate and delay time of the network transmission link are sampled in real time at every preset first sampling interval to obtain a signal strength sample set, a packet loss rate sample set and a delay time sample set in the sliding time window; the average value of the signal strength sample set is calculated to obtain the average signal strength, the average value of the packet loss rate sample set is calculated to obtain the average packet loss rate, and the average value of the delay time sample set is calculated to obtain the average delay time.

[0014] According to a preset signal strength reference range, a forward correlation mapping is used to normalize the average signal strength into a signal strength index; according to a preset packet loss rate reference range, a preset first reverse correlation mapping is used to normalize the average packet loss rate into a packet loss rate index; according to a preset delay time reference range, a preset second reverse correlation mapping is used to normalize the average delay time into a delay time index; the signal strength index, the packet loss rate index and the delay time index are multiplied by preset weight coefficients respectively and then summed to obtain a comprehensive network environment index; the comprehensive network environment index is compared with a preset first scoring threshold and a preset second scoring threshold to obtain a network environment quality level within a sliding time window.

[0015] The average resource utilization of the network transmission link is counted at intervals of a preset second period, and the first scoring threshold and the second scoring threshold are dynamically adjusted according to the average resource utilization; the second period is an integer multiple of the first period.

[0016] Furthermore, the method of comparing the network environment comprehensive index with the preset first scoring threshold and the second scoring threshold to obtain the network environment quality level within the sliding time window includes: When the comprehensive network environment index is greater than or equal to the first scoring threshold, the network environment quality level is level one; when the comprehensive network environment index is less than the first scoring threshold and greater than or equal to the second scoring threshold, the network environment quality level is level two; when the comprehensive network environment index is less than the second scoring threshold, the network environment quality level is level three.

[0017] Furthermore, the method of obtaining a dynamic interactive equipment operation and maintenance digital twin environment by using satellite positioning technology and three-dimensional visualization technology based on the cleaned data with spatial association includes: Beidou satellite positioning technology and geographic surveying and mapping technology are used to construct a three-dimensional spatial basic model of the railway operation and maintenance area; personnel location data are obtained based on the positioning devices worn by the operation and maintenance personnel; the three-dimensional spatial basic model is spatially aligned with the cleaned data and personnel location data with spatial association to obtain a basic visualization environment with geographic reference.

[0018] Extract valid restored data from the cleaned data with spatial association; classify the data corresponding to temperature data, humidity data, smoke concentration data, and video image data in the valid restored data as valid environmental monitoring data; classify the data corresponding to equipment operating parameters, alarm information, and historical operation and maintenance records in the valid restored data as valid equipment status data; use WebGL three-dimensional visualization technology to establish an association mapping between the valid equipment status data and the geometric entities of the corresponding equipment in the three-dimensional space basic model to obtain a digital twin model of the equipment; map the effective environmental monitoring data to the three-dimensional space basic model through visualization technology to obtain an environmental data visualization layer; map the personnel location data to the three-dimensional space basic model to obtain dynamic personnel icon data.

[0019] The equipment digital twin model, environmental data visualization layer, and dynamic personnel icon data are combined to obtain a dynamic and interactive equipment operation and maintenance digital twin environment.

[0020] Furthermore, the method of using time series feature extraction and multi-factor correlation analysis algorithm to identify potential equipment faults based on pre-set historical fault modes and cleaned data with spatial correlation, and diagnosing, offline processing and repairing equipment with potential faults includes: The cleaned data with spatial association are sorted according to time series, the time series features of effective environmental monitoring data and effective equipment status data are extracted, the time series features are matched with the feature templates in the preset historical fault mode library, and the matching similarity is calculated.

[0021] The device corresponding to the data whose matching similarity is greater than a preset first fault threshold is marked as a potential fault device; the potential fault device indicates a device that has been identified to have a potential fault; based on the computer room space location information and personnel location data corresponding to the potential fault device, the operation and maintenance personnel who is closest to the potential fault device is matched and recorded as the first operation and maintenance personnel; the computer room space location information corresponding to the potential fault device is pushed to the first operation and maintenance personnel; and the first operation and maintenance personnel performs offline processing and repair on the potential fault device.

[0022] Furthermore, the method of sorting the cleaned data with spatial association according to time series, extracting the time series features of effective environmental monitoring data and effective equipment status data, matching the time series features with feature templates in a preset historical fault mode library, and calculating the matching similarity includes: The cleaned data with spatial correlation are sorted according to the time series, and the preset second sliding time window is used to extract the temperature change rate, humidity change rate, smoke concentration change rate, video image gray value change rate, equipment operation parameter fluctuation rate, and periodic spectrum as time series features.

[0023] The time series features are standardized to obtain standardized time series features; based on the standardized time series features and the feature templates in the historical fault mode library, the Euclidean distance algorithm and the pre-set third reverse correlation function are used to perform similarity calculation to obtain the first similarity; the temperature change rate, the humidity change rate, the smoke concentration change rate, and the equipment operation parameter fluctuation rate are associated with pattern recognition to obtain the second similarity; the second similarity represents the degree of conformity between the relationship between the temperature change rate, the humidity change rate, the smoke concentration change rate, and the equipment operation parameter fluctuation rate and the physical law.

[0024] A weighted average of the second similarity and the first similarity is taken as the matching similarity.

[0025] Furthermore, the method of performing correlation pattern recognition on the temperature change rate, the humidity change rate, the smoke concentration change rate, and the equipment operation parameter fluctuation rate to obtain the second similarity includes: According to the corresponding time segment in the seasonal feature matching library to which the current time belongs, obtain the preset benchmark functional relationship between the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate; calculate the degree of deviation of the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate from the benchmark functional relationship, and obtain a quantitative indicator characterizing the degree to which the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate conform to physical laws, which is recorded as the second similarity; the value of the second similarity is between 0 and 1; the seasonal feature matching library is obtained by pre-setting, and represents the typical functional relationship between the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate in the pre-divided time segment.

[0026] Based on the same inventive concept, on the other hand, the present invention also provides an intelligent equipment operation and maintenance system based on railway operation and maintenance, and the system includes: a data acquisition module, a data cleaning module, a digital twin module and a fault identification module connected in sequence.

[0027] The data acquisition module is used to collect multi-dimensional operation and maintenance perception data, including equipment environment data and operation status data; the equipment environment data includes temperature data, humidity data, smoke concentration data, and video image data; the operation status data includes equipment operation parameters, alarm information, and historical operation and maintenance records; the equipment operation parameters include equipment CPU usage.

[0028] The data cleaning module is used to clean the multi-dimensional operation and maintenance perception data, perform adaptive multi-level encryption according to the sensitivity level and importance of the data to obtain a secure data packet, and use a dynamic transmission protocol switching algorithm to transmit the secure data packet to the main station according to the stability and bandwidth conditions of the network transmission link, and restore and verify the data quality of the secure data packet received by the main station to obtain cleaned data with spatial correlation.

[0029] The digital twin module is used to obtain a dynamic and interactive equipment operation and maintenance digital twin environment based on the cleaned data with spatial association, using satellite positioning technology and three-dimensional visualization technology.

[0030] The fault identification module is used to use time series feature extraction and multi-factor correlation analysis algorithms to identify potential equipment faults based on pre-set historical fault patterns and cleaned data with spatial correlation, and to diagnose, offline process and repair equipment identified to have potential faults.

[0031] (3) Beneficial effects Compared with the prior art, the present invention has the following beneficial effects: By using satellite positioning technology and three-dimensional visualization technology, a dynamic and interactive digital twin environment for equipment operation and maintenance is obtained, realizing interactive perception of equipment status.

[0032] By adopting time series feature extraction and multi-factor correlation analysis algorithms, automatic identification of potential equipment failures can be achieved, reducing the labor cost of equipment operation and maintenance and improving operation and maintenance efficiency.

[0033] Through the dynamic transmission protocol switching algorithm based on the network environment quality level, a data transmission scheme that adapts to different network environments is obtained, achieving efficient and reliable data transmission under complex and changeable network conditions, and ensuring the transmission reliability and efficiency of key monitoring data. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a flowchart of an equipment intelligent operation and maintenance method based on railway operation and maintenance according to Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the module composition of an equipment intelligent operation and maintenance system based on railway operation and maintenance according to Example 2 of the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] Before giving examples, it is necessary to explain the application scenarios of the present invention, which are applied to data processing and potential fault identification of railway monitoring equipment. There are 51 machine room facilities distributed along the 468.212-kilometer-long line of the Handan-Huanghuai Railway, in which various monitoring equipment are installed to ensure the safety and stability of railway operation. These monitoring devices mainly include temperature sensors, humidity sensors, smoke detectors, and high-definition cameras, etc., which are distributed in different machine rooms and collect machine room environmental data in real time. The monitoring data also includes the operating status data of the monitoring equipment, such as parameters such as the CPU usage rate of the processor connected to these monitoring devices, as well as the alarm information and historical maintenance records generated by the equipment. These data are collectively referred to as multi-dimensional operation and maintenance perception data. For example, in the machine room of a certain railway section, the temperature sensor records the temperature value in the machine room once a minute; the humidity sensor records the humidity value once a minute; the smoke detector continuously monitors the smoke concentration value in the air; and the high-definition camera continuously captures the video image in the machine room. At the same time, the monitoring equipment itself is also generating operating status data, including changes in CPU usage, equipment alarm information (such as abnormal conditions such as high temperature warnings, network connection interruptions, etc.), and historical maintenance records (such as the time of the last replacement of parts, the staff involved in maintenance, etc.). These originally collected multi-dimensional perception data usually have various problems, such as missing data, duplication or abnormal values, and the network environment of the machine room along the railway is different, which will affect the transmission of data. This embodiment aims to clean the monitoring data, optimize the data transmission strategy according to the confidentiality of the data and the network environment, and finally identify the fault of the railway monitoring equipment based on the obtained data. It should be noted that the failure of the equipment described in this embodiment does not refer to problems such as abnormal equipment monitoring data and missing monitoring data, but refers to hidden faults such as damage to the sensor device or processor of the equipment. In the field of railway monitoring equipment operation and maintenance, when the data collected by the monitoring equipment is abnormal or deviates from the true value, it cannot be directly considered that the monitoring equipment has a fault, because the abnormal monitoring data may also be caused by the limitations of the signal processing algorithm or by environmental interference or network transmission interference, and does not bring the table equipment sensor device or processor itself. There is a fault. However, when the abnormal characteristics of the monitoring data meet certain rules, it may mean two situations: one is that the sensor device or processor and other components of the monitoring equipment are suffering from cumulative damage. When the accumulated damage exceeds the critical point, the device may fail. The second is that the sensor device or processor and other components of the monitoring equipment have failed, but because the abnormal monitoring data may not be caused by the monitoring equipment itself, the operation and maintenance personnel do not know whether the failure exists. In this embodiment, the above two situations are collectively referred to as potential failures. After the potential failure is identified, the operation and maintenance personnel need to go to the site to dismantle, diagnose the failure and repair the monitoring equipment with the potential failure.

[0037] Example 1: Figure 1 As shown, this embodiment provides an equipment intelligent operation and maintenance method based on railway operation and maintenance, and the method includes the following steps: S1, collect multi-dimensional operation and maintenance perception data, including equipment environment data and operation status data; the equipment environment data includes temperature data, humidity data, smoke concentration data, and video image data; the operation status data includes equipment operation parameters, alarm information, and historical operation and maintenance records; the equipment operation parameters include equipment CPU usage.

[0038] S2, cleans the multi-dimensional operation and maintenance perception data, performs adaptive multi-level encryption according to the sensitivity level and importance of the data to obtain a secure data packet, and uses a dynamic transmission protocol switching algorithm to transmit the secure data packet to the main station according to the stability and bandwidth conditions of the network transmission link, restores and verifies the data quality of the secure data packet received by the main station, and obtains cleaned data with spatial correlation.

[0039] S3, based on the cleaned data with spatial correlation, uses satellite positioning technology and three-dimensional visualization technology to obtain a dynamic and interactive equipment operation and maintenance digital twin environment.

[0040] S4 uses time series feature extraction and multi-factor correlation analysis algorithms to identify potential equipment failures based on pre-set historical failure modes and cleaned data with spatial correlation, and diagnoses, processes offline, and repairs equipment with potential failures.

[0041] For example, multi-dimensional operation and maintenance perception data is collected. The collected multi-dimensional perception data usually has various problems, such as missing data, duplication or outliers.

[0042] The multi-dimensional operation and maintenance perception data is cleaned, and adaptive multi-level encryption is performed according to the sensitivity level and importance of the data to obtain a secure data packet. According to the stability and bandwidth conditions of the network transmission link, a dynamic transmission protocol switching algorithm is used to transmit the secure data packet to the main station, and the secure data packet received by the main station is restored and the data quality is verified to obtain cleaned data with spatial association. The obtained cleaned data with spatial association is more accurate than the multi-dimensional operation and maintenance perception data, and is also associated with the spatial location information of the railway machine room, which is convenient for subsequent visual display and analysis.

[0043] Based on the cleaned data, Beidou satellite positioning technology and WebGL 3D visualization technology are used to build a dynamic and interactive equipment operation and maintenance digital twin environment. In the equipment operation and maintenance digital twin environment, each machine room and each device along the railway has its corresponding 3D model, and the equipment status and environmental data are presented in an intuitive way. Operation and maintenance personnel can monitor the equipment operation in real time through the digital twin environment and quickly locate possible problems.

[0044] In order to identify potential equipment failures, we use time series feature extraction and multi-factor association analysis algorithms to perform data mining on equipment status. By matching with pre-set historical failure patterns, we can identify potential equipment failures and immediately notify the nearest maintenance personnel to go to the site to diagnose, process offline, and repair the equipment.

[0045] Furthermore, the method of performing data cleaning on the multi-dimensional operation and maintenance perception data, performing adaptive multi-level encryption according to the sensitivity level and importance of the data to obtain a secure data packet, transmitting the secure data packet to the main station using a dynamic transmission protocol switching algorithm according to the stability and bandwidth conditions of the network transmission link, restoring and verifying the data quality of the secure data packet received by the main station, and obtaining the cleaned data with spatial association includes: According to the null values, duplicate values ​​and outliers existing in the multi-dimensional operation and maintenance perception data, a statistical filtering algorithm and an intelligent value filling method are used to obtain the second temperature data, second humidity data, second smoke concentration data, second video image data, second equipment operation parameters, second alarm information, and second historical operation and maintenance records with unified format and complete data; according to the sensitivity level and importance of the data, the second temperature data, second humidity data, and second smoke concentration data are divided into three-level sensitive data, the second video image data and the second historical operation and maintenance records are divided into two-level sensitive data, and the second equipment operation parameters and the second alarm information are divided into first-level sensitive data; an adaptive multi-level encryption method is used, RSA asymmetric encryption is used for the first-level sensitive data to obtain first-level encrypted data, AES symmetric encryption is used for the second-level sensitive data to obtain second-level encrypted data, and lightweight encryption is used for the third-level sensitive data to obtain third-level encrypted data; the first-level encrypted data, the second-level encrypted data, and the third-level encrypted data are combined to obtain a secure data packet.

[0046] According to the stability and bandwidth conditions of the network transmission link, a dynamic transmission protocol switching algorithm is used to transmit the security data packet to the main station; based on the security data packet received by the main station, a decryption algorithm is used to calculate the restored original data information; according to the preset data quality verification algorithm, the integrity rate of the restored original data information is evaluated; data with an integrity rate less than a preset integrity rate threshold is intelligently supplemented to obtain valid restored data; based on the valid restored data and the spatial location information of the computer room, a geographic coordinate mapping algorithm is used to obtain cleaned data with spatial association that includes valid restored data and spatial location.

[0047] For example, the multi-dimensional operation and maintenance perception data from various machine rooms along the railway have various quality problems and need to be systematically processed before they can be effectively used. The first problem is data missing, duplication and outliers. For example, in the temperature data collection process of a certain machine room, due to network transmission reasons, the temperature records for several consecutive hours may have null values; in another machine room, due to the retransmission mechanism in the network transmission, the humidity data at the same time is recorded repeatedly; and in the third machine room, due to external electromagnetic interference, the smoke detector occasionally outputs abnormal readings that are significantly deviated from the normal range.

[0048] First, data cleaning is performed using statistical filtering algorithms and intelligent value-adding methods. For example, for missing temperature data, the temperature change pattern of the computer room in the same historical period is analyzed, combined with the meteorological conditions of the day, and time series interpolation technology is used to fill the gaps; for repeated humidity data, redundant records are identified and removed by comparing timestamps and numerical consistency; for abnormal readings of smoke detectors, data that are obviously beyond the normal range are marked and replaced with reasonable estimates based on the outlier detection method in statistics. Through the statistical filtering algorithm and intelligent value-adding method, the second temperature data, second humidity data, second smoke concentration data, second video image data, second equipment operating parameters, second alarm information, and second historical operation and maintenance records with unified format and complete data are obtained.

[0049] Then, based on the data security requirements, the cleaned data is protected in a graded manner. According to the sensitivity level and importance of the data, the second temperature data, the second humidity data, and the second smoke concentration data are classified as three-level sensitive data, because the second temperature data, the second humidity data, and the second smoke concentration data are important, but the degree of harm caused by leakage is relatively low; the second video image data and the second historical operation and maintenance records are classified as secondary sensitive data, because these data may contain information such as equipment layout and maintenance strategy, which will lead to more serious consequences after leakage; the second equipment operation parameters and the second alarm information are classified as first-level sensitive data, because these data are directly related to the operating status and safety alarms of railway equipment, and are crucial to the safety of railway operation. The importance of the first-level sensitive data, the second-level sensitive data, and the third-level sensitive data decreases in turn. For sensitive data of different levels, an adaptive multi-level encryption method is used for protection. For the most sensitive first-level data (equipment operating parameters and alarm information), RSA asymmetric encryption technology is used to ensure security during transmission through public key encryption and private key decryption; for second-level sensitive data (video images and historical maintenance records), AES symmetric encryption technology is used to ensure security while taking into account processing efficiency; for third-level sensitive data (temperature, humidity and smoke concentration data), lightweight encryption methods with low computational overhead, such as PRESENT encryption methods, are used to optimize processing performance while ensuring basic security. After encryption processing, these three types of data form first-level encrypted data, second-level encrypted data, and third-level encrypted data, and then are combined into secure data packets.

[0050] Since the network environments of the computer rooms along the Handan-Huanghuai Railway vary greatly, it is necessary to dynamically adjust the transmission strategy according to the actual network conditions. By continuously monitoring the stability and bandwidth conditions of the network links from each computer room to the central station, corresponding transmission protocols are adopted for different network quality levels. In computer rooms in plain areas with good network conditions, all levels of encrypted data can be transmitted simultaneously; while in computer rooms in mountainous areas with poor network conditions, the transmission of first-level sensitive data is prioritized, while other data is temporarily stored locally and transmitted after network conditions improve. This dynamically adjusted transmission strategy ensures the timely arrival of key data while also taking into account the integrity of other data.

[0051] When the data packet arrives at the central station, the original data is restored through the corresponding decryption algorithm and the data quality is evaluated. For data that may be damaged during transmission, if the integrity rate is lower than the preset threshold, it is repaired through intelligent value-adding technology. For example, for a certain video data, if some frames are lost, it can be reconstructed by interpolation through the content of the previous and next frames; for some missing equipment operation parameter records, interpolation estimation is performed in combination with historical data and current environmental conditions, so as to obtain more complete and reliable effective restoration data.

[0052] Finally, based on the effective restoration of the data and the spatial location information of the computer room, the geographic coordinate mapping algorithm is used to obtain the cleaned data with spatial association, so that the data not only contains the numerical values ​​of various monitoring parameters, but also has geographic location attributes, laying the foundation for subsequent three-dimensional visualization and spatial analysis.

[0053] Furthermore, the method of transmitting the security data packet to the master station by using a dynamic transmission protocol switching algorithm according to the stability and bandwidth conditions of the network transmission link includes: The signal strength, packet loss rate and delay time of the network transmission link are detected in real time with the preset first period as the sliding time window, and the network environment evaluation algorithm is used to obtain the network environment quality level within the sliding time window; the network environment quality level is one of the first quality level, the second quality level and the third quality level; when the network environment quality level is the first quality level, the first level encrypted data, the second level encrypted data and the third level encrypted data are transmitted simultaneously by the standard data transmission protocol; when the network environment quality level is the second quality level, the first level encrypted data is first transmitted by the high-reliability transmission protocol, and after the transmission is completed, the second level encrypted data and the third level encrypted data are transmitted in sequence; when the network environment quality level is the third quality level, the first level encrypted data is transmitted by the lightweight transmission protocol, and the second level encrypted data and the third level encrypted data are temporarily stored in the local cache, and the network environment quality level is upgraded to the second quality level or the first quality level before being transmitted.

[0054] For example, in the actual operating environment of the Handan-Huanghuai Railway, the network environments of the 51 computer rooms distributed along the line are quite different. For example, computer rooms located in the urban plain area usually have stable network connections, high signal strength, low packet loss rate, and short delay time; while computer rooms located in mountainous areas or far away from communication base stations often face network problems such as weak signals, high packet loss rate, and long delay time, and even temporary network interruptions under severe weather conditions. In order to cope with complex network environments, this embodiment adopts a dynamic transmission protocol switching algorithm to optimize the data transmission strategy.

[0055] First, the network transmission link from each computer room to the main station is monitored in real time with the preset first period (such as 5 minutes) as the sliding time window. In this sliding time window, three key parameters are continuously collected through network detection technology: signal strength, packet loss rate and delay time. For example, in a computer room located in a mountainous area, its network signal strength fluctuates at a low level, the packet loss rate is sometimes high and sometimes low, and the delay time often exceeds the normal range. Based on a comprehensive analysis of signal strength, packet loss rate and delay time, the network environment evaluation algorithm is used to obtain the network environment quality level within the sliding time window.

[0056] According to the evaluation results, the network environment quality is divided into three levels: the first level quality indicates that the network environment is good and suitable for high-speed transmission of large amounts of data; the second level quality indicates that the network environment is average and reliable but not necessarily high-speed data transmission is possible; the third level quality indicates that the network environment is poor and only suitable for transmitting a small amount of critical data. When the network environment of a computer room is evaluated as the first level quality, standard data transmission protocols (such as TCP / IP protocols) are used to transmit all types of encrypted data at the same time. In this case, the first level encrypted data (including equipment operating parameters and alarm information), the second level encrypted data (including video images and historical maintenance records) and the third level encrypted data (including temperature, humidity and smoke concentration data) can be quickly and reliably transmitted to the main station, ensuring that the monitoring center can fully grasp the equipment status and environmental conditions of the computer room. When the network environment is evaluated as the second level quality, a more cautious transmission strategy is adopted. At this time, a high-reliability transmission protocol (such as the MQTT protocol with stronger error correction capabilities and retransmission mechanisms) is used to transmit the first level encrypted data first, ensuring that the most critical equipment operating parameters and alarm information can reach the main station accurately and without error. After the first level encrypted data is successfully transmitted, the second level encrypted data and the third level encrypted data are transmitted in sequence. Although this step-by-step transmission method increases the overall transmission time, it can ensure that the most important information will not be lost due to network congestion or transmission errors when the network conditions are poor. When the network environment is evaluated as level 3 quality, it faces extremely poor network conditions. At this time, only lightweight transmission protocols (such as UDP protocol) are used to transmit level 1 encrypted data. Although this type of protocol does not guarantee the complete arrival of data, due to the small resource consumption of lightweight transmission protocols, it is more suitable for transmitting a small amount of critical data under extremely poor network conditions. At the same time, the level 2 encrypted data and level 3 encrypted data are temporarily stored in the local cache and wait for the network environment quality to be improved to level 2 or level 1 before transmission. In this way, even under the worst network conditions, the most basic equipment status information can be transmitted back to the main station in time, providing necessary guarantees for railway operation safety. By adopting a transmission strategy based on dynamic adjustment of network environment quality, the timeliness and reliability of monitoring data transmission can be maximized in a complex and changing network environment.

[0057] Furthermore, the method of detecting the signal strength, packet loss rate and delay time of the network transmission link in real time using the preset first period as the sliding time window, and using the network environment evaluation algorithm to obtain the network environment quality level within the sliding time window includes: In the sliding time window, the signal strength, packet loss rate and delay time of the network transmission link are sampled in real time at every preset first sampling interval to obtain a signal strength sample set, a packet loss rate sample set and a delay time sample set in the sliding time window; the average value of the signal strength sample set is calculated to obtain the average signal strength, the average value of the packet loss rate sample set is calculated to obtain the average packet loss rate, and the average value of the delay time sample set is calculated to obtain the average delay time.

[0058] According to a preset signal strength reference range, a forward correlation mapping is used to normalize the average signal strength into a signal strength index; according to a preset packet loss rate reference range, a preset first reverse correlation mapping is used to normalize the average packet loss rate into a packet loss rate index; according to a preset delay time reference range, a preset second reverse correlation mapping is used to normalize the average delay time into a delay time index; the signal strength index, the packet loss rate index and the delay time index are multiplied by preset weight coefficients respectively and then summed to obtain a comprehensive network environment index; the comprehensive network environment index is compared with a preset first scoring threshold and a preset second scoring threshold to obtain a network environment quality level within a sliding time window.

[0059] The average resource utilization of the network transmission link is counted at intervals of a preset second period, and the first scoring threshold and the second scoring threshold are dynamically adjusted according to the average resource utilization; the second period is an integer multiple of the first period.

[0060] Exemplarily, the first cycle is set to 5 minutes, and the sliding time window is 5 minutes. Within the 5-minute sliding time window, the signal strength, packet loss rate and delay time of the network transmission link are sampled in real time every preset first sampling interval (set to 10 seconds). Taking a computer room located in a mountainous area as an example, within a 5-minute sliding time window, a total of 30 signal strength samples, 30 packet loss rate samples and 30 delay time samples were collected, forming a signal strength sample set, a packet loss rate sample set and a delay time sample set respectively. From these sample sets, the average signal strength, average packet loss rate and average delay time within the time window are obtained by taking the average value.

[0061] However, these three parameters cannot be directly compared or integrated due to their different units and meanings. For example, the unit of signal strength is decibel milliwatts, and the larger the value, the stronger the signal; the packet loss rate is expressed as a percentage, ranging from zero to one hundred, and the smaller the value, the better the network quality; the unit of delay time is milliseconds, usually ranging from tens to hundreds, and the smaller the value, the better the network quality. Therefore, three different mapping methods are used to convert these parameters into standardized index values. For average signal strength, a forward correlation mapping is used to map the signal strength to a signal strength index between 0 and 1, and the stronger the signal, the higher the corresponding index value. For example, when the signal strength is higher than negative sixty decibel milliwatts, the signal strength is mapped to a signal strength index close to 1; when the signal strength is lower than negative ninety decibel milliwatts, the signal strength is mapped to an index value close to 0. For average packet loss rate and average delay time, since they are negatively correlated with network quality, that is, the smaller the value, the better the quality, a reverse correlation mapping is used. Specifically, through the pre-set first reverse association mapping, the packet loss rate is converted into a packet loss rate index, and the lower the packet loss rate, the higher the corresponding index value; through the pre-set second reverse association mapping, the delay time is converted into a delay time index, and the shorter the delay time, the higher the corresponding index value. Through the above processing, the three parameters are standardized into index values ​​between 0 and 1, and all follow the unified meaning of "the larger the value, the better the network quality". According to the relative importance of these three indexes in network quality assessment, they are assigned pre-set weight coefficients. For example, a higher weight is given to the packet loss rate index because the impact of packet loss on data transmission is particularly significant; while relatively lower weights are given to the signal strength index and the delay time index. By multiplying these three indices by their respective weight coefficients and summing them up, a comprehensive network environment index that comprehensively reflects the quality of the network environment is obtained.

[0062] In order to convert the continuous network environment comprehensive index into a discrete quality level, a first scoring threshold and a second scoring threshold are set. The network environment quality level within the sliding time window is obtained by comparing the network environment comprehensive index with the preset first scoring threshold and second scoring threshold. For example, when the comprehensive index is higher than the first scoring threshold, it is determined to be a first-level quality; when the comprehensive index is lower than the first scoring threshold but higher than the second scoring threshold, it is determined to be a second-level quality; when the comprehensive index is lower than the second scoring threshold, it is determined to be a third-level quality.

[0063] However, fixed scoring thresholds cannot cope with complex and changeable network environments. For example, during periods of high overall network load, even a relatively good network environment may not be able to meet the requirements of transmission timeliness and reliability. Therefore, the average resource utilization of the network transmission link is counted at intervals of a pre-set second period (set to an integer multiple of the first period, such as 1 hour), and the first scoring threshold and the second scoring threshold are dynamically adjusted according to this indicator. For example, when it is detected that the overall network resource utilization is high, a pre-set threshold adjustment amount with a positive value and positively correlated with the average resource utilization is superimposed on the first scoring threshold and the second scoring threshold, thereby increasing the first scoring threshold and the second scoring threshold to avoid network congestion; conversely, when the network resource utilization is low, a pre-set threshold adjustment amount with a negative value and an absolute value positively correlated with the average resource utilization is superimposed on the first scoring threshold and the second scoring threshold, thereby reducing the first scoring threshold and the second scoring threshold, thereby improving the network resource utilization.

[0064] Furthermore, the method of comparing the network environment comprehensive index with the preset first scoring threshold and the second scoring threshold to obtain the network environment quality level within the sliding time window includes: When the comprehensive network environment index is greater than or equal to the first scoring threshold, the network environment quality level is level one; when the comprehensive network environment index is less than the first scoring threshold and greater than or equal to the second scoring threshold, the network environment quality level is level two; when the comprehensive network environment index is less than the second scoring threshold, the network environment quality level is level three.

[0065] Exemplarily, the first scoring threshold and the second scoring threshold are preset according to historical data analysis and actual transmission requirements. For example, the first scoring threshold is set to 0.8 and the second scoring threshold is set to 0.5, thereby dividing the network environment quality into three different levels.

[0066] When the network environment comprehensive index is greater than or equal to the first scoring threshold, it indicates that the current network environment is in good condition, with high signal strength, low packet loss rate, and short delay time, which is suitable for high-speed transmission of large amounts of data, so it is judged as first-level quality. For example, the network environment comprehensive index of a computer room in a plain area is 0.9, which exceeds the first scoring threshold of 0.8, so it is judged as first-level quality. At this time, the computer room can use standard data transmission protocols to transmit all levels of encrypted data at the same time without worrying about network congestion or transmission failure. When the network environment comprehensive index is less than the first scoring threshold but greater than or equal to the second scoring threshold, it indicates that the current network environment is general, and there may be signal fluctuations, occasional packet loss or increased delay, which is not suitable for the simultaneous transmission of large amounts of data, but it can still ensure basic data transmission reliability, so it is judged as second-level quality. For example, the network environment comprehensive index of a computer room in a hilly area is 0.6, which is judged as second-level quality. Therefore, the computer room uses a high-reliability transmission protocol to preferentially transmit first-level encrypted data (key equipment operating parameters and alarm information), and then transmits other data in sequence. When the network environment comprehensive index is less than the second scoring threshold, it indicates that the current network environment is poor, and there may be problems such as weak signal, large packet loss or severe delay. It is only suitable for transmitting a small amount of extremely critical data, so it is judged as level 3 quality. For example, the network environment comprehensive index of a certain mountain computer room under severe weather conditions dropped to 0.3, which is lower than the second scoring threshold of 0.5, and is judged as level 3 quality. In this extreme case, the computer room will only use a lightweight transmission protocol to transmit first-level encrypted data, and temporarily store other data locally, waiting for network conditions to improve before transmitting.

[0067] Through the hierarchical judgment mechanism, the corresponding data transmission strategy can be adopted according to the actual network environment, which not only ensures the timely transmission of key data, but also avoids network congestion or transmission failure caused by rashly transmitting large amounts of data when network conditions are poor, thereby maximizing the efficiency and reliability of data transmission. This technology effectively improves the robustness of remote monitoring data transmission of equipment in the machine room of the Handan-Huanghua Railway distributed in complex terrain.

[0068] Furthermore, the method of obtaining a dynamic interactive equipment operation and maintenance digital twin environment by using satellite positioning technology and three-dimensional visualization technology based on the cleaned data with spatial association includes: Beidou satellite positioning technology and geographic surveying and mapping technology are used to construct a three-dimensional spatial basic model of the railway operation and maintenance area; personnel location data are obtained based on the positioning devices worn by the operation and maintenance personnel; the three-dimensional spatial basic model is spatially aligned with the cleaned data and personnel location data with spatial association to obtain a basic visualization environment with geographic reference.

[0069] Extract valid restored data from the cleaned data with spatial association; classify the data corresponding to temperature data, humidity data, smoke concentration data, and video image data in the valid restored data as valid environmental monitoring data; classify the data corresponding to equipment operating parameters, alarm information, and historical operation and maintenance records in the valid restored data as valid equipment status data; use WebGL three-dimensional visualization technology to establish an association mapping between the valid equipment status data and the geometric entities of the corresponding equipment in the three-dimensional space basic model to obtain a digital twin model of the equipment; map the effective environmental monitoring data to the three-dimensional space basic model through visualization technology to obtain an environmental data visualization layer; map the personnel location data to the three-dimensional space basic model to obtain dynamic personnel icon data.

[0070] The equipment digital twin model, environmental data visualization layer, and dynamic personnel icon data are combined to obtain a dynamic and interactive equipment operation and maintenance digital twin environment.

[0071] For example, in order to intuitively display the equipment operation status and environmental data, a highly realistic three-dimensional visualization environment is constructed. First, Beidou satellite positioning technology is used to accurately locate each machine room along the railway to obtain its geographic coordinate information. At the same time, through geographic surveying and mapping technology, the terrain data around the railway line is collected, and then a three-dimensional space basic model that is highly consistent with the actual geographical environment is constructed. The three-dimensional space basic model not only includes the railway line, but also includes each machine room building and internal equipment layout along the line. In order to achieve real-time monitoring of personnel positions, each operation and maintenance personnel is equipped with an intelligent terminal device with a built-in Beidou positioning module. These devices periodically send location information to the main station to obtain accurate personnel location data. Through the position mapping algorithm, the real-time updated personnel location data is spatially aligned with the three-dimensional space basic model to synchronize the personnel position in the model with reality. At the same time, the cleaned data with spatial association is also associated with the three-dimensional space basic model through spatial coordinate mapping to obtain a basic visualization environment with geographic reference.

[0072] From the cleaned data with spatial association, effective restored data is extracted and classified according to data type. Temperature data, humidity data, smoke concentration data and video image data are classified as effective environmental monitoring data, which reflect the environmental conditions in the computer room; equipment operating parameters, alarm information and historical operation and maintenance records are classified as effective equipment status data, which directly reflect the working status of the equipment. Using WebGL three-dimensional visualization technology, an association mapping is established between the effective equipment status data and the geometric entities of the corresponding equipment in the three-dimensional space basic model to obtain the equipment digital twin model. The effective environmental monitoring data is mapped to the three-dimensional space basic model through visualization technology to obtain the environmental data visualization layer. The personnel location data is mapped to the three-dimensional space basic model to obtain dynamic personnel icon data. Finally, a complete dynamic interactive equipment operation and maintenance digital twin environment is constructed.

[0073] Furthermore, the method of using time series feature extraction and multi-factor correlation analysis algorithm to identify potential equipment faults based on pre-set historical fault modes and cleaned data with spatial correlation, and diagnosing, offline processing and repairing equipment with potential faults includes: The cleaned data with spatial association are sorted according to time series, the time series features of effective environmental monitoring data and effective equipment status data are extracted, the time series features are matched with the feature templates in the preset historical fault mode library, and the matching similarity is calculated.

[0074] The device corresponding to the data whose matching similarity is greater than a preset first fault threshold is marked as a potential fault device; the potential fault device indicates a device that has been identified to have a potential fault; based on the computer room space location information and personnel location data corresponding to the potential fault device, the operation and maintenance personnel who is closest to the potential fault device is matched and recorded as the first operation and maintenance personnel; the computer room space location information corresponding to the potential fault device is pushed to the first operation and maintenance personnel; and the first operation and maintenance personnel performs offline processing and repair on the potential fault device.

[0075] Exemplarily, when the abnormal characteristics of the spatially correlated cleaned data meet certain rules, it may mean two situations: one is that the sensor device or processor and other components of the monitoring equipment are suffering from cumulative damage. When the accumulated damage exceeds the critical point, the device may fail. The second is that the sensor device or processor and other components of the monitoring equipment have failed, but because the abnormal monitoring data may not be caused by the monitoring equipment itself, the operation and maintenance personnel do not know whether the failure exists. In this embodiment, the above two situations are collectively referred to as potential failures. After the potential failure is identified, the operation and maintenance personnel need to go to the site to dismantle, diagnose and repair the monitoring equipment with the potential failure.

[0076] First, the cleaned data with spatial association are sorted according to the time series. For example, the cleaned data with spatial association of each monitoring device in a certain computer room are arranged in time order to form a continuous data sequence. From the continuous data sequence, the time series features of the effective environmental monitoring data and the effective equipment status data are extracted. The time series features include not only the data values ​​at each moment, but also the patterns of data changes over time, such as rising trends, periodic fluctuations or sudden changes, which are specifically reflected in the rate of change, periodic spectrum, mutation amount, etc. The extracted time series features are matched with the feature templates in the pre-set historical fault mode library to calculate the matching similarity. The historical fault mode library is established based on the long-term accumulated maintenance records and fault case analysis, and contains the data features of various types of equipment under potential fault conditions. For example, when a certain model of temperature sensor has a potential fault, it exhibits the following characteristics: the CPU usage of the processor connected to the temperature sensor drops abnormally and the temperature data is displayed as a fixed value, or the CPU usage of the processor connected to the temperature sensor rises abnormally and the temperature data fluctuates chaotically. The matching similarity is obtained by calculating the similarity between the cleaned data with spatial association and these historical fault modes. When the matching similarity of a certain device exceeds a preset first fault threshold, it means that the device is likely to have a potential fault, and the device is marked as a potential fault device.

[0077] According to the spatial location information of the computer room corresponding to the potential faulty equipment and the location data of the currently on-duty maintenance personnel, the maintenance personnel closest to the equipment and with the corresponding skills are automatically matched and recorded as the first operation and maintenance personnel. For example, when the temperature sensor in a computer room is determined to be a potential faulty device, the location matching algorithm finds that there is currently an operation and maintenance personnel inspecting the nearby lines, so he is designated as the first operation and maintenance personnel. The key information such as the spatial location information of the computer room corresponding to the potential faulty equipment, the equipment type, the predicted fault type and the processing suggestions are pushed to the mobile terminal device of the first operation and maintenance personnel. After receiving the task, the first operation and maintenance personnel goes to the designated location to diagnose, offline inspect and repair the potential faulty equipment. The diagnosis, offline inspection and repair include equipment status detection, parameter testing, parts replacement or software reset, etc. The specific operations depend on the type of fault.

[0078] Furthermore, the method of sorting the cleaned data with spatial association according to time series, extracting the time series features of effective environmental monitoring data and effective equipment status data, matching the time series features with feature templates in a preset historical fault mode library, and calculating the matching similarity includes: The cleaned data with spatial correlation are sorted according to the time series, and the preset second sliding time window is used to extract the temperature change rate, humidity change rate, smoke concentration change rate, video image gray value change rate, equipment operation parameter fluctuation rate, and periodic spectrum as time series features.

[0079] The time series features are standardized to obtain standardized time series features; based on the standardized time series features and the feature templates in the historical fault mode library, the Euclidean distance algorithm and the pre-set third reverse correlation function are used to perform similarity calculation to obtain the first similarity; the temperature change rate, the humidity change rate, the smoke concentration change rate, and the equipment operation parameter fluctuation rate are associated with pattern recognition to obtain the second similarity; the second similarity represents the degree of conformity between the relationship between the temperature change rate, the humidity change rate, the smoke concentration change rate, and the equipment operation parameter fluctuation rate and the physical law.

[0080] A weighted average of the second similarity and the first similarity is taken as the matching similarity.

[0081] Exemplarily, the cleaned data with spatial association are sorted according to the time series to form a continuous data sequence. The time series data is segmented using a preset second sliding time window. The second sliding time window is usually set to a range of several hours to one day, depending on the equipment type and parameter characteristics. Time series features are extracted from the data in the second sliding time window, and the time series features include temperature change rate, humidity change rate, smoke concentration change rate, video image gray value change rate, equipment operation parameter fluctuation rate, and periodic spectrum. The periodic spectrum is calculated by Fourier decomposition. The temperature change rate reflects the speed at which the temperature increases or decreases; the humidity change rate reflects the speed at which the ambient humidity increases or decreases; the smoke concentration change rate reflects the speed at which the abnormal gas concentration increases or decreases; the video image gray value change rate reflects the degree of change of the environmental video image; the equipment operation parameter fluctuation rate reflects the change speed of the CPU usage of the processor connected to the monitoring device; the numerical mutation rate reflects the mutation of the monitoring data; and the periodic spectrum reflects the periodic change law of the monitoring data at different time scales.

[0082] In order to make different types of timing features comparable, the extracted timing features are standardized. The standardization process comprehensively considers the normal range of each timing feature and converts each timing feature into a dimensionless standard value, that is, a standardized timing feature. The standardization process includes algorithms such as normalization. The standardized timing feature is compared with the feature template in the historical fault mode library. The historical fault mode library is established based on long-term accumulated equipment failure cases and contains typical manifestation modes of various potential failures. For example, if the parameters of the moving average filter algorithm in the temperature sensor firmware are improperly set, the periodic noise may be amplified, resulting in characteristic subharmonics in the periodic spectrum of the monitored temperature value. If the ADC chip of the temperature sensor is damaged, the reference voltage drifts, or the sampling clock is unstable, the monitored temperature change rate will be too large. Based on the standardized timing features and the feature template in the historical fault mode library, the Euclidean distance algorithm is used to calculate the similarity. The Euclidean distance algorithm calculates the straight-line distance between two points in a multidimensional space. The smaller the distance, the higher the similarity. The Euclidean distance is input into the pre-set third reverse correlation function to obtain the first similarity. The larger the value of the first similarity is, the greater the similarity between the standardized time series feature and the feature template in the historical fault mode library is, that is, the greater the possibility of a potential fault. The temperature change rate, humidity change rate, and smoke concentration change rate are associated with pattern recognition to obtain the second similarity. The second similarity is a quantitative indicator that characterizes the degree to which the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate conform to physical laws. The larger the second similarity is, the less the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate conform to physical laws, that is, the greater the possibility of a potential fault. The weighted average of the second similarity and the first similarity is taken as the matching similarity.

[0083] Furthermore, the method of performing correlation pattern recognition on the temperature change rate, the humidity change rate, the smoke concentration change rate, and the equipment operation parameter fluctuation rate to obtain the second similarity includes: According to the corresponding time segment in the seasonal feature matching library to which the current time belongs, obtain the preset benchmark functional relationship between the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate; calculate the degree of deviation of the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate from the benchmark functional relationship, and obtain a quantitative indicator characterizing the degree to which the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate conform to physical laws, which is recorded as the second similarity; the value of the second similarity is between 0 and 1; the seasonal feature matching library is obtained by pre-setting, and represents the typical functional relationship between the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate in the pre-divided time segment.

[0084] For example, under normal circumstances, there is a specific relationship between the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operation parameter fluctuation rate, and this relationship is recorded as a reference function relationship. For example, when the temperature change rate is large, in order to maintain the acquisition accuracy, the equipment operation parameter fluctuation rate, that is, the equipment CPU usage rate change rate, will also be large. Therefore, the correlation between the temperature change rate and the equipment operation parameter fluctuation rate can be measured by the reference function relationship between the temperature change rate and the equipment operation parameter fluctuation rate. The reference function relationship is also related to the season. For example, the temperature change rate and the humidity change rate show a relatively obvious positive correlation in summer, while this relationship is not obvious in winter. The humidity change rate and the smoke concentration change rate show a negative correlation within a certain range. The degree of deviation of the temperature change rate, the humidity change rate, and the smoke concentration change rate from the reference function relationship is calculated to obtain a quantitative indicator that characterizes the degree to which the temperature change rate, the humidity change rate, and the smoke concentration change rate conform to the physical law, which is recorded as the second similarity. The greater the second similarity, the greater the degree of deviation of the temperature change rate, the humidity change rate, and the smoke concentration change rate from the reference function relationship. The seasonal feature matching library is obtained by presetting.

[0085] Embodiment 2: Based on the same inventive concept, Figure 2 As shown, this embodiment also provides an equipment intelligent operation and maintenance system based on railway operation and maintenance, and the system includes: a data acquisition module, a data cleaning module, a digital twin module and a fault identification module connected in sequence.

[0086] The data acquisition module is used to collect multi-dimensional operation and maintenance perception data, including equipment environment data and operation status data; the equipment environment data includes temperature data, humidity data, smoke concentration data, and video image data; the operation status data includes equipment operation parameters, alarm information, and historical operation and maintenance records; the equipment operation parameters include equipment CPU usage.

[0087] The data cleaning module is used to clean the multi-dimensional operation and maintenance perception data, perform adaptive multi-level encryption according to the sensitivity level and importance of the data to obtain a secure data packet, and use a dynamic transmission protocol switching algorithm to transmit the secure data packet to the main station according to the stability and bandwidth conditions of the network transmission link, and restore and verify the data quality of the secure data packet received by the main station to obtain cleaned data with spatial correlation.

[0088] The digital twin module is used to obtain a dynamic and interactive equipment operation and maintenance digital twin environment based on the cleaned data with spatial association, using satellite positioning technology and three-dimensional visualization technology.

[0089] The fault identification module is used to use time series feature extraction and multi-factor correlation analysis algorithms to identify potential equipment faults based on pre-set historical fault patterns and cleaned data with spatial correlation, and to diagnose, offline process and repair equipment identified to have potential faults.

[0090] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0091] Finally, it should be noted that: Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent equipment operation and maintenance method based on railway operation and maintenance, characterized in that: The method comprises the following steps: S1, collect multi-dimensional operation and maintenance perception data, including equipment environment data and operation status data; the equipment environment data includes temperature data, humidity data, smoke concentration data, and video image data; the operation status data includes equipment operation parameters, alarm information, and historical operation and maintenance records; the equipment operation parameters include equipment CPU usage; S2, performing data cleaning on the multi-dimensional operation and maintenance perception data, performing adaptive multi-level encryption according to the sensitivity level and importance of the data to obtain a secure data packet, using a dynamic transmission protocol switching algorithm to transmit the secure data packet to the master station according to the stability and bandwidth conditions of the network transmission link, restoring and verifying the data quality of the secure data packet received by the master station, and obtaining cleaned data with spatial association; S3, based on the cleaned data with spatial correlation, uses satellite positioning technology and 3D visualization technology to obtain a dynamic and interactive equipment operation and maintenance digital twin environment; S4 uses time series feature extraction and multi-factor correlation analysis algorithms to identify potential equipment failures based on pre-set historical failure modes and cleaned data with spatial correlation, and diagnoses, processes offline, and repairs equipment with potential failures.

2. The intelligent equipment operation and maintenance method based on railway operation and maintenance as claimed in claim 1, characterized in that: The method of performing data cleaning on the multi-dimensional operation and maintenance perception data, performing adaptive multi-level encryption according to the sensitivity level and importance of the data to obtain a security data packet, using a dynamic transmission protocol switching algorithm to transmit the security data packet to the main station according to the stability and bandwidth conditions of the network transmission link, restoring and verifying the data quality of the security data packet received by the main station, and obtaining the cleaned data with spatial association includes: According to the null values, duplicate values ​​and outliers existing in the multi-dimensional operation and maintenance perception data, a statistical filtering algorithm and an intelligent value-adding method are used to obtain second temperature data, second humidity data, second smoke concentration data, second video image data, second equipment operating parameters, second alarm information, and second historical operation and maintenance records with unified format and complete data; according to the sensitivity level and importance of the data, the second temperature data, second humidity data, and second smoke concentration data are divided into three-level sensitive data, the second video image data and the second historical operation and maintenance records are divided into two-level sensitive data, and the second equipment operating parameters and the second alarm information are divided into first-level sensitive data; an adaptive multi-level encryption method is used, RSA asymmetric encryption is used for the first-level sensitive data to obtain first-level encrypted data, AES symmetric encryption is used for the second-level sensitive data to obtain second-level encrypted data, and lightweight encryption is used for the third-level sensitive data to obtain third-level encrypted data; the first-level encrypted data, the second-level encrypted data, and the third-level encrypted data are combined to obtain a secure data packet; According to the stability and bandwidth conditions of the network transmission link, a dynamic transmission protocol switching algorithm is used to transmit the security data packet to the main station; based on the security data packet received by the main station, a decryption algorithm is used to calculate the restored original data information; according to the preset data quality verification algorithm, the integrity rate of the restored original data information is evaluated; data with an integrity rate less than a preset integrity rate threshold is intelligently supplemented to obtain valid restored data; based on the valid restored data and the spatial location information of the computer room, a geographic coordinate mapping algorithm is used to obtain cleaned data with spatial association that includes valid restored data and spatial location.

3. The intelligent equipment operation and maintenance method based on railway operation and maintenance as claimed in claim 2, characterized in that: The method of transmitting the security data packet to the master station by using a dynamic transmission protocol switching algorithm according to the stability and bandwidth conditions of the network transmission link includes: The signal strength, packet loss rate and delay time of the network transmission link are detected in real time with the preset first period as the sliding time window, and the network environment evaluation algorithm is used to obtain the network environment quality level within the sliding time window; the network environment quality level is one of the first quality level, the second quality level and the third quality level; when the network environment quality level is the first quality level, the first level encrypted data, the second level encrypted data and the third level encrypted data are transmitted simultaneously by the standard data transmission protocol; when the network environment quality level is the second quality level, the first level encrypted data is first transmitted by the high-reliability transmission protocol, and after the transmission is completed, the second level encrypted data and the third level encrypted data are transmitted in sequence; when the network environment quality level is the third quality level, the first level encrypted data is transmitted by the lightweight transmission protocol, and the second level encrypted data and the third level encrypted data are temporarily stored in the local cache, and the network environment quality level is upgraded to the second quality level or the first quality level before being transmitted.

4. The intelligent equipment operation and maintenance method based on railway operation and maintenance as claimed in claim 3, characterized in that: The method of using a preset first period as a sliding time window to detect the signal strength, packet loss rate and delay time of a network transmission link in real time and using a network environment evaluation algorithm to obtain the network environment quality level within the sliding time window includes: In the sliding time window, the signal strength, packet loss rate and delay time of the network transmission link are sampled in real time at every preset first sampling interval to obtain a signal strength sample set, a packet loss rate sample set and a delay time sample set in the sliding time window; the average value of the signal strength sample set is calculated to obtain the average signal strength, the average value of the packet loss rate sample set is calculated to obtain the average packet loss rate, and the average value of the delay time sample set is calculated to obtain the average delay time; According to a preset signal strength reference range, a forward correlation mapping is used to normalize the average signal strength into a signal strength index; according to a preset packet loss rate reference range, a preset first reverse correlation mapping is used to normalize the average packet loss rate into a packet loss rate index; according to a preset delay time reference range, a preset second reverse correlation mapping is used to normalize the average delay time into a delay time index; the signal strength index, the packet loss rate index and the delay time index are multiplied by a preset weight coefficient respectively and then summed to obtain a network environment comprehensive index; the network environment comprehensive index is compared with a preset first scoring threshold and a preset second scoring threshold to obtain a network environment quality level within a sliding time window; The average resource utilization of the network transmission link is counted at intervals of a preset second period, and the first scoring threshold and the second scoring threshold are dynamically adjusted according to the average resource utilization; the second period is an integer multiple of the first period.

5. The intelligent equipment operation and maintenance method based on railway operation and maintenance as claimed in claim 4, characterized in that: The method of comparing the network environment comprehensive index with the preset first scoring threshold and the second scoring threshold to obtain the network environment quality level within the sliding time window includes: When the comprehensive network environment index is greater than or equal to the first scoring threshold, the network environment quality level is level one; when the comprehensive network environment index is less than the first scoring threshold and greater than or equal to the second scoring threshold, the network environment quality level is level two; when the comprehensive network environment index is less than the second scoring threshold, the network environment quality level is level three.

6. The intelligent equipment operation and maintenance method based on railway operation and maintenance as claimed in claim 5, characterized in that: The method for obtaining a dynamic interactive equipment operation and maintenance digital twin environment by using satellite positioning technology and three-dimensional visualization technology based on the cleaned data with spatial association includes: Beidou satellite positioning technology and geographic surveying and mapping technology are used to construct a three-dimensional spatial basic model of the railway operation and maintenance area; personnel location data is obtained based on the positioning devices worn by the operation and maintenance personnel; the three-dimensional spatial basic model is spatially aligned with the cleaned data and personnel location data with spatial association to obtain a basic visualization environment with geographic reference; Extract valid restored data from the cleaned data with spatial association; classify the data corresponding to temperature data, humidity data, smoke concentration data, and video image data in the valid restored data as valid environmental monitoring data; classify the data corresponding to equipment operating parameters, alarm information, and historical operation and maintenance records in the valid restored data as valid equipment status data; use WebGL three-dimensional visualization technology to establish an association mapping between the valid equipment status data and the geometric entities of the corresponding equipment in the three-dimensional space basic model to obtain a digital twin model of the equipment; map the effective environmental monitoring data to the three-dimensional space basic model through visualization technology to obtain an environmental data visualization layer; map the personnel location data to the three-dimensional space basic model to obtain dynamic personnel icon data; The equipment digital twin model, environmental data visualization layer, and dynamic personnel icon data are combined to obtain a dynamic and interactive equipment operation and maintenance digital twin environment.

7. The intelligent equipment operation and maintenance method based on railway operation and maintenance as claimed in claim 6, characterized in that: The method of using time series feature extraction and multi-factor correlation analysis algorithm to identify potential equipment faults based on pre-set historical fault modes and cleaned data with spatial correlation, and diagnosing, offline processing and repairing equipment with potential faults includes: Sort the cleaned data with spatial association according to time series, extract the time series features of effective environmental monitoring data and effective equipment status data, match the time series features with the feature templates in the preset historical fault mode library, and calculate the matching similarity; The device corresponding to the data whose matching similarity is greater than a preset first fault threshold is marked as a potential fault device; the potential fault device indicates a device that has been identified to have a potential fault; based on the computer room space location information and personnel location data corresponding to the potential fault device, the operation and maintenance personnel who is closest to the potential fault device is matched and recorded as the first operation and maintenance personnel; the computer room space location information corresponding to the potential fault device is pushed to the first operation and maintenance personnel; and the first operation and maintenance personnel performs offline processing and repair on the potential fault device.

8. The intelligent equipment operation and maintenance method based on railway operation and maintenance as claimed in claim 7, characterized in that: The method of sorting the cleaned data with spatial association according to time series, extracting the time series features of effective environmental monitoring data and effective equipment status data, matching the time series features with feature templates in a preset historical fault mode library, and calculating the matching similarity includes: The cleaned data with spatial correlation are sorted according to the time series, and the temperature change rate, humidity change rate, smoke concentration change rate, video image gray value change rate, equipment operation parameter fluctuation rate, and periodic spectrum are extracted as time series features using the preset second sliding time window; The time series features are standardized to obtain standardized time series features; based on the standardized time series features and the feature templates in the historical fault mode library, a Euclidean distance algorithm and a preset third reverse correlation function are used to perform similarity calculation to obtain a first similarity; the temperature change rate, the humidity change rate, the smoke concentration change rate, and the equipment operation parameter fluctuation rate are identified by correlation patterns to obtain a second similarity; the second similarity indicates the degree of conformity between the relationship between the temperature change rate, the humidity change rate, the smoke concentration change rate, and the equipment operation parameter fluctuation rate and the physical law; A weighted average of the second similarity and the first similarity is taken as the matching similarity.

9. The intelligent equipment operation and maintenance method based on railway operation and maintenance as claimed in claim 8, characterized in that: The method of performing correlation pattern recognition on the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operation parameter fluctuation rate to obtain the second similarity includes: According to the corresponding time segment in the seasonal feature matching library to which the current time belongs, obtain the preset benchmark functional relationship between the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate; calculate the degree of deviation of the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate from the benchmark functional relationship, and obtain a quantitative indicator characterizing the degree to which the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate conform to physical laws, which is recorded as the second similarity; the value of the second similarity is between 0 and 1; the seasonal feature matching library is obtained by pre-setting, and represents the typical functional relationship between the temperature change rate, humidity change rate, smoke concentration change rate, and equipment operating parameter fluctuation rate in the pre-divided time segment.

10. An intelligent equipment operation and maintenance system based on railway operation and maintenance, used to execute the method according to any one of claims 1 to 9, characterized in that: The system includes: a data acquisition module, a data cleaning module, a digital twin module and a fault identification module connected in sequence; The data acquisition module is used to collect multi-dimensional operation and maintenance perception data, including equipment environment data and operation status data; the equipment environment data includes temperature data, humidity data, smoke concentration data, and video image data; the operation status data includes equipment operation parameters, alarm information, and historical operation and maintenance records; the equipment operation parameters include equipment CPU usage; The data cleaning module is used to clean the multi-dimensional operation and maintenance perception data, perform adaptive multi-level encryption according to the sensitivity level and importance of the data to obtain a security data packet, and adopt a dynamic transmission protocol switching algorithm to transmit the security data packet to the main station according to the stability and bandwidth conditions of the network transmission link, and restore and verify the data quality of the security data packet received by the main station to obtain cleaned data with spatial association; The digital twin module is used to obtain a dynamic interactive equipment operation and maintenance digital twin environment based on the cleaned data with spatial association using satellite positioning technology and three-dimensional visualization technology; The fault identification module is used to use time series feature extraction and multi-factor correlation analysis algorithms to identify potential equipment faults based on pre-set historical fault patterns and cleaned data with spatial correlation, and to diagnose, offline process and repair equipment identified to have potential faults.

Images