An underground cable pipeline anomaly monitoring method and system

Abstract

The present application discloses an underground cable pipeline anomaly monitoring method and system. The method includes the following steps: S1: Construct a three-dimensional grid based on the regional soil distribution, and obtain a set of acquisition point data based on the three-dimensional grid and the underground cable pipeline data; S2: Obtain the acquisition distribution corresponding to each type of sensor to be deployed based on the type of sensor to be deployed and the set of acquisition point data; S3: Obtain the time-series data of the underground cable pipeline collected by each sensor based on the acquisition distribution, perform clustering analysis on the time-series data of the underground cable pipeline, and obtain anomaly data points; S4: Obtain the anomaly location according to the anomaly data points and the acquisition distribution, and output the underground cable pipeline anomaly information according to the anomaly location. The beneficial effects of the present application: While considering the physical location of the cable pipeline, it also takes into account the soil environmental characteristics, improves the accuracy of sensor deployment in the pipeline, improves the analysis efficiency of anomaly data points, and facilitates the timely investigation of underground cable pipeline anomalies.

Description

Technical Field

[0001] This application relates to the technical field of underground cable anomaly monitoring, and particularly to a method and system for monitoring underground cable duct anomalies. Background Art

[0002] With the acceleration of urbanization and the continuous growth of power demand, underground cable ducts, as an important infrastructure for urban power transmission, their safety and stability are directly related to the normal operation of the city and the quality of life of residents. However, due to the complex and changeable underground environment, cable ducts are prone to be affected by various factors during operation, such as soil corrosion, external force damage, insulation aging, etc., resulting in changes in the duct state and even causing safety accidents. Therefore, it is particularly important to monitor underground cable ducts in real time and detect anomalies.

[0003] Traditional cable duct monitoring methods often rely on manual inspections and regular detections. This method is not only inefficient but also prone to missed detections and misjudgments. In addition, manual inspections are also restricted by environmental and human factors, making it difficult to achieve full coverage and real-time monitoring of cable ducts. The environment where underground cable ducts are located is complex and changeable, affected by various factors such as soil humidity, temperature, and pressure, which will all affect the state of cable ducts. At the same time, there are numerous underground pipelines that are intertwined with each other, increasing the difficulty and complexity of monitoring. And when an anomaly occurs in the cable duct, it is necessary to quickly and accurately locate the anomaly position and transmit the information to relevant personnel for processing in real time.

[0004] Therefore, how to comprehensively and accurately monitor the state of cable ducts and the surrounding environment, timely discover and locate potential safety hazards, and improve the operation efficiency and safety of the power system are problems that need to be urgently solved by those skilled in the art.

[0005] In related technologies, sensors are used to monitor underground cable ducts. When the data collected by the sensors shows anomalies, it is assumed that the corresponding position of the underground cable duct is abnormal, or a large amount of data obtained by the sensors is processed and mined to discover potential fault signs and abnormal patterns from the data. In this case, although manual inspections are not required and the efficiency of anomaly detection is improved, once the environment changes and has an impact, due to the sensor settings being affected by duct characteristics, it is impossible to effectively distinguish or even give an alarm for the impact of environmental anomalies.

[0006] Patent "Cable Fault Detection Method Based on Artificial Neural Network", Publication Number: CN115795360A, Publication Date: March 14, 2023, specifically discloses that the method includes the following steps: S1. Collect sample data when the cable fails to obtain real-time fault samples; S2. Normalize the real-time fault samples to obtain normalized samples; S3. Improve the artificial neural network using the Dropout algorithm to construct a cable fault detection model; S4. Use the gradient descent method to train the cable fault detection model to obtain a trained cable fault detection model; S5. Input the normalized samples into the trained cable fault detection model to output the fault detection result. This solution uses an artificial neural network for cable fault training. Once the environmental state fluctuates, accurate fault detection cannot be achieved. At the same time, the output of abnormal coordinates depends on the data collection of sensors. Once the sensor data collection is incorrect, the abnormal coordinates cannot be accurately output.

[0007] Patent "Intelligent Grid Operation and Maintenance System Security Online Real-Time Monitoring Method and Storage Medium", Publication Number: CN113253051A, Publication Date: May 12, 2021, specifically discloses that the method includes: obtaining the road length corresponding to the underground cable project in this area and dividing it into sections; obtaining the burial duration of the underground cables corresponding to each cable section; detecting the internal environmental parameters of the underground cable ducts for each cable section; detecting the basic parameters of the outer sheaths of the underground cables for each cable section; obtaining the operation parameters of the underground cables for each cable section during each collection period; respectively analyzing the burial duration, internal environmental parameters of the underground cable ducts, basic parameters of the outer sheaths of the underground cables, and operation parameters of the underground cables corresponding to each cable section. This solution also has high requirements for the accuracy of sensors. Once a sensor has a problem, it is very likely to affect the final abnormal identification. Summary of the Invention

[0008] In view of the technical problem in the prior art that when a certain sensor is abnormal, it is easy to mislead the abnormal analysis, this application provides an underground cable duct abnormal monitoring method and system. By constructing a three-dimensional grid, the pre-arrangement of sensors is associated with the environment, and in the later stage, the abnormal points are checked through the clustering analysis of each sensor. Through the distribution of sensors on the three-dimensional grid, the abnormal position is obtained, so as to avoid the loss of abnormal coordinates or the misleading of the abnormal coordinate check due to the abnormality of a certain sensor, and improve the accuracy and efficiency of abnormal analysis at the same time.

[0009] To achieve the above technical objectives, a technical solution provided by this application is an underground cable pipeline anomaly monitoring method, which includes the following steps: S1: Construct a three-dimensional grid based on the regional soil distribution, and obtain a set of collection point data based on the three-dimensional grid and the underground cable pipeline data; S2: Obtain the collection distribution corresponding to each type of sensor to be deployed based on the type of sensor to be deployed and the set of collection point data; S3: Obtain the time-series data of the underground cable pipeline collected by each sensor based on the collection distribution, perform clustering analysis on the time-series data of the underground cable pipeline, and obtain abnormal data points; S4: Obtain the abnormal location based on the abnormal data points and the collection distribution, and output the underground cable pipeline anomaly information according to the abnormal location.

[0010] Further, the construction of the three-dimensional grid based on the regional soil distribution includes: constructing horizontal grids at a preset interval along the horizontal direction; constructing vertical grids according to the soil structure information; and constructing a three-dimensional grid with the horizontal grids and the vertical grids.

[0011] Further, the construction of the horizontal grids at a preset interval along the horizontal direction includes: obtaining an interval range according to expert experience, obtaining an average interval based on the layout of the underground cable pipeline in the horizontal direction, the minimum cost, and the interval range, and constructing the horizontal grids with the average interval as the preset interval.

[0012] Further, the obtaining of the set of collection point data based on the three-dimensional grid and the underground cable pipeline data includes: obtaining a set of grid intersection points based on the three-dimensional grid, obtaining the layout point angle based on the pipeline shape, and obtaining the set of collection point data with the set of grid intersection points and the layout point angle.

[0013] Further, the obtaining of the collection distribution corresponding to each type of sensor to be deployed based on the type of sensor to be deployed and the set of collection point data includes: obtaining the type of sensor to be deployed based on the data requirements to be collected, obtaining the collection parameter characteristics corresponding to the type of sensor to be deployed, and obtaining the collection distribution that meets the collection parameter characteristic requirements from the set of collection point data corresponding to each type of sensor to be deployed.

[0014] Further, the performing of clustering analysis on the time-series data of the underground cable pipeline to obtain abnormal data points includes: dividing the time-series data of the underground cable pipeline into several subsequences containing continuous observation values according to a preset parameter content; performing clustering analysis on each subsequence to obtain the clustering result of each subsequence; and obtaining the abnormal data points based on the distance algorithm and the clustering result.

[0015] Further, the obtaining of the abnormal position based on the abnormal data points and the acquisition distribution includes: obtaining a corresponding sensor set according to each abnormal data point; obtaining a sensor position point set from the sensor set; obtaining an abnormal area from the sensor position point set and the acquisition distribution, and acquiring the sensor monitoring data corresponding to the abnormal area; performing weighted averaging on the sensor monitoring data based on the relative position relationship of the sensors to obtain the abnormal position.

[0016] Further, the outputting of the underground cable pipeline abnormal information according to the abnormal position further includes: obtaining historical underground cable pipeline data, and constructing an abnormal recognition model according to the historical underground cable pipeline data; performing key feature recognition according to the sensors corresponding to the abnormal position and the underground cable pipeline time series data corresponding to the sensors, and obtaining the underground cable pipeline abnormal information based on the key features and the abnormal recognition model.

[0017] Further, S3 further includes: S31: obtaining the current offset physical quantity based on the historical underground cable pipeline offset data and the current regional soil distribution, and constructing an offset grid with the current offset physical quantity and a three-dimensional grid; S32: obtaining the underground cable pipeline time series data collected by each sensor based on the acquisition distribution, and performing clustering analysis on the underground cable pipeline time series data based on the offset grid to obtain abnormal data points.

[0018] Further, S31 further includes: obtaining the correlation between the soil distribution and the offset physical quantity based on the historical underground cable pipeline offset data; obtaining the current offset physical quantity according to the regional soil distribution and the correlation between the soil distribution and the offset physical quantity; constructing an offset grid with the current offset physical quantity and a three-dimensional grid.

[0019] Further, S32 further includes: obtaining the underground cable pipeline time series data collected by each sensor based on the acquisition distribution; performing a first clustering on each sensor based on the offset grid and the acquisition distribution; performing a second clustering on the underground cable pipeline time series data collected by the sensors in the same cluster in the first clustering; obtaining abnormal data points based on the second clustering.

[0020] Further, S32 further includes: obtaining the sensors to be checked according to a preset threshold of the first clustering; judging whether the sensors to be checked are abnormal based on the historical underground cable pipeline monitoring data and the current underground cable pipeline time series data of the sensors to be checked, and if so, outputting sensor abnormal information.

[0021] Further, it further includes: S5: obtaining the offset grid data within a preset time series segment, obtaining the offset fluctuation trend of each grid based on the offset grid data, and judging whether there is an offset abnormality in the underground cable pipeline according to the offset fluctuation trend of each grid and the grid offset threshold, and if so, outputting the underground cable pipeline offset abnormal information.

[0022] Another technical solution provided by this application is an underground cable pipeline anomaly monitoring system for implementing the method as described above, including: a sensor layout module that constructs a three-dimensional grid based on the regional soil distribution and obtains the acquisition distribution corresponding to each type of sensor to be laid; a data acquisition module for obtaining the time-series data of the underground cable pipeline collected by each sensor; and a data analysis module that performs clustering analysis based on the time-series data of the underground cable pipeline to obtain anomaly data points and analyzes and outputs the anomaly information of the underground cable pipeline.

[0023] Advantages of this application:

[0024] 1. Construct a three-dimensional grid according to the regional soil distribution, combine pipeline information to obtain the sensor acquisition point data set, take into account the physical location of the cable pipeline while considering the soil environment characteristics, improve the accuracy of sensor layout in the pipeline, and then improve the accuracy of time-series data acquisition of the underground cable pipeline, improve the accuracy of anomaly data point analysis. At the same time, there is no need to conduct complex misinformation data screening, improve the efficiency of anomaly data point analysis, and facilitate timely investigation of underground cable pipeline anomalies.

[0025] 2. Obtain the preset interval through the average interval, minimum cost, and interval range together, which can not only ensure that the horizontal grid meets the sensor layout requirements, but also ensure that there must be an intersection of the horizontal grid layer and the vertical grid layer at the end of the underground cable pipeline layout area for sensor layout, avoiding errors in anomaly analysis caused by the omission of end data and further improving the accuracy of anomaly analysis.

[0026] 3. The data requirements to be collected can be obtained according to the pipeline type. Different pipeline types may have different anomalies. At this time, different data are required for anomaly analysis, and the corresponding types of sensors to be laid are obtained to avoid interference from unnecessary data.

[0027] 4. Divide according to the preset parameter content into several subsequences containing continuous observed values, and perform clustering analysis on each subsequence to obtain data points deviating from the clustering center. It not only maintains unity in time series, avoids deviation interference caused by time series differences, but also obtains short-term anomaly data points based on the clustering of subsequences of continuous observed values. At the same time, anomaly analysis is carried out in both the long term and the short term, and the mutual interference between long-term and short-term data is avoided, improving the accuracy of data anomaly analysis.

[0028] 5. When performing clustering analysis, add the deviation between the default detection point of the sensor and the actual sensor detection to compensate for the sensor detection position error. Without the need for manual correction of the sensor, the anomaly data points can still be obtained through the offset difference between the three-dimensional grid and the offset grid, ensuring the accuracy of anomaly analysis while reducing the need for sensor adjustment.

[0029] 6. By performing clustering twice, it is possible to comprehensively consider the spatial distribution of sensors and the characteristics of time-series data, thereby more accurately identifying abnormal data points and avoiding the influence of sensors of the same type at different spatial positions on the analysis of abnormal data. Description of the Drawings

[0030] Figure 1 This is a schematic flowchart of a method for monitoring anomalies in underground cable ducts according to the present application.

[0031] Figure 2 This is a schematic diagram of cluster analysis of a method for monitoring anomalies in underground cable ducts according to the present application.

[0032] Figure 3 This is a schematic diagram of the analysis of abnormal states of a method for monitoring anomalies in underground cable ducts according to the present application. Detailed Embodiments

[0033] To make the objectives, technical solutions, and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only the best embodiments of the present application, only for explaining the present application, and do not limit the protection scope of the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present application.

[0034] As Figure 1 shown, as the first embodiment of the present application, a method for monitoring anomalies in underground cable ducts includes the following steps:

[0035] S1: Construct a three-dimensional grid based on the regional soil distribution, and obtain a set of acquisition point data based on the three-dimensional grid and the underground cable duct data;

[0036] S2: Obtain the acquisition distribution corresponding to each type of sensor to be deployed based on the type of sensor to be deployed and the set of acquisition point data;

[0037] S3: Obtain the time-series data of the underground cable ducts collected by each sensor based on the acquisition distribution, perform cluster analysis on the time-series data of the underground cable ducts, and obtain abnormal data points;

[0038] S4: Obtain the abnormal location based on the abnormal data points and the acquisition distribution, and output the abnormal information of the underground cable ducts according to the abnormal location.

[0039] In this embodiment, a three-dimensional grid is constructed according to the regional soil distribution, and the data set of sensor acquisition points is obtained by combining pipeline information. While considering the physical location of the cable pipeline, the characteristics of the soil environment are also taken into account, so as to improve the accuracy of sensor layout in the pipeline, and further improve the accuracy of time-series data acquisition of underground cable pipelines, improve the accuracy of abnormal data point analysis, and at the same time, there is no need to conduct complex misinformation data investigation, improve the efficiency of abnormal data point analysis, and facilitate the timely investigation of underground cable pipeline anomalies.

[0040] Among them, step S1 further includes:

[0041] Based on the layout of the underground cable pipeline, the area to be analyzed is obtained, and the regional soil distribution is obtained from the area to be analyzed.

[0042] The layout of the underground cable pipeline at least includes the pipeline direction, pipeline length, and pipeline burial depth. The area with underground cable pipelines is obtained based on the layout of the underground cable pipeline as the area to be analyzed, and the soil distribution in the area to be analyzed is obtained, which is used as the regional soil distribution. The construction of unnecessary three-dimensional grids is reduced. While improving the accuracy of sensor layout, the layout of unnecessary sensors is avoided, the monitoring cost is reduced, the acquisition of unnecessary time-series data of underground cable pipelines is reduced, the amount of data for abnormal analysis is reduced, and the efficiency of abnormal data point analysis is further improved.

[0043] The soil distribution at least includes soil structure information. Constructing a three-dimensional grid based on the regional soil distribution includes:

[0044] Grid lines are constructed based on the change of soil structure information, and a three-dimensional grid is constructed with the grid lines.

[0045] Grid lines are constructed at the positions where the soil structure information changes, so that different soil structures are in different grid areas, which is convenient for subsequent investigation of the influence of soil structure.

[0046] In some cases, the soil structure information includes soil types. According to different soil types, grid layers with different spacings are constructed. In areas where the soil type changes greatly, such as at the junction of soil and rock, the density of the grid layers is increased.

[0047] Generally, the area where the underground cable pipeline is laid is usually in the form of layered filling, that is, the soil types in the horizontal direction are basically the same, and there are only differences in soil types in the vertical direction. At this time, constructing a three-dimensional grid based on the regional soil distribution includes:

[0048] Horizontal grids are constructed at preset intervals along the horizontal direction;

[0049] Vertical grids are constructed according to the soil structure information;

[0050] A three-dimensional grid is constructed with the horizontal grids and the vertical grids.

[0051] In the horizontal direction, along the alignment and length of the cable duct, the area around the duct is divided into a number of horizontal grids at a preset interval; in the vertical direction, according to the duct burial depth and soil structure, vertical grid layers of different heights are constructed. In some cases, the preset interval is the optimal horizontal interval for sensor layout obtained based on expert experience. In other cases, constructing the horizontal grids at the preset interval in the horizontal direction includes:

[0052] Obtain the interval range according to expert experience, calculate the average interval based on the layout of the underground cable ducts in the horizontal direction, the minimum cost, and the interval range, and use the average interval as the preset interval to construct the horizontal grids.

[0053] Obtain the minimum interval and the maximum interval of the grid distribution in the horizontal direction according to expert experience to get the interval range. Based on the layout of the underground cable ducts in the horizontal direction, obtain the duct alignment and length in the horizontal direction to get the grid layout area in the horizontal direction. Calculate the horizontal grids that meet the uniform distribution based on the minimum cost and the interval range. Divide the grid layout interval in the horizontal direction evenly to get the set of interval parameters that meet the even division. Obtain the optimal interval parameters based on the minimum cost and the interval range. Consider the case where the interval from the end point to the previous grid is the average interval parameter as meeting the even division. The minimum cost is the maximum interval parameter, that is, the larger the horizontal grid spacing, the fewer sensors are arranged in the horizontal direction, and the lower the cost. On the contrary, the smaller the horizontal grid spacing, the more sensors are arranged in the horizontal direction, and the higher the cost. The maximum average interval that meets the interval range can be obtained based on the minimum cost and the interval range, and the horizontal grids are constructed with the maximum average interval. The preset interval is jointly obtained through the average interval, the minimum cost, and the interval range, which can ensure that the horizontal grids meet the sensor layout requirements and that there must be an intersection of a horizontal grid layer and a vertical grid layer at the end point of the underground cable duct layout area for sensor layout, avoiding errors in abnormal analysis caused by the omission of end point data and further improving the accuracy of abnormal analysis.

[0054] Obtaining the set of acquisition point data based on the three-dimensional grids and the underground cable duct data includes:

[0055] Obtain the set of grid intersection points based on the three-dimensional grids, obtain the layout point angles based on the duct shape, and obtain the set of acquisition point data based on the set of grid intersection points and the layout point angles.

[0056] Underground cable duct data includes at least the duct shape. Taking the intersection points of the horizontal grid and the vertical grid as the set of grid intersection points, the duct shape is used to finely distribute the placement point angles to ensure that the sensors cover all directions around the duct. For example, usually the duct shape is circular, and in this case, it is only necessary to deploy sensors at the horizontal angle, vertical angle, and 45° angle of the duct. However, if the duct has a rectangular seal, and the duct shape is rectangular at this time, sensors need to be deployed at the horizontal angle, vertical angle, 22.5° angle, 45° angle, and 75° angle of the duct. Specific grid intersection points are selected as the acquisition points according to the duct shape, so as to adapt to the actual situations of different ducts, ensure the completeness of the acquisition of duct information, and improve the accuracy of anomaly analysis.

[0057] In step S2, obtaining the acquisition distribution corresponding to each type of sensor to be deployed based on the type of sensor to be deployed and the acquisition point data set includes:

[0058] Obtaining the type of sensor to be deployed based on the data requirements to be acquired, obtaining the acquisition parameter characteristics corresponding to the type of sensor to be deployed, and obtaining the acquisition distribution that meets the requirements of the acquisition parameter characteristics from the acquisition point data set corresponding to each type of sensor to be deployed.

[0059] The requirements for data to be collected can be obtained according to the pipeline type. Different pipeline types may generate different anomalies. At this time, different data requirements are used for anomaly analysis, and the corresponding types of sensors to be deployed are obtained to avoid the interference of unnecessary data. In this embodiment, the types of sensors to be deployed include strain gauges, vibration sensors, and temperature sensors for monitoring the state of the pipeline itself, as well as humidity sensors and gas sensors for monitoring the pipeline environment. The distribution positions of different types of sensors are not the same. For example, strain gauges are installed on the inner wall of the pipeline to monitor the stress changes of the pipeline and prevent fractures caused by overload or fatigue. The installation position of the strain gauge is the collection point on the inner wall of the pipeline to ensure comprehensive monitoring of the stress changes of the entire inner wall of the pipeline. Vibration sensors need to detect vibrations, and the vibrations at the bends and joints of the pipeline are more intense. At this time, the installation positions of the vibration sensors are the collection points evenly distributed around the pipeline, especially at the bends and joints, to detect abnormal vibrations and warn of possible mechanical failures or external interferences. Temperature sensors are used to monitor the temperature changes inside and around the pipeline to prevent the impact of overheating or overcooling on the cable performance. The installation positions of the temperature sensors are the combination of collection points with a spacing equal to the sensing distance of the temperature sensor. Considering the influence of the environmental temperature gradient, the temperature sensors are installed near the grid points to ensure accurate capture of temperature changes. The installation positions of the humidity sensors are the collection points near the interface between the soil and the pipeline and the collection points in the key water level fluctuation areas to monitor the soil humidity changes and prevent corrosion or short circuits caused by excessive humidity. The installation positions of the gas sensors are the collection points in the areas at risk of gas leakage to monitor the ambient gas concentration in real time (such as the detection of harmful gases such as methane and hydrogen sulfide) to ensure the safety of personnel and equipment.

[0060] In step S3, it further includes:

[0061] Based on the acquisition distribution, the time-series data of the underground cable pipeline collected by each sensor is obtained, and data preprocessing is performed on the time-series data of the underground cable pipeline;

[0062] Perform clustering analysis on the time-series data of the underground cable pipeline after data preprocessing to obtain abnormal data points.

[0063] Among them, performing data preprocessing on the time-series data of the underground cable pipeline includes:

[0064] Perform data cleaning on the time-series data of the underground cable pipeline;

[0065] Perform data normalization on the time-series data of the underground cable pipeline;

[0066] Perform denoising on the time-series data of the underground cable pipeline.

[0067] Data cleaning includes identifying and processing outliers, handling missing values, etc. Data normalization is to bring data to the same scale for subsequent comparison. The denoising process can use filtering techniques to reduce or eliminate noise in the data, improve the signal-to-noise ratio of the data, and thus more accurately reflect the true state of the pipeline and its surrounding environment.

[0068] As Figure 2 shown, performing clustering analysis on the time series data of underground cable pipelines to obtain abnormal data points includes:

[0069] Dividing the time series data of underground cable pipelines into several subsequences containing continuous observations according to the preset parameter content;

[0070] Performing clustering analysis on each subsequence to obtain the clustering result of each subsequence;

[0071] Obtaining abnormal data points based on the distance algorithm and the clustering result.

[0072] The preset parameter content is the number of parameters contained in a preset subsequence. The time series data of underground cable pipelines is divided into several subsequences, and each subsequence contains a fixed number of continuous observations. Determine the number of clustering centers K, and apply the K-means clustering algorithm to each subsequence. The data points of each subsequence will be assigned to K clusters to obtain the clustering result of each subsequence. It can be understood that the K-means clustering algorithm is adopted in this embodiment. In other embodiments, clustering algorithms such as K-medoids, DBSCAN, and OPTICS can also be used. For example, in the K-means clustering algorithm, each cluster is defined by the average value of its members, while in the K-medoids clustering algorithm, each cluster is defined by the representative point. In the DBSCAN and OPTICS clustering algorithms, clusters are formed according to the density distribution in the data space.

[0073] In this embodiment, the clustering algorithm is K-means and the distance algorithm is the Euclidean distance. According to the clustering result, calculate the distance from each data point to its nearest cluster center based on the Euclidean distance and compare it with the preset distance threshold. Points exceeding the preset distance threshold are regarded as abnormal data points. The preset distance threshold is the average value of the distances from all points to the cluster center plus the standard deviation. That is, at this time, obtaining abnormal data points based on the distance algorithm and the clustering result includes:

[0074] Calculating the shortest distance from each data point in each subsequence to the clustering center based on the Euclidean distance and the clustering result;

[0075] Obtaining the preset distance threshold based on the average distance from each data point in each subsequence to the clustering center and the standard deviation;

[0076] Taking the data points with the minimum distance greater than the preset distance threshold as abnormal data points.

[0077] By collecting the information of each sensor according to the time sequence to form the time-sequence data of the underground cable pipeline, and then dividing it into several subsequences containing continuous observed values according to the preset parameter content, and performing clustering analysis on each subsequence, data points deviating from the clustering center can be obtained. This not only maintains unity in time sequence and avoids deviation interference caused by time-sequence differences, but also obtains short-term abnormal data points based on the clustering of subsequences of continuous observed values. At the same time, abnormal analysis is carried out in both the long term and the short term, and the mutual interference between long-term and short-term data is avoided, improving the accuracy of data abnormal analysis.

[0078] The abnormal positions obtained according to the abnormal data points and the acquisition distribution include:

[0079] Obtain the corresponding sensor according to the abnormal data point;

[0080] Take the sensor position corresponding to the corresponding sensor and the acquisition distribution as the abnormal position.

[0081] Each abnormal data point must correspond to a sensor that collects this data point. Obtain the corresponding sensor, and according to the acquisition distribution, obtain the sensor position, and thus obtain the position where the abnormality occurs as the abnormal position.

[0082] In some other embodiments, the abnormal positions obtained according to the abnormal data points and the acquisition distribution include:

[0083] Obtain the corresponding set of sensors according to each abnormal data point;

[0084] Obtain a set of sensor position points based on the set of sensors;

[0085] Obtain the abnormal area based on the set of sensor position points and the acquisition distribution, and take the abnormal area as the abnormal position.

[0086] In this case, the area surrounded by the sensors corresponding to all abnormal data points is used as the abnormal position. That is, after identifying the abnormal data points, these points are associated with their respective sensors. Since each sensor has clear position information in the three-dimensional grid, the position information of the abnormal data points is combined with that of the sensors to initially determine the abnormal area where the abnormal data points occur. This area is usually a spatial range centered on the sensors associated with the abnormal data points, which facilitates the staff to check the abnormal positions at this time.

[0087] In this embodiment, the abnormal positions obtained according to the abnormal data points and the acquisition distribution further include:

[0088] Obtain the sensor monitoring data corresponding to the abnormal area;

[0089] Perform weighted averaging on the sensor monitoring data based on the relative position relationship of the sensors to obtain the abnormal location.

[0090] After initially determining the abnormal area, further utilize the monitoring data of all relevant sensors within this area. These data include different types of physical quantities (such as temperature, humidity, vibration, etc.), which jointly reflect the comprehensive state of the abnormal area. That is, use the time-series data of the underground cable pipeline containing the monitoring data of the abnormal area as the sensor monitoring data corresponding to the abnormal area. Through multi-source data fusion technology, integrate and fuse these data from different sources and of different types to improve the accuracy and reliability of abnormal detection. Based on multi-source data fusion, use a three-dimensional positioning algorithm to calculate the precise three-dimensional coordinates of the abnormal point. The weighted average method is a simple and effective positioning method. According to the distance between the data points monitored by each sensor and the cluster center and their relative position relationship, different weights are assigned to the data of each sensor. First, calculate the distance of each sensor data point relative to its cluster center, and then calculate the weight of each sensor based on these distances. The weight is usually inversely proportional to the distance, that is, the closer the sensor, the greater the weight:

[0091] ;

[0092] Among them, represents the distance weight of the th sensor, represents the distance from the th sensor to the abnormal data point (or cluster center), is a positive number.

[0093] Finally, use the weighted average method to perform weighted summation on the position information (three-dimensional coordinates) of each sensor to obtain the estimated three-dimensional coordinates of the abnormal point as the abnormal location, further improving the accuracy of abnormal analysis and facilitating the accurate investigation of the abnormal point by the staff.

[0094] As Figure 3 shown, the output of the underground cable pipeline abnormal information according to the abnormal location also includes:

[0095] Obtain historical underground cable pipeline data and construct an abnormal recognition model based on the historical underground cable pipeline data;

[0096] Perform key feature recognition according to the sensors corresponding to the abnormal location and the time-series data of the underground cable pipeline corresponding to the sensors, and obtain the underground cable pipeline abnormal information based on the key features and the abnormal recognition model.

[0097] Historical underground cable pipeline data includes at least historical underground cable pipeline anomaly data. The data under abnormal pipeline conditions is used as underground cable pipeline anomaly data, such as monitoring data in cases of abnormal pipeline temperature changes, abnormal ambient humidity changes, gas leakage, abnormal pipeline vibrations, and abnormal pipeline deformations. The collected historical anomaly data is preprocessed, and the preprocessed historical underground cable pipeline anomaly data is used to train a neural network model to obtain an anomaly recognition model. It can be understood that the anomaly recognition model can be pre-constructed and directly called when abnormal data points are found, or the anomaly recognition model can be updated immediately with the anomaly recognition result after a certain anomaly recognition to ensure the accuracy of the anomaly recognition model.

[0098] Extract key features from the positional relationship between the abnormal location and surrounding sensors, the correlation relationship between the monitoring data at the abnormal location and the monitoring data of surrounding sensors, and the temporal sequence influence relationship, etc. The key features include time series features, statistical features, and correlation features. Time series features include the change trend of data (such as rising, falling, stable), the change rate (i.e., the amount of data change per unit time), and periodic patterns, which are used by users to identify dynamic change patterns in the data. Statistical features include mean, standard deviation, skewness, and kurtosis, etc., which are used to understand the overall distribution characteristics of the data. Extract features that can reflect the correlation between the monitoring data at the abnormal location and the monitoring data of its surrounding sensors, such as the correlation coefficient and mutual information between different types of sensor data. Input the identified key features into the anomaly recognition model to obtain the corresponding abnormal state. At this time, the abnormal location and abnormal state of the underground cable pipeline are output as underground cable pipeline anomaly information to the display terminal, facilitating the staff to intuitively obtain the abnormal situation of the underground cable pipeline, so as to promptly conduct anomaly investigation and resolution.

[0099] In some other cases, a three-dimensional chart of the underground cable pipeline is pre-stored on the display terminal. When the abnormal location and abnormal state of the underground cable pipeline are received, the display terminal performs parsing and processing, and displays the abnormal location and abnormal state in an intuitive and easy-to-understand manner on the three-dimensional chart of the underground cable pipeline. When displaying, methods such as using map markings for the abnormal location, using charts to display the change trend of the abnormal state, and using alarm lights or sounds to prompt can be used to alert relevant personnel.

[0100] As the second embodiment of this application, step S3 further includes:

[0101] S31: Obtain the current offset physical quantity based on the historical underground cable pipeline offset data and the current regional soil distribution, and construct an offset grid with the current offset physical quantity and a three-dimensional grid;

[0102] S32: Obtain the time-series data of the underground cable ducts collected by each sensor based on the acquisition distribution, perform clustering analysis on the time-series data of the underground cable ducts based on the offset grid, and obtain abnormal data points.

[0103] In this embodiment, the current offset physical quantity that affects the offset of the underground cable duct under the current regional soil distribution is obtained through the relationship between the historical offset of the underground cable affected by physical quantities, the soil distribution, and the relationship between the soil distribution and the offset physical quantity. Thus, the three-dimensional grid is offset-corrected based on the current offset physical quantity to obtain the corresponding offset grid. Therefore, when performing clustering analysis, the deviation between the default detection points of the sensors and the actual sensor detections is added to compensate for the sensor detection position error. Without the need for manual correction of the sensors, abnormal data points can still be obtained through the offset difference between the three-dimensional grid and the offset grid, ensuring the accuracy of abnormal analysis while reducing the need for sensor adjustment.

[0104] Among them, step S31 further includes:

[0105] Obtain the correlation between the soil distribution and the offset physical quantity based on the historical offset data of the underground cable ducts;

[0106] Obtain the current offset physical quantity according to the regional soil distribution and the correlation between the soil distribution and the offset physical quantity;

[0107] Construct an offset grid with the current offset physical quantity and the three-dimensional grid.

[0108] The historical offset data of the underground cable ducts includes at least the historical soil distribution of the underground cable ducts, the historical offset amount of the underground cable ducts, the historical offset direction of the underground cable ducts, and the historical offset time of the underground cable ducts. The offset amount of the underground cable duct, the offset direction of the underground cable duct, and the offset time of the underground cable duct are used as the offset physical quantities, and the influence of the soil distribution of the underground cable duct on the offset direction and offset amount of the underground cable duct is used as the weight vector. In some cases, the offset time of the underground cable duct also varies with different soil distributions of the underground cable duct. For example, in some soil distributions, the offset time of the underground cable duct often occurs during the rainy season. Obtain the correlation between the soil distribution and the offset physical quantity based on the historical offset data of the underground cable ducts, and obtain the offset physical quantity under the current regional soil distribution, so as to perform deviation correction on the three-dimensional grid to obtain the offset grid after the three-dimensional grid corresponding to the offset underground cable duct.

[0109] Correspondingly, step S32 further includes:

[0110] Obtain the time-series data of the underground cable ducts collected by each sensor based on the acquisition distribution;

[0111] Perform a clustering on each sensor based on the offset grid and the acquisition distribution;

[0112] Perform secondary clustering on the time-series data of the underground cable ducts collected by each sensor in the same cluster during the first clustering;

[0113] Obtain abnormal data points based on the secondary clustering.

[0114] Through the correspondence between the offset grid and the 3D grid, the offset distribution corresponding to the acquisition distribution can be obtained. According to the offset distribution, sensors that are adjacent in space and similar in function are grouped into the same cluster. Then, secondary clustering is performed on the time-series data of the underground cable ducts collected by each sensor in the same cluster. The steps for performing secondary clustering are as follows:

[0115] Divide the time-series data of the underground cable ducts collected by each sensor in the same cluster into several subsequences containing continuous observation values according to the preset parameter content;

[0116] Perform clustering analysis on each subsequence to obtain the clustering result of each subsequence;

[0117] Obtain abnormal data points based on the distance algorithm and the clustering result.

[0118] Thus, potential anomalies in the time-series data are mined. The secondary clustering is analyzed based on the similarity of the time-series data, such as the shape and change trend of the time series. After the secondary clustering is completed, abnormal data points that are significantly different from most data points are identified. Through the two-stage clustering, the spatial distribution of the sensors and the characteristics of the time-series data can be integrated, so as to more accurately identify abnormal data points and avoid the influence of the same type of sensors at different spatial positions on the abnormal data analysis.

[0119] In this embodiment, the first clustering can adopt clustering algorithms such as K-means, hierarchical clustering, DBSCAN, etc., and the secondary clustering can adopt clustering algorithms suitable for time-series data such as dynamic time warping, hidden Markov model, autoregressive model, etc.

[0120] In some other cases, there may be sensors that deviate from all clustering centers during the first clustering. At this time, step S32 further includes:

[0121] Obtain the sensors to be checked according to the preset threshold of the first clustering;

[0122] Based on the historical monitoring data of the underground cable ducts of the sensors to be checked and the current time-series data of the underground cable ducts, determine whether the sensors to be checked are abnormal. If so, output sensor abnormal information.

[0123] In this case, if the current time-series data of the underground cable pipeline conforms to the fluctuation curve of the historical monitoring data of the underground cable pipeline, it is considered that the sensor to be checked is normal. If the current time-series data of the underground cable pipeline does not conform to the fluctuation curve of the historical monitoring data of the underground cable pipeline, it is considered that the sensor to be checked is abnormal, and then the sensor abnormality information is output, which is convenient for the staff to check whether the sensor has self-offset or failure.

[0124] In this embodiment, an underground cable pipeline abnormality monitoring method further includes:

[0125] S5: Obtain the offset grid data within a preset time series segment, obtain the offset fluctuation trend of each grid based on the offset grid data, and judge whether there is an offset abnormality in the underground cable pipeline according to the offset fluctuation trend of each grid and the grid offset threshold. If so, output the underground cable pipeline offset abnormality information.

[0126] By the offset fluctuation trend of each grid within the preset time series segment, judge whether it exceeds the grid offset threshold, so as to judge whether the offset of the underground cable pipeline needs to be processed, improving the efficiency and accuracy of checking the abnormality of the underground cable pipeline. In some cases, the grid offset threshold is the offset difference threshold of the opposite grid lines. That is, if the offset difference of the opposite grid lines gradually becomes larger, the corresponding grid will show a situation of becoming larger or smaller, which indicates that the soil environment of the underground cable pipeline has undergone a large migration change or the underground cable pipeline has deformed. At this time, it is necessary to prompt the staff to check the actual environment of the underground cable pipeline, and locate the offset abnormality by outputting the corresponding grid coordinates, so as to facilitate the staff to check.

[0127] As the third embodiment of this application, an underground cable pipeline abnormality monitoring system includes:

[0128] A sensor layout module, which constructs a three-dimensional grid based on the regional soil distribution and obtains the acquisition distribution corresponding to each type of sensor to be laid out;

[0129] A data acquisition module, which is used to obtain the time-series data of the underground cable pipeline collected by each sensor;

[0130] A data analysis module, which performs clustering analysis on the time-series data of the underground cable pipeline to obtain abnormal data points and analyzes and outputs the underground cable pipeline abnormality information.

[0131] In this embodiment, a three-dimensional grid is constructed by the sensor layout module to obtain the acquisition distribution corresponding to each type of sensor to be laid, facilitating the pre-layout of sensors. After obtaining the time-series data of the underground cable pipeline collected by each sensor, the data analysis module performs clustering analysis on the time-series data of the underground cable pipeline collected by each sensor and obtains the three-dimensional coordinates where the abnormal data points are located based on the three-dimensional grid. While improving the accuracy of abnormal analysis of the underground cable pipeline, it can directly obtain the abnormal coordinates based on the three-dimensional grid, improving the efficiency of abnormal analysis.

[0132] The data acquisition module at least includes: strain gauges, vibration sensors, and temperature sensors for monitoring the state of the pipeline itself, as well as humidity sensors and gas sensors for monitoring the pipeline environment.

[0133] The data analysis module at least includes: an abnormal data judgment unit, an abnormal position determination unit, and an abnormal state judgment unit. The abnormal data judgment unit is used to perform clustering analysis on the preprocessed time-series data of the underground cable pipeline to identify abnormal data points in the time-series data of the underground cable pipeline; the abnormal position determination unit is used to combine the monitoring data of multiple sensors and use multi-source data fusion and three-dimensional positioning algorithms to locate the abnormal data points to obtain the abnormal position; the abnormal state judgment unit is used to combine the abnormal position and the change situation of the sensor monitoring data at the abnormal data point to judge the abnormal state of the abnormal data point.

[0134] In some other embodiments, an underground cable pipeline abnormal monitoring system further includes:

[0135] A display module that interacts with the user and displays the abnormal position and abnormal state of the underground cable pipeline.

[0136] As the fourth embodiment of this application, a computer-readable storage medium is used to store computer programs or instructions. When the computer programs or instructions are executed by a processing device, the above-mentioned method for monitoring abnormal conditions of an underground cable pipeline is implemented. The computer-readable storage medium can be any available medium that a computing device can store or a data storage device such as a data center that contains one or more available media. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid-state drive), etc.

[0137] The above-mentioned specific implementation manners are the preferred implementation manners of a method and a system for monitoring abnormal conditions of an underground cable pipeline in this application, and do not limit the specific implementation scope of this application. The scope of this application includes but is not limited to this specific implementation manner. Any equivalent changes made according to the shape and structure of this application are within the protection scope of this application.

Claims

1. A method for monitoring abnormalities of underground cable pipelines, characterized in that: The steps include: S1: Construct a three-dimensional grid based on the regional soil distribution, and obtain a collection point data set based on the three-dimensional grid and underground cable pipeline data; S2: based on the type of sensor to be deployed and the data set of the collection points, obtaining the collection distribution corresponding to each type of sensor to be deployed; S3: acquiring the underground cable pipeline time series data collected by each sensor based on the collection distribution, performing cluster analysis on the underground cable pipeline time series data, and obtaining abnormal data points; S4: Obtain the abnormal position according to the abnormal data points and the collection distribution, and output the abnormal information of the underground cable pipeline according to the abnormal position; The acquisition of the collection point data set based on the three-dimensional grid and the underground cable pipeline data includes: Construct a horizontal grid at preset intervals along the horizontal direction; Construct vertical grids based on soil structure information; Construct a three-dimensional grid with horizontal and vertical grids; According to the pipeline shape, the corresponding grid intersection in the three-dimensional grid is selected as the collection point.

2. The method for monitoring abnormalities of underground cable pipelines according to claim 1, characterized in that: The step of constructing a horizontal grid at preset intervals along the horizontal direction comprises: The interval is obtained based on expert experience, the average interval is obtained based on the horizontal underground cable pipeline layout, minimum cost and interval, and the horizontal grid is constructed using the average interval as the preset interval.

3. The method for monitoring abnormalities of underground cable pipelines according to claim 1, characterized in that: The acquisition of the collection point data set based on the three-dimensional grid and the underground cable pipeline data includes: A grid intersection set is obtained based on the three-dimensional grid, the layout point angle is obtained based on the pipeline shape, and the collection point data set is obtained based on the grid intersection set and the layout point angle.

4. The method for monitoring abnormalities of underground cable pipelines according to claim 1, characterized in that: The step of obtaining the collection distribution corresponding to each type of sensor to be deployed based on the type of sensor to be deployed and the collection point data set includes: Based on the demand for data to be collected, the type of sensor to be deployed is obtained, the collection parameter characteristics corresponding to the sensor type to be deployed are obtained, and the collection distribution that meets the collection parameter characteristic requirements is obtained in the collection point data set corresponding to each sensor type to be deployed.

5. The method for monitoring abnormalities of underground cable pipelines according to claim 1, characterized in that: The performing cluster analysis on the underground cable pipeline time series data to obtain abnormal data points includes: Divide the underground cable pipeline time series data into several subsequences containing continuous observation values ​​according to the preset parameter content; Perform cluster analysis on each subsequence to obtain the clustering results of each subsequence; Abnormal data points are obtained based on the distance algorithm and clustering results.

6. The method for monitoring abnormalities of underground cable pipelines according to claim 1, characterized in that: The obtaining of the abnormal position according to the abnormal data point and the collection distribution includes: Obtain the corresponding sensor set according to each abnormal data point; Obtain a sensor position point set using a sensor set; The abnormal area is obtained by using the sensor location point set and collection distribution, and the sensor monitoring data corresponding to the abnormal area is obtained; Based on the relative position relationship of the sensors, the sensor monitoring data is weighted averaged to obtain the abnormal position.

7. The method for monitoring abnormalities of underground cable pipelines according to claim 1, characterized in that: The outputting of underground cable pipeline abnormal information according to the abnormal position also includes: Obtain historical underground cable pipeline data, and build an anomaly recognition model based on the historical underground cable pipeline data; Key feature recognition is performed according to the sensors corresponding to the abnormal positions and the underground cable pipeline time series data corresponding to the sensors, and the abnormal information of the underground cable pipeline is obtained based on the key features and the abnormal recognition model.

8. The method for monitoring abnormalities of underground cable pipelines according to claim 1, characterized in that: The S3 further includes: S31: obtaining a current offset physical quantity based on historical underground cable pipeline offset data and soil distribution in the current area, and constructing an offset grid using the current offset physical quantity and a three-dimensional grid; S32: acquiring underground cable pipeline time series data collected by each sensor based on the collection distribution, performing cluster analysis on the underground cable pipeline time series data based on the offset grid, and acquiring abnormal data points.

9. The method for monitoring abnormalities of underground cable pipelines according to claim 8, characterized in that: The S31 further includes: Obtain the correlation between soil distribution and displacement physical quantity based on historical underground cable pipeline displacement data; Obtain the current offset physical quantity according to the regional soil distribution and the correlation between the soil distribution and the offset physical quantity; Constructs an offset mesh using the current offset physics and the 3D mesh.

10. The method for monitoring abnormalities of underground cable pipelines according to claim 8, characterized in that: The S32 further includes: Acquire underground cable pipeline time series data collected by each sensor based on the collection distribution; Perform a clustering of each sensor based on the offset grid and the acquisition distribution; Perform secondary clustering based on the underground cable pipeline time series data collected by each sensor in the same cluster in the primary clustering; Abnormal data points are obtained based on secondary clustering.

11. The method for monitoring abnormalities of underground cable pipelines according to claim 10, characterized in that: The S32 further includes: The sensor to be checked is obtained according to a preset threshold of clustering; Based on the historical underground cable pipeline monitoring data of the sensor to be checked and the current underground cable pipeline time series data, it is determined whether the sensor to be checked is abnormal. If so, the sensor abnormality information is output.

12. The method for monitoring abnormalities of underground cable pipelines according to claim 8, characterized in that: Also includes: S5: Obtain the offset grid data within a preset time segment, obtain the offset fluctuation trend of each grid based on the offset grid data, and determine whether the underground cable pipeline has an offset anomaly based on the offset fluctuation trend of each grid and the grid offset threshold. If so, output the underground cable pipeline offset anomaly information.

13. An underground cable pipeline abnormality monitoring system, used to implement the method according to any one of claims 1 to 12, characterized in that: include: The sensor deployment module builds a three-dimensional grid based on the regional soil distribution and obtains the collection distribution corresponding to each sensor type to be deployed; A data acquisition module is used to obtain the underground cable pipeline time series data collected by each sensor; The data analysis module performs cluster analysis based on the underground cable pipeline time series data to obtain abnormal data points, and analyzes and outputs the abnormal information of the underground cable pipeline.

Images