A multi-source data analysis method for underground cable exploration

Abstract

The present application relates to the technical field of underground cable detection, and relates to a multi-source data analysis method for underground cable detection.The method comprises the following steps: S1, collecting the buried position of the underground cable, then detecting the installation depth of the underground cable according to the buried position, and obtaining the installation environment of the cable; S2, collecting standard electromagnetic data of the cable, extracting fault features and normal features from the standard electromagnetic data, and monitoring real-time electromagnetic data of the underground cable.The present application can more comprehensively understand the operation of the cable in different positions by monitoring the real-time electromagnetic data of the underground cable, dividing monitoring nodes according to the installation depth and environmental differences of different positions, and analyzing in combination with historical electromagnetic data, thereby avoiding the limitation of a single monitoring point.

Description

Technical Field

[0001] This invention relates to the field of underground cable detection technology, and more specifically, to a multi-source data analysis method for underground cable detection. Background Technology

[0002] In the management and maintenance of underground cables, existing technologies mainly focus on accurately detecting cable locations, monitoring cable operating status, and promptly identifying faults to ensure stable power and communication transmission.

[0003] Currently, electromagnetic induction detection may be affected by interference from other electrical equipment, and ground penetrating radar may be affected by underground metal objects, thus affecting the accurate judgment of the cable burial location and installation depth. Secondly, existing technologies also have shortcomings in fault diagnosis. When abnormal situations occur, it is difficult to quickly and accurately determine the fault type and location, which increases the difficulty and time cost of maintenance. Therefore, a multi-source data analysis method for underground cable detection is proposed. Summary of the Invention

[0004] The purpose of this invention is to provide a multi-source data analysis method for underground cable detection, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, a multi-source data analysis method for underground cable detection is provided, comprising the following steps:

[0006] S1. Collect the location of underground cables, then detect the installation depth of underground cables based on the location of the underground cables, and at the same time obtain the installation environment of the cables.

[0007] S2. Collect standard electromagnetic data of the cable, extract fault features and normal features from the standard electromagnetic data, and monitor the real-time electromagnetic data of the underground cable.

[0008] S3. Based on the installation depth and environment, the underground cable is distributed into multiple monitoring nodes. Historical electromagnetic data is extracted according to the location of the monitoring nodes, and faults are screened out from the historical electromagnetic data, retaining only the historical electromagnetic data under normal operating conditions.

[0009] S4. Based on the historical electromagnetic data retained in S3, time segments are performed. Then, the historical electromagnetic data of different time periods are compared to obtain the difference rate between the historical electromagnetic data of different time periods and the standard electromagnetic data. The difference rate is then combined to make real-time difference rate prediction.

[0010] S5. Compare the real-time electromagnetic data of each monitoring node with fault features and normal features. Based on the comparison results, identify any real-time electromagnetic data that cannot be identified as abnormal features. Then, compare the abnormal features with the predicted real-time difference rate and normal features. Based on the comparison results, determine the fault of the abnormal feature.

[0011] S6. Establish a separate normal feature recognition library for each monitoring node. When S5 identifies that the abnormal feature belongs to the non-faulty category, the abnormal feature is input into the normal feature recognition library as a normal feature.

[0012] As a further improvement to this technical solution, S1 collects the construction drawings of the underground cable from the cable data terminal, extracts the burial location of the cable from the construction drawings, and then uses a data acquisition device to detect the installation depth of the cable according to the burial location. At the same time, the installation environment of the cable is collected during the depth detection process, thereby obtaining the installation depth and installation environment of the cable.

[0013] As a further improvement to this technical solution, step S2 is as follows:

[0014] S2.1 Obtain the model parameters of the underground cable, and at the same time perform cable operation simulation based on the model parameters, and use the electromagnetic data fed back by the cable in the simulation results as the standard electromagnetic data;

[0015] S2.2. Standard electromagnetic data is divided into normal operation data and fault operation data. Then, feature extraction is performed on the normal operation data and fault operation data respectively to obtain fault features and normal features.

[0016] S2.3 Monitor the real-time electromagnetic data of underground cables.

[0017] As a further improvement to this technical solution, step S3 is as follows:

[0018] S3.1 Based on the differences in installation depth and installation environment of underground cables at different locations, underground cables are divided into multiple monitoring nodes, and the same locations are allocated according to the geographical location of the monitoring nodes and historical electromagnetic data.

[0019] S3.2. Filter the historical electromagnetic data of each monitoring node into normal operation status and fault operation status, delete the historical electromagnetic data in the fault operation status, so that only the normal operation status remains in the historical electromagnetic data of each monitoring node.

[0020] As a further improvement to this technical solution, the calculation formula for S3 is as follows:

[0021]

[0022] in, h represents the average installation depth. i Here, n represents the individual installation depth values ​​obtained at different measurement points along the underground cable route, and h represents the total number of measurement points. shallow h is the limit value for shallow burial depth. medium This represents the limit value for the burial depth. This is a shallowly buried section. For the buried section, For deeply buried sections;

[0023]

[0024] Where W is the known soil moisture measurement value, W0 is the baseline soil moisture, and k w The soil moisture influence coefficient;

[0025]

[0026] Where E represents the electric field strength generated by the interference source at the location of the underground cable, d is the distance between the underground cable and the surrounding interference sources, Q is the charge carried by the interference source, and k is the electrostatic constant. Then, based on the calculated value of E, different threshold values ​​are set. th1 E th2 E, etc., to classify environments such as those near strong interference sources (E≥E) th1 ), near weak interference source environment (E th2 <E<E th1 Different interference environment categories, such as ) are used to divide monitoring nodes.

[0027] As a further improvement to this technical solution, step S4 is as follows:

[0028] S4.1. Divide the historical electromagnetic data retained in S3.2 into time segments. At the same time, perform maximum and minimum difference analysis between the historical electromagnetic data and the standard electromagnetic data in different time periods to obtain the maximum and minimum difference values ​​between the historical electromagnetic data and the standard electromagnetic data in different time periods. Then, form an interval between the maximum and minimum difference values ​​as the difference rate corresponding to different time periods.

[0029] S4.2 Sort the difference rates corresponding to different time periods according to time and perform real-time difference rate prediction, and obtain the predicted difference rate based on the prediction results.

[0030] As a further improvement to this technical solution, the calculation formula for S4 is as follows:

[0031]

[0032] Among them, H ijLet n be the value of the j-th data point in the historical electromagnetic data within the i-th time interval. i Let be the number of historical electromagnetic data within the i-th time interval. This represents the mean of historical electromagnetic data within the i-th time interval;

[0033]

[0034] Among them, S j Let n be the value of the j-th data point in the standard electromagnetic data. s This refers to the total number of data points included in the standard electromagnetic data. The average value of standard electromagnetic data;

[0035]

[0036] Among them, D i To represent the difference between historical electromagnetic data and standard electromagnetic data within the i-th time interval, and simultaneously find the maximum difference value D. max and minimum difference D min ;

[0037]

[0038] Among them, R i Let R be the difference rate corresponding to the i-th time interval, and then the difference rate interval is [R]. min R max ];

[0039]

[0040] Among them, R predicted To predict the variance rate for the next period, where n is the number of periods, and R is the variance rate over the past n periods. i-n+1 R i-n+2 ,…,R i .

[0041] As a further improvement to this technical solution, step S5 is as follows:

[0042] S5.1 Compare the real-time electromagnetic data of each monitoring node with the fault characteristics. When the real-time electromagnetic data contains fault characteristics, it is determined that the underground cable of the monitoring node has a fault. Conversely, when the real-time electromagnetic data does not contain fault characteristics, it is compared with normal characteristics. When all the characteristics of the real-time electromagnetic data meet the normal characteristics, it is determined that the underground cable is in normal working condition. Conversely, when there are electromagnetic characteristics in all the characteristics of the real-time electromagnetic data that cannot be identified by normal characteristics, the electromagnetic characteristics are extracted and regarded as abnormal characteristics.

[0043] S5.2. The abnormal feature is compared with the predicted real-time difference rate and the normal feature for similarity. When the difference rate between the abnormal feature and the normal feature is greater than the predicted real-time difference rate, the abnormal feature is determined to be a fault feature. Conversely, when the difference rate between the abnormal feature and the normal feature is less than the predicted real-time difference rate, the abnormal feature is determined to be a normal feature.

[0044] As a further improvement to this technical solution, S6 establishes a separate normal feature recognition library for each monitoring node, inputs the abnormal features identified as normal features in S5.2 into the normal feature recognition library, and when the same abnormal feature appears in subsequent recognition, the abnormal feature is determined to be a normal feature. Similarly, abnormal features identified as fault features are added to the fault feature recognition library.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] 1. In this multi-source data analysis method for underground cable detection, real-time electromagnetic data of underground cables is monitored, and monitoring nodes are divided according to the installation depth and environmental differences at different locations. Combined with historical electromagnetic data analysis, this multi-dimensional monitoring and analysis method can provide a more comprehensive understanding of the operation of cables at different locations and avoid the limitations of a single monitoring point.

[0047] 2. In this multi-source data analysis method for underground cable detection, historical electromagnetic data from different time periods are compared and analyzed with standard electromagnetic data to obtain the difference rate and make real-time difference rate predictions. In this way, the changing trend of cable operating status can be detected in advance, and early warning signals can be issued before the fault occurs, providing maintenance personnel with enough time to take preventive measures.

[0048] 3. In this multi-source data analysis method for underground cable detection, when abnormal features are detected, the system compares the predicted real-time difference rate with normal features to further determine the nature of the abnormal features. If the abnormal features are identified as fault features, the system can issue a fault alarm in a timely manner, prompting maintenance personnel to take swift action. At the same time, abnormal features identified as normal features are input into the normal feature recognition library, and abnormal features identified as fault features are added to the fault feature recognition library, continuously enriching and improving the system's feature library and enhancing the accuracy and efficiency of subsequent fault judgment. Attached Figure Description

[0049] Figure 1 This is an overall flowchart of the present invention;

[0050] Figure 2 This is a flowchart illustrating the real-time electromagnetic data monitoring of underground cables according to the present invention.

[0051] Figure 3The flowchart of this invention ensures that only the normal operating state is stored in the historical electromagnetic data of each monitoring node;

[0052] Figure 4 This is a flowchart illustrating the process of obtaining the prediction difference rate based on the prediction results according to the present invention.

[0053] Figure 5 This is a flowchart illustrating how the present invention compares the real-time electromagnetic data of each monitoring node with fault characteristics. Detailed Implementation

[0054] 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, and 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.

[0055] Please see Figures 1-5 As shown, the purpose of this embodiment is to provide a multi-source data analysis method for underground cable detection, including the following steps:

[0056] S1. Collect the location of underground cables, then detect the installation depth of underground cables based on the location of the underground cables, and at the same time obtain the installation environment of the cables.

[0057] S1 collects construction drawings of underground cables from the cable data terminal, extracts the cable burial location from the construction drawings, and then uses a data acquisition device to detect the cable installation depth based on the burial location. Simultaneously, during the depth detection process, the cable installation environment is collected, thereby obtaining the cable installation depth and the cable installation environment. The specific working steps are as follows:

[0058] Collect construction drawings and extract the location of buried cables: Determine the storage location of the construction drawings, which may be the archives of relevant departments, project management systems, or specific databases. Then, apply for access to the construction drawings and use measurement tools or software functions to determine the coordinates, direction, and other buried cable location information on the drawings.

[0059] Using a data acquisition device to detect cable installation depth: Based on the buried cable location information, select a suitable data acquisition device, such as an electromagnetic induction detector or ground penetrating radar. Place the data acquisition device on the ground near the buried cable location, then start the data acquisition device to perform the detection operation and record the data obtained during the detection process, including parameters such as signal strength and time delay. Finally, analyze the data according to the device's algorithm and data processing method to determine the cable installation depth.

[0060] Data collection on cable installation environment: While conducting in-depth exploration, use appropriate sensors or equipment to collect information on the cable installation environment. For example, use soil moisture sensors to measure soil moisture and use thermometers to measure soil temperature. Ensure that the sensors are placed in a reasonable location to accurately reflect the environmental conditions around the cable.

[0061] S2. Collect standard electromagnetic data of the cable, extract fault features and normal features from the standard electromagnetic data, and monitor the real-time electromagnetic data of the underground cable.

[0062] The steps in S2 are as follows:

[0063] S2.1 Obtain the model parameters of the underground cable, and simultaneously perform cable operation simulation based on the model parameters. Use the electromagnetic data fed back by the cable in the simulation results as the standard electromagnetic data. The specific steps are as follows:

[0064] To obtain underground cable model parameters: consult the project archives, design drawings, and other documents of the cable laying project. These documents usually clearly indicate the cable model parameters.

[0065] Cable operation simulation: Select a professional power system simulation software, such as ETAP or PSCAD, and then input the obtained cable model parameters into the simulation software, including conductor material, cross-sectional area, insulation material, rated voltage, etc. Define the simulated operating scenario, such as the magnitude of the applied current, voltage level, ambient temperature, etc. At the same time, start the software to perform simulation calculations and observe the cable's operating status under different conditions.

[0066] View simulation results: After the simulation software finishes running, view the output data, which will include electromagnetic data fed back by the cable, such as electromagnetic field strength, magnetic field direction, induced electromotive force, etc.

[0067] Extract standard electromagnetic data: Select electromagnetically relevant data from the simulation results and mark these electromagnetic data as standard electromagnetic data for subsequent comparison and analysis with actual measured data.

[0068] S2.2. Standard electromagnetic data is divided into normal operation data and fault operation data. Then, feature extraction is performed on the normal operation data and fault operation data respectively to obtain fault features and normal features. The specific work steps are as follows:

[0069] Determine classification criteria: Based on the actual operating status of the cable and known fault conditions, determine the criteria for distinguishing between normal operating data and fault operating data. For example, data collected when the cable shows no obvious abnormalities can be marked as normal operating data, while data collected when the cable experiences faults (such as short circuits, open circuits, grounding, etc.) can be marked as fault operating data.

[0070] Feature extraction from normal operation data: The normal operation data is cleaned to remove noise and outliers to ensure data quality. Statistical measures such as mean, variance, and standard deviation are calculated to reflect the central tendency and dispersion of the data. At the same time, the peak value, valley value, and peak factor of the data are analyzed to understand the amplitude variation characteristics of the data. Then, Fourier transform is performed on the data to convert the time domain data to the frequency domain and calculate the spectral characteristics of the frequency domain data, such as the dominant frequency, harmonic content, and bandwidth.

[0071] Feature extraction of fault operation data: Similar to the preprocessing of normal operation data, the fault operation data is cleaned and denoised, the changes in data before and after the fault occur are observed, and the features of the abrupt change points, such as the abrupt change amplitude and rate of change, are extracted. Then, the characteristics of the data during the fault duration are analyzed, such as the oscillation frequency and decay time. After that, the fault data is analyzed in the frequency domain to find the spectral features that are different from those of normal operation data and to determine the specific frequency components or frequency changes caused by the fault.

[0072] S2.3 Monitor the real-time electromagnetic data of underground cables.

[0073] S3. Based on the installation depth and environment, the underground cable is distributed into multiple monitoring nodes. Historical electromagnetic data is extracted according to the location of the monitoring nodes, and faults are screened out from the historical electromagnetic data, retaining only the historical electromagnetic data under normal operating conditions.

[0074] The steps in S3 are as follows:

[0075] S3.1. Based on the differences in installation depth and environment at different locations of underground cables, the underground cables are divided into multiple monitoring nodes. These nodes are then assigned to the same locations based on their geographical location and historical electromagnetic data, using the following formula:

[0076]

[0077] in, h represents the average installation depth. i Here, n represents the individual installation depth values ​​obtained at different measurement points along the underground cable route, and h represents the total number of measurement points. shallow This is the limit value for shallow burial depth, such as 1 meter, h medium This is the limit value for the burial depth, such as 2 meters;

[0078] This is a shallowly buried section. For the buried section, For deeply buried sections;

[0079]

[0080] Where W is the known soil moisture measurement value, W0 is the baseline soil moisture, and k w Let k be the soil moisture influence coefficient. w A value close to 1 indicates that the current soil moisture is close to the baseline moisture level. kw When the value is much greater than or less than 1, it indicates a large difference in humidity, which can be used to classify different environmental categories to assist in the classification of monitoring nodes;

[0081]

[0082] Where E represents the electric field strength generated by the interference source at the location of the underground cable, d is the distance between the underground cable and the surrounding interference sources, Q is the charge carried by the interference source, and k is the electrostatic constant. Then, based on the calculated value of E, different threshold values ​​are set. th1 E th2 E, etc., to classify environments such as those near strong interference sources (E≥E) th1 ), near weak interference source environment (E th2 <E<E th1 Different interference environment categories, such as ) are used to divide monitoring nodes.

[0083] S3.2. The historical electromagnetic data of each monitoring node is categorized into normal operation and fault operation states. Historical electromagnetic data in a fault operation state is deleted, ensuring that only data from the normal operation state remains in the historical electromagnetic data of each monitoring node. The specific steps are as follows:

[0084] Determine the criteria for judging normal and fault operation status: Analyze the characteristics of historical electromagnetic data, such as the range of variation of parameters such as magnetic field strength, electric field strength, and induced electromotive force, and combine them with the actual operation of the cable and experience to determine the threshold or characteristic indicators that distinguish between normal operation status and fault operation status.

[0085] Screening historical electromagnetic data: For the historical electromagnetic data of each monitoring node, check them one by one. According to the established judgment criteria, determine the operating status of each data point. If the parameter value of the data point is within the normal range, it is marked as normal operating status; if it is outside the normal range, it is marked as fault operating status. Then, the data marked as fault operating status is recorded separately for subsequent analysis.

[0086] Delete fault operation status data: For each monitoring node, iterate through all data points marked as fault operation status and delete these data points from the historical electromagnetic dataset of that monitoring node to ensure that the dataset after deletion only contains data of normal operation status.

[0087] S4. Based on the historical electromagnetic data retained in S3, time segments are performed. Then, the historical electromagnetic data of different time periods are compared to obtain the difference rate between the historical electromagnetic data of different time periods and the standard electromagnetic data. The difference rate is then combined to make real-time difference rate prediction.

[0088] The steps in S4 are as follows:

[0089] S4.1. Divide the historical electromagnetic data retained in S3.2 into time segments. Simultaneously, perform maximum and minimum difference analysis between the historical electromagnetic data and the standard electromagnetic data for different time periods to obtain the maximum and minimum difference values ​​between the historical electromagnetic data and the standard electromagnetic data for different time periods. Then, form an interval based on the maximum and minimum difference values ​​as the difference rate corresponding to different time periods. The specific calculation formula is as follows:

[0090]

[0091] Among them, H ij Let n be the value of the j-th data point in the historical electromagnetic data within the i-th time interval. i Let be the number of historical electromagnetic data within the i-th time interval. This represents the mean of historical electromagnetic data within the i-th time interval;

[0092]

[0093] Among them, S j Let n be the value of the j-th data point in the standard electromagnetic data. s This refers to the total number of data points included in the standard electromagnetic data. The average value of standard electromagnetic data;

[0094]

[0095] Among them, D i To represent the difference between historical electromagnetic data and standard electromagnetic data within the i-th time interval, and simultaneously find the maximum difference value D. max and minimum difference D min ;

[0096]

[0097] Among them, R i Let R be the difference rate corresponding to the i-th time interval, and then the difference rate interval is [R]. min R max ].

[0098] S4.2 Sort the difference rates corresponding to different time periods by time and predict the difference rates in real time. Obtain the predicted difference rates based on the prediction results, organize the difference rate data corresponding to different time periods, and ensure that each difference rate is associated with the corresponding time interval. Sort the difference rate data in chronological order, which can be determined by timestamps or other time identifiers. The specific calculation formula is as follows:

[0099]

[0100] Among them, R predicted To predict the variance rate for the next period, where n is the number of periods, and R is the variance rate over the past n periods. i-n+1 R i-n+2 ,…,R i。

[0101] S5. Compare the real-time electromagnetic data of each monitoring node with fault features and normal features. Based on the comparison results, identify any real-time electromagnetic data that cannot be identified as abnormal features. Then, compare the abnormal features with the predicted real-time difference rate and normal features. Based on the comparison results, determine the fault of the abnormal feature.

[0102] The steps in S5 are as follows:

[0103] S5.1. Compare the real-time electromagnetic data of each monitoring node with fault characteristics. If the real-time electromagnetic data contains fault characteristics, it is determined that the underground cable of that monitoring node is faulty. Conversely, if the real-time electromagnetic data does not contain fault characteristics, it is compared with normal characteristics. If all characteristics of the real-time electromagnetic data meet the normal characteristics, the underground cable is determined to be in normal working condition. Conversely, if there are electromagnetic characteristics in the real-time electromagnetic data that cannot be identified by normal characteristics, these electromagnetic characteristics are extracted and treated as abnormal characteristics. The specific steps are as follows:

[0104] Data preparation and feature extraction: First, real-time electromagnetic data of each monitoring node is acquired, and corresponding feature values ​​are extracted from the real-time electromagnetic data to form a feature set of real-time electromagnetic data.

[0105] Fault Feature Comparison: Take out the pre-determined fault feature set. The fault feature set is a set of various features that can characterize cable faults, which are obtained by analyzing and summarizing the fault operation data. Compare the feature set of real-time electromagnetic data with each feature in the fault feature set in turn. For example, check whether there is a situation in the real-time electromagnetic data that matches a certain fault feature in the fault feature set. For example, if the real-time magnetic field strength change rate exceeds the magnetic field strength change rate threshold set in the fault features, it means that the corresponding fault feature has appeared.

[0106] Determining the existence of a fault: If, during the comparison process, a feature matching the fault feature set is found in the feature set of the real-time electromagnetic data, then according to the determination rules, it can be determined that the underground cable of the monitoring node has a fault.

[0107] Normal feature preparation: Obtain the normal feature set obtained by extracting and summarizing the normal operation data. The normal feature set contains the typical characteristics of various electromagnetic data of the cable under normal operation, such as normal average range and stable main frequency. Similarly, it is necessary to ensure that the real-time electromagnetic data has completed feature extraction and formed the corresponding feature set.

[0108] Normal feature comparison: Compare the feature set of real-time electromagnetic data with each feature in the normal feature set one by one. For example, check whether the mean of real-time electromagnetic data falls within the mean range specified by the normal features, and whether the harmonic content in the frequency domain meets the corresponding range requirements in the normal features.

[0109] Determining normal operation: If all features of the real-time electromagnetic data can be found to match and meet the requirements in the normal feature set after comparison, that is, all features fall within the reasonable range defined by the normal features, then it can be determined that the underground cable corresponding to the monitoring node is in normal working condition.

[0110] Anomaly detection: During the comparison of real-time electromagnetic data with normal features, if it is found that some features of the real-time electromagnetic data cannot be covered by any feature in the normal feature set, that is, they exceed the reasonable range defined by the normal features, it means that an anomaly has occurred.

[0111] Anomaly feature extraction: Once such anomalies are detected, the corresponding electromagnetic features that do not conform to normal characteristics are extracted from the real-time electromagnetic data feature set and recorded separately as anomaly features for subsequent analysis.

[0112] S5.2. The abnormal feature is compared with the predicted real-time difference rate and the normal feature for similarity. When the difference rate between the abnormal feature and the normal feature is greater than the predicted real-time difference rate, the abnormal feature is determined to be a fault feature. Conversely, when the difference rate between the abnormal feature and the normal feature is less than the predicted real-time difference rate, the abnormal feature is determined to be a normal feature.

[0113] S6. Establish a separate normal feature recognition library for each monitoring node. When S5 identifies that the abnormal feature belongs to the non-faulty category, the abnormal feature is input into the normal feature recognition library as a normal feature.

[0114] S6 establishes a separate normal feature recognition library for each monitoring node, inputting abnormal features identified as normal features in S5.2 into the normal feature recognition library. In subsequent recognition, when the same abnormal feature appears, it is determined that the abnormal feature is a normal feature. Similarly, abnormal features identified as fault features are added to the fault feature recognition library. The specific working steps are as follows:

[0115] Establish a normal feature identification database and a fault feature identification database: Create two independent databases or data storage structures for each monitoring node. One database is used to store normal features, called the normal feature identification database; the other database is used to store fault features, called the fault feature identification database.

[0116] Handling abnormal features identified as normal features: For abnormal features identified as normal features in step S5.2, record their detailed information and add these features to the normal feature identification library of the corresponding monitoring node. In the subsequent identification process, when the same abnormal feature appears again, the normal feature identification library will be queried. If the feature exists in the library, the abnormal feature is determined to be a normal feature. The steps for handling abnormal features identified as fault features are the same.

[0117] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-source data analysis method for underground cable detection, characterized in that: Includes the following steps: S1. Collect the location of underground cables, then detect the installation depth of underground cables based on the location of the underground cables, and at the same time obtain the installation environment of the cables. S2. Collect standard electromagnetic data of the cable, extract fault features and normal features from the standard electromagnetic data, and monitor the real-time electromagnetic data of the underground cable. S3. Based on the installation depth and environment, the underground cable is distributed into multiple monitoring nodes. Historical electromagnetic data is extracted according to the location of the monitoring nodes, and faults are screened out from the historical electromagnetic data, retaining only the historical electromagnetic data under normal operating conditions. S4. Based on the historical electromagnetic data retained in S3, time segments are performed. Then, the historical electromagnetic data of different time periods are compared to obtain the difference rate between the historical electromagnetic data of different time periods and the standard electromagnetic data. The difference rate is then combined to make real-time difference rate prediction. S5. Compare the real-time electromagnetic data of each monitoring node with fault features and normal features. Based on the comparison results, identify any real-time electromagnetic data that cannot be identified as abnormal features. Then, compare the abnormal features with the predicted real-time difference rate and normal features. Based on the comparison results, determine the fault of the abnormal feature. S6. Establish a separate normal feature recognition library for each monitoring node. When S5 identifies that the abnormal feature belongs to the non-faulty category, the abnormal feature is input into the normal feature recognition library as a normal feature. In step S5, the fault determination of the abnormal feature based on the comparison result specifically includes: The abnormal feature is compared with the predicted real-time difference rate and the normal feature. When the difference rate between the abnormal feature and the normal feature is greater than the predicted real-time difference rate, the abnormal feature is determined to be a fault feature. Conversely, when the difference rate between the abnormal feature and the normal feature is less than the predicted real-time difference rate, the abnormal feature is determined to be a normal feature.

2. The multi-source data analysis method for underground cable detection according to claim 1, characterized in that: S1 collects the construction drawings of the underground cable from the cable data terminal, extracts the burial location of the cable from the construction drawings, and then uses a data acquisition device to detect the installation depth of the cable according to the burial location. At the same time, the installation environment of the cable is collected during the depth detection process, thereby obtaining the installation depth and installation environment of the cable.

3. The multi-source data analysis method for underground cable detection according to claim 1, characterized in that: The steps in S2 are as follows: S2.1 Obtain the model parameters of the underground cable, and at the same time perform cable operation simulation based on the model parameters, and use the electromagnetic data fed back by the cable in the simulation results as the standard electromagnetic data; S2.

2. Standard electromagnetic data is divided into normal operation data and fault operation data. Then, feature extraction is performed on the normal operation data and fault operation data respectively to obtain fault features and normal features. S2.3 Monitor the real-time electromagnetic data of underground cables.

4. The multi-source data analysis method for underground cable detection according to claim 1, characterized in that: The steps in S3 are as follows: S3.1 Based on the differences in installation depth and installation environment of underground cables at different locations, underground cables are divided into multiple monitoring nodes, and the same locations are allocated according to the geographical location of the monitoring nodes and historical electromagnetic data. S3.

2. Filter the historical electromagnetic data of each monitoring node into normal operation status and fault operation status, delete the historical electromagnetic data in the fault operation status, so that only the normal operation status remains in the historical electromagnetic data of each monitoring node.

5. The multi-source data analysis method for underground cable detection according to claim 1, characterized in that: The formula for calculating S3 is as follows: in, h represents the average installation depth. i Here, n represents the individual installation depth values ​​obtained at different measurement points along the underground cable route, and h represents the total number of measurement points. shallow h is the limit value for shallow burial depth. medium This represents the limit value for the burial depth. ≤h shallow For the shallow buried section, h medium < ≤h medium For the buried section, >h medium For deeply buried sections; Where W is the known measured soil moisture value, W0 is the baseline soil moisture, and k w Let k be the soil moisture influence coefficient. w When k approaches 1, it indicates that the current soil moisture is close to the reference moisture level; when k... w When the value is much greater than or much less than 1, different environmental categories are defined to assist in the classification of monitoring nodes; Where E represents the electric field strength generated by the interference source at the location of the underground cable, d is the distance between the underground cable and the surrounding interference sources, Q is the charge carried by the interference source, and k is the electrostatic constant. Then, based on the calculated value of E, different threshold values ​​are set. th1 E th2 , E≥E th1 The interference environment is classified as an environment close to a strong interference source, and E is... th2 <E<E th1 The interference environment is classified into environments close to weak interference sources, and different interference environment categories are used to classify monitoring nodes.

6. The multi-source data analysis method for underground cable detection according to claim 1, characterized in that: The steps in S4 are as follows: S4.

1. Divide the historical electromagnetic data retained in S3.2 into time segments. At the same time, perform maximum and minimum difference analysis between the historical electromagnetic data and the standard electromagnetic data in different time periods to obtain the maximum and minimum difference values ​​between the historical electromagnetic data and the standard electromagnetic data in different time periods. Then, form an interval between the maximum and minimum difference values ​​as the difference rate corresponding to different time periods. S4.2 Sort the difference rates corresponding to different time periods according to time and perform real-time difference rate prediction, and obtain the predicted difference rate based on the prediction results.

7. The multi-source data analysis method for underground cable detection according to claim 1, characterized in that: The formula for calculating S4 is as follows: Among them, H ij Let n be the value of the j-th data point in the historical electromagnetic data within the i-th time interval. i Let be the number of historical electromagnetic data within the i-th time interval. This represents the mean of historical electromagnetic data within the i-th time interval; Among them, S j Let n be the value of the j-th data point in the standard electromagnetic data. s This refers to the total number of data points included in the standard electromagnetic data. The average value of standard electromagnetic data; Among them, D i To represent the difference between historical electromagnetic data and standard electromagnetic data within the i-th time interval, and simultaneously find the maximum difference value D. max and minimum difference D min ; Among them, R i Let R be the difference rate corresponding to the i-th time interval, and then the difference rate interval is [R]. min R max ]; Among them, R predicted To predict the variance rate for the next period, where n is the number of periods, and R is the variance rate over the past n periods. i-n+1 R i-n+2 ,…,R i .

8. The multi-source data analysis method for underground cable detection according to claim 1, characterized in that: The steps in S5 are as follows: S5.1 Compare the real-time electromagnetic data of each monitoring node with the fault characteristics. When the real-time electromagnetic data contains fault characteristics, it is determined that the underground cable of the monitoring node is faulty. Conversely, when the real-time electromagnetic data does not contain fault characteristics, it is compared with normal characteristics. When all the characteristics of the real-time electromagnetic data meet the normal characteristics, it is determined that the underground cable is in normal working condition. Conversely, when there are electromagnetic characteristics in all the characteristics of the real-time electromagnetic data that cannot be identified by normal characteristics, the electromagnetic characteristics are extracted and regarded as abnormal characteristics.

9. The multi-source data analysis method for underground cable detection according to claim 1, characterized in that: S6 establishes a separate normal feature recognition library for each monitoring node, inputs the abnormal features identified as normal features in S5.2 into the normal feature recognition library, and when the same abnormal feature appears in subsequent recognition, the abnormal feature is determined to be a normal feature. Similarly, abnormal features identified as fault features are added to the fault feature recognition library.

Images