Multi-source data fusion underground pipe gallery real-time state dynamic monitoring method and system
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA COAL RES INST
- Filing Date
- 2025-07-14
- Publication Date
- 2026-06-02
Smart Images
Figure CN120499529B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of condition monitoring technology, specifically to a method and system for real-time dynamic monitoring of the condition of underground utility tunnels using multi-source data fusion. Background Technology
[0002] Real-time status monitoring of underground utility tunnels typically relies on a fixed monitoring frequency. This fixed-frequency monitoring method cannot flexibly respond to changes in the tunnel's status, especially when risks occur, as it cannot promptly increase the monitoring frequency to obtain more real-time data for analysis. Due to the lack of a dynamic adjustment mechanism, the monitoring system cannot capture key data in a timely manner when risks suddenly arise, leading to delays in risk identification and early warning response, thereby affecting the safety and operational efficiency of the utility tunnel. Summary of the Invention
[0003] This application provides a method and system for real-time dynamic monitoring of underground utility tunnels based on multi-source data fusion, which addresses the technical problem that the real-time monitoring frequency of underground utility tunnels is fixed in the existing technology and cannot be dynamically adjusted according to changes in risk.
[0004] In view of the above problems, this application provides a method and system for real-time dynamic monitoring of the status of underground utility tunnels by multi-source data fusion.
[0005] The first aspect of this application provides a method for real-time dynamic monitoring of the status of underground utility tunnels through multi-source data fusion, the method comprising:
[0006] A set of monitoring points for the underground utility tunnel is acquired. A multi-source status monitoring terminal set monitors the set of monitoring points at a preset monitoring frequency, and the monitoring data is transmitted to a data transmission center, which has a data transfer storage device with window constraints. Real-time risk identification is performed based on the data transmission center to determine a set of real-time risk received data and a set of risk monitoring points. Causal tracing is performed based on the set of risk monitoring points and the set of real-time risk received data to determine a cluster of tracing clues. Using the tracing clue clusters as indexes, clues are extracted from the data transfer storage device to determine a cluster of tracing received data sequences. Time-series iterative trend risk authentication is performed on each of the tracing received data sequence clusters to determine a cluster of risk trend features. Based on the risk trend feature clusters, a first dynamic monitoring frequency and a second dynamic monitoring frequency are determined, and both are distributed to status monitoring terminals for dual-frequency synchronous status monitoring of the monitoring point set.
[0007] A second aspect of this application provides a real-time dynamic monitoring system for the status of underground utility tunnels based on multi-source data fusion, the system comprising:
[0008] The monitoring module acquires a set of monitoring points for the underground utility tunnel, monitors the set of monitoring points using a multi-source status monitoring terminal set at a preset monitoring frequency, and transmits the monitoring data to a data transmission center, which has a data transfer storage device with window constraints. The risk identification module performs real-time risk identification based on the data transmission center, determining the real-time risk received data set and the risk monitoring point set. The causal tracing module performs causal tracing based on the risk monitoring point set and the real-time risk received data set, determining tracing clue clusters. The clue extraction module uses the tracing clue clusters as indexes to extract clues from the data transfer storage device, determining tracing received data sequence clusters. The risk authentication module performs time-series iterative trend risk authentication on the tracing received data sequence clusters, determining risk trend feature clusters. The dual-frequency synchronous monitoring module determines a first dynamic monitoring frequency and a second dynamic monitoring frequency based on the risk trend feature clusters, and distributes both the first and second dynamic monitoring frequencies to status monitoring terminals for dual-frequency synchronous status monitoring of the monitoring point set.
[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0010] This application acquires a set of monitoring points for underground utility tunnels, monitors the set of monitoring points using a multi-source status monitoring terminal set at a preset monitoring frequency, and transmits the monitoring data to a data transmission center, wherein the data transmission center has a data transfer storage device with window constraints; real-time risk identification is performed based on the data transmission center to determine a set of real-time risk received data and a set of risk monitoring points; causal tracing is performed based on the set of risk monitoring points and the set of real-time risk received data to determine a cluster of tracing clues; using the clusters of tracing clues as indexes, clues are extracted from the data transfer storage device to determine a cluster of tracing received data sequences; time-series iterative trend risk authentication is performed on the clusters of tracing received data sequences to determine a cluster of risk trend features; a first dynamic monitoring frequency and a second dynamic monitoring frequency are determined based on the clusters of risk trend features, and both the first dynamic monitoring frequency and the second dynamic monitoring frequency are distributed to status monitoring terminals to perform dual-frequency synchronous status monitoring of the set of monitoring points. This application addresses the technical problem in existing technologies where the real-time status monitoring frequency of underground utility tunnels is fixed and cannot be dynamically adjusted according to changes in risk. By integrating multi-source monitoring data for causal tracing and trend risk authentication, and generating a dynamic monitoring frequency accordingly, the technical effect of achieving adaptive adjustment of the monitoring frequency according to the risk status is realized, thereby improving monitoring accuracy and response timeliness. Attached Figure Description
[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0012] Figure 1 A schematic diagram of the process for real-time dynamic monitoring of the status of underground utility tunnels using multi-source data fusion, provided in an embodiment of this application.
[0013] Figure 2 A schematic diagram of the structure of a multi-source data fusion-based real-time dynamic monitoring system for underground utility tunnels provided in this application embodiment.
[0014] Explanation of reference numerals in the attached diagram: Monitoring module 11, Risk identification module 12, Causal tracing module 13, Clue extraction module 14, Risk authentication module 15, Dual-frequency synchronous monitoring module 16. Detailed Implementation
[0015] This application provides a method and system for real-time dynamic monitoring of underground utility tunnels by fusing multi-source data. It addresses the technical problem that the real-time monitoring frequency of underground utility tunnels is fixed in the existing technology and cannot be dynamically adjusted according to changes in risk. By fusing multi-source monitoring data to perform causal tracing and trend risk authentication, and generating a dynamic monitoring frequency accordingly, the technical effect of achieving adaptive adjustment of the monitoring frequency according to the risk status is achieved, thereby improving the monitoring accuracy and response timeliness.
[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.
[0017] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.
[0018] Example 1, as Figure 1 As shown, this application provides a method for real-time dynamic monitoring of the status of underground utility tunnels based on multi-source data fusion, the method comprising:
[0019] Step S100: Obtain the set of monitoring points for the underground utility tunnel, use a set of multi-source status monitoring terminals to monitor the set of monitoring points at a preset monitoring frequency, and transmit the monitoring data to the data transmission center, wherein the data transmission center has a data transfer storage device with window constraints.
[0020] In this embodiment, a predetermined set of monitoring points for the underground utility tunnel is first obtained. These monitoring points refer to sensors or devices installed at various key locations within the tunnel to collect various status data in real time, such as temperature, humidity, and gas concentration. Then, a multi-source status monitoring terminal set, comprising multiple different types of monitoring devices (such as gas sensors, temperature sensors, and video surveillance), is used to periodically or continuously monitor these points according to a preset monitoring frequency. The monitoring frequency refers to the time interval between data collection by the devices.
[0021] Data collected through multi-source monitoring terminals is transmitted to the data transmission center. The data transmission center is responsible for centrally receiving, processing, and storing the monitoring data. To ensure the efficiency and real-time nature of data processing, the data transmission center uses a data transfer storage device with window constraints. The window constraint of this storage device means that there are certain time or data volume limitations when storing data, thereby ensuring the efficiency and timeliness of data transmission and processing.
[0022] Step S200: Based on the data transmission center, perform real-time risk identification to determine the real-time risk reception data set and the risk monitoring point set.
[0023] In this embodiment, when performing real-time risk identification based on a data transmission center, firstly, a set of real-time received data from multiple sources is extracted from the data transmission center. Then, a multi-source standard state data set is used to identify deviations in the extracted real-time data, identifying anomalies or deviations in the data. Next, risk identification is performed on the identified deviation data according to a preset risk deviation threshold, filtering out real-time received data that may pose a risk. Finally, the set of risky real-time received data and the set of risk monitoring points are determined.
[0024] Furthermore, in the method provided in the application embodiments, real-time risk identification based on the data transmission center, determining the real-time risk received data set and the risk monitoring point set, further includes:
[0025] The multi-source real-time received data set is extracted from the data transmission center. The deviation of the multi-source real-time received data set is identified using the multi-source standard state data set to determine the multi-source real-time received data deviation set. The multi-source real-time received data deviation set is then risk-identified according to a preset risk deviation threshold to determine the risk real-time received data set and the risk monitoring point set, wherein each multi-source real-time received data set corresponds to one monitoring point.
[0026] In this embodiment, a multi-source real-time received data set is first extracted from the data transmission center. This data comes from multiple monitoring terminals in the underground utility tunnel, such as gas sensors and temperature sensors. These monitoring terminals are used to collect environmental status information at each monitoring point within the tunnel in real time. Next, this real-time received data is compared with a multi-source standard status data set. The standard data set is based on historical data or preset standard reference data under normal operating conditions of the utility tunnel. By comparing the data, deviations are identified by subtracting the corresponding multi-source real-time received data from the multi-source standard status data, calculating the difference between the real-time data and the standard data at each monitoring point, thereby obtaining a multi-source real-time received data deviation set.
[0027] Next, risk identification is performed on the multi-source real-time received data deviation set according to the preset risk deviation thresholds set by technical experts. When the deviation data exceeds the set threshold, it is determined that there is a risk at that monitoring point, thus forming a risk real-time received data set, which includes all real-time received data whose deviation data exceeds the threshold. At the same time, a risk monitoring point set is determined, that is, the set of monitoring points associated with these deviation data.
[0028] Step S300: Perform causal tracing based on the set of risk monitoring points and the set of real-time risk data to determine the cluster of tracing clues.
[0029] In this embodiment, when performing causal tracing based on the risk monitoring point set and the real-time risk data set, the first step is to obtain a set of historical risk logs for the underground utility tunnel. These logs contain records of various risk events that have occurred in the tunnel in the past. Next, a causal tracing vector set is constructed based on the risk monitoring point set and the real-time risk data set. This set forms a causal relationship network of potential risks by associating real-time and historical data. Then, the causal tracing vector set is approximated and matched within the historical risk log set of the underground utility tunnel to obtain a set of matching historical risk log clusters related to the real-time data. Finally, tracing clues are extracted from each log set in these matching historical risk log clusters to obtain a cluster of tracing clues.
[0030] Furthermore, in the method provided in the application embodiments, the process of determining the cluster of tracing clues by performing causal tracing based on the set of risk monitoring points and the set of real-time risk received data further includes:
[0031] Obtain a set of historical risk logs for underground utility tunnels; construct a set of causal tracing vectors based on the set of risk monitoring points and the set of real-time risk received data; perform approximate matching of the causal tracing vector set in the set of historical risk logs for underground utility tunnels to obtain a cluster of matching historical risk logs for underground utility tunnels; extract tracing clues from each set of matching historical risk logs for underground utility tunnels in the cluster of matching historical risk logs for underground utility tunnels to obtain the cluster of tracing clues.
[0032] In this embodiment of the application, the historical risk log set of the underground utility tunnel is first obtained from the historical database. These logs record risk events and fault information that occurred in the history of the utility tunnel, including equipment failure, environmental anomalies, system failures, etc.
[0033] Next, a set of causal origination vectors is constructed based on the set of risk monitoring points and the set of real-time risk data. Each causal origination vector consists of a risk monitoring point and a set of real-time risk data. For example, a monitoring point might be a gas sensor inside an underground utility tunnel, and the real-time data is the gas concentration value detected by that sensor. By subtracting the real-time data from the standard data, the deviation is calculated, and these deviation values are compared with the data from the monitoring point to construct a set of causal origination vectors. For example, if the deviation of the real-time gas concentration data exceeds a set threshold, a corresponding causal origination vector will be generated.
[0034] Then, an approximate matching method is used to match the constructed causal source vector set within the historical risk log set of the underground utility tunnel. To perform this operation, a similarity measurement algorithm, such as cosine similarity, is employed to calculate the similarity between the current causal source vector and each event in the historical log. Only when the causal source vector and an event in the historical log reach a preset matching threshold in terms of time, space, and features is the event considered to have significant similarity to the current risk state. Through similarity calculation, matching clusters of historical risk logs for the underground utility tunnel are obtained; these clusters contain all historical risk logs similar to the current monitoring data.
[0035] Subsequently, the first set of historical risk logs for matching underground utility tunnels is extracted from the historical risk log cluster, i.e., the historical logs that match the current monitoring data. Next, the set of associated monitoring points is extracted from this log, i.e., all monitoring points associated with this historical event. These sets of associated monitoring points are then merged by performing a union operation, combining monitoring points from different logs to obtain the first set of tracing clues. These clue sets reveal potential risk sources and their associated monitoring points. Finally, by traversing the entire historical risk log cluster for matching underground utility tunnels and extracting tracing clues from all matching logs, a complete set of tracing clues is obtained.
[0036] Furthermore, in the method provided in the application embodiment, extracting tracing clues from each set of historical risk logs of the matched underground utility tunnel in the historical risk log cluster to obtain the tracing clue cluster, further includes:
[0037] Extract a first set of historical risk logs for matching underground utility tunnels from the historical risk log clusters of matching underground utility tunnels; extract a set of associated monitoring points from the first set of historical risk logs for matching underground utility tunnels; perform a union operation on the set of associated monitoring points, and obtain a first set of tracing clues based on the processing result; traverse the historical risk log clusters of matching underground utility tunnels to extract tracing clues, and obtain the set of tracing clues.
[0038] In this embodiment, a first set of matching historical risk logs for underground utility tunnels is first extracted from the matching historical risk log clusters. This set is randomly selected from the already obtained historical risk log clusters.
[0039] Next, the associated monitoring point set is extracted from the first set of historical risk logs for underground utility tunnels. This step identifies the monitoring points associated with each log entry by analyzing risk events in the historical logs. For example, if the historical logs record a gas leak event, the gas sensor monitoring points associated with that event are extracted, and these monitoring points are grouped into an associated monitoring point set.
[0040] Next, a union operation is performed on the set of associated monitoring points. The purpose of this step is to merge monitoring points mentioned in different log entries, eliminating duplicates. The union operation ensures that a single set containing all relevant monitoring points is obtained. For example, if multiple log entries mention gas sensor A, that sensor will only appear once in the final set, without duplicates. Based on the merged set of associated monitoring points, the first set of source tracing clues is then obtained. This step uses data analysis and feature extraction methods to identify monitoring points and their behaviors that may be closely related to the occurrence of risk events by analyzing the historical data, deviation values, and trends of each monitoring point. The characteristics of these monitoring points (such as abnormal fluctuations, changes exceeding thresholds, etc.) are used to construct the source tracing clue set, which includes all monitoring points related to the current risk and provides a basis for subsequent risk source tracing analysis.
[0041] Finally, for each historical log entry in the matching underground utility tunnel historical risk log cluster, tracing clues are extracted to identify potential risk sources and causal chains within each log entry. Using causal reasoning methods, the causal relationships between different events are determined by analyzing the time series data and monitoring point data in the logs. For example, if abnormal data from a gas sensor coincides with a equipment failure event, it is inferred that there might be a causal relationship between these two events, and this is extracted as a tracing clue to ultimately obtain a cluster of tracing clues.
[0042] Step S400: Using the traceability clue cluster as an index, extract clues from the data transfer memory to determine the traceability received data sequence cluster.
[0043] In this embodiment, the traceability clue cluster is used as an index to extract clues from the data transfer memory, quickly locating the data set related to the traceability clues. By using the monitoring points and event information specified in the traceability clue cluster, corresponding historical and real-time data are retrieved from the data transfer memory. These data are closely related to the key monitoring points or events extracted from the traceability clue cluster, including sensor data, device status, environmental parameters, etc.
[0044] After retrieving relevant data, the data sequence cluster for tracing the source is determined. This step involves organizing the data extracted from the transfer storage into a data sequence cluster in chronological order, forming a data time series related to the tracing clues. This sequence cluster shows the trend of data changes over time, helping to analyze the behavior patterns and abnormal fluctuations of monitoring points, and further revealing potential risk factors. For example, data read from the gas accumulation sensor at different time periods can be used to construct a complete time series.
[0045] Step S500: Perform time-series iterative trend risk authentication on the traceability received data sequence clusters respectively to determine the risk trend feature clusters.
[0046] In this embodiment, a first source-received data sequence is first extracted from the source-received data sequence cluster. Then, a time-series iterative correlation trend analysis is performed on this sequence in chronological order to identify changing trends and potential risks within the data sequence. By analyzing the patterns of data changes over time, data trend features, such as abnormal fluctuations or continuous growth trends, are extracted. Subsequently, a pre-trained risk trend identifier is invoked to authenticate the extracted trend features. If the data trend meets preset risk criteria, the trend feature is added to the risk trend feature cluster. Finally, all source-received data sequence clusters are traversed, and a time-series iterative trend analysis is performed on each sequence, which is then authenticated using the risk trend identifier, ultimately yielding a complete risk trend feature cluster.
[0047] Furthermore, in the method provided in the application embodiments, performing time-series iterative trend risk authentication on the traceability received data sequence clusters to determine risk trend feature clusters, further includes:
[0048] Extract the first source-received data sequence from the source-received data sequence cluster; perform time-series iterative correlation trend analysis on the first source-received data sequence in chronological order to determine the first source-received data trend characteristics; call the risk trend identifier to perform risk authentication on the first source-received data trend characteristics, and if the authentication is successful, add the first source-received data trend characteristics to the risk trend feature cluster; traverse the source-received data sequence cluster to perform time-series iterative trend analysis, and use the risk trend identifier to perform risk authentication on the analysis results to obtain the risk trend feature cluster.
[0049] In this embodiment, a first source tracing received data sequence is randomly selected from the source tracing received data sequence cluster. This sequence contains real-time monitoring data related to source tracing clues. These data reflect the status and changes of risk monitoring points, such as temperature and gas concentration.
[0050] Next, a temporal iterative correlation trend analysis is performed on the first source-tracing received data sequence in chronological order. This process begins with feature extraction of the first source-tracing received data sequence, including extracting corresponding temporal and spatial trend features from both temporal and spatial scales. Then, similarity identification is performed on these features; by comparing the similarity between temporal and spatial trend features, a first trend feature similarity set is determined. Next, a first iterative correlation matrix is constructed based on the similarity set, and this matrix is used to perform feature convolution on the spatial trend features, thereby generating the final trend features of the first source-tracing received data.
[0051] Subsequently, a risk trend identifier is invoked to perform risk authentication on the extracted trend features of the first source tracing received data. The risk trend identifier is a pre-trained machine learning model whose training data includes historical source tracing received data trend features and labels annotated by technical experts. These labels are categorized as "pass" (indicating the data trend conforms to a normal risk pattern) and "fail" (indicating the data trend does not conform to a known risk pattern). By training with this data, the identifier learns the relationship between different trend features and risk events. During the authentication process, the trend features of the first source tracing received data are input into the risk trend identifier for risk authentication, and the output is either "pass" or "fail." If the authentication passes, it means the current trend matches the characteristics of a potential risk, and this trend feature is added to the risk trend feature cluster.
[0052] Finally, all data sequences in the received data sequence cluster are traversed, and time-series iterative trend analysis is performed on each one. A risk trend identifier is then used to authenticate the analysis results of each data sequence. Each data sequence undergoes time-series analysis to extract its trend features, and then the identifier determines whether it meets the risk criteria. If it meets the criteria, the trend feature is added to the risk trend feature cluster. Through this method, the final risk trend feature cluster is obtained.
[0053] Furthermore, in the method provided in the application embodiments, performing time-series iterative correlation trend analysis on the first source tracing received data sequence in chronological order to determine the trend characteristics of the first source tracing received data further includes:
[0054] Feature extraction is performed on the first source-tracing received data sequence according to time and space scales to obtain the time trend features and spatial trend features of the first source-tracing received data; similarity identification is performed on the time trend features and spatial trend features of the first source-tracing received data to determine a first trend feature similarity set; a first iterative correlation matrix is constructed based on the first trend feature similarity set, and feature convolution is performed on the spatial trend features of the first source-tracing received data using the first iterative correlation matrix to generate the first source-tracing received data trend features.
[0055] Furthermore, the method provided in the application embodiments also includes:
[0056] The first trend feature similarity set is normalized, and the processing result is filled into an initially empty diagonal matrix to generate the first iterative correlation matrix.
[0057] In this embodiment, feature extraction is first performed on the first traceability received data sequence according to both temporal and spatial scales. During this process, a one-dimensional convolution operation is used to extract the temporal trend features of the first traceability received data sequence in the temporal dimension, capturing patterns of data change over time, such as gradual increases, decreases, or periodic fluctuations. Simultaneously, a two-dimensional convolution operation is used to extract the spatial trend features of the first traceability received data sequence in both temporal and spatial dimensions, thereby identifying the spatial relationships and trends between monitoring points. Through this step, features reflecting temporal and spatial variation patterns are obtained, resulting in the temporal trend features and spatial trend features of the first traceability received data.
[0058] Next, similarity identification is performed on the extracted temporal and spatial trend features of the first source-received data. In this process, methods such as cosine similarity or Euclidean distance are used to calculate the similarity between each temporal and spatial trend feature, determining which data sequences exhibit similar patterns of change in both time and space. This process yields a first trend feature similarity set, which contains similarity values between all temporal and spatial features.
[0059] Based on the first trend feature similarity set, the first iterative association matrix is then constructed. During this process, the similarity values in the similarity set are normalized to ensure that all similarity values are within the same scale range (e.g., 0 to 1). The normalized similarity values are then filled into an initially empty diagonal matrix to generate the first iterative association matrix.
[0060] Finally, the spatial trend features of the first source-tracing received data are subjected to feature convolution using the first iteration correlation matrix. Feature convolution combines the similarity information in the correlation matrix with the spatial trend features in a weighted manner. Specifically, the convolution operation uses the similarity values of the correlation matrix as weights to perform a weighted average of the spatial trend features, thereby strengthening the features of monitoring points with similar patterns in the space. This weighted convolution operation can better capture the potential relationships between data, enhance the correlation of features, and ultimately generate more accurate trend features of the first source-tracing received data.
[0061] Step S600: Determine the first dynamic monitoring frequency and the second dynamic monitoring frequency based on the risk trend feature cluster, and distribute both the first dynamic monitoring frequency and the second dynamic monitoring frequency to the status monitoring terminal to perform dual-frequency synchronous status monitoring on the monitoring point set.
[0062] In this embodiment, risk scoring is first performed based on risk trend feature clusters to obtain a risk score set. These scores reflect the severity and priority of each risk trend feature, and the potential risk level is determined by assessing the changing trend of each feature. Next, dynamic monitoring frequencies are matched based on the maximum and minimum values of the risk score set to generate a first dynamic monitoring frequency and a second dynamic monitoring frequency. The first dynamic monitoring frequency corresponds to high-risk points and is typically sampled at high frequencies for rapid response to changes, while the second dynamic monitoring frequency corresponds to low-risk points and uses lower-frequency monitoring to reduce data redundancy and sampling burden.
[0063] The two frequencies are then distributed to the status monitoring terminals to ensure that different types of monitoring points are sampled appropriately according to their risk levels. This step enables synchronous status monitoring at dual frequencies, allowing different monitoring points to be monitored in real time according to the set dynamic frequencies, ensuring the efficiency and accuracy of data collection.
[0064] Furthermore, in the method provided in the application embodiments, determining the first dynamic monitoring frequency and the second dynamic monitoring frequency based on the risk trend feature cluster further includes:
[0065] Risk scoring is performed based on the risk trend feature cluster to obtain a risk score set; dynamic monitoring frequency matching is performed based on the maximum and minimum values of the risk score set to generate a first dynamic monitoring frequency and a second dynamic monitoring frequency.
[0066] In this embodiment, risk scoring is first performed based on risk trend feature clusters to obtain a risk score set. Specifically, each trend feature in the risk trend feature cluster is scored. The scoring process depends on the magnitude of change, the rate of change, the duration, and the degree of anomaly. For example, if the gas concentration at a monitoring point changes significantly over a period of time, a higher score is assigned to that feature according to preset rules. For instance, if the gas concentration changes by more than 10% within one minute, the feature is assigned 10 points; if the change is between 1% and 5%, it is assigned 5 points; and if the change is less than 1%, it is assigned 1 point. Similarly, the rate of change and duration are scored according to set thresholds. If the rate of change is fast, such as a fluctuation exceeding 5% per second, the score is higher (e.g., 10 points); if the rate is slow, the score is lower (e.g., 3 points). By scoring based on the degree of anomaly, if the data deviates significantly from the normal range and lasts for a long time, the anomaly score is higher. Finally, these scores are weighted and summed. The weights of each scoring factor are set by technical experts according to actual needs, resulting in a total score for each risk feature. Through these steps, a risk score set is obtained.
[0067] Next, dynamic monitoring frequency matching is performed based on the maximum and minimum values of the risk score set. First, all risk scores are normalized, that is, the score values are converted to the range [0, 1]. After normalization, the maximum and minimum risk scores represent the highest and lowest risk scores among the monitoring points, respectively. The normalized maximum and minimum score values are matched with a preset rule table. In the preset rule table, corresponding monitoring frequencies are set according to different risk score values. For example, the rule table may specify that when the risk score is close to 1 (i.e., high risk), the monitoring frequency is once per second; when the risk score is close to 0 (i.e., low risk), the monitoring frequency is once per minute. According to the settings of the rule table, these normalized score values are matched with the corresponding frequencies to obtain the first dynamic monitoring frequency and the second dynamic monitoring frequency.
[0068] Furthermore, the method provided in the application embodiments also includes:
[0069] Based on the maximum value in the risk score set, the size of the window constraint of the data transfer storage is optimized and adjusted.
[0070] In this embodiment, when optimizing the window constraint size of the data transfer storage based on the maximum value in the risk score set, the maximum risk score is first extracted and normalized to ensure that the score value is within a uniform range (between 0 and 1). After normalization, the normalized maximum risk score value is converted into a window size through a linear mapping. Specifically, window size = minimum window size + normalized maximum risk score × (maximum window size - minimum window size). This method completes the optimization and adjustment of the window constraint size of the data transfer storage.
[0071] In summary, the embodiments of this application have at least the following technical effects:
[0072] This application acquires a set of monitoring points for underground utility tunnels, monitors the set of monitoring points using a multi-source status monitoring terminal set at a preset monitoring frequency, and transmits the monitoring data to a data transmission center, wherein the data transmission center has a data transfer storage device with window constraints; real-time risk identification is performed based on the data transmission center to determine a set of real-time risk received data and a set of risk monitoring points; causal tracing is performed based on the set of risk monitoring points and the set of real-time risk received data to determine a cluster of tracing clues; using the clusters of tracing clues as indexes, clues are extracted from the data transfer storage device to determine a cluster of tracing received data sequences; time-series iterative trend risk authentication is performed on the clusters of tracing received data sequences to determine a cluster of risk trend features; a first dynamic monitoring frequency and a second dynamic monitoring frequency are determined based on the clusters of risk trend features, and both the first dynamic monitoring frequency and the second dynamic monitoring frequency are distributed to status monitoring terminals to perform dual-frequency synchronous status monitoring of the set of monitoring points. This application addresses the technical problem in existing technologies where the real-time status monitoring frequency of underground utility tunnels is fixed and cannot be dynamically adjusted according to changes in risk. By integrating multi-source monitoring data for causal tracing and trend risk authentication, and generating a dynamic monitoring frequency accordingly, the technical effect of achieving adaptive adjustment of the monitoring frequency according to the risk status is realized, thereby improving monitoring accuracy and response timeliness.
[0073] Example 2, based on the same inventive concept as the multi-source data fusion method for real-time dynamic monitoring of underground utility tunnels in the aforementioned examples, such as... Figure 2 As shown, this application provides a real-time dynamic monitoring system for underground utility tunnels based on multi-source data fusion. The system and method embodiments in this application are based on the same inventive concept. The system includes:
[0074] The monitoring module 11 is used to acquire a set of monitoring points for the underground utility tunnel, monitor the set of monitoring points using a set of multi-source status monitoring terminals at a preset monitoring frequency, and transmit the monitoring data to a data transmission center, wherein the data transmission center has a data transfer storage device with window constraints; the risk identification module 12 is used to perform real-time risk identification based on the data transmission center, and determine the real-time risk received data set and the risk monitoring point set; the causal tracing module 13 is used to perform causal tracing based on the risk monitoring point set and the real-time risk received data set, and determine the tracing clue cluster; the clue extraction module 14 is used to extract clues from the data transfer storage device using the tracing clue cluster as an index, and determine the tracing received data sequence cluster; the risk authentication module 15 is used to perform time-series iterative trend risk authentication on the tracing received data sequence clusters respectively, and determine the risk trend feature cluster; the dual-frequency synchronous monitoring module 16 is used to determine a first dynamic monitoring frequency and a second dynamic monitoring frequency based on the risk trend feature cluster, and distribute the first dynamic monitoring frequency and the second dynamic monitoring frequency to the status monitoring terminal, and perform dual-frequency synchronous status monitoring on the set of monitoring points.
[0075] Furthermore, the system is also used to implement the following functions:
[0076] The multi-source real-time received data set is extracted from the data transmission center. The deviation of the multi-source real-time received data set is identified using the multi-source standard state data set to determine the multi-source real-time received data deviation set. The multi-source real-time received data deviation set is then risk-identified according to a preset risk deviation threshold to determine the risk real-time received data set and the risk monitoring point set, wherein each multi-source real-time received data set corresponds to one monitoring point.
[0077] Furthermore, the system is also used to implement the following functions:
[0078] Obtain a set of historical risk logs for underground utility tunnels; construct a set of causal tracing vectors based on the set of risk monitoring points and the set of real-time risk received data; perform approximate matching of the causal tracing vector set in the set of historical risk logs for underground utility tunnels to obtain a cluster of matching historical risk logs for underground utility tunnels; extract tracing clues from each set of matching historical risk logs for underground utility tunnels in the cluster of matching historical risk logs for underground utility tunnels to obtain the cluster of tracing clues.
[0079] Furthermore, the system is also used to implement the following functions:
[0080] Extract a first set of historical risk logs for matching underground utility tunnels from the historical risk log clusters of matching underground utility tunnels; extract a set of associated monitoring points from the first set of historical risk logs for matching underground utility tunnels; perform a union operation on the set of associated monitoring points, and obtain a first set of tracing clues based on the processing result; traverse the historical risk log clusters of matching underground utility tunnels to extract tracing clues, and obtain the set of tracing clues.
[0081] Furthermore, the system is also used to implement the following functions:
[0082] Extract the first source-received data sequence from the source-received data sequence cluster; perform time-series iterative correlation trend analysis on the first source-received data sequence in chronological order to determine the first source-received data trend characteristics; call the risk trend identifier to perform risk authentication on the first source-received data trend characteristics, and if the authentication is successful, add the first source-received data trend characteristics to the risk trend feature cluster; traverse the source-received data sequence cluster to perform time-series iterative trend analysis, and use the risk trend identifier to perform risk authentication on the analysis results to obtain the risk trend feature cluster.
[0083] Furthermore, the system is also used to implement the following functions:
[0084] Feature extraction is performed on the first source-tracing received data sequence according to time and space scales to obtain the time trend features and spatial trend features of the first source-tracing received data; similarity identification is performed on the time trend features and spatial trend features of the first source-tracing received data to determine a first trend feature similarity set; a first iterative correlation matrix is constructed based on the first trend feature similarity set, and feature convolution is performed on the spatial trend features of the first source-tracing received data using the first iterative correlation matrix to generate the first source-tracing received data trend features.
[0085] Furthermore, the system is also used to implement the following functions:
[0086] The first trend feature similarity set is normalized, and the processing result is filled into an initially empty diagonal matrix to generate the first iterative correlation matrix.
[0087] Furthermore, the system is also used to implement the following functions:
[0088] Risk scoring is performed based on the risk trend feature cluster to obtain a risk score set; dynamic monitoring frequency matching is performed based on the maximum and minimum values of the risk score set to generate a first dynamic monitoring frequency and a second dynamic monitoring frequency.
[0089] Furthermore, the system is also used to implement the following functions:
[0090] Based on the maximum value in the risk score set, the size of the window constraint of the data transfer storage is optimized and adjusted.
[0091] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.
[0092] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.
[0093] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.
Claims
1. A method for real-time dynamic monitoring of the status of underground utility tunnels based on multi-source data fusion, characterized in that, The method includes: The system acquires a set of monitoring points for the underground utility tunnel, uses a set of multi-source status monitoring terminals to monitor the set of monitoring points at a preset monitoring frequency, and transmits the monitoring data to a data transmission center. The data transmission center has a data transfer storage device with window constraints. Real-time risk identification based on the data transmission center, determining the real-time risk received data set and the risk monitoring point set, includes: extracting a multi-source real-time received data set from the data transmission center; using a multi-source standard state data set to identify deviations in the multi-source real-time received data set to determine a multi-source real-time received data deviation set; and performing risk identification on the multi-source real-time received data deviation set according to a preset risk deviation threshold to determine the real-time risk received data set and the risk monitoring point set, wherein each multi-source real-time received data corresponds to one monitoring point. Based on the set of risk monitoring points and the set of real-time risk received data, causal tracing is performed to determine tracing clue clusters, including: obtaining a set of historical risk logs for underground utility tunnels; constructing a set of causal tracing vectors based on the set of risk monitoring points and the set of real-time risk received data; performing approximate matching of the causal tracing vector set in the set of historical risk logs for underground utility tunnels to obtain matching historical risk log clusters for underground utility tunnels; and extracting tracing clues from each matching historical risk log set in the matching historical risk log clusters for underground utility tunnels to obtain the tracing clue clusters. Using the aforementioned traceability clue cluster as an index, the data transfer storage is used to extract clues and determine the traceability received data sequence cluster; Perform time-series iterative trend risk authentication on the data sequence clusters received for tracing the source, and determine the risk trend feature clusters; Based on the risk trend feature cluster, a first dynamic monitoring frequency and a second dynamic monitoring frequency are determined, and both the first dynamic monitoring frequency and the second dynamic monitoring frequency are distributed to the status monitoring terminal to perform dual-frequency synchronous status monitoring on the set of monitoring points. The determination of the first dynamic monitoring frequency and the second dynamic monitoring frequency based on the risk trend feature cluster includes: Risk scoring is performed based on the aforementioned risk trend feature clusters to obtain a risk score set; Based on the maximum and minimum values of the risk score set, dynamic monitoring frequency matching is performed to generate a first dynamic monitoring frequency and a second dynamic monitoring frequency. Specifically, the window constraint size of the data transfer storage is optimized and adjusted based on the maximum value in the risk score set.
2. The method for real-time dynamic monitoring of underground utility tunnel status based on multi-source data fusion as described in claim 1, characterized in that, For each set of historical risk logs of the matched underground utility tunnel in the historical risk log cluster, source tracing clues are extracted to obtain the source tracing clue cluster, including: Extract the first matching set of historical risk logs for underground utility tunnels from the matching cluster of historical risk logs for underground utility tunnels; Extract the set of associated monitoring points from the first set of historical risk logs of the matched underground utility tunnel; The set of associated monitoring points is subjected to union processing, and the first set of tracing clues is obtained based on the processing result; Tracing clues are extracted by traversing the historical risk log clusters of the matching underground utility tunnels to obtain the traceability clue clusters.
3. The method for real-time dynamic monitoring of underground utility tunnel status based on multi-source data fusion as described in claim 1, characterized in that, The time-series iterative trend risk authentication is performed on each of the traceability received data sequence clusters to determine risk trend feature clusters, including: Extract the first source-tracing received data sequence from the source-tracing received data sequence cluster; Perform time-series iterative correlation trend analysis on the first source-tracing received data sequence in chronological order to determine the trend characteristics of the first source-tracing received data; The risk trend identifier is invoked to perform risk authentication on the trend features of the first source tracing received data. If the authentication is successful, the trend features of the first source tracing received data are added to the risk trend feature cluster. The data sequence clusters received from the source are traversed for time-series iterative trend analysis, and the analysis results are verified for risk using a risk trend identifier to obtain the risk trend feature clusters.
4. The method for real-time dynamic monitoring of underground utility tunnel status based on multi-source data fusion as described in claim 3, characterized in that, Perform time-series iterative correlation trend analysis on the first source-tracing received data sequence in chronological order to determine the trend characteristics of the first source-tracing received data, including: Feature extraction is performed on the first source tracing received data sequence according to the time scale and the spatial scale to obtain the time trend features and spatial trend features of the first source tracing received data. Similarity identification is performed on the time trend features and spatial trend features of the first source tracing received data to determine the first trend feature similarity set; A first iterative correlation matrix is constructed based on the first trend feature similarity set, and the first iterative correlation matrix is used to perform feature convolution on the spatial trend features of the first source-received data to generate the trend features of the first source-received data.
5. The method for real-time dynamic monitoring of underground utility tunnel status based on multi-source data fusion as described in claim 4, characterized in that, The first trend feature similarity set is normalized, and the processing result is filled into an initially empty diagonal matrix to generate the first iterative correlation matrix.
6. A real-time dynamic monitoring system for underground utility tunnels based on multi-source data fusion, characterized in that: The system is used to execute the real-time dynamic monitoring method for underground utility tunnels based on multi-source data fusion as described in any one of claims 1-5, and the system includes: The monitoring module is used to acquire the set of monitoring points in the underground utility tunnel, monitor the set of monitoring points using a set of multi-source status monitoring terminals at a preset monitoring frequency, and transmit the monitoring data to the data transmission center. The data transmission center has a data transfer memory with window constraints. The risk identification module is used to perform real-time risk identification based on the data transmission center, and to determine the risk real-time received data set and the risk monitoring point set. This includes: extracting a multi-source real-time received data set from the data transmission center; using a multi-source standard state data set to identify deviations in the multi-source real-time received data set to determine a multi-source real-time received data deviation set; and performing risk identification on the multi-source real-time received data deviation set according to a preset risk deviation threshold to determine the risk real-time received data set and the risk monitoring point set, wherein each multi-source real-time received data corresponds to one monitoring point. The causal tracing module is used to perform causal tracing based on the risk monitoring point set and the real-time risk data reception set, and to determine the tracing clue clusters. This includes: obtaining a set of historical risk logs for underground utility tunnels; constructing a set of causal tracing vectors based on the risk monitoring point set and the real-time risk data reception set; performing approximate matching of the causal tracing vector set in the set of historical risk logs for underground utility tunnels to obtain matching historical risk log clusters for underground utility tunnels; and extracting tracing clues from each matching historical risk log set in the matching historical risk log clusters for underground utility tunnels to obtain the tracing clue clusters. The clue extraction module is used to extract clues from the data transfer storage using the traceability clue cluster as an index, and to determine the traceability received data sequence cluster; The risk authentication module is used to perform time-series iterative trend risk authentication on the traceability received data sequence clusters respectively, and determine the risk trend feature clusters; The dual-frequency synchronous monitoring module is used to determine the first dynamic monitoring frequency and the second dynamic monitoring frequency based on the risk trend feature cluster, and to distribute the first dynamic monitoring frequency and the second dynamic monitoring frequency to the status monitoring terminal to perform dual-frequency synchronous status monitoring on the set of monitoring points. The determination of the first dynamic monitoring frequency and the second dynamic monitoring frequency based on the risk trend feature cluster includes: Risk scoring is performed based on the aforementioned risk trend feature clusters to obtain a risk score set; Based on the maximum and minimum values of the risk score set, dynamic monitoring frequency matching is performed to generate a first dynamic monitoring frequency and a second dynamic monitoring frequency. Specifically, the window constraint size of the data transfer storage is optimized and adjusted based on the maximum value in the risk score set.