Cable Channel Detection Method and System Based on Distributed Optical Fiber Sensing Technology

Abstract

The present application provides a cable channel detection method and system based on distributed optical fiber sensing technology, which relates to the technical field of cable detection, and includes: configuring a monitoring area and an uncovered area according to distribution coordinates and monitoring accuracy; using a distributed optical fiber sensor to perform cable monitoring to establish a monitoring data set; performing collaborative anomaly recognition on the monitoring data set to establish a first anomaly recognition result; inputting the monitoring data set into a distance-temperature interpolation prediction network, performing interpolation prediction on the uncovered area, using the interpolation prediction result to perform fault risk prediction, and establishing a second anomaly recognition result; after performing anomaly fusion on the first and second anomaly recognition results, reporting cable anomalies. Through the present application, the technical problem that the traditional detection method fails to comprehensively cover all monitoring areas, resulting in monitoring blind spots and being unable to effectively detect the uncovered area, can be solved, the monitoring blind spot problem can be effectively eliminated, and full-area and blind-spot-free anomaly monitoring and fault warning can be realized.

Description

Technical Field

[0001] This application relates to the technical field of cable detection, and particularly to a cable channel detection method and system based on distributed optical fiber sensing technology. Background Art

[0002] In recent years, distributed optical fiber sensing technology has been widely applied in the field of cable monitoring. By using optical fiber as the sensing medium and leveraging the scattering effects inside the optical fiber (such as Rayleigh scattering, Brillouin scattering, etc.), this technology can achieve continuous monitoring of cable channels. Compared with traditional point sensors, distributed optical fiber sensing technology has a higher spatial resolution and real-time monitoring ability, and can obtain information such as temperature, stress, and vibration over a long distance, providing a more comprehensive monitoring coverage.

[0003] However, although distributed optical fiber sensing technology can provide a wide monitoring range, in actual application scenarios, it still faces some challenges. Especially in cable systems, some areas may have blind spots where effective monitoring cannot be achieved due to factors such as improper sensor layout, environmental impact, or equipment configuration. These monitoring blind spots may pose potential fault hazards. If not discovered in time, they may lead to equipment damage or outage, thus increasing the operation and maintenance costs and the risk of system failures. Summary of the Invention

[0004] The purpose of this application is to provide a cable channel detection method and system based on distributed optical fiber sensing technology, so as to solve the technical problem that the traditional cable channel detection method based on distributed optical fiber sensing technology fails to fully cover all monitoring areas, resulting in monitoring blind spots and being unable to effectively detect the uncovered areas.

[0005] In view of the above problems, this application provides a cable channel detection method and system based on distributed optical fiber sensing technology.

[0006] In a first aspect, the present application provides a cable channel detection method based on distributed optical fiber sensing technology, which is implemented through a cable channel detection system based on distributed optical fiber sensing technology, including: obtaining the distribution coordinates of distributed optical fiber sensors, configuring a monitoring area and an uncovered area according to the distribution coordinates and monitoring accuracy; using the distributed optical fiber sensors to perform cable monitoring, establishing a monitoring data set, the monitoring data set including temperature data, vibration data, and stress data, and the monitoring data set being marked with monitoring position identifiers; using a fault identification channel to perform collaborative anomaly identification on the monitoring data set, establishing a first anomaly identification result; inputting the monitoring data set into a distance-temperature interpolation prediction network, performing interpolation prediction on the uncovered area, and establishing an interpolation prediction result; using the interpolation prediction result to perform fault risk prediction, establishing a second anomaly identification result; after anomaly fusion of the first anomaly identification result and the second anomaly identification result, reporting cable anomalies.

[0007] Optionally, the cable channel detection method based on distributed optical fiber sensing technology further includes: using the dynamic weight optimization layer of the distance-temperature interpolation prediction network to calculate the dynamic weights of interpolation points based on distance and temperature as follows: ; where represents the dynamic weight of the unknown point predicted based on the known point , represents the distance between the unknown point and the known point , represents the position of the th point, represents the spatial distance smoothing parameter, is the distance exponential factor, represents the temperature of the known point , represents the average temperature of the known points within the neighborhood window, is the smoothing parameter for temperature difference, is the total number of known points, is the index of any known point; performing interpolation prediction based on the monitoring data set according to the calculation result of the dynamic weights of interpolation points: ; where is the interpolation prediction feature, is the feature vector of the th known point, , is the stress eigenvalue, is the vibration eigenvalue.

[0008] Optionally, the cable channel detection method based on distributed optical fiber sensing technology further includes: performing fault risk prediction through a fault probability prediction model as follows: ; where represents the fault probability of the unknown point , represents the interpolation prediction feature mode, is the mode normalization parameter, represents the inner product of the interpolation prediction feature and its neighborhood average feature , is the inner product normalization parameter, represents the historical change amount, is the historical change normalization parameter; generating a second abnormal recognition result according to the fault probability of the unknown point .

[0009] Optionally, the cable channel detection method based on distributed optical fiber sensing technology further includes: after classifying the monitoring data set by the preprocessing layer in the fault recognition channel based on the monitoring position identifier and the acquisition time of the monitoring data set, performing outlier extraction to establish a preprocessing data set; using the feature extraction layer in the fault recognition channel to extract the change rate, peak value, and window abnormal features of the preprocessing data set to establish a feature extraction data set; performing feature space correlation analysis based on the feature extraction data set through the spatial collaborative analysis layer to establish a first collaborative abnormal recognition result; performing feature time correlation analysis based on the feature extraction data set through the time collaborative analysis layer to establish a second collaborative abnormal recognition result; after fusing the first collaborative abnormal recognition result and the second collaborative abnormal recognition result, outputting a first abnormal recognition result.

[0010] Optionally, the cable channel detection method based on distributed optical fiber sensing technology further includes: obtaining an abnormal fusion result, and obtaining abnormal type, abnormal position, and abnormal value information based on the abnormal fusion result; obtaining the current working parameters of the cable, and generating first auxiliary verification information based on the working parameters; obtaining the position environment information of the cable, and generating second auxiliary verification information according to the position environment information; using the abnormal type, the abnormal position, the abnormal value information, the first auxiliary verification information, and the second auxiliary verification information to perform abnormal trigger level recognition; using the abnormal trigger level recognition result to report cable abnormality.

[0011] Optionally, the cable channel detection method based on distributed optical fiber sensing technology further includes: obtaining the model information and usage information of the distributed optical fiber sensor; performing accuracy loss identification according to the model information and the usage information, and establishing an accuracy loss identification result; reconstructing the monitoring accuracy based on the accuracy loss identification result and the model information, and configuring the monitoring area and the uncovered area by using the reconstructed monitoring accuracy.

[0012] Optionally, the cable channel detection method based on distributed optical fiber sensing technology further includes: establishing an accumulated weighted coefficient for anomalies; when a cable is abnormal at any position and has not been repaired, accumulating the cable anomalies according to the accumulated weighted coefficient to establish an accumulation result; and performing cable anomaly management according to the accumulation result.

[0013] In a second aspect, the present application further provides a cable channel detection system based on distributed optical fiber sensing technology for executing the cable channel detection method based on distributed optical fiber sensing technology as described in the first aspect, including: a region configuration module for obtaining the distribution coordinates of the distributed optical fiber sensor and configuring the monitoring area and the uncovered area according to the distribution coordinates and the monitoring accuracy; a monitoring data establishment module for using the distributed optical fiber sensor to perform cable monitoring and establishing a monitoring data set, where the monitoring data set includes temperature data, vibration data, and stress data, and the monitoring data set is marked with monitoring position identifiers; a first anomaly identification module for using a fault identification channel to perform collaborative anomaly identification on the monitoring data set and establishing a first anomaly identification result; an interpolation prediction module for inputting the monitoring data set into a distance-temperature interpolation prediction network to perform interpolation prediction on the uncovered area and establishing an interpolation prediction result; a second anomaly identification module for using the interpolation prediction result to perform fault risk prediction and establishing a second anomaly identification result; and an anomaly fusion module for performing anomaly fusion on the first anomaly identification result and the second anomaly identification result and then reporting cable anomalies.

[0014] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0015] By obtaining the distribution coordinates of the distributed optical fiber sensor, the monitoring area and the uncovered area are configured according to the distribution coordinates and the monitoring accuracy; then the distributed optical fiber sensor is used to monitor the cable, and a monitoring data set is established. The monitoring data set includes temperature data, vibration data, and stress data, and the monitoring data set is marked with monitoring position identifiers; then a collaborative anomaly recognition of the monitoring data set is performed using a fault recognition channel to establish a first anomaly recognition result; on the other hand, the monitoring data set is input into a distance-temperature interpolation prediction network to perform interpolation prediction of the uncovered area and establish an interpolation prediction result; further, the interpolation prediction result is used for fault risk prediction to establish a second anomaly recognition result; finally, after anomaly fusion of the first anomaly recognition result and the second anomaly recognition result, a cable anomaly is reported; it can effectively eliminate the problem of monitoring blind spots in cable channel detection, timely discover potential anomalies in the uncovered area, thereby realizing full-area and blind-spot-free anomaly monitoring and fault warning, and ensuring the safe and stable operation of cable equipment.

[0016] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically illustrates the specific embodiments of the present application. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become easily understandable through the following description. Brief Description of the Drawings

[0017] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings described below are only exemplary, and for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the provided drawings.

[0018] Figure 1 It is a schematic flowchart of a cable channel detection method based on distributed optical fiber sensing technology of the present application;

[0019] Figure 2 It is a schematic flowchart of establishing a first anomaly recognition result in a cable channel detection method based on distributed optical fiber sensing technology of the present application;

[0020] Figure 3 It is a schematic structural diagram of a cable channel detection system based on distributed optical fiber sensing technology of the present application.

[0021] Description of the reference numerals:

[0022] Area configuration module 01, monitoring data establishment module 02, first anomaly recognition module 03, interpolation prediction module 04, second anomaly recognition module 05, anomaly fusion module 06. Detailed implementation

[0023] By providing a cable channel detection method and system based on distributed optical fiber sensing technology, this application solves the technical problem that the traditional cable channel detection method based on distributed optical fiber sensing technology fails to comprehensively cover all monitoring areas, resulting in monitoring blind spots and being unable to effectively detect the uncovered areas. It can effectively eliminate the monitoring blind spot problem in cable channel detection, timely discover potential anomalies in the uncovered areas, thereby achieving full-area and blind-spot-free anomaly monitoring and fault warning, and ensuring the safe and stable operation of cable equipment.

[0024] Next, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited by the example embodiments described here. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of this application. Additionally, it should be noted that for the sake of description, only parts related to this application are shown in the accompanying drawings rather than all of them.

[0025] Embodiment 1, please refer to the attached Figure 1 , this application provides a cable channel detection method based on distributed optical fiber sensing technology, which is applied to a cable channel detection system based on distributed optical fiber sensing technology, and specifically includes the following steps:

[0026] S100: Obtain the distribution coordinates of the distributed optical fiber sensor, and configure the monitoring area and the uncovered area according to the distribution coordinates and the monitoring accuracy.

[0027] Furthermore, step S100 of this application further includes:

[0028] S110: Obtain the model information and usage information of the distributed optical fiber sensor; S120: Perform accuracy loss recognition according to the model information and the usage information, and establish an accuracy loss recognition result; S130: Reconstruct the monitoring accuracy based on the accuracy loss recognition result and the model information, and configure the monitoring area and the uncovered area by using the reconstructed monitoring accuracy.

[0029] Specifically, first, obtain the model information, usage information, and location coordinates of all distributed fiber optic sensors in the distributed fiber optic sensor. Among them, the model information includes sensing type, sensitivity range, measurement accuracy, applicable environment, etc. These information help to understand the performance and applicable range of the sensor; the usage information includes installation time, working environment (such as temperature, humidity, electromagnetic interference, etc.), whether it has been calibrated, whether it has been maintained, etc. These information can reflect the actual usage of the sensor and the change of sensing accuracy; the distribution coordinates refer to the spatial position coordinates of the sensor, indicating its specific position in the cable channel or equipment. Through these coordinates, the monitoring coverage area of the sensor can be clarified and data support can be provided for subsequent area configuration.

[0030] Next, perform accuracy loss identification based on the model information and the usage information, that is, based on the model information and usage information of the sensor, analyze the factors that may cause accuracy loss, such as fiber optic aging caused by long-term use, the impact of temperature change on the performance of the sensor, long-term non-maintenance of the equipment, etc. Based on these factors, use statistical methods or machine learning algorithms to identify the sensors with accuracy loss and the degree of their influence. For example, according to the usage of the sensor, select the key features that may affect accuracy loss (such as usage time, temperature fluctuation, environmental interference, sensor model, etc.), then adopt the regression analysis method, combine historical data or experimental data to train the regression model, and use the trained regression model to predict the accuracy loss of each sensor. The model will output the corresponding accuracy loss value according to the input features (such as the working environment and usage duration of the sensor), and establish the accuracy loss identification result. Through regression analysis, a quantitative evaluation of the accuracy loss can be provided for each sensor, and it can be known in real time which sensors' monitoring accuracy has changed, providing a basis for optimizing the configuration of the monitoring area.

[0031] Then, based on the accuracy loss identification result and the model information, the monitoring accuracy of each sensor is reconstructed. When each distributed fiber optic sensor is installed, an initial monitoring accuracy is set, which is usually provided by the sensor manufacturer and obtained based on its specifications and test results. The initial monitoring accuracy is typically the accuracy value under ideal conditions, representing the optimal accuracy that the sensor can achieve without any external factor interference. Then, subtract the accuracy loss identification result from the initial monitoring accuracy of each sensor to obtain the actual monitoring accuracy. The reconstructed actual monitoring accuracy reflects the precision of the sensor in the actual working environment. Further, use the reconstructed actual monitoring accuracy and the distribution coordinates of the distributed fiber optic sensors to configure the monitored area and the uncovered area. First, calculate the range of the effective monitoring area of each sensor according to its actual monitoring accuracy. The higher the accuracy, the larger the area that the sensor can monitor. For example, a sensor with a smaller accuracy loss can monitor a larger range of cable areas, while a sensor with a larger accuracy loss may only be suitable for monitoring a smaller area. Then, based on the accuracy and distribution coordinates of the sensors, calculate the circular or polygonal monitoring area of each sensor. For example, if the monitoring accuracy of the sensor is r, the monitoring area of the sensor is a circular area centered at the sensor position with a radius of r. On the other hand, by calculating the monitoring areas of all sensors and combining the overall layout of the cable channel, identify the areas that are not monitored by any sensor, which are called uncovered areas. Uncovered areas are usually dead corners in the cable channel or areas with low sensor accuracy and cannot be effectively detected by existing sensors.

[0032] By combining the actual monitoring accuracy of the distributed fiber optic sensors to configure the monitored area and the uncovered area, the monitoring blind spots can be effectively identified and support can be provided for the effective detection of subsequent monitoring blind spots.

[0033] S200: Use the distributed fiber optic sensors to conduct cable monitoring and establish a monitoring data set, where the monitoring data set includes temperature data, vibration data, and stress data, and the monitoring data set is marked with monitoring location identifiers.

[0034] Specifically, distribute fiber optic sensors along the cable line. These sensors can monitor the entire cable in real time and transmit the collected raw data to the central processing system in real time through the sensor network or data acquisition module to obtain a monitoring data set. Among them, the monitoring data set includes temperature data, vibration data, and stress data. The temperature data is used to identify the temperature change of the cable channel and analyze overheating phenomena, especially when the cable load is heavy; the vibration data can identify the situation where the cable is affected by external physical interference, such as mechanical vibration, equipment vibration, or abnormal vibration generated by improper operation of the cable itself; the stress data is used to identify the stress condition of the cable and detect stress problems such as tension and compression that may occur during long-term use of the cable; among them, each data record is marked with a monitoring position identifier to clarify the specific position of the data point in the cable system. This position identifier can be defined based on the layout coordinates of the sensor, physical distance, or cable segment number.

[0035] S300: Use the fault identification channel to perform collaborative anomaly identification on the monitoring data set and establish a first anomaly identification result.

[0036] Furthermore, as Figure 2 shown, step S300 of this application further includes:

[0037] S310: After classifying the monitoring data set based on the monitoring position identifier and the acquisition time of the monitoring data set through the preprocessing layer in the fault identification channel, perform outlier extraction to establish a preprocessing data set; S320: Use the feature extraction layer in the fault identification channel to extract the change rate, peak value, and window anomaly features of the preprocessing data set to establish a feature extraction data set; S330: Perform feature space correlation analysis based on the feature extraction data set through the spatial collaborative analysis layer to establish a first collaborative anomaly identification result; S340: Perform feature time correlation analysis based on the feature extraction data set through the time collaborative analysis layer to establish a second collaborative anomaly identification result. S350: After fusing the first collaborative anomaly identification result and the second collaborative anomaly identification result, output the first anomaly identification result.

[0038] Specifically, first, through the preprocessing layer in the fault identification channel, the monitoring data set is classified according to the monitoring position identifier and the acquisition time of the monitoring data set. That is, according to the monitoring position identifier of each data point, the data set is grouped by position. Each position can represent different parts of the cable or different sensor positions. After classification, it can help determine which areas of the equipment may be abnormal; by classifying the monitoring data according to the acquisition time, data fluctuations in different time periods can be identified, which can further help analyze anomalies in time; through the combined classification of time and space, the monitoring data set can be segmented into multiple data subsets, and each data subset represents the cable state within a specific position and time range, facilitating detailed analysis for different regions and time periods. Then, outliers are extracted from multiple data subsets. Outliers refer to those data points that deviate from the normal data trend and may indicate cable faults or potential problems, including extreme data exceeding the preset threshold (such as temperature exceeding the safe range), abnormal data fluctuations (such as drastic changes in vibration sensor data, indicating that the cable has been severely impacted or mechanically damaged), data that remains stable for a long time (such as stress data remaining unchanged for a long time, indicating that the sensor may be faulty), etc. Through classification and outlier extraction, effective abnormal monitoring data is extracted to form a new preprocessing data set.

[0039] Then, the change rate, peak value, and window anomaly features of the preprocessed data set are extracted through the feature extraction layer in the fault identification channel. Among them, the change rate refers to the change amplitude between a certain data point and its previous and subsequent data points. By calculating the change rates of temperature, stress, or vibration, it is possible to identify whether the cable has significant fluctuations during a certain period, including the temperature change rate (identifying cable temperature fluctuations, too rapid temperature changes may indicate a fault), the vibration change rate (sudden vibration changes may indicate that the cable is subjected to external force impact or damage), and the stress change rate (excessive stress changes may indicate uneven stress on the cable or mechanical damage); the peak value refers to the highest value in a segment of data, usually representing the extreme manifestation of a certain abnormal state, including the temperature peak value (if the temperature value exceeds the set threshold within a certain period, it may mean that the equipment is overheating), the vibration peak value (abnormal vibration peak values may indicate that the cable has suffered excessive mechanical impact, which may cause damage), and the stress peak value (peak value changes in stress may indicate that the cable is under excessive stress and at risk of damage); window anomaly refers to abnormal fluctuations or data mutations within a specific time period. Usually, the data window refers to the set time period or a data interval of a fixed length. Calculate the statistics (such as mean, variance, etc.) within this interval and compare them with the expected values to identify whether there are anomalies in the data. For example, for each monitoring parameter, use a sliding window (such as 30 seconds, 1 minute, etc.) to calculate statistics such as the mean, variance, and standard deviation of the data within the window and compare them with the preset normal range. By comparing the statistics of each window with the set threshold, detect whether there are anomalies, including temperature window anomaly (abnormal temperature fluctuations within a specific time period may indicate overheating problems with the cable), vibration window anomaly (abnormal vibrations within a time window may reflect mechanical failures or external force interference with the cable), and stress window anomaly (if the stress data fluctuates too much within a time window, it may indicate uneven stress on the cable or potential faults); by calculating features such as change rate, peak value, and window anomaly, a new feature extraction data set is established. This data set contains the feature values of all sensors at different monitoring positions, providing data support for subsequent anomaly identification.

[0040] Furthermore, the feature extraction dataset is subjected to feature space correlation analysis through the spatial collaboration analysis layer. Through spatial collaboration analysis, the association relationships between different monitoring points (sensors) are identified, that is, whether multiple sensors show similar abnormal features at the same time, or whether there is a concentrated abnormal pattern in the same spatial region. Each record in the dataset contains the features of the monitoring data (such as temperature change rate, vibration peak value, stress window anomaly, etc.), and each feature has a corresponding monitoring position identifier. Spatial collaboration analysis determines the collaborative effect between different sensors by calculating the correlation or similarity between these feature values. For example, the correlation coefficient (such as the Pearson correlation coefficient) is used to measure the spatial correlation between sensors. A higher correlation may indicate that these sensors are in the same fault area. For adjacent sensors or sensors within the same region, the similarity between the features of these sensors can be calculated to determine whether they jointly indicate a certain potential fault. According to the results of the spatial correlation analysis, it is judged whether there is a spatially concentrated abnormal pattern. If the abnormal values of multiple sensors appear in the same area and these abnormalities show collaborative changes in space, a potential spatial anomaly can be identified and the first collaborative anomaly recognition result is generated. The first collaborative anomaly recognition result can mark the abnormal area with spatial correlation in the cable channel. This area may indicate that there is a certain common fault or anomaly in the cable, which requires the attention of maintenance personnel. For example, if multiple temperature sensors and stress sensors simultaneously show abnormal fluctuations in a certain area, this may mean that the cable in this area has problems such as overheating or excessive stress.

[0041] On the other hand, the feature time correlation analysis is performed on the feature extraction data set by the time collaborative analysis layer. Through time collaboration analysis, the correlation between the time series data of multiple monitoring points is identified, and whether the data synchronously exhibits anomalies in time is analyzed, so as to identify persistent faults, short-term sudden faults, or long-term trend changes, etc. Among them, the timestamp of each data point and the monitoring features (such as temperature, vibration, stress, etc.) together constitute the time series data. By analyzing these time series data, abnormal patterns in the time dimension are identified. Time series analysis methods, such as the autocorrelation function, can be used to analyze the synchrony between time series data, and calculate the abnormal change trend within a time period, such as whether the monitoring data such as temperature and stress have changed drastically at a certain time point, or whether they have continuously exceeded the normal range within a period of time. If multiple sensors exhibit anomalies within the same time period, or the abnormal fluctuations in different time periods have similar patterns, a time anomaly can be identified, and a second collaborative anomaly recognition result is generated. The second collaborative anomaly recognition result can identify those regions or sensors that show a consistent abnormal trend in time. These time anomalies may indicate continuous or sudden fault problems in the cable system. For example, if multiple sensors (such as temperature, vibration, stress, etc.) show synchronous abnormal fluctuations within a certain time period, then this time period will be identified as a potential fault occurrence period, a second collaborative anomaly recognition result will be generated, and the possible fault type will be speculated.

[0042] Finally, the first collaborative anomaly recognition result (based on spatial correlation) and the second collaborative anomaly recognition result (based on time correlation) are fused. For each anomaly recognition point (region or time period), different weights can be assigned to the spatial anomaly and the time anomaly according to the priorities of spatial and time anomalies in practical applications. According to the weight configuration result, weighted calculations are performed on the first collaborative anomaly recognition result and the second collaborative anomaly recognition result, and the first anomaly recognition result is output.

[0043] By fusing the anomaly recognition results in space and time, the feature extraction data set can be comprehensively analyzed from both spatial and time dimensions, so as to accurately locate the fault area and fault time of the cable, and provide accurate decision-making support for subsequent operation and maintenance.

[0044] S400: Input the monitoring data set into the distance-temperature interpolation prediction network, perform the interpolation prediction of the uncovered area, and establish the interpolation prediction result.

[0045] Furthermore, step S400 of the present application further includes:

[0046] S410: Use the dynamic weight optimization layer of the distance-temperature interpolation prediction network to calculate the dynamic weights of the interpolation points based on distance and temperature as follows: ; where Characterize based on known points Predicted unknown points Of the dynamic weight Characterize unknown points And known points Of the distance Characterize the Position of the point Characterize the spatial distance smoothing parameter Is the distance exponential factor Characterize known points Of the temperature Characterize the average temperature of the known points within the neighborhood window Is the smoothing parameter for temperature difference Is the total number of known points Is the index of any known point

[0047] Specifically, in the distance-temperature interpolation prediction network, the dynamic weight optimization level is used to calculate the dynamic weight of the interpolation points. This process adjusts the weights of the interpolation points in the uncovered area based on factors such as distance and temperature. That is, according to the formula and the dynamic weight optimization layer of the network, the dynamic weight of each interpolation point is calculated. This weight is determined by the distance and temperature information of the known points. The weight of the interpolation point reflects its similarity to the known data. Through the eigenvalue (such as temperature, stress, vibration, etc.) of the known data points and the calculated weight, interpolation calculation is performed to fill the data in the uncovered area, so as to improve the accuracy and reliability of the interpolation prediction

[0048] In the interpolation point weight calculation formula Characterize based on known points Predicted unknown points Of the dynamic weight, indicating the influence degree of the known point On the predicted point The weight is dynamically adjusted by considering the distance and temperature difference. The closer the distance or the smaller the temperature difference of the known point, the greater the influence on the unknown point Characterize unknown points And known points Of the distance (spatial position distance); Characterize the Position of the point Characterize the spatial distance smoothing parameter, which is used to control the influence degree of distance on the weight, and can adjust the attenuation speed of the weight with the increase of distance, thereby affecting the accuracy of interpolation. A smaller Will cause the influence of points with a longer distance on the interpolation result to decay rapidly, while a larger Will cause points with a longer distance to still have a greater influence on the interpolation; Is the distance exponential factor Characterize known points The temperature represents the average temperature of known points within the neighborhood window, that is, the average value of the temperatures of all known points within a certain time window. This value is used to measure the overall temperature level of the cable area and serves as a benchmark for comparison with the temperature differences of known points. is the smoothing parameter for temperature differences, used to control the degree of influence of temperature differences on the interpolation result. A larger will result in a smaller influence of temperature differences, while a smaller will enhance the influence of temperature differences. is the total number of known points, is the index of any known point, representing the unique identifier of each known point in the interpolation calculation.

[0049] S420: Perform interpolation prediction based on the monitoring data set according to the interpolation point dynamic weight calculation result; ; where is the interpolation prediction feature, is the feature vector of the th known point, , is the stress eigenvalue, is the vibration eigenvalue.

[0050] Specifically, then perform a weighted calculation on the monitoring data set according to the interpolation point dynamic weight calculation result to obtain the interpolation prediction result to fill the data in the uncovered area. The key to interpolation prediction lies in comprehensively considering the features of known points by calculating the weight of each interpolation point, so as to predict the features (such as temperature, stress, vibration, etc.) of the uncovered area.

[0051] In the interpolation prediction calculation function, is the interpolation prediction feature, which is the result of predicting by weighted averaging the features of known points, including the temperature eigenvalue, vibration eigenvalue, and stress eigenvalue; is the feature vector of the th known point, containing various monitoring data (such as temperature, stress, vibration, etc.) of this point, , is the stress eigenvalue, is the vibration eigenvalue; that is, the features of unknown points are predicted by weighted averaging according to the dynamic weights of interpolation points and the features of known points. By weighted averaging the features of known points (such as temperature, stress, vibration, etc.), various eigenvalue features of unknown points can be accurately predicted, filling the missing data in the uncovered area, so as to achieve effective detection of the uncovered area.

[0052] S500: Use the interpolation prediction result to perform fault risk prediction and establish a second abnormal recognition result.

[0053] Furthermore, step S500 of the present application further includes:

[0054] S510: Conduct fault risk prediction through a fault probability prediction model as follows: ; where represents the fault probability of the unknown point , represents the interpolation prediction feature mode, is the mode normalization parameter, represents the inner product of the interpolation prediction feature and its neighborhood average feature , is the inner product normalization parameter, represents the historical change amount, is the historical change normalization parameter; S520: Generate a second anomaly recognition result according to the fault probability of the unknown point .

[0055] Specifically, construct a fault probability prediction function. The fault probability prediction function calculates the fault probability of the unknown point and generates a second anomaly recognition result by combining data such as interpolation prediction features, historical change amounts, and neighborhood average features; in the fault probability prediction function, represents the fault probability of the unknown point . The greater the fault probability, the higher the fault risk of the unknown point is represented; represents the interpolation prediction feature mode, which is the mode normalization parameter, indicating the mode normalization of the interpolation prediction feature. Since different features may have different dimensions and ranges, performing mode normalization helps eliminate dimension differences and enables fair comparison of each feature in the calculation; represents the inner product of the interpolation prediction feature and its neighborhood average feature . The inner product reflects the similarity between the interpolation prediction feature and the neighborhood average feature, and can reveal the similarity and correlation of the unknown point in space. is the inner product normalization parameter, that is, in order to eliminate the influence of the feature mode, the inner product result is normalized; represents the historical change amount, indicating the historical change trend and anomaly degree of the unknown point . It is the difference between the historical feature value of the unknown point minus the average value of the historical feature values of the neighborhood points, and can reflect whether there are large changes or precursors of faults in the cable system; is the historical change normalization parameter.

[0056] Then, use the fault probability prediction function to conduct fault risk prediction on the interpolation prediction result, obtain the fault probabilities of several unknown points , and use several unknown points The failure probability is used as the second abnormal recognition result. By generating the second abnormal recognition result, efficient abnormal prediction for the uncovered area can be achieved, potential abnormalities in the uncovered area can be detected in a timely manner, and the problem of monitoring blind spots in cable channel detection can be effectively eliminated.

[0057] S600: After fusing the first abnormal recognition result and the second abnormal recognition result, report cable abnormalities.

[0058] Furthermore, step S600 of this application further includes:

[0059] S610: Obtain the abnormal fusion result, and obtain the abnormal type, abnormal location, and abnormal value information based on the abnormal fusion result; S620: Obtain the current working parameters of the cable, and generate the first auxiliary verification information based on the working parameters; S630: Obtain the position environment information of the cable, and generate the second auxiliary verification information according to the position environment information; S640: Use the abnormal type, the abnormal location, the abnormal value information, the first auxiliary verification information, and the second auxiliary verification information to identify the abnormal trigger level; S650: Report cable abnormalities using the abnormal trigger level recognition result.

[0060] Specifically, first, fuse the first abnormal recognition result and the second abnormal recognition result, that is, integrate the abnormal recognition results of the monitored area and the uncovered area into the comprehensive abnormal recognition result of the cable channel area to obtain the abnormal fusion result; then, extract abnormal features according to the abnormal fusion result to obtain the abnormal type (such as overheating abnormality), abnormal location, and abnormal value (such as abnormal temperature) information.

[0061] Then obtain the current working parameters of the cable, including but not limited to load, voltage, etc. These parameters can reflect the current working state of the cable and help determine whether the cable is within the normal working range. For example, the load refers to the current or power load of the cable, and an excessive load may cause cable failure; the voltage refers to the voltage across the cable, and excessive voltage fluctuations may indicate cable failure; then, by analyzing the working parameters of the cable, generate the first auxiliary verification information, which is usually used to confirm whether the cable is in a normal working state. If the working parameters exceed the normal range, it indicates that the cable may have a fault or abnormality. For example, if the load is higher than the safe range that the cable can withstand, a load abnormality verification information can be generated. On the other hand, obtain the position environment information of the cable. The environmental information has an important impact on its working state, including but not limited to environmental temperature (environmental temperature affects the heat dissipation and load-bearing capacity of the cable), humidity (a high-humidity environment may accelerate the aging of the cable), external mechanical interference, etc.; then generate the second auxiliary verification information according to the position environment information. For example, if the cable is installed in an environment with high temperature, high humidity, or strong mechanical vibration, this may accelerate the aging of the cable or increase the probability of failure.

[0062] Next, evaluate according to the abnormal type, the abnormal position, the abnormal value information, the first auxiliary verification information, and the second auxiliary verification information, assign different weights according to the importance of each abnormal factor, and then comprehensively score each item of data by the weighted scoring method to determine the abnormal trigger level; finally, report the cable abnormality according to the abnormal trigger level. By evaluating the abnormal type, the abnormal position, the abnormal value information, the first auxiliary verification information, and the second auxiliary verification information, the abnormal trigger level of the cable system can be accurately determined, so that the cable monitoring system can provide accurate responses and early warnings in the face of various abnormalities.

[0063] Furthermore, step S600 of the present application further includes:

[0064] S660: Establish an accumulated weighted coefficient for the abnormality; S670: When the cable is abnormal at any position and has not been repaired, accumulate the cable abnormality according to the accumulated weighted coefficient to establish an accumulated result; S680: Manage the cable abnormality according to the accumulated result.

[0065] Specifically, first, establish an accumulated weighted coefficient for the abnormality. The weighted coefficient reflects the severity of the abnormality and is usually determined by factors such as the abnormal type, the abnormal position, and the abnormal value. Each time an abnormality occurs, a weight value is assigned according to its trigger level and abnormal type, which can be set according to the actual situation. The accumulated weighted coefficient is the accumulation of the weighted coefficients each time an abnormality is triggered. That is, when an abnormality occurs at a certain position of the cable, the weighted coefficient of the abnormality is recorded and added to the total weighted coefficient of that position. Next, if the cable is abnormal at any position and has not been repaired, that is, the abnormal state continues to be unprocessed, then accumulate the cable abnormality according to the accumulated weighted coefficient of that position, that is, record the abnormalities at all cable positions and accumulate the weighted coefficients. The accumulated weighted coefficient results of all positions form an accumulated result, and whether each position needs to be urgently processed is judged through this result. Finally, according to the working environment, importance, and working load of the cable, set different thresholds for the accumulated weighted coefficient. When the accumulated weighted coefficient reaches the threshold, it indicates that the cable has a persistent abnormal risk and needs to be repaired or replaced.

[0066] In summary, a cable channel detection method based on distributed optical fiber sensing technology provided by the present application has the following technical effects:

[0067] By obtaining the distribution coordinates of the distributed optical fiber sensor, the monitoring area and the uncovered area are configured according to the distribution coordinates and the monitoring accuracy; then the distributed optical fiber sensor is used to monitor the cable, and a monitoring data set is established. The monitoring data set includes temperature data, vibration data, and stress data, and the monitoring data set is marked with a monitoring position identifier; then the fault identification channel is used to perform collaborative anomaly identification on the monitoring data set to establish a first anomaly identification result; on the other hand, the monitoring data set is input into the distance-temperature interpolation prediction network to perform interpolation prediction on the uncovered area and establish an interpolation prediction result; further, the interpolation prediction result is used to perform fault risk prediction to establish a second anomaly identification result; finally, after anomaly fusion of the first anomaly identification result and the second anomaly identification result, a cable anomaly is reported. It can effectively eliminate the monitoring blind area problem in cable channel detection, timely discover potential anomalies in the uncovered area, so as to realize full-area and blind-spot-free anomaly monitoring and fault warning, and ensure the safe and stable operation of cable equipment.

[0068] Embodiment 2. Based on a cable channel detection method based on distributed optical fiber sensing technology in the foregoing embodiment, with the same inventive concept, the present application also provides a cable channel detection system based on distributed optical fiber sensing technology. Please refer to the attached Figure 3 , including:

[0069] A region configuration module 01, configured to obtain the distribution coordinates of the distributed optical fiber sensor, and configure the monitoring area and the uncovered area according to the distribution coordinates and the monitoring accuracy; a monitoring data establishment module 02, configured to use the distributed optical fiber sensor to monitor the cable and establish a monitoring data set. The monitoring data set includes temperature data, vibration data, and stress data, and the monitoring data set is marked with a monitoring position identifier; a first anomaly identification module 03, configured to perform collaborative anomaly identification on the monitoring data set by using a fault identification channel to establish a first anomaly identification result; an interpolation prediction module 04, configured to input the monitoring data set into a distance-temperature interpolation prediction network, perform interpolation prediction on the uncovered area, and establish an interpolation prediction result; a second anomaly identification module 05, configured to use the interpolation prediction result to perform fault risk prediction to establish a second anomaly identification result; an anomaly fusion module 06, configured to perform anomaly fusion on the first anomaly identification result and the second anomaly identification result, and then report a cable anomaly.

[0070] Further, the cable channel detection system based on distributed optical fiber sensing technology is also used for: calculating the dynamic weight of the interpolation point based on distance and temperature by using the dynamic weight optimization layer of the distance-temperature interpolation prediction network, as follows: ; where represents the unknown point predicted based on the known point ​The dynamic weight, characterizes the unknown point and the known point distance, characterizes the position of the point, characterizes the spatial distance smoothing parameter, is the distance exponential factor, characterizes the known point temperature, characterizes the average temperature of the known points within the neighborhood window, is the smoothing parameter for temperature difference, is the total number of known points, is the index of any known point; interpolation prediction based on the monitoring data set is performed according to the calculation result of the dynamic weight of the interpolation point: ; where is the interpolation prediction feature, is the feature vector of the known point, , is the stress eigenvalue, is the vibration eigenvalue.

[0071] Furthermore, the cable channel detection system based on distributed optical fiber sensing technology is also used for: performing fault risk prediction through a fault probability prediction model, as follows: ; where characterizes the fault probability of the unknown point, characterizes the interpolation prediction feature modulus, is the modulus normalization parameter, characterizes the inner product of the interpolation prediction feature and its neighborhood average feature , is the inner product normalization parameter, characterizes the historical change amount, is the historical change normalization parameter; a second anomaly recognition result is generated according to the fault probability of the unknown point.

[0072] Further, the cable channel detection system based on distributed optical fiber sensing technology is also used for: after classifying the monitoring data set by the preprocessing layer in the fault identification channel based on the monitoring position identifier and the acquisition time of the monitoring data set, performing outlier extraction to establish a preprocessing data set; using the feature extraction layer in the fault identification channel to extract the change rate, peak value, and window anomaly features of the preprocessing data set to establish a feature extraction data set; performing feature space correlation analysis based on the feature extraction data set through the spatial collaborative analysis layer to establish a first collaborative anomaly recognition result; performing feature time correlation analysis based on the feature extraction data set through the time collaborative analysis layer to establish a second collaborative anomaly recognition result; after fusing the first collaborative anomaly recognition result and the second collaborative anomaly recognition result, outputting a first anomaly recognition result.

[0073] Further, the cable channel detection system based on distributed optical fiber sensing technology is also used for: obtaining an anomaly fusion result, and obtaining anomaly type, anomaly location, and anomaly value information based on the anomaly fusion result; obtaining the current working parameters of the cable, and generating first auxiliary verification information based on the working parameters; obtaining the position environment information of the cable, and generating second auxiliary verification information according to the position environment information; using the anomaly type, the anomaly location, the anomaly value information, the first auxiliary verification information, and the second auxiliary verification information to perform anomaly trigger level recognition; using the anomaly trigger level recognition result to report cable anomalies.

[0074] Further, the cable channel detection system based on distributed optical fiber sensing technology is also used for: obtaining the model information and usage information of the distributed optical fiber sensor; performing accuracy loss recognition according to the model information and the usage information to establish an accuracy loss recognition result; reconstructing the monitoring accuracy based on the accuracy loss recognition result and the model information, and configuring the monitoring area and the uncovered area using the reconstructed monitoring accuracy.

[0075] Further, the cable channel detection system based on distributed optical fiber sensing technology is also used for: establishing an accumulated weighting coefficient for anomalies; when a cable is abnormal at any position and has not been repaired, then accumulating the cable anomalies according to the accumulated weighting coefficient to establish an accumulated result; performing cable anomaly management according to the accumulated result.

[0076] The various embodiments in this specification are described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The cable channel detection method and specific examples in the foregoing Embodiment 1 based on distributed optical fiber sensing technology are equally applicable to the cable channel detection system based on distributed optical fiber sensing technology in this embodiment. Through the detailed description of the cable channel detection method based on distributed optical fiber sensing technology above, those skilled in the art can clearly understand the cable channel detection system based on distributed optical fiber sensing technology in this embodiment. Therefore, for the sake of brevity of the specification, it will not be elaborated herein. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For the relevant parts, please refer to the description in the method section.

[0077] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0078] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is also intended to include these changes and modifications.

Claims

1. A cable channel detection method based on distributed optical fiber sensing technology, characterized in that: Methods include: Obtaining distribution coordinates of distributed optical fiber sensors, and configuring monitoring areas and uncovered areas according to the distribution coordinates and monitoring accuracy; Using the distributed optical fiber sensor to monitor the cable, establishing a monitoring data set, wherein the monitoring data set includes temperature data, vibration data, and stress data, and the monitoring data set is provided with a monitoring location identifier; Using the fault identification channel to perform collaborative anomaly identification of the monitoring data set to establish a first anomaly identification result includes: After classifying the monitoring data set based on the monitoring location identifier and the acquisition time of the monitoring data set through the preprocessing layer in the fault identification channel, performing outlier extraction and establishing a preprocessing data set; Using the feature extraction layer in the fault identification channel to extract the change rate, peak value, and window abnormality features of the preprocessed data set, and establish a feature extraction data set; Performing feature space correlation analysis based on the feature extraction data set through a spatial collaborative analysis layer to establish a first collaborative anomaly recognition result; Performing feature time correlation analysis based on the feature extraction data set through a time collaborative analysis layer to establish a second collaborative anomaly recognition result; After fusing the first collaborative anomaly recognition result and the second collaborative anomaly recognition result, outputting a first anomaly recognition result; Inputting the monitoring data set into the distance-temperature interpolation prediction network, performing interpolation prediction of the uncovered area, and establishing an interpolation prediction result, including: The dynamic weight optimization level of the distance-temperature interpolation prediction network is used to calculate the dynamic weight of the interpolation points based on distance and temperature, as follows: ； in, Characterization based on known points Predicted unknown points The dynamic weight of Characterizing unknown points With known points The distance Characterization The location of the point, Characterizes the spatial distance smoothing parameter, is the distance exponential factor, Characterizing known points The temperature, Characterizes the average temperature of known points within the neighborhood window, is the smoothing parameter of the temperature difference, is the total number of known points, is the index of any known point; According to the calculation results of the dynamic weights of the interpolation points, an interpolation prediction based on the monitoring data set is performed: ； in, is the interpolation prediction feature, For the The feature vectors of known points, , is the stress characteristic value, is the vibration characteristic value; Using the interpolation prediction result to perform fault risk prediction and establish a second abnormality identification result; After abnormal fusion of the first abnormal recognition result and the second abnormal recognition result, a cable abnormality is reported.

2. The cable channel detection method based on distributed optical fiber sensing technology as claimed in claim 1, characterized in that: The method of using the interpolation prediction result to perform fault risk prediction and establish a second abnormality identification result includes: Failure risk prediction is performed through the failure probability prediction model as follows: ； in, Characterizing unknown points The probability of failure, Characterize the interpolation prediction characteristic mode, is the modulus normalization parameter, Characterizing interpolation prediction features and their neighborhood average features The inner product of is the inner product normalization parameter, Represents historical changes, Normalize parameters for historical changes; According to the unknown point The failure probability generates a second abnormality identification result.

3. The cable channel detection method based on distributed optical fiber sensing technology according to claim 1, characterized in that: After performing abnormal fusion on the first abnormal recognition result and the second abnormal recognition result, the method includes: Acquire an abnormal fusion result, and acquire abnormal type, abnormal position and abnormal value information based on the abnormal fusion result; Acquire current working parameters of the cable, and generate first auxiliary verification information based on the working parameters; Acquire location environment information of the cable, and generate second auxiliary verification information according to the location environment information; Identify an abnormality trigger level using the abnormality type, the abnormality location, the abnormal value information, the first auxiliary verification information, and the second auxiliary verification information; The abnormal trigger level identification result is used to report the cable abnormality.

4. The cable channel detection method based on distributed optical fiber sensing technology according to claim 1, characterized in that: The configuring the monitoring area and the uncovered area according to the distribution coordinates and the monitoring accuracy includes: Obtaining model information and usage information of the distributed optical fiber sensor; Perform precision loss identification according to the model information and the usage information, and establish a precision loss identification result; The monitoring accuracy is reconstructed based on the accuracy loss identification result and the model information, and the monitoring area and the uncovered area are configured using the reconstructed monitoring accuracy.

5. The cable channel detection method based on distributed optical fiber sensing technology according to claim 1, characterized in that: The method also includes: Establishing cumulative weighting coefficients for anomalies; When a cable at any position is abnormal and has not been repaired, the cable abnormality is accumulated according to the cumulative weighted coefficient to establish a cumulative result; Cable abnormality management is performed according to the accumulated results.

6. A cable channel detection system based on distributed optical fiber sensing technology, characterized in that: The steps for implementing the cable channel detection method based on distributed optical fiber sensing technology as described in any one of claims 1 to 5 include: An area configuration module, used to obtain the distribution coordinates of the distributed optical fiber sensor, and configure the monitoring area and the uncovered area according to the distribution coordinates and the monitoring accuracy; A monitoring data establishment module, used to use the distributed optical fiber sensor to perform cable monitoring and establish a monitoring data set, wherein the monitoring data set includes temperature data, vibration data, and stress data, and the monitoring data set has a monitoring location identifier; A first anomaly identification module, configured to use a fault identification channel to perform collaborative anomaly identification of the monitoring data set and establish a first anomaly identification result; An interpolation prediction module, used for inputting the monitoring data set into a distance-temperature interpolation prediction network, performing interpolation prediction of the uncovered area, and establishing an interpolation prediction result; A second abnormality identification module, used to use the interpolation prediction result to perform fault risk prediction and establish a second abnormality identification result; The abnormality fusion module is used to report the cable abnormality after abnormality fusion of the first abnormality recognition result and the second abnormality recognition result.

Images