An Internet of Things-based dynamic risk prediction method and system for integrated pipe corridors

Abstract

The present invention provides an integrated pipe gallery dynamic risk prediction method and system based on the Internet of Things, which relates to the technical field of data processing. The method includes: acquiring historical integrated pipe gallery monitoring data; performing principal component analysis on the historical integrated pipe gallery monitoring data to determine the eigenvalues and eigenvectors of the historical integrated pipe gallery monitoring data; determining the main axis direction and the half-axis length of the main axis direction according to the eigenvectors and eigenvalues, and establishing a minimum ellipsoid describing the normal operation state of the integrated pipe gallery in combination with the chi-square test and the law of large numbers; sending the minimum ellipsoid to the edge side; collecting the current integrated pipe gallery monitoring data through a data acquisition terminal; uploading the current integrated pipe gallery monitoring data to the edge side, and determining the abnormal pipe gallery characteristics by means of the Shapley value in the case of abnormal current integrated pipe gallery monitoring data; outputting the abnormal current integrated pipe gallery monitoring data and the abnormal pipe gallery characteristics, and issuing a risk warning. The accuracy and timeliness of integrated pipe gallery risk prediction are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method and system for dynamically predicting the risks of an integrated utility tunnel based on the Internet of Things. Background Art

[0002] The Internet of Things refers to a technical system that connects devices, systems, and environments in the physical world through sensors, devices, and networks to achieve information collection, transmission, processing, and sharing. Through intelligent perception and interconnection, devices can operate automatically and interact in real time. An integrated utility tunnel refers to a centralized tunnel space built underground in a city for placing various municipal pipelines (such as electricity, communication, tap water, gas, heating, etc.) to achieve unified management and maintenance and avoid repeated excavation on the ground.

[0003] As a key infrastructure in the city, once problems such as pipeline leakage, collapse, and overload occur in the integrated utility tunnel, it will have a serious impact on urban operations such as water supply, power supply, and communication. Through risk prediction methods, potential problems in the operation of the utility tunnel can be discovered in advance, the probability of failures can be reduced, the safety and reliability of the utility tunnel can be improved, and at the same time, maintenance costs and interference to urban life can be reduced, ensuring the normal operation of the city and the safety of residents.

[0004] However, the existing risk prediction of integrated utility tunnels often uses fixed values to independently predict the risk of the operating factors of the integrated utility tunnel. If the value exceeds the fixed value, there is a risk; if it does not exceed, there is no risk. This risk prediction method does not consider the potential coupling relationships among the operating factors. The risk prediction method analyzed only through single factors and fixed thresholds cannot consider the associations and changing trends among multiple factors, has a low risk prediction accuracy, is prone to false alarms and missed reports, and cannot discover potential risks, affecting the safe operation of the integrated utility tunnel. Summary of the Invention

[0005] In order to solve the technical problem that the existing risk prediction of integrated utility tunnels often uses fixed values to independently predict the risk of the operating factors of the integrated utility tunnel. If the value exceeds the fixed value, there is a risk; if it does not exceed, there is no risk. This risk prediction method does not consider the potential coupling relationships among the operating factors. The risk prediction method analyzed only through single factors and fixed thresholds cannot consider the associations and changing trends among multiple factors, has a low risk prediction accuracy, is prone to false alarms and missed reports, and cannot discover potential risks, affecting the safe operation of the integrated utility tunnel, the present invention provides a method and system for dynamically predicting the risks of an integrated utility tunnel based on the Internet of Things.

[0006] The technical solutions provided by the embodiments of the present invention are as follows:

[0007] In the first aspect

[0008] A method for dynamic risk prediction of an integrated utility tunnel based on the Internet of Things provided by an embodiment of the present invention is applied to a risk prediction platform of the integrated utility tunnel. The risk prediction platform includes a data acquisition terminal, an edge terminal, and a cloud end that are connected in sequence. The data acquisition terminal is connected to the cloud end. The method includes:

[0009] S1: Obtain historical integrated utility tunnel monitoring data including different tunnel characteristics through the cloud end;

[0010] S2: Perform principal component analysis on the historical integrated utility tunnel monitoring data to determine the eigenvalues and eigenvectors of the historical integrated utility tunnel monitoring data;

[0011] S3: Determine the main axis direction and the half-axis length of the main axis direction according to the eigenvectors and eigenvalues, and establish a minimum ellipsoid describing the normal operation state of the integrated utility tunnel in combination with the chi-square test and the law of large numbers;

[0012] S4: Send the minimum ellipsoid to the edge terminal;

[0013] S5: Collect current integrated utility tunnel monitoring data through the data acquisition terminal;

[0014] S6: Upload the current integrated utility tunnel monitoring data to the edge terminal, and determine whether the current integrated utility tunnel monitoring data is abnormal in combination with the minimum ellipsoid and the half-axis length of the main axis direction. If so, enter step S7; otherwise, return to step S5;

[0015] S7: Determine the abnormal tunnel characteristics in the abnormal current integrated utility tunnel monitoring data through the Shapley value;

[0016] S8: Output the abnormal current integrated utility tunnel monitoring data and the abnormal tunnel characteristics, and issue a risk warning.

[0017] In a second aspect

[0018] A dynamic risk prediction system for an integrated utility tunnel based on the Internet of Things provided by an embodiment of the present invention includes:

[0019] A processor;

[0020] A memory, on which computer-readable instructions are stored. When the computer-readable instructions are executed by the processor, the dynamic risk prediction method for the integrated utility tunnel based on the Internet of Things as described in the first aspect is implemented.

[0021] In a third aspect

[0022] A computer-readable storage medium provided by an embodiment of the present invention has a computer program stored thereon. When the program is executed by the processor, the dynamic risk prediction method for the integrated utility tunnel based on the Internet of Things as described in the first aspect is implemented.

[0023] The beneficial effects brought by the technical solutions provided by the embodiments of the present invention at least include:

[0024] (1) In the present invention, principal component analysis is performed on the historical monitoring data of the comprehensive utility tunnel to determine the eigenvalues and eigenvectors of the historical monitoring data of the comprehensive utility tunnel. The main axis direction and the semi-axis length of the main axis direction are determined based on the eigenvectors and eigenvalues, and the minimum ellipsoid describing the normal operation state of the comprehensive utility tunnel is established by combining the chi-square test and the law of large numbers. Key features are extracted through principal component analysis, and the model is dynamically adjusted by combining the chi-square test and the law of large numbers to comprehensively capture the data distribution characteristics of the normal tunnel data, thereby reversely improving the accuracy and adaptability of risk prediction, effectively reducing false alarms and missed alarms, adapting to the changes in the complex tunnel environment, and ensuring operation safety. Then, the current monitoring data of the comprehensive utility tunnel is used as a whole, that is, associated data, to determine whether it is abnormal by using the minimum ellipsoid and the semi-axis length of the main axis direction, rather than making separate and independent judgments. If the current monitoring data of the comprehensive utility tunnel is not located within the minimum ellipsoid, it is directly determined to be abnormal. If it is located within the minimum ellipsoid, the semi-axis length of the main axis direction is further used to determine and verify whether it is abnormal, which can quickly process a large amount of data, with low latency and high risk prediction accuracy. Finally, the Shapley value is used to accurately quantify the contribution of each feature to the abnormality, which can accurately locate the abnormal tunnel features, improve the pertinence and interpretability of risk prediction, effectively reduce the false alarm range at the same time, provide a clear direction for subsequent risk intervention, and further enhance the efficiency and accuracy of the dynamic risk prediction of the comprehensive utility tunnel.

[0025] (2) In the present invention, an Internet of Things risk prediction platform including a data acquisition terminal, an edge side, and a cloud side is also combined. The data processing ability near the edge is fully utilized to improve the prediction real-time performance, and the powerful data computing ability of the cloud side is used to accurately and quickly establish the minimum ellipsoid describing the normal operation state of the comprehensive utility tunnel, further improving the risk prediction accuracy and real-time performance of the comprehensive utility tunnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0027] Figure 1 It is a schematic flow chart of a method for dynamic risk prediction of a comprehensive utility tunnel based on the Internet of Things provided by an embodiment of the present invention;

[0028] Figure 2 It is a schematic structural diagram of a risk prediction platform provided by an embodiment of the present invention;

[0029] Figure 3Schematic diagram of a dynamic risk prediction system for an integrated utility tunnel based on the Internet of Things provided by an embodiment of the present invention. Detailed implementation manners

[0030] The technical solutions in the present invention will be described below with reference to the accompanying drawings.

[0031] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as an "example" in the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of the word "example" is intended to present concepts in a specific manner. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" may be both, or either one of the two.

[0032] To make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0033] Refer to the attached specification Figure 1 , which shows a schematic flowchart of a dynamic risk prediction method for an integrated utility tunnel based on the Internet of Things provided by an embodiment of the present invention.

[0034] Refer to the attached specification Figure 2 , which shows a schematic diagram of the structure of a risk prediction platform provided by an embodiment of the present invention.

[0035] An embodiment of the present invention provides a dynamic risk prediction method for an integrated utility tunnel based on the Internet of Things, which is applied to a risk prediction platform for an integrated utility tunnel. The risk prediction platform includes a data acquisition terminal, an edge side and a cloud end connected in sequence, and the data acquisition terminal is connected to the cloud end.

[0036] Among them, the data acquisition terminal is composed of multi-modal sensors deployed in the utility tunnel, and is responsible for collecting monitoring data such as the temperature, humidity, and vibration of the utility tunnel in real time. These terminal devices upload the data to the edge side or the cloud end through the network. The edge side is deployed on a local computing device close to the data acquisition terminal, and is responsible for quickly processing and preliminarily analyzing the data, reducing the data transmission delay, and is suitable for scenarios that require real-time feedback. The cloud end is used as the core computing and storage platform, and is responsible for high-complexity model training, risk prediction and storage analysis of global data. It is connected to the edge side to achieve overall data management and centralized analysis. The dynamic risk prediction platform for the integrated utility tunnel based on the Internet of Things collects real-time utility tunnel data through the data acquisition terminal, the edge side quickly analyzes and reduces the transmission delay, and the cloud end is responsible for comprehensive storage and complex model calculation, realizing an efficient and multi-level dynamic risk prediction system.

[0037] The processing flow of the dynamic risk prediction method for the integrated utility tunnel based on the Internet of Things may include the following steps:

[0038] S1: Obtain historical integrated utility tunnel monitoring data including different tunnel characteristics from the cloud.

[0039] It can be understood that obtaining historical integrated utility tunnel monitoring data including different tunnel characteristics from the cloud aims to comprehensively collect multi-dimensional information on the operation status of the tunnel, providing a data basis for subsequent analysis. These characteristics include key indicators such as temperature, humidity, gas concentration, etc., which can reflect the environmental and equipment operation conditions of the tunnel. Through comprehensive data collection, the accuracy and reliability of the risk prediction model are ensured.

[0040] In a possible implementation, the tunnel characteristics include temperature, humidity, gas concentration, light intensity, pressure value, vibration amount, displacement amount, cable voltage, and cable current.

[0041] It should be noted that through real-time monitoring and analysis of these data, the operation status and potential abnormal changes of the tunnel can be accurately reflected, providing key multi-dimensional data support for dynamic risk prediction, thereby improving the safety and prediction accuracy of the system.

[0042] In a possible implementation, after S1, it further includes:

[0043] Perform preprocessing on the historical integrated utility tunnel monitoring data, including data cleaning and data normalization.

[0044] It should be noted that by performing data cleaning and normalization preprocessing on the historical integrated utility tunnel monitoring data, outliers can be removed, missing data can be filled, and the dimension can be unified, ensuring data quality and comparability between features, thereby providing more accurate and standardized input for subsequent analysis and modeling, and improving the reliability and effectiveness of risk prediction.

[0045] S2: Perform principal component analysis on the historical integrated utility tunnel monitoring data to determine the eigenvalues and eigenvectors of the historical integrated utility tunnel monitoring data.

[0046] Among them, principal component analysis (PCA, Principal Component Analysis) is a dimensionality reduction technique used to extract the direction with the largest variance in the data through linear transformation, map high-dimensional data to a low-dimensional space, and retain the main information of the original data as much as possible. It determines the main direction (eigenvector) and its importance (eigenvalue) of the data through eigenvalue decomposition or singular value decomposition, thereby reducing redundant information and revealing the potential structure between variables. By extracting the main features from the historical monitoring data through principal component analysis, determining the eigenvalues and eigenvectors, and capturing the coupling relationship between key features, the data complexity is reduced, providing a more accurate and concise input basis for the subsequent risk prediction model.

[0047] In a possible implementation, S2 specifically includes:

[0048] S201: Establish a data matrix of historical monitoring data of the integrated utility tunnel:

[0049]

[0050] where X represents the data matrix, represents the j-th utility tunnel eigenvalue of the i-th sample data, , , n represents the total number of sample data, and d represents the total number of utility tunnel characteristics.

[0051] S202: Solve the covariance matrix of the data matrix:

[0052]

[0053] where, represents the covariance matrix, x represents the elements in the data matrix, represents the mean value of the utility tunnel characteristics, that is, the mean value of the column elements of the data matrix. The superscript T represents transpose, represents the probability distribution of, W represents the diagonal weight matrix, represents the weight of the j-th utility tunnel characteristic, diag represents the diagonal elements, and log represents the logarithmic function.

[0054] It should be noted that by introducing information entropy to calculate the weights and dynamically adjusting the characteristic weights of the covariance matrix, the characteristics with large amounts of information will have more influence in the eigenvalue decomposition. Ultimately, the optimized covariance matrix can more accurately reflect the data distribution, which helps to improve the effect of principal component analysis. If the value range of a certain characteristic is very large (such as showing a uniform distribution), its information entropy is high, indicating that this characteristic contains more information. If the value range of a certain characteristic is small or concentrated (such as a constant value), its information entropy is low, indicating that the amount of information of this characteristic is small or it is a redundant characteristic. The characteristics with higher weights (characteristics with large amounts of information) occupy more important positions in the covariance matrix. The characteristics with lower weights (characteristics with little information or noise characteristics) are weakened, thereby improving the accuracy of principal component decomposition.

[0055] S203: Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues and eigenvectors:

[0056]

[0057] where, represents the eigenvector matrix, represents the eigenvalue matrix, and The eigenvector and eigenvalue representing the j-th utility tunnel feature in a one-to-one correspondence.

[0058] Specifically, first, by establishing a data matrix of historical utility tunnel monitoring data, the system structurally represents the multi-dimensional feature data in the sample. Then, by calculating the covariance matrix of the data matrix, the correlation and distribution characteristics among the features are analyzed. To improve the accuracy of the covariance matrix, weights based on information entropy are introduced. By calculating the information entropy of each feature, the weights of the features are dynamically adjusted. Features with high information entropy represent rich information and contribute more significantly to the covariance matrix, while the influence of features with low information entropy or noise features is weakened, thus optimizing the reflection of the covariance matrix on the data distribution. Finally, by performing eigenvalue decomposition on the covariance matrix, an eigenvalue matrix and an eigenvector matrix are obtained. The eigenvalues reflect the importance of the data in the direction of the corresponding eigenvectors. These decomposition results are used to extract the key features of the data, reduce the dimension, and provide a mathematical basis for subsequent modeling. This process not only improves the accuracy and efficiency of data analysis, accurately extracts key features and eliminates redundant information, thereby improving the accuracy and efficiency of risk prediction, but also dynamically adapts to changes in the importance of different features.

[0059] S3: Determine the main axis direction and the half-axis length of the main axis direction according to the eigenvector and eigenvalue, and establish a minimum ellipsoid describing the normal operation state of the utility tunnel in combination with the chi-square test and the law of large numbers.

[0060] Among them, the main axis direction is defined by the eigenvector of the principal component analysis, representing the direction in which the data changes the most in a certain feature dimension. Each eigenvector corresponds to a main axis direction, pointing to the main trend of the data distribution. The half-axis length is determined by the square root of the eigenvalue, representing the distribution range of the data in the corresponding main axis direction. The larger the eigenvalue, the longer the corresponding half-axis, and the more significant the change of the data. The chi-square test is used to evaluate whether the data deviates from the expected distribution. In this application, it is used to judge whether the data points are within the range of the normal distribution, and the boundaries are set through the chi-square distribution to define the normal and abnormal. The law of large numbers indicates that when the sample size tends to infinity, the sample mean approaches the population mean. In this method, the law of large numbers is used to dynamically adjust the boundary range of the model to adapt to changes in the data scale. The minimum ellipsoid is a geometric model used to enclose the vast majority of points in the dataset. Its center is the mean of the data, the half-axis direction is defined by the eigenvector, and the half-axis length is determined by the eigenvalue, which is used to describe the normal distribution range of the data.

[0061] It should be noted that by determining the main axis direction through the eigenvector, determining the half-axis length through the eigenvalue, establishing the minimum ellipsoid in combination with the chi-square test and the law of large numbers, dynamically adjusting the boundary to adapt to the data distribution, it is used to accurately describe the normal operation state of the utility tunnel, providing a reliable reference standard for anomaly detection, thereby significantly improving the accuracy and dynamics of risk prediction.

[0062] In a possible implementation, the minimum ellipsoid is specifically:

[0063]

[0064] where represents the mean value of the utility tunnel features, that is, the mean value of the column elements of the data matrix, which is the center of the minimum ellipsoid, represents the semi-axis length of the major axis corresponding to in the minimum ellipsoid, R represents the maximum radius describing the boundary range of the minimum ellipsoid, represents the significance level, represents the chi-square distribution with a degree of freedom of d and a confidence level of 1 - , represents the sample data variance.

[0065] Specifically, by establishing a minimum ellipsoid that is dynamically adjusted based on the chi-square distribution and the sample variance, the normal operating state range of the utility tunnel can be accurately described, while dynamically adapting to changes in the data distribution, avoiding false alarms and missed alarms, thereby significantly improving the accuracy and reliability of dynamic risk prediction.

[0066] Optionally, the significance level can be 0.05 or 0.01.

[0067] Among them, the determination of R combines the law of large numbers, and the purpose is to dynamically relax or tighten the boundary. Specifically, when n is very large, will approach 0, and the adjustment factor tends to 1. In this case, the boundary is close to the classical chi-square distribution value . When n is very small, is larger, and the adjustment factor is significantly greater than 1. This will cause to be amplified, thereby increasing the boundary range and better accommodating normal points. This is to prevent too many misjudgments (i.e., normal points are judged as abnormal points) when the sample size is too small and the boundary range is too small. In addition, when the variance of the data is large (i.e., the data distribution is more dispersed), is larger, and the adjustment factor will also increase. This can appropriately relax the boundary and avoid excluding too many normal points. When the variance of the data is small, is smaller, the adjustment factor tends to 1, and the boundary range will contract accordingly, more strictly defining anomalies. In addition, when the variance of the data is large (i.e., the data distribution is more dispersed), is larger, and the adjustment factor will also increase. This can appropriately relax the boundary and avoid excluding too many normal points. When the variance of the data is small, is smaller, the adjustment factor tends to 1, and the boundary range will contract accordingly, more strictly defining anomalies.

[0068] S4: Send the minimum ellipsoid to the edge device.

[0069] It should be noted that sending the minimum ellipsoid model constructed in the cloud to the edge device enables the edge device to quickly perform anomaly judgment on real-time monitoring data locally, reducing data transmission latency and computational pressure, and achieving efficient real-time risk prediction.

[0070] S5: Collect the current integrated pipe gallery monitoring data through the data acquisition terminal.

[0071] Among them, the current integrated pipe gallery monitoring data is the environmental and equipment operation status information collected in real time by sensors, reflecting the operation status of the pipe gallery at the current moment. The current integrated pipe gallery monitoring data refers to the real-time collected and to-be-detected operation data of the pipe gallery, and these data need to be judged whether they are abnormal by comparing with the normal state (described by the minimum ellipsoid). This is the basic input data for dynamic detection and risk prediction.

[0072] S6: Upload the current integrated pipe gallery monitoring data to the edge device, and judge whether the current integrated pipe gallery monitoring data is abnormal by combining the minimum ellipsoid and the semi-axis lengths of the main axis directions. If so, go to step S7; otherwise, return to step S5.

[0073] It should be noted that by uploading the current to-be-detected integrated pipe gallery monitoring data to the edge device, the minimum ellipsoid model and its semi-axis lengths of the main axis directions are used to judge the abnormality of the data. Combining the minimum ellipsoid, it can be overall judged whether the data conforms to the multi-dimensional distribution characteristics of normal operation. Combining the semi-axis lengths of the main axis directions, it can be further verified whether the data in each direction exceeds the normal range. This method combines global judgment and local verification, which can not only quickly identify significant anomalies but also reduce the false positive rate, comprehensively improving the accuracy and robustness of risk prediction.

[0074] In a possible implementation, judging whether the current integrated pipe gallery monitoring data is abnormal by the minimum ellipsoid in S6 specifically includes:

[0075] S601: Judge whether the current integrated pipe gallery monitoring data is located within the minimum ellipsoid, that is, whether it satisfies the following formula. If so, go to step S602; otherwise, go to step S605:

[0076]

[0077] Among them, represents the current integrated pipe gallery monitoring data.

[0078] S602: Calculate the projection values of the current integrated pipe gallery monitoring data in the main axis directions of the minimum ellipsoid:

[0079]

[0080] Among them, represents the current monitoring data of the utility tunnel in the main axis direction of the projection value.

[0081] S603: Determine whether each projection value is less than the semi - major axis length of the maximum main axis direction of the minimum ellipsoid, that is, whether it satisfies the following formula. If it satisfies, go to step S604; otherwise, go to step S605:

[0082]

[0083] Among them, represents the semi - major axis length of the maximum main axis direction of the minimum ellipsoid.

[0084] S604: Output the current monitoring data of the utility tunnel as normal data.

[0085] S605: Output the current monitoring data of the utility tunnel as abnormal data.

[0086] It should be noted that through the hierarchical judgment method based on the minimum ellipsoid, the dynamic risk prediction of the utility tunnel can gradually screen the abnormality of the monitoring data. First, it judges whether the overall data falls within the minimum ellipsoid, and then uses the projection value in the main axis direction to further accurately verify the abnormality. This method combines global and local judgments, avoids misjudgment of a single standard, and improves the accuracy of prediction. At the same time, through hierarchical screening, it reduces the computational requirements for anomaly analysis, significantly reduces the diagnostic delay and computational resource consumption while ensuring the detection accuracy, enabling it to efficiently handle the complex monitoring data environment of the utility tunnel.

[0087] S7: Determine the abnormal utility tunnel characteristics in the current abnormal monitoring data of the utility tunnel through the Shapley value.

[0088] Among them, the Shapley Value is an index based on cooperative game theory, used to measure the contribution of each feature to the overall result. In the risk prediction of the utility tunnel, the Shapley Value quantifies the abnormal contribution degree of each feature by calculating the influence of each feature on the anomaly detection result, helping to accurately locate the source of the problem. Using the Shapley Value to calculate the contribution of each utility tunnel feature to the anomaly detection result, accurately locate the most abnormal feature, provide a clear basis for risk tracing, and further improve the pertinence and efficiency of risk prediction.

[0089] In a possible implementation, S7 specifically includes:

[0090] S701: Calculate the Shapley value of each utility tunnel feature:

[0091]

[0092] Among them, represents the Shapley value of the j-th utility tunnel feature, N represents the feature set including all utility tunnel features, represents the mean value of the utility tunnel features of the j-th utility tunnel feature, represents the feature set excluding the utility tunnel feature j, represents belonging to the feature subset of represents the number of features in S, and! represents the factorial, represents the feature subset including the utility tunnel feature j, and respectively represent the score index values for different utility tunnel features.

[0093] S702: Output the utility tunnel feature corresponding to the maximum Shapley value as the abnormal utility tunnel feature:

[0094]

[0095] Among them, represents the utility tunnel feature j when taking the maximum Shapley value, represents the abnormal utility tunnel feature index.

[0096] It should be noted that by calculating the Shapley values of each utility tunnel feature and outputting the abnormal feature with the highest contribution, the comprehensive utility tunnel dynamic risk prediction method can accurately quantify the contribution of each feature to the abnormality and accurately locate the key features causing the abnormality. This method improves the interpretability and pertinence of anomaly detection, helps reduce the false alarm range and provides a clear intervention direction, thereby greatly improving the accuracy and practicality of risk prediction, and at the same time providing data support for subsequent optimization and management.

[0097] S8: Output the current comprehensive utility tunnel monitoring data of the abnormality and the abnormal utility tunnel feature, and issue a risk warning.

[0098] In a possible implementation manner, after S8, it further includes:

[0099] Update the minimum ellipsoid at preset time intervals.

[0100] It can be understood that updating the minimum ellipsoid at preset time intervals can enable the model to dynamically adapt to the long-term changes of utility tunnel monitoring data, ensure the accuracy and applicability of the prediction results, thereby enhancing the sensitivity and robustness of the system to environmental changes, and effectively reducing misjudgments or missed judgments caused by data distribution drift.

[0101] It should be noted that those skilled in the art can set the size of the preset time interval according to actual needs, and the present invention does not limit it here.

[0102] In the actual application process, first, historical monitoring data of the integrated utility tunnel is collected, and multi-dimensional features such as temperature, humidity, and vibration are extracted to provide a data basis for analysis. Principal component analysis is used to reduce the dimension of historical data, extract eigenvalues and eigenvectors, and capture the correlations between key features. A minimum ellipsoid model is constructed based on the eigenvalues and eigenvectors, and the boundary is dynamically adjusted by combining the chi-square test and the law of large numbers to accurately describe the normal operating state of the utility tunnel. This model is sent to the edge side to support local real-time anomaly detection. The data of the utility tunnel to be detected is collected in real time as the input for dynamic risk detection. The current monitoring data is uploaded to the edge side, and it is judged whether the data is abnormal by combining the minimum ellipsoid model and the semi-axis length, and potential risks are quickly identified. If an anomaly is detected, the contribution degree of the abnormal feature is calculated using the Shapley value to accurately locate the problem feature. Finally, the abnormal data and features are output and a warning is issued. Through this process, real-time, efficient, and accurate dynamic risk prediction is achieved.

[0103] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include:

[0104] (1) In the present invention, principal component analysis is performed on the historical monitoring data of the integrated utility tunnel to determine the eigenvalues and eigenvectors of the historical monitoring data of the integrated utility tunnel. The main axis direction and the semi-axis length of the main axis direction are determined according to the eigenvectors and eigenvalues, and a minimum ellipsoid describing the normal operating state of the integrated utility tunnel is established by combining the chi-square test and the law of large numbers. Key features are extracted through principal component analysis, and the model is dynamically adjusted by combining the chi-square test and the law of large numbers to comprehensively capture the data distribution characteristics of normal utility tunnel data, thereby reversely improving the accuracy and adaptability of risk prediction, effectively reducing false alarms and missed alarms, adapting to changes in the complex utility tunnel environment, and ensuring operating safety. Then, the current monitoring data of the integrated utility tunnel is used as a whole, that is, associated data, to judge whether it is abnormal by using the minimum ellipsoid and the semi-axis length of the main axis direction, rather than making separate independent judgments. If the current monitoring data of the integrated utility tunnel is directly determined to be abnormal when it is not located in the minimum ellipsoid, and if it is located in the minimum ellipsoid, the semi-axis length of the main axis direction is used to further judge and verify whether it is abnormal. It can quickly process a large amount of data, with low latency and high risk prediction accuracy. Finally, the Shapley value is used to accurately quantify the contribution of each feature to the anomaly, which can accurately locate the abnormal utility tunnel features, improve the pertinence and interpretability of risk prediction, effectively reduce the false alarm range at the same time, provide a clear direction for subsequent risk intervention, and further enhance the efficiency and accuracy of dynamic risk prediction of the integrated utility tunnel.

[0105] (2) In the present invention, an Internet of Things risk prediction platform including a data acquisition terminal, an edge side, and a cloud side is also combined. The data processing ability of the edge side is fully utilized to improve the real-time performance of prediction, and the powerful data computing ability of the cloud side is used to accurately and quickly establish a minimum ellipsoid describing the normal operating state of the integrated utility tunnel, further improving the risk prediction accuracy and real-time performance of the integrated utility tunnel.

[0106] Refer to the attached instruction manual Figure 3 , which shows a schematic structural diagram of an integrated pipe gallery dynamic risk prediction system provided by the present invention based on the Internet of Things.

[0107] The present invention also provides an integrated pipe gallery dynamic risk prediction system 20 based on the Internet of Things, which is applied to the above-mentioned integrated pipe gallery dynamic risk prediction method based on the Internet of Things, and includes:

[0108] A processor 201.

[0109] A memory 202, on which computer-readable instructions are stored. When the computer-readable instructions are executed by the processor 201, the integrated pipe gallery dynamic risk prediction method based on the Internet of Things as in the method embodiment is realized.

[0110] The integrated pipe gallery dynamic risk prediction system 20 provided by the present invention can execute the above-mentioned integrated pipe gallery dynamic risk prediction method based on the Internet of Things and achieve the same or similar technical effects. To avoid repetition, the present invention will not elaborate further.

[0111] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include:

[0112] (1) In the present invention, principal component analysis is performed on historical integrated pipe gallery monitoring data to determine the eigenvalues and eigenvectors of the historical integrated pipe gallery monitoring data. According to the eigenvectors and eigenvalues, the main axis direction and the half-axis length of the main axis direction are determined, and a minimum ellipsoid describing the normal operation state of the integrated pipe gallery is established in combination with the chi-square test and the law of large numbers. Key features are extracted through principal component analysis, and the model is dynamically adjusted in combination with the chi-square test and the law of large numbers to comprehensively capture the data distribution characteristics of normal pipe gallery data, thereby reversely improving the accuracy and adaptability of risk prediction, effectively reducing false alarms and missed reports, adapting to changes in complex pipe gallery environments, and ensuring operation safety. Then, the current integrated pipe gallery monitoring data is used as a whole, that is, associated data, to determine whether it is abnormal by using the minimum ellipsoid and the half-axis length of the main axis direction, rather than making separate independent judgments. If the current integrated pipe gallery monitoring data is not located in the minimum ellipsoid, it is directly determined to be abnormal. If it is located in the minimum ellipsoid, the half-axis length of the main axis direction is used to further judge and verify whether it is abnormal, which can quickly process a large amount of data, with low latency and high risk prediction accuracy. Finally, the Shapley value is used to accurately quantify the contribution of each feature to the abnormality, which can accurately locate the abnormal pipe gallery features, improve the pertinence and interpretability of risk prediction, effectively reduce the false alarm range at the same time, provide a clear direction for subsequent risk intervention, and further enhance the efficiency and accuracy of integrated pipe gallery dynamic risk prediction.

[0113] (2) In the present invention, an Internet of Things risk prediction platform including a data acquisition terminal, an edge side, and a cloud side is also incorporated. The data near-processing ability of the edge side is fully utilized to improve the prediction real-time performance, and the powerful data computing ability of the cloud side is used to accurately and quickly establish the minimum ellipsoid describing the normal operation state of the integrated pipe gallery, further improving the risk prediction accuracy and real-time performance of the integrated pipe gallery.

[0114] It should be understood that the processor in the embodiment of the present invention may be a central processing unit (CPU), and this processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or this processor may also be any conventional processor, etc.

[0115] It should also be understood that the memory in the embodiment of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchlink DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0116] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that the computer can access, or a data storage device such as a server or data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, or magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0117] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context.

[0118] In the present invention, "at least one" means one or more, and "a plurality" means two or more. "At least one of the following" or its similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple.

[0119] It should be understood that in various embodiments of the present invention, the magnitudes of the sequence numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0120] Those of ordinary skill in the art will realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0121] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the devices, apparatuses, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0122] In several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0123] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0124] In addition, the functional units in various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0125] When the above-described function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or a part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0126] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the method for dynamically predicting risks of an integrated pipe gallery based on the Internet of Things as described in the method embodiment.

[0127] A computer-readable storage medium provided by the present invention can implement the steps and effects of the method for dynamically predicting risks of an integrated pipe gallery based on the Internet of Things in the above method embodiment. To avoid repetition, the present invention will not elaborate further.

[0128] The beneficial effects brought by the technical solution provided by the embodiments of the present invention at least include:

[0129] (1) In the present invention, principal component analysis is performed on the historical monitoring data of the integrated utility tunnel to determine the eigenvalues and eigenvectors of the historical monitoring data of the integrated utility tunnel. The principal axis direction and the half-length of the principal axis direction semi-axis are determined based on the eigenvectors and the eigenvalues. A minimum ellipsoid describing the normal operation state of the integrated utility tunnel is established by combining the chi-square test and the law of large numbers. Key features are extracted through principal component analysis, and the model is dynamically adjusted by combining the chi-square test and the law of large numbers to comprehensively capture the data distribution characteristics of the normal tunnel data. Furthermore, the accuracy and adaptability of risk prediction are improved in reverse, effectively reducing false alarms and missed alarms, adapting to the changes in the complex tunnel environment, and ensuring operation safety. Then, the current monitoring data of the integrated utility tunnel is used as a whole, that is, associated data, to determine whether it is abnormal by using the minimum ellipsoid and the half-length of the principal axis direction semi-axis, rather than making separate independent judgments. If the current monitoring data of the integrated utility tunnel is not located within the minimum ellipsoid, it is directly determined to be abnormal. If it is located within the minimum ellipsoid, the half-length of the principal axis direction semi-axis is used to further judge and verify whether it is abnormal, which can quickly process a large amount of data, with low latency and high accuracy of risk prediction. Finally, the Shapley value is used to accurately quantify the contribution of each feature to the abnormality, which can accurately locate the abnormal tunnel features, improve the pertinence and interpretability of risk prediction, effectively reduce the false alarm range at the same time, provide a clear direction for subsequent risk intervention, and further enhance the efficiency and accuracy of the dynamic risk prediction of the integrated utility tunnel.

[0130] (2) In the present invention, an Internet of Things risk prediction platform including a data acquisition terminal, an edge side, and a cloud side is also combined. The data processing ability near the edge is fully utilized to improve the prediction real-time performance, and the powerful data computing ability of the cloud side is used to accurately and quickly establish a minimum ellipsoid describing the normal operation state of the integrated utility tunnel, further improving the risk prediction accuracy and real-time performance of the integrated utility tunnel.

[0131] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claimed rights.

[0132] The following points need to be explained:

[0133] (1) The attached drawings of the embodiments of the present invention only relate to the structures involved in the embodiments of the present invention, and other structures can refer to the general design.

[0134] (2) For clarity, in the drawings used to describe the embodiments of the present invention, the thickness of layers or regions is enlarged or reduced, that is, these drawings are not drawn to actual scale. It can be understood that when an element such as a layer, film, region, or substrate is referred to as being "on" or "under" another element, the element can be "directly" on or under the other element or there can be intervening elements.

[0135] (3) Without conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0136] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. An integrated pipe gallery dynamic risk prediction method based on the Internet of Things, characterized in that, A risk prediction platform applied to the utility tunnel, the risk prediction platform including a data acquisition terminal, an edge terminal, and a cloud end that are connected in sequence, the data acquisition terminal being connected to the cloud end; the method includes: S1: Obtain historical utility tunnel monitoring data including different tunnel characteristics through the cloud end; S2: Perform principal component analysis on the historical utility tunnel monitoring data to determine the eigenvalues and eigenvectors of the historical utility tunnel monitoring data; S3: Determine the main axis direction and the half-axis length of the main axis direction according to the eigenvectors and the eigenvalues, and establish a minimum ellipsoid describing the normal operation state of the utility tunnel in combination with the chi-square test and the law of large numbers; S4: Send the minimum ellipsoid to the edge terminal; S5: Collect current utility tunnel monitoring data through the data acquisition terminal; S6: Upload the current utility tunnel monitoring data to the edge terminal, and determine whether the current utility tunnel monitoring data is abnormal in combination with the minimum ellipsoid and the half-axis length of the main axis direction. If so, enter step S7; otherwise, return to step S5; S7: Determine the abnormal tunnel characteristics in the abnormal current utility tunnel monitoring data through the Shapley value; S8: Output the abnormal current utility tunnel monitoring data and the abnormal tunnel characteristics, and issue a risk warning; Among them, the S2 specifically includes: S201: Establish a data matrix of the historical utility tunnel monitoring data: ； Among them, X represents the data matrix, represents the i th j pipe gallery eigenvalue of the th sample data, n represents the total number of sample data, d represents the total number of pipe gallery features; S202: Solve the covariance matrix of the data matrix: ； Among them, represents the covariance matrix, x represents the elements in the data matrix, represents the mean of the utility tunnel features, i.e., the mean of the column elements of the data matrix. The subscript T represents the transpose, represents the probability distribution of W represents the diagonal weight matrix, represents the j weight of the i-th utility tunnel feature. diag represents the diagonal elements, and log represents the logarithmic function; S203: Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalues and the eigenvectors: ； Among them, represents the eigenvector matrix, represents the eigenvalue matrix, and represent the eigenvector and eigenvalue of the j th utility tunnel feature in one-to-one correspondence; Among them, the minimum ellipsoid is specifically: ； Among them, represents the mean value of the utility tunnel features, that is, the mean value of the column elements of the data matrix, which is the center of the minimum ellipsoid, represents the semi-axis length of the corresponding major axis direction in the minimum ellipsoid, R represents the maximum radius describing the boundary range of the minimum ellipsoid, represents the significance level, represents based on the degree of freedom d and the confidence level is 1 - chi-square distribution, represents the sample data variance; Among them, the determination of whether the current utility tunnel monitoring data is abnormal through the minimum ellipsoid in S6 specifically includes: S601: Determine whether the current utility tunnel monitoring data is located within the minimum ellipsoid, that is, whether it satisfies the following formula. If it satisfies, enter step S602; otherwise, enter step S605: ； Among them, represents the current monitoring data of the utility tunnel; S602: Calculate the projection values of the current utility tunnel monitoring data in the respective main axis directions of the minimum ellipsoid: ； Among them, represents the current monitoring data of the utility tunnel in the main axis direction of the projected value; S603: Determine whether each projection value is less than the maximum half-axis length of the main axis direction of the minimum ellipsoid, that is, whether it satisfies the following formula. If it satisfies, enter step S604; otherwise, enter step S605: ； Among them, represents the semi-major axis length of the maximum principal axis direction of the minimum ellipsoid; S604: Output the current utility tunnel monitoring data as normal data; S605: Output the current utility tunnel monitoring data as abnormal data.

2. The integrated pipe gallery dynamic risk prediction method based on the Internet of Things according to claim 1, characterized in that, The tunnel characteristics include temperature, humidity, gas concentration, light intensity, pressure value, vibration amount, displacement amount, cable voltage, and cable current.

3. The integrated pipe gallery dynamic risk prediction method based on the Internet of Things according to claim 1, characterized in that After the S1, it further includes: Perform preprocessing on the historical utility tunnel monitoring data including data cleaning and data normalization.

4. The integrated pipe gallery dynamic risk prediction method based on the Internet of Things according to claim 1, characterized in that The S7 specifically includes: S701: Calculate the Shapley value of each tunnel characteristic: ； Among them, represents the Shapley value of the j th utility tunnel feature, N represents the feature set including all utility tunnel features, represents the average value of the utility tunnel features of the j th utility tunnel feature, represents the feature set excluding the utility tunnel feature j , represents the feature subset belonging to , represents S the number of features,! represents factorial, represents the feature subset including the utility tunnel feature j , and respectively represent the score index values for different utility tunnel features; S702: Output the tunnel characteristic corresponding to the maximum Shapley value as the abnormal tunnel characteristic: ； Among them, represents the utility tunnel feature when the Shapley value is the largest j , represents the abnormal utility tunnel feature index.

5. The integrated pipe gallery dynamic risk prediction method based on the Internet of Things according to claim 1, wherein After the S8, it further includes: Update the minimum ellipsoid at preset time intervals.

6. An integrated pipe gallery dynamic risk prediction system based on the Internet of Things, characterized in that, Includes: Processor; A memory, on which computer-readable instructions are stored, and when the computer-readable instructions are executed by the processor, the method for dynamically predicting risks of an integrated utility tunnel based on the Internet of Things as described in any one of claims 1 to 5 is implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, the method for dynamically predicting risks of an integrated utility tunnel based on the Internet of Things as described in any one of claims 1 to 5 is implemented.

Images