Rail transit field section operation and maintenance data security evaluation method

Abstract

The application relates to the technical field of rail transit operation and maintenance data security evaluation, in particular to a rail transit field section operation and maintenance data security evaluation method, which solves the problems of data fragmentation, static threshold limitations and insufficient deep verification in traditional evaluation through multi-level correlation analysis and dynamic verification processing of equipment operation, personnel operation and environment working condition data. The method comprises the following steps: acquiring real-time operation and maintenance comprehensive data set based on a multi-source data acquisition system; performing continuous fluctuation screening processing to obtain a screening result; and generating an operation and maintenance data security monitoring result. Through dynamic threshold adjustment and cross-module collaborative verification, combined with multi-level correlation analysis and real-time scene data cross comparison, the data tampering identification precision and abnormal detection capability are improved, and full-process automatic control is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of rail transit operation and maintenance data security management, in particular to a rail transit field section operation and maintenance data security evaluation method. BACKGROUND

[0002] With the intelligent and digital transformation of the rail transit industry, the amount of field section operation and maintenance data is growing explosively, covering multi-dimensional data such as vehicle vibration, traction motor current, personnel operation trajectory, and equipment communication log. There are some core problems in current operation and maintenance data security evaluation that need to be solved. Data fragmentation processing is characterized by independent analysis of vehicle mechanical data (such as vibration and wheel-rail adhesion state), electrical data (such as motor current), and personnel control data (such as forbidden zone approach behavior), without establishing a security correlation logic across types of data, which may lead to insufficient identification of potential data tampering or abnormal conduction risks. The limitations of static threshold are reflected in the fact that traditional evaluation relies on fixed threshold to determine data security (such as vibration amplitude exceeding threshold to determine abnormality), which is not adaptive to data fluctuations in dynamic conditions such as rain and snow, and may lead to misjudgment or missed judgment. The problem of lack of deep verification mechanism is that the evaluation of the authenticity and integrity of operation and maintenance data is mostly limited to surface checking (such as data format compliance), without building a multi-level cross-verification model combining historical data and real-time scenarios, making it difficult to detect hidden data tampering or abnormal injection. In addition, the risk warning lag is also a shortcoming of existing systems, which is mostly passive after the occurrence of data security incidents, lacks the ability to predict based on data trends, and is difficult to block the risk conduction path in advance. To address the above problems, the present application proposes an operation and maintenance data security evaluation method that integrates multi-source data collection, deep learning analysis, dynamic threshold evolution, and cross-module collaborative verification, realizing the whole process of safety control from data collection to risk warning. SUMMARY

[0003] In view of the deficiencies of the prior art, the present application provides a rail transit field section operation and maintenance data security evaluation method based on multi-source data fusion, which solves the problems of data fragmentation, static threshold limitations, lack of deep verification, and risk warning lag in traditional evaluation through multi-level correlation analysis and dynamic verification of equipment operation, personnel operation, and environmental condition data.

[0004] To achieve the above purpose, the present application provides a rail transit field section operation and maintenance data security evaluation method based on multi-source data fusion, comprising:

[0005] S1, obtaining real-time operation and maintenance data based on a multi-source data collection system, and performing data filtering to obtain a real-time operation and maintenance data set;

[0006] S2, performing continuity fluctuation screening processing according to the real-time operation and maintenance data set to obtain a continuity fluctuation screening result of the real-time operation and maintenance data;

[0007] S3, generating operation and maintenance data security monitoring results by using the continuity fluctuation screening results.

[0008] Preferably, the real-time operation and maintenance comprehensive data set obtained by filtering and processing real-time operation and maintenance data based on a multi-source data acquisition system comprises:

[0009] S1-1, obtaining equipment operation data, personnel operation data and environmental working condition data as real-time operation and maintenance data based on a multi-source data acquisition system;

[0010] S1-2, obtaining operation parameters of the corresponding multi-source data acquisition system according to the real-time operation and maintenance data;

[0011] S1-3, calculating an output signal-to-noise ratio value according to the operation parameters of the multi-source data acquisition system;

[0012] S1-4, using the operation parameters of the multi-source data acquisition system as auxiliary feature labels of the real-time operation and maintenance data;

[0013] S1-5, using the output signal-to-noise ratio value as an interference feature label of the real-time operation and maintenance data;

[0014] S1-6, using the real-time operation and maintenance data, the auxiliary feature labels and the interference feature labels of the real-time operation and maintenance data as a real-time operation and maintenance comprehensive data set; wherein the operation parameters of the multi-source data acquisition system include sampling frequency, data transmission rate and equipment operation state.

[0015] Further, the continuity fluctuation screening result of the real-time operation and maintenance data obtained by performing continuity fluctuation screening processing on the real-time operation and maintenance comprehensive data set comprises:

[0016] S2-1, obtaining a real-time operation and maintenance improvement data set by data cleaning using the real-time operation and maintenance comprehensive data set;

[0017] S2-2, establishing a continuous time period operation and maintenance improvement data set using the real-time operation and maintenance improvement data set;

[0018] S2-3, obtaining the continuity fluctuation screening result of the real-time operation and maintenance data by performing continuity fluctuation screening processing on the continuous time period operation and maintenance improvement data set.

[0019] Further, the real-time operation and maintenance improvement data set obtained by data cleaning using the real-time operation and maintenance comprehensive data set comprises:

[0020] S2-1-1, obtaining a real-time operation and maintenance initial improvement data set by data comprehensive processing of the real-time operation and maintenance data of the real-time operation and maintenance comprehensive data set based on a data preprocessing tool;

[0021] S2-1-2, determine whether the auxiliary feature labels of the real-time operation and maintenance initial improvement data set and the real-time operation and maintenance comprehensive data set at the current time are consistent, if yes, directly execute S2-1-4, otherwise, execute S2-1-3;

[0022] S2-1-3, determine whether the execution times of S2-1-2 at the current time are greater than 1, if yes, abandon the processing, otherwise, return to S2-1-1;

[0023] S2-1-4, determine whether the interference feature labels of the real-time operation and maintenance initial improvement data set and the real-time operation and maintenance comprehensive data set at the current time are consistent, if yes, use the real-time operation and maintenance initial improvement data set as the real-time operation and maintenance improvement data set, otherwise, execute S2-1-5;

[0024] S2-1-5, determine whether the execution times of S2-1-4 at the current time are greater than 1, if yes, abandon the processing, otherwise, return to S2-1-1; wherein the data comprehensive processing includes abnormal data elimination, data alignment processing and noise suppression processing.

[0025] Further, using the real-time operation and maintenance improvement data set to establish a continuous period operation and maintenance improvement data set includes:

[0026] Using the corresponding time of the real-time operation and maintenance improvement data set as a standard median time t;

[0027] Respectively acquiring the real-time operation and maintenance improvement data sets at t-1 time and t+1 time;

[0028] Using the real-time operation and maintenance data of the real-time operation and maintenance improvement data sets at t-1 time, the standard median time t and t+1 time to establish a real-time operation and maintenance data change trend;

[0029] Using the auxiliary feature labels of the real-time operation and maintenance improvement data sets at t-1 time, the standard median time t and t+1 time to establish an auxiliary feature label change trend;

[0030] Using the interference feature labels of the real-time operation and maintenance improvement data sets at t-1 time, the standard median time t and t+1 time to establish an interference feature label change trend;

[0031] Using the real-time operation and maintenance data change trend, the auxiliary feature label change trend and the interference feature label change trend as the continuous period operation and maintenance improvement data set.

[0032] Further, using the continuous period operation and maintenance improvement data set to perform continuous fluctuation screening processing to obtain a continuous fluctuation screening result of the real-time operation and maintenance data includes:

[0033] S2-3-1, determine whether the real-time operation data change trend of the operation and maintenance improvement data set and the auxiliary feature label change trend are consistent from the t-1 moment to the standard median moment t, if yes, execute S2-3-2, otherwise, return to S2-1-1;

[0034] S2-3-2, determine whether the auxiliary feature label change trend of the operation and maintenance improvement data set and the interference feature label are consistent from the t-1 moment to the standard median moment t, if yes, execute S2-3-3, otherwise, return to S1-2;

[0035] S2-3-3, determine whether the real-time operation data change trend of the operation and maintenance improvement data set and the auxiliary feature label change trend are consistent from the standard median moment t to the t+1 moment, if yes, execute S2-3-4, otherwise, return to S2-1-1;

[0036] S2-3-4, determine whether the auxiliary feature label change trend of the operation and maintenance improvement data set and the interference feature label are consistent from the standard median moment t to the t+1 moment, if yes, keep the current continuous period operation and maintenance improvement data set, and the continuity fluctuation screening processing result is normal, otherwise, keep the current continuous period operation and maintenance improvement data set, and the continuity fluctuation screening processing result is abnormal.

[0037] Further, the operation and maintenance data security monitoring result is generated by using the continuity fluctuation screening result, which comprises:

[0038] S3-1, establish a correlation analysis model of real-time operation data according to the real-time operation data;

[0039] S3-2, compare the correlation analysis model with the continuity fluctuation screening result to obtain feature label data of the real-time operation data;

[0040] S3-3, generate the operation and maintenance data security monitoring result by using the feature label data of the real-time operation data.

[0041] Further, the correlation analysis model of real-time operation data is established according to the real-time operation data, which comprises:

[0042] An initial correlation model is established based on the historical data storage according to the real-time operation data and the standard median moment t;

[0043] A derived event set is obtained according to the initial correlation model in sequence corresponding to subsequent correlation events;

[0044] A derived time set is obtained according to the derived event set corresponding to the event generation moment;

[0045] The initial correlation model, the derived event set and the derived time set are used as the correlation analysis model of the real-time operation data.

[0046] Furthermore, by comparing the results of the correlation analysis model with the continuous fluctuation screening results, the characteristic marker data of the real-time operation and maintenance data are obtained, including:

[0047] S3-2-1. Determine whether the continuous fluctuation screening result is normal. If so, execute S3-2-2. Otherwise, use the derivative time set of the correlation analysis model to obtain the corresponding subset of the derivative event set at time t+1 corresponding to the continuous fluctuation screening result as the initial comparison label, and directly execute S3-2-3.

[0048] S3-2-2 Determine whether the trend of the continuous fluctuation screening result corresponding to the continuous period operation and maintenance improvement dataset and the initial correlation model of the correlation analysis model are consistent. If so, use the continuous fluctuation screening result corresponding to the real-time operation and maintenance improvement dataset and the correlation analysis model as the feature label data of the real-time operation and maintenance data. Otherwise, use the initial correlation model with inconsistent trends as the feature label data of the real-time operation and maintenance data.

[0049] S3-2-3. Determine whether the initial comparison label is consistent with the real-time operation and maintenance data at time t+1. If so, use the continuous fluctuation screening result to select the corresponding real-time operation and maintenance improvement dataset and the correlation analysis model as the feature label data of the real-time operation and maintenance data. Otherwise, use the initial comparison label as the feature label data of the real-time operation and maintenance data. Among them, the trend consistency means that the changes in the data content of the operation and maintenance improvement dataset and the initial correlation model in the continuous time period correspond to the time of data change.

[0050] Furthermore, generating operation and maintenance data security monitoring results using feature markers from real-time operation and maintenance data includes:

[0051] S3-3-1. Determine whether the feature marker data of the real-time operation and maintenance data exists in the real-time operation and maintenance improvement dataset and correlation analysis model. If so, the operation and maintenance data security monitoring result is output normally, and the current real-time operation and maintenance improvement dataset and correlation analysis model are retained. Otherwise, execute S3-3-2.

[0052] S3-3-2. Determine whether the feature marker data of the real-time operation and maintenance data corresponds to the initial association model with inconsistent trends. If so, the operation and maintenance data security monitoring result is an abnormal output, and the initial association model with inconsistent current trends is used as the abnormal label. Otherwise, the initial association model corresponding to the initial comparison label is used as the abnormal label.

[0053] Compared with the closest existing technology, the present invention has the following advantages:

[0054] A targeted and reliable multi-source data fusion mechanism was established. Combined with dynamic threshold adjustment and cross-module collaborative verification, it addresses the problems of data fragmentation, limitations of static thresholds, and insufficient in-depth verification in traditional evaluation methods. Through multi-level correlation analysis and real-time scenario data cross-comparison, the accuracy of data tampering identification and the ability to detect abnormal data were improved, and automated management of the entire process from data collection to risk warning was achieved. Attached Figure Description

[0055] Figure 1 This is a flowchart of a method for security assessment of rail transit depot operation and maintenance data provided by the present invention. Detailed Implementation

[0056] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0057] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Example 1:

[0059] This invention provides a method for security assessment of rail transit depot operation and maintenance data, such as... Figure 1 As shown, it includes:

[0060] S1. Obtain real-time operation and maintenance data from a multi-source data acquisition system and perform data filtering to obtain a comprehensive real-time operation and maintenance dataset.

[0061] S2. Perform continuous fluctuation filtering on the real-time operation and maintenance comprehensive dataset to obtain the continuous fluctuation filtering results of the real-time operation and maintenance data.

[0062] S3. Utilize the continuous fluctuation screening results to generate operation and maintenance data security monitoring results.

[0063] S1 specifically includes:

[0064] S1-1. Acquire equipment operation data, personnel operation data, and environmental condition data based on a multi-source data acquisition system as real-time operation and maintenance data;

[0065] S1-2. Obtain the operating parameters of the corresponding multi-source data acquisition system based on real-time operation and maintenance data;

[0066] S1-3. Calculate the output signal-to-noise ratio value based on the operating parameters of the multi-source data acquisition system;

[0067] S1-4. Use the operating parameters of the multi-source data acquisition system as auxiliary feature labels for real-time operation and maintenance data;

[0068] S1-5. Use the output signal-to-noise ratio value as an interference feature label for real-time operation and maintenance data;

[0069] S1-6. Use real-time operation and maintenance data, auxiliary feature labels and interference feature labels of real-time operation and maintenance data as a comprehensive real-time operation and maintenance dataset;

[0070] The operating parameters of a multi-source data acquisition system include sampling frequency, data transmission rate, and equipment operating status.

[0071] S2 specifically includes:

[0072] S2-1. Use the real-time operation and maintenance comprehensive dataset to perform data cleaning to obtain the real-time operation and maintenance improvement dataset;

[0073] S2-2. Establish a continuous time period operation and maintenance improvement dataset using the real-time operation and maintenance improvement dataset;

[0074] S2-3. Continuous fluctuation filtering is performed on the continuous time period operation and maintenance improvement dataset to obtain the continuous fluctuation filtering results of real-time operation and maintenance data.

[0075] S2-1 specifically includes:

[0076] S2-1-1. Using the real-time operation and maintenance data from the real-time operation and maintenance comprehensive dataset, the data is comprehensively processed based on the data preprocessing tool to obtain the initial improved dataset for real-time operation and maintenance.

[0077] S2-1-2. Determine whether the auxiliary feature labels of the current real-time operation and maintenance initial improvement dataset and the real-time operation and maintenance comprehensive dataset are consistent. If they are, proceed directly to S2-1-4; otherwise, proceed to S2-1-3.

[0078] S2-1-3. Determine if the number of times S2-1-2 has been executed at the current time is greater than 1. If yes, abandon the process; otherwise, return to S2-1-1.

[0079] S2-1-4. Determine whether the interference feature labels of the current real-time operation and maintenance initial improvement dataset and the real-time operation and maintenance comprehensive dataset are consistent. If they are, use the real-time operation and maintenance initial improvement dataset as the real-time operation and maintenance improvement dataset. Otherwise, execute S2-1-5.

[0080] S2-1-5. Determine if the number of times S2-1-4 has been executed at the current time is greater than 1. If yes, abandon the process; otherwise, return to S2-1-1.

[0081] The comprehensive data processing includes outlier removal, data alignment, and noise suppression.

[0082] S2-2 specifically includes:

[0083] S2-2-1. Use real-time operation and maintenance to improve the corresponding time of the dataset as the standard median time t;

[0084] S2-2-2, Obtain the real-time operation and maintenance improvement datasets at time t-1 and t+1 respectively;

[0085] S2-2-3. Establish the real-time operation and maintenance data change trend using the real-time operation and maintenance improvement dataset at time t-1, standard median time t, and time t+1.

[0086] S2-2-4. Establish the trend of auxiliary feature label changes by using the auxiliary feature labels of the real-time operation and maintenance improvement dataset at time t-1, standard median time t, and time t+1.

[0087] S2-2-5. Establish the trend of interference feature label changes by using the interference feature labels of the real-time operation and maintenance improvement dataset at time t-1, standard median time t, and time t+1.

[0088] S2-2-6. Use the real-time operation and maintenance data change trend, auxiliary feature label change trend and interference feature label change trend as a continuous time period operation and maintenance improvement dataset.

[0089] S2-3 specifically includes:

[0090] S2-3-1. Determine whether the real-time operation and maintenance data change trend of the operation and maintenance improvement dataset from time t-1 to the standard median time t is consistent with the auxiliary feature label change trend. If yes, execute S2-3-2; otherwise, return to S2-1-1.

[0091] S2-3-2. Determine whether the trend of auxiliary feature labels in the continuous time period from time t-1 to the standard median time t is consistent with the trend of interference feature labels. If yes, execute S2-3-3; otherwise, return to S1-2.

[0092] S2-3-3: Determine whether the real-time operation and maintenance data change trend of the operation and maintenance improvement dataset during the continuous period from the standard median time t to time t+1 is consistent with the auxiliary feature label change trend. If yes, execute S2-3-4; otherwise, return to S2-1-1.

[0093] S2-3-4. Determine whether the trend of auxiliary feature labels in the continuous period of operation and maintenance improvement dataset from the median time t to t+1 is consistent with the trend of interference feature labels. If so, retain the current continuous period of operation and maintenance improvement dataset and the result of continuous fluctuation screening is normal. Otherwise, retain the current continuous period of operation and maintenance improvement dataset and the result of continuous fluctuation screening is abnormal.

[0094] S3 specifically includes:

[0095] S3-1. Establish a correlation analysis model for real-time operation and maintenance data based on real-time operation and maintenance data;

[0096] S3-2. The characteristic marker data of real-time operation and maintenance data is obtained by comparing the results of continuous fluctuation screening with the correlation analysis model.

[0097] S3-3. Generate security monitoring results for operation and maintenance data using feature markers from real-time operation and maintenance data.

[0098] S3-1 specifically includes:

[0099] S3-1-1. Utilize real-time operation and maintenance data to establish an initial correlation model based on the historical data repository according to the standard median time t.

[0100] S3-1-2. Based on the initial association model, sequentially obtain the corresponding subsequent association events to establish a set of derived events;

[0101] S3-1-3. Obtain the corresponding event generation time from the derived event set and establish a derived time set;

[0102] S3-1-4. Use the initial correlation model, the derived event set, and the derived time set as the correlation analysis model for real-time operation and maintenance data.

[0103] S3-2 specifically includes:

[0104] S3-2-1. Determine whether the continuous fluctuation screening result is normal. If so, execute S3-2-2. Otherwise, use the derivative time set of the correlation analysis model to obtain the corresponding subset of the derivative event set at time t+1 corresponding to the continuous fluctuation screening result as the initial comparison label, and directly execute S3-2-3.

[0105] S3-2-2 Determine whether the trend of the continuous fluctuation screening result corresponding to the continuous period operation and maintenance improvement dataset and the initial correlation model of the correlation analysis model are consistent. If so, use the continuous fluctuation screening result corresponding to the real-time operation and maintenance improvement dataset and the correlation analysis model as the feature label data of the real-time operation and maintenance data. Otherwise, use the initial correlation model with inconsistent trends as the feature label data of the real-time operation and maintenance data.

[0106] S3-2-3. Determine whether the initial comparison label is consistent with the real-time operation and maintenance data at time t+1. If so, use the real-time operation and maintenance improvement dataset corresponding to the continuous fluctuation screening result and the correlation analysis model as the feature label data of the real-time operation and maintenance data. Otherwise, use the initial comparison label as the feature label data of the real-time operation and maintenance data.

[0107] Among them, the consistent trend means that the changes in the data content of the continuous time period operation and maintenance improvement dataset and the initial correlation model correspond to each other with the time of data change.

[0108] S3-3 specifically includes:

[0109] S3-3-1. Determine whether the feature marker data of the real-time operation and maintenance data exists in the real-time operation and maintenance improvement dataset and correlation analysis model. If so, the operation and maintenance data security monitoring result is output normally, and the current real-time operation and maintenance improvement dataset and correlation analysis model are retained. Otherwise, execute S3-3-2.

[0110] S3-3-2. Determine whether the feature marker data of the real-time operation and maintenance data corresponds to the initial association model with inconsistent trends. If so, the operation and maintenance data security monitoring result is an abnormal output, and the initial association model with inconsistent current trends is used as the abnormal label. Otherwise, the initial association model corresponding to the initial comparison label is used as the abnormal label.

[0111] In practical applications, this invention first acquires equipment operation data, personnel operation data, and environmental condition data as real-time maintenance data through a multi-source data acquisition system 1. The multi-source data acquisition system 1 is the foundational module of the entire process, achieving real-time data acquisition through connections with multiple sensors, monitoring equipment, and operating terminals. These sensors and terminals are installed at different locations within the rail transit depot, such as along the track, in station control rooms, and equipment rooms, to ensure comprehensive data coverage. The operating parameters of the multi-source data acquisition system 1 include sampling frequency, data transmission rate, and equipment operating status; these parameters directly affect the quality and real-time performance of the acquired data. To ensure data reliability, the signal-to-noise ratio (SNR) is calculated based on the operating parameters during acquisition, and this value is added as an interference feature label to the real-time maintenance data. Furthermore, the operating parameters of the multi-source data acquisition system 1 themselves also serve as auxiliary feature labels for verification and filtering in subsequent data processing steps.

[0112] After data acquisition, the real-time operation and maintenance data is sent to the data filtering and processing module, where it undergoes preliminary processing to generate Real-time Operation and Maintenance Comprehensive Dataset 2. This process includes several sub-steps. First, data preprocessing tools are used to remove outliers, align data, and suppress noise in the real-time operation and maintenance data, resulting in the initial improved dataset. Then, the auxiliary feature labels and interference feature labels of the current-time initial improved dataset and the Real-time Operation and Maintenance Comprehensive Dataset 2 are compared to determine data consistency. If the consistency check fails, processing is repeated until the conditions are met or the maximum number of processing iterations is reached. This process ensures that the data in Real-time Operation and Maintenance Comprehensive Dataset 2 has high accuracy and completeness, laying the foundation for subsequent analysis.

[0113] Next, based on the real-time operation and maintenance comprehensive dataset 2, continuous fluctuation filtering is performed to generate continuous fluctuation filtering result 3. The core of this step is to establish a continuous time period operation and maintenance improvement dataset and perform fluctuation analysis on it. Specifically, firstly, taking the standard median time t as the center, real-time operation and maintenance improvement datasets at time t-1 and t+1 are obtained respectively. Then, using the data at these three times, the real-time operation and maintenance data change trend, auxiliary feature label change trend, and interference feature label change trend are constructed to form a continuous time period operation and maintenance improvement dataset. On this basis, a series of logical judgments are used to filter out data with abnormal fluctuations. For example, it is judged whether the real-time operation and maintenance data change trend from time t-1 to the standard median time t is consistent with the auxiliary feature label change trend. If they are consistent, it is further judged whether the auxiliary feature label change trend matches the interference feature label; otherwise, it is returned to reprocessing. Similarly, the same logical judgment is also required for the data from the standard median time t to t+1. Finally, based on the filtering results, the continuous fluctuation filtering result 3 is marked as normal or abnormal.

[0114] After completing the continuous fluctuation screening, the process moves to the generation stage of operation and maintenance data security monitoring result 4. The key to this stage is using correlation analysis model 5 to compare the continuous fluctuation screening result 3 and extract characteristic marker data from the real-time operation and maintenance data. The correlation analysis model 5 is constructed based on a historical data repository. An initial correlation model is established using historical data at the standard median time t, and further expanded with derivative event sets and derivative time sets. During the comparison process, it is first determined whether the continuous fluctuation screening result 3 is normal. If normal, the trend consistency between the continuous time period operation and maintenance improvement dataset and the initial correlation model is directly compared; if abnormal, a subset of derivative events corresponding to time t+1 is obtained from the derivative time set as the initial comparison label, and the initial comparison label is further compared with the real-time operation and maintenance data at time t+1. Finally, characteristic marker data for real-time operation and maintenance data is generated based on the comparison results.

[0115] Finally, the operation and maintenance data security monitoring result 4 is generated based on the feature marker data. This process includes two logical judgments. First, it is determined whether there is a match between the real-time operation and maintenance improvement dataset and the correlation analysis model 5 in the feature marker data. If a match exists, the operation and maintenance data security monitoring result 4 is determined to be a normal output, and the current real-time operation and maintenance improvement dataset and correlation analysis model 5 are retained; otherwise, proceed to the next judgment. Second, it is determined whether the feature marker data corresponds to an initial correlation model with inconsistent trends. If a match exists, the operation and maintenance data security monitoring result 4 is determined to be an abnormal output, and the current initial correlation model with inconsistent trends is recorded as an abnormal label; if no match exists, the initial correlation model corresponding to the initial comparison label is used as the abnormal label.

[0116] The above steps, through the coordinated operation of a multi-source data acquisition system 1, a real-time comprehensive operation and maintenance dataset 2, continuous fluctuation screening results 3, operation and maintenance data security monitoring results 4, and a correlation analysis model 5, achieve a safety assessment of rail transit depot operation and maintenance data. The multi-source data acquisition system 1 is responsible for comprehensive data collection; the real-time comprehensive operation and maintenance dataset 2 ensures data accuracy and completeness; the continuous fluctuation screening results 3 enables detailed analysis of data fluctuations; and the correlation analysis model 5 enhances anomaly detection capabilities through in-depth mining of historical data. These modules are closely linked through data flow and logical judgments, collectively forming a complete closed-loop assessment system.

[0117] To enable those skilled in the art to fully understand and implement this invention, the specific implementation principle of this invention is further explained below in conjunction with a specific application scenario.

[0118] In actual operation and maintenance scenarios of rail transit depots, assuming a subway depot is undergoing routine operation and maintenance, the system needs to collect and assess multi-source data in real time, including vehicle vibration data, traction motor current data, and personnel operation trajectories. First, multiple sensors and monitoring devices in the multi-source data acquisition system 1 acquire real-time data from vibration sensors along the track, traction motor current monitoring modules, and personnel positioning terminals. These devices operate at preset sampling frequencies and transmission rates to ensure data can be quickly and synchronously uploaded to the system. Simultaneously, the system calculates and outputs a signal-to-noise ratio (SNR) value based on the equipment's operating status and appends this value as an interference feature label to the real-time operation and maintenance data to identify potential noise interference. Furthermore, the operating parameters of the multi-source data acquisition system (such as sampling frequency and transmission rate) are also recorded as auxiliary feature labels for subsequent data verification and filtering.

[0119] Subsequently, the real-time operation and maintenance data is sent to the data filtering and processing module to generate Real-time Operation and Maintenance Comprehensive Dataset 2. During this process, the system uses data preprocessing tools to remove outliers, align data, and suppress noise in the original data, resulting in a preliminarily improved dataset. For example, if the amplitude of traction motor current data at a given moment exceeds the normal range and does not conform to historical data trends, it will be marked as outlier and removed. Simultaneously, the system performs a consistency check on the auxiliary and interference feature labels of the current moment's initial improved real-time operation and maintenance dataset and Real-time Operation and Maintenance Comprehensive Dataset 2. If the check fails, the data cleaning steps are repeated until the conditions are met or the maximum number of processing iterations is reached. This process ensures that the data in Real-time Operation and Maintenance Comprehensive Dataset 2 has high accuracy and completeness, providing a reliable foundation for subsequent analysis.

[0120] Next, based on the real-time operation and maintenance comprehensive dataset 2, the system performs continuous fluctuation filtering to generate continuous fluctuation filtering result 3. Specifically, the system uses the standard median time t as the center and obtains real-time operation and maintenance improvement datasets at times t-1 and t+1 respectively. It then uses the data from these three times to construct real-time operation and maintenance data change trends, auxiliary feature label change trends, and interference feature label change trends. For example, when analyzing vibration data for a certain track segment, if the vibration amplitude change trend from time t-1 to the standard median time t is found to be consistent with the auxiliary feature label change trend, but the auxiliary feature label change trend does not match the interference feature label, the system will return to reprocessing. Similarly, the system also needs to perform the same logical judgment on the data from the standard median time t to t+1. Finally, based on the filtering results, the continuous fluctuation filtering result 3 is marked as normal or abnormal. For example, if the vibration data fluctuation amplitude at a certain time significantly deviates from the historical average and fails the consistency check, the data will be marked as abnormal.

[0121] After completing the continuous fluctuation screening, the system enters the generation stage of operation and maintenance data security monitoring result 4. The key to this stage is to use the correlation analysis model 5 to compare the continuous fluctuation screening result 3. The correlation analysis model 5 builds an initial correlation model based on the historical data repository and further expands the derivative event set and derivative time set. For example, when the continuous fluctuation screening result 3 is normal, the system directly compares the trend consistency between the continuous time period operation and maintenance improvement dataset and the initial correlation model. If the trend is consistent, the data is determined to be normal; otherwise, the initial correlation model with inconsistent trends is used as an anomaly label. If the continuous fluctuation screening result 3 is abnormal, the system uses the derivative time set to obtain the derivative event subset corresponding to time t+1 as the initial comparison label, and further compares the initial comparison label with the real-time operation and maintenance data at time t+1. Finally, feature label data of real-time operation and maintenance data is generated based on the comparison results. For example, if the traction motor current data at a certain time has a significant deviation from the historical data trend and cannot pass the comparison, the data will be marked as abnormal.

[0122] Finally, the system generates the operation and maintenance data security monitoring result 4 based on the feature marker data. The system first determines whether there is a match between the real-time operation and maintenance improvement dataset and the correlation analysis model 5 in the feature marker data. If a match exists, the operation and maintenance data security monitoring result 4 is determined to be a normal output, and the current real-time operation and maintenance improvement dataset and correlation analysis model 5 are retained; otherwise, it further determines whether the feature marker data corresponds to an initial correlation model with inconsistent trends. If a match exists, the operation and maintenance data security monitoring result 4 is determined to be an abnormal output, and the initial correlation model with inconsistent trends is recorded as an anomaly label; if no match exists, the initial correlation model corresponding to the initial comparison label is used as the anomaly label. For example, in a real-world scenario, if the system detects a significant deviation between the vibration data of a certain track section and the historical data trend, and the data cannot be compared through the correlation analysis model, then the data will be marked as abnormal, triggering a risk warning mechanism.

[0123] The above steps, through the coordinated operation of a multi-source data acquisition system 1, a real-time comprehensive operation and maintenance dataset 2, continuous fluctuation screening results 3, operation and maintenance data security monitoring results 4, and a correlation analysis model 5, achieve a safety assessment of operation and maintenance data for rail transit depots. The modules are closely linked through data flow and logical judgment, forming a complete closed-loop assessment system. For example, in a rainy or snowy weather scenario, the system can dynamically adjust the threshold range of vibration data based on real-time environmental conditions, avoiding misjudgments caused by fixed thresholds. Simultaneously, through in-depth mining of historical data and cross-comparison of real-time scenario data, the system can effectively identify concealed data tampering or abnormal injection, thereby improving the accuracy of data tampering identification and the ability to detect abnormal data. Furthermore, the system can predict data anomalies based on data trends before they occur, proactively blocking risk transmission paths and achieving fully automated control from data acquisition to risk warning.

[0124] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0125] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0126] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0127] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for security assessment of rail transit depot operation and maintenance data, characterized in that, include: S1. Obtain real-time operation and maintenance data from a multi-source data acquisition system and perform data filtering to obtain a comprehensive real-time operation and maintenance dataset. S2. Perform continuous fluctuation filtering on the real-time operation and maintenance comprehensive dataset to obtain the continuous fluctuation filtering results of the real-time operation and maintenance data. S2-1. Use the real-time operation and maintenance comprehensive dataset to perform data cleaning to obtain the real-time operation and maintenance improvement dataset; S2-2. Establish a continuous time period operation and maintenance improvement dataset using the real-time operation and maintenance improvement dataset; S2-2-1. Use real-time operation and maintenance to improve the corresponding time of the dataset as the standard median time t; S2-2-2, Obtain the real-time operation and maintenance improvement datasets at time t-1 and t+1 respectively; S2-2-3. Establish the real-time operation and maintenance data change trend using the real-time operation and maintenance improvement dataset at time t-1, standard median time t, and time t+1. S2-2-4. Establish the trend of auxiliary feature label changes by using the auxiliary feature labels of the real-time operation and maintenance improvement dataset at time t-1, standard median time t, and time t+1. S2-2-5. Establish the trend of interference feature label changes by using the interference feature labels of the real-time operation and maintenance improvement dataset at time t-1, standard median time t, and time t+1. S2-2-6. Use the real-time operation and maintenance data change trend, auxiliary feature label change trend and interference feature label change trend as a continuous time period operation and maintenance improvement dataset; S2-3. Continuous fluctuation filtering is performed on the continuous time period operation and maintenance improvement dataset to obtain the continuous fluctuation filtering results of real-time operation and maintenance data. S3. Utilize the continuous fluctuation screening results to generate operation and maintenance data security monitoring results.

2. The method for security assessment of rail transit depot operation and maintenance data as described in claim 1, characterized in that, The real-time operation and maintenance comprehensive dataset obtained by acquiring real-time operation and maintenance data through a multi-source data acquisition system and performing data filtering processing includes: S1-1. Acquire equipment operation data, personnel operation data, and environmental condition data based on a multi-source data acquisition system as real-time operation and maintenance data; S1-2. Obtain the operating parameters of the corresponding multi-source data acquisition system based on real-time operation and maintenance data; S1-3. Calculate the output signal-to-noise ratio value based on the operating parameters of the multi-source data acquisition system; S1-4. Use the operating parameters of the multi-source data acquisition system as auxiliary feature labels for real-time operation and maintenance data; S1-5. Use the output signal-to-noise ratio value as an interference feature label for real-time operation and maintenance data; S1-6. Use real-time operation and maintenance data, auxiliary feature labels and interference feature labels of real-time operation and maintenance data as a comprehensive real-time operation and maintenance dataset; The operating parameters of a multi-source data acquisition system include sampling frequency, data transmission rate, and equipment operating status.

3. The method for security assessment of rail transit depot operation and maintenance data as described in claim 1, characterized in that, Data cleaning using the real-time operations and maintenance comprehensive dataset yielded a real-time operations and maintenance improvement dataset, which includes: S2-1-1. Using the real-time operation and maintenance data from the real-time operation and maintenance comprehensive dataset, the data is comprehensively processed based on the data preprocessing tool to obtain the initial improved dataset for real-time operation and maintenance. S2-1-2. Determine whether the auxiliary feature labels of the current real-time operation and maintenance initial improvement dataset and the real-time operation and maintenance comprehensive dataset are consistent. If they are, proceed directly to S2-1-4; otherwise, proceed to S2-1-3. S2-1-3. Determine if the number of times S2-1-2 has been executed at the current time is greater than 1. If yes, abandon the process; otherwise, return to S2-1-1. S2-1-4. Determine whether the interference feature labels of the current real-time operation and maintenance initial improvement dataset and the real-time operation and maintenance comprehensive dataset are consistent. If they are, use the real-time operation and maintenance initial improvement dataset as the real-time operation and maintenance improvement dataset. Otherwise, execute S2-1-5. S2-1-5. Determine if the number of times S2-1-4 has been executed at the current time is greater than 1. If yes, abandon the process; otherwise, return to S2-1-1. The comprehensive data processing includes outlier removal, data alignment, and noise suppression.

4. The method for security assessment of rail transit depot operation and maintenance data as described in claim 1, characterized in that, The continuous fluctuation filtering results obtained by using the continuous time period operation and maintenance improvement dataset for continuous fluctuation filtering include: S2-3-1. Determine whether the real-time operation and maintenance data change trend of the operation and maintenance improvement dataset from time t-1 to the standard median time t is consistent with the auxiliary feature label change trend. If yes, execute S2-3-2; otherwise, return to S2-1-1. S2-3-2. Determine whether the trend of auxiliary feature labels in the continuous time period from time t-1 to the standard median time t is consistent with the trend of interference feature labels. If yes, execute S2-3-3; otherwise, return to S1-2. S2-3-3: Determine whether the real-time operation and maintenance data change trend of the operation and maintenance improvement dataset during the continuous period from the standard median time t to time t+1 is consistent with the auxiliary feature label change trend. If yes, execute S2-3-4; otherwise, return to S2-1-1. S2-3-4. Determine whether the trend of auxiliary feature labels in the continuous period of operation and maintenance improvement dataset from the median time t to t+1 is consistent with the trend of interference feature labels. If so, retain the current continuous period of operation and maintenance improvement dataset and the result of continuous fluctuation screening is normal. Otherwise, retain the current continuous period of operation and maintenance improvement dataset and the result of continuous fluctuation screening is abnormal.

5. The method for security assessment of rail transit depot operation and maintenance data as described in claim 1, characterized in that, The results of operation and maintenance data security monitoring generated by using continuous fluctuation screening results include: S3-1. Establish a correlation analysis model for real-time operation and maintenance data based on real-time operation and maintenance data; S3-2. The characteristic marker data of real-time operation and maintenance data is obtained by comparing the results of continuous fluctuation screening with the correlation analysis model. S3-3. Generate security monitoring results for operation and maintenance data using feature markers from real-time operation and maintenance data.

6. The method for security assessment of rail transit depot operation and maintenance data as described in claim 5, characterized in that, Establishing a correlation analysis model for real-time operation and maintenance data based on real-time operation and maintenance data includes: An initial correlation model is established based on a historical data repository using real-time operation and maintenance data and the standard median time t. Based on the initial association model, obtain the corresponding subsequent associated events to establish a set of derived events; Based on the derived event set, obtain the corresponding event generation time and establish a derived time set; The initial correlation model, the derived event set, and the derived time set are used as the correlation analysis model for real-time operation and maintenance data.

7. The method for security assessment of rail transit depot operation and maintenance data as described in claim 6, characterized in that, The characteristic marker data of real-time operation and maintenance data obtained by comparing the results with the continuous fluctuation screening results using the correlation analysis model include: S3-2-1. Determine whether the continuous fluctuation screening result is normal. If so, execute S3-2-2. Otherwise, use the derivative time set of the correlation analysis model to obtain the corresponding subset of the derivative event set at time t+1 corresponding to the continuous fluctuation screening result as the initial comparison label, and directly execute S3-2-3. S3-2-2 Determine whether the trend of the continuous fluctuation screening result corresponding to the continuous period operation and maintenance improvement dataset and the initial correlation model of the correlation analysis model are consistent. If so, use the continuous fluctuation screening result corresponding to the real-time operation and maintenance improvement dataset and the correlation analysis model as the feature label data of the real-time operation and maintenance data. Otherwise, use the initial correlation model with inconsistent trends as the feature label data of the real-time operation and maintenance data. S3-2-3. Determine whether the initial comparison label is consistent with the real-time operation and maintenance data at time t+1. If so, use the real-time operation and maintenance improvement dataset corresponding to the continuous fluctuation screening result and the correlation analysis model as the feature label data of the real-time operation and maintenance data. Otherwise, use the initial comparison label as the feature label data of the real-time operation and maintenance data. Among them, the consistent trend means that the changes in the data content of the continuous time period operation and maintenance improvement dataset and the initial correlation model correspond to each other with the time of data change.

8. The method for security assessment of rail transit depot operation and maintenance data as described in claim 7, characterized in that, The generation of operation and maintenance data security monitoring results using feature data of real-time operation and maintenance data includes: S3-3-1. Determine whether the feature marker data of the real-time operation and maintenance data exists in the real-time operation and maintenance improvement dataset and correlation analysis model. If so, the operation and maintenance data security monitoring result is output normally, and the current real-time operation and maintenance improvement dataset and correlation analysis model are retained. Otherwise, execute S3-3-2. S3-3-2. Determine whether the feature marker data of the real-time operation and maintenance data corresponds to the initial association model with inconsistent trends. If so, the operation and maintenance data security monitoring result is an abnormal output, and the initial association model with inconsistent current trends is used as the abnormal label. Otherwise, the initial association model corresponding to the initial comparison label is used as the abnormal label.

Images