Intelligent monitoring method and system for hydrogeological parameters

Abstract

This invention discloses an intelligent monitoring method and system for hydrogeological parameters, comprising: when performing hydrogeological environment monitoring tasks, setting up multiple monitoring locations and installing sensors within the monitoring area to collect hydrogeological data from each monitoring location; using the initial hydrogeological data from all monitoring locations as sample data to classify all monitoring locations; obtaining the subclass and membership degree of each monitoring location at the monitoring time based on the sample data, and then calculating the membership distance between every two monitoring locations; classifying all monitoring locations based on the membership distance and geographical distance between every two monitoring locations; and setting up an edge computing device at the center of all monitoring locations in each cluster to subsequently determine in real time whether there is a faulty sensor based on the real-time hydrogeological data of all monitoring locations in the cluster, and then issuing a maintenance work order through a cloud platform. This invention improves the accuracy of faulty sensor identification.

Description

Technical Field

[0001] This invention relates to the field of data processing technology. More specifically, this invention relates to an intelligent monitoring method and system for hydrogeological parameters. Background Technology

[0002] With social development and technological progress, the demand for hydrological environment monitoring is increasing. Traditional monitoring methods usually rely on manual periodic sampling and laboratory analysis. This method is not only time-consuming and labor-intensive, but also has slow data updates and is difficult to reflect the dynamic changes in the hydrological environment in a timely manner. Therefore, traditional monitoring methods can no longer meet the requirements of high precision, real-time performance and wide coverage of modern hydrological environment monitoring.

[0003] To overcome these limitations, modern hydrological and environmental monitoring has gradually introduced advanced information technology and automated equipment, constructing a "cloud-edge-device" architecture. This architecture combines sensors (devices), edge computing (edge), and cloud computing (cloud) to achieve efficient acquisition, real-time processing, stable transmission, and intelligent management of hydrogeological data, providing strong support for water resource protection, environmental governance, and disaster prevention.

[0004] For the edge computing layer in the "cloud-edge-device" architecture, it is mainly responsible for collecting, processing and analyzing hydrogeological data from sensors. In addition, if we want to analyze the hydrogeological data collected by each edge computing device from multiple monitoring locations to determine whether there are faulty sensors, then the hydrogeological data collected by each edge computing device needs to be highly comparable.

[0005] Therefore, by using hydrogeological data from monitoring locations, all monitoring locations are classified, and then edge computing devices are installed at the center of all monitoring locations in each cluster. However, since hydrogeological data is dynamic and affected by various factors, directly using hydrogeological data from each monitoring location for classification cannot capture the correlation and trend information of hydrogeological data over a long period of time. This results in low comparability of hydrogeological data from all monitoring locations in each cluster, which in turn affects the accuracy of identifying faulty sensors and seriously impacts hydrological environmental monitoring tasks. Summary of the Invention

[0006] To address the technical problem that directly using hydrogeological data from various monitoring locations for classification fails to capture the correlation and trend information of hydrogeological data over a long period, resulting in low comparability of hydrogeological data from all monitoring locations in each cluster and thus affecting the accuracy of fault sensor identification, this invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides an intelligent monitoring method for hydrogeological parameters, comprising: when performing a hydrogeological environment monitoring task, setting up multiple monitoring locations and installing sensors within the monitoring area to collect hydrogeological data at each monitoring location at each monitoring time; using the hydrogeological data of all monitoring locations at the initial installation stage as sample data to classify all monitoring locations; for each cluster in the classification results, setting up an edge computing device at the center of all monitoring locations in the cluster, for subsequent real-time determination of whether there is a faulty sensor based on the real-time hydrogeological data of all monitoring locations in the cluster, and then issuing a maintenance work order through a cloud platform; wherein, the process of classifying all monitoring locations is as follows: Hydrogeological data from all monitoring locations collected at each monitoring time in the sample data are used as cluster samples for each monitoring time. Clustering is performed on the cluster samples for each monitoring time to obtain the clustering results for each monitoring time. Based on the data distance between each monitoring location and each subclass in the clustering results for each monitoring time, the subclass to which each monitoring location belongs at each monitoring time and the membership degree with each subclass are obtained. Based on the subclass to which each monitoring location belongs at each monitoring time and the membership degree with each subclass, the membership distance between every two monitoring locations is calculated. The sum of the membership distance and the geographical distance between every two monitoring locations is used as the cluster distance between every two monitoring locations. All monitoring locations are classified based on the cluster distance between every two monitoring locations.

[0008] Preferably, in the clustered samples, each sample point consists of all hydrogeological data from each monitoring location; therefore, each sample point is... A point in 3D space, This indicates the number of types of all hydrogeological data at the monitoring location.

[0009] Preferably, obtaining the subclass of each monitoring location at the monitoring time includes: taking any monitoring location as the target location; obtaining the data distance between the target location and all subclasses based on the clustering results at the monitoring time; sorting all subclasses in descending order of the data distance between the target location and each subclass; and calculating the distance between the target location and each sorted subclass based on the data distance between the target location and each sorted subclass. The preference of each subclass as a member subclass of the target position; among which... The range of values ​​is , To monitor the number of all subclasses contained in the clustering results at a given time point; iterate through... The range of values, until the previous The optimality of each subclass as a member subclass of the target position Stop traversing, and move the previous... Each subclass serves as the subordinate subclass of the target location at the monitoring time. This is the preset first threshold.

[0010] Preferably, the former The formula for calculating the preference of a subclass as a member subclass of the target position is: In the formula, For the future The preference of each subclass as a member subclass of the target position. For the target position and the sorted front The variance of the data distance between each subclass. For the target position and the sorted front The variance of the data distance between each subclass.

[0011] Preferably, the membership degree of the monitoring location with each subclass at the monitoring time includes: using the normalized result of the data distance between the target location and each subclass at the monitoring time as the membership degree of the target location with each subclass at the monitoring time.

[0012] Preferably, the method for calculating the membership distance between any two monitoring locations is as follows: for any two monitoring locations, at the monitoring time... The monitoring location with the fewest number of subclasses is denoted as the monitoring location. Record another monitoring location as the monitoring location. Then the monitoring location and monitoring location During monitoring time Membership distance , For monitoring location During monitoring time The number of all subordinate subclasses, For monitoring location During monitoring time The Each subclass Subclass With monitoring location During monitoring time The optimal distance to the member subclass; the monitoring location and monitoring location The sum of the membership distances at all monitoring times is used as the monitoring location. and monitoring location The membership distance.

[0013] Preferably, the optimal distance is calculated as follows: calculate the monitoring location. During monitoring time Each subclass and its subclass The data distance between two subclasses refers to the data distance between the center points of the two subclasses; if the monitoring location During monitoring time Among all the subordinate subclasses, there exists a subclass that is related to the subordinate subclass. Data distance less than The subclass of the class will be related to the subclass of the class. Data distance less than Subclasses of a class, as subclasses The same subclass, and , For monitoring location During monitoring time and its subclass membership degree For monitoring location During monitoring time and its subclass The sum of the membership degrees of all their subclasses; otherwise, the membership degree of the subclasses. There are no subclasses, and ; ; Indicates a subclass The maximum distance from the center point among all sample points. This is the preset second threshold.

[0014] Preferably, the step of determining in real time whether there are faulty sensors based on real-time hydrogeological data of all monitoring locations in the cluster includes: for any edge computing device, constructing a box plot for real-time hydrogeological data of all monitoring locations in the cluster corresponding to the edge computing device, the obtained box plot including upper quartiles and lower quartiles; setting upper and lower limits based on the upper and lower quartiles, and recording sensors at monitoring locations whose real-time hydrogeological data exceeds the upper or lower limit as faulty sensors; in addition, the edge computing device transmits the real-time hydrogeological data of all monitoring locations in its corresponding cluster and the geographical locations of all faulty sensors to the cloud platform.

[0015] Preferably, the step of issuing maintenance work orders through the cloud platform includes: on the cloud platform, for any given sensor, if the sensor is continuously... The sensor was identified as faulty at every moment. For the preset quantity, it will be continuous The first moment of each time point is taken as the time when the faulty sensor begins to malfunction, and a maintenance work order is issued to the sensor. The maintenance work order includes the geographical location of the faulty sensor and the time when the malfunction began, and the maintenance work order is assigned to the maintenance personnel with the nearest geographical location.

[0016] Secondly, the present invention provides an intelligent monitoring system for hydrogeological parameters, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned intelligent monitoring method for hydrogeological parameters is implemented.

[0017] By adopting the above technical solution, a computer program for intelligent monitoring of hydrogeological parameters is generated and stored in a memory so that it can be loaded and executed by a processor. A terminal device can then be made based on the memory and processor for convenient use.

[0018] The beneficial effects of this invention are as follows:

[0019] This invention uses initial hydrogeological data as sample data to provide a more stable and accurate reference for subsequent classification of all monitoring locations. Furthermore, by obtaining the subclass of each monitoring location at each monitoring time and its degree of membership with each subclass, the temporal correlation of hydrogeological data is captured. Then, based on the subclass of each monitoring location at each monitoring time and its degree of membership with each subclass, the membership distance between every two monitoring locations is calculated, capturing the long-term correlation and trend information of hydrogeological data. By combining membership distance and geographical distance, and comprehensively considering the data similarity and spatial distribution characteristics of monitoring locations over a long period, this invention not only improves the comparability of hydrogeological data at monitoring locations but also provides a scientific basis for the rational deployment of edge computing devices. This improves the accuracy of fault sensor identification and ensures the efficient operation and reliability of the hydrological environment monitoring system. Attached Figure Description

[0020] The above and other objects, features, and advantages of the present invention will become readily apparent from the following detailed description of exemplary embodiments, accompanied by the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0021] Figure 1 This is a flowchart illustrating an intelligent monitoring method for hydrogeological parameters according to the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0024] This invention discloses an intelligent monitoring method for hydrogeological parameters, referring to... Figure 1 This includes steps S1 to S4:

[0025] S1. When performing hydrological and environmental monitoring tasks, multiple monitoring locations are set up and sensors are installed in the monitoring area to collect hydrogeological data of each monitoring location at each monitoring time.

[0026] When performing hydrological and environmental monitoring tasks, setting up a "cloud-edge-device" architecture can effectively realize the collection, processing, transmission, and management of hydrogeological data. Here, "cloud" refers to the cloud platform, which is responsible for storing, analyzing, and managing data transmitted from the edge computing layer; "edge" refers to the edge computing layer, which consists of multiple edge computing devices, namely edge computing servers or gateways, responsible for collecting, processing, and analyzing hydrogeological data from sensors; and "device" refers to the sensor layer, which is responsible for collecting hydrogeological data from each monitoring area.

[0027] In the "cloud-edge-device" architecture, the sensor layer is primarily responsible for collecting hydrogeological data from various monitoring areas. Hydrogeological data is crucial for researching and managing water resources, assessing environmental impacts, and predicting natural disasters. It exhibits diverse data types and requires different acquisition methods. The following are common types of hydrogeological data and their corresponding sensors:

[0028] 1. Water level data includes, but is not limited to, groundwater levels and river and lake levels. Groundwater levels are collected by groundwater level gauges (such as vibrating wire level gauges and capacitive level gauges) to monitor changes in groundwater levels in real time. River and lake levels are usually monitored using ultrasonic level gauges, radar level gauges, or pressure level gauges. These sensors can adapt to different water environments and provide high-precision water level data.

[0029] 2. Flow velocity data includes, but is not limited to, river flow velocity and groundwater flow velocity. River flow velocity is usually measured using electromagnetic or ultrasonic current meters, while groundwater flow velocity is usually determined using groundwater velocity sensors.

[0030] 3. Flow data can be directly measured using flow meters (such as electromagnetic flow meters and ultrasonic flow meters).

[0031] 4. Water quality parameters include, but are not limited to, pH value, conductivity, dissolved oxygen, turbidity, and chemical oxygen demand (COD). Among them, pH value is obtained using a pH sensor to monitor the acidity and alkalinity of the water body in real time; conductivity refers to the electrical conductivity of the water body measured by a conductivity sensor, which can reflect the ion concentration in the water; dissolved oxygen refers to the content of dissolved oxygen in the water body, which is crucial for assessing the ecological status of the water body and is measured by a dissolved oxygen sensor; turbidity refers to the content of suspended particulate matter in the water body measured by a turbidity sensor, which is used to reflect the clarity of the water body; and COD can be determined by a COD sensor, which reflects the content of organic matter in the water body.

[0032] 5. Pressure data can be measured directly using a pressure sensor.

[0033] 6. Temperature data can be measured directly using temperature sensors, which helps to understand the thermal characteristics of groundwater.

[0034] Specifically, multiple monitoring locations are set up within the monitoring area, and sensors are installed at each monitoring location; hydrogeological data at each monitoring location at each monitoring time are collected through the sensors; all hydrogeological data are normalized to unify the dimensions, and the normalization method includes, but is not limited to, maximum and minimum value normalization. The hydrogeological data in the subsequent process refers to the normalized hydrogeological data.

[0035] The time interval for collecting data can be set according to the actual application scenario and requirements. In this embodiment, the time interval is set to 10 minutes.

[0036] S2. Use the hydrogeological data of all monitoring locations during the initial installation as sample data. Based on the sample data, obtain the subclass of each monitoring location at each monitoring time and the degree of membership with each subclass.

[0037] It should be noted that the edge computing layer in the "cloud-edge-device" architecture is mainly responsible for collecting, processing, and analyzing hydrogeological data from sensors. It determines whether there are faulty sensors by comparing the hydrogeological data collected by sensors at different monitoring locations. Therefore, for each edge computing device responsible for collecting hydrogeological data from multiple monitoring locations, the hydrogeological data from these monitoring locations needs to be highly comparable to accurately determine whether there are faulty sensors through data comparison. Therefore, when setting up edge computing devices, all monitoring locations are classified to ensure that the hydrogeological data of all monitoring locations in each cluster are highly comparable. Then, the edge computing device is set at the center of all monitoring locations in each cluster.

[0038] It should be further noted that all sensors have just been deployed and debugged, and are in optimal working condition. They are able to collect complete hydrogeological data from each monitoring location. Furthermore, the hydrogeological conditions are relatively stable, and the sensors have not yet been affected by external interference or environmental changes that may occur during long-term monitoring. Therefore, the missing rate of the hydrogeological data collected at this time is low, the records of each parameter are relatively complete, and they can better reflect the hydrogeological characteristics of the monitoring location under normal conditions. They are highly representative and provide comprehensive basic information for classification.

[0039] Specifically, the hydrogeological data of all monitoring locations during the initial installation phase will be used as sample data to classify all monitoring locations; the initial installation phase refers to the period one week after the sensors at each monitoring location are installed and debugged.

[0040] It should be noted that by classifying all monitoring locations using hydrogeological data, and then installing edge computing devices at the center of all monitoring locations within each cluster, the hydrogeological data becomes less comparable due to the dynamic nature of the data and its susceptibility to various factors. Directly using the hydrogeological data from each monitoring location for classification fails to capture the correlation and trend information of the hydrogeological data over a long period. This results in a lack of comparability of the hydrogeological data from all monitoring locations within each cluster, which in turn affects the accuracy of identifying faulty sensors and seriously impacts the hydrological environment monitoring task.

[0041] Therefore, before classifying all monitoring locations, this embodiment clusters the clustered samples at each monitoring time to obtain the clustering results for each monitoring time. Based on the data distance between each monitoring location and each subclass in the clustering results of the monitoring time, the subclass to which each monitoring location belongs at the monitoring time and the degree of membership with each subclass are obtained. Through cluster analysis, the inherent temporal correlation of hydrogeological data at different monitoring times is captured.

[0042] Specifically, the hydrogeological data from all monitoring locations collected at each monitoring time in the sample data are used as the cluster sample for each monitoring time; in the cluster sample, each sample point consists of all the hydrogeological data from each monitoring location, therefore, each sample point is... A point in 3D space, This indicates the number of types of all hydrogeological data at the monitoring location.

[0043] Furthermore, for any clustered sample at any monitoring time: cluster the clustered samples at the monitoring time to obtain the clustering result at the monitoring time; based on the data distance between each monitoring location and each subclass in the clustering result at the monitoring time, obtain the subclass to which each monitoring location belongs at the monitoring time and the degree of membership with each subclass.

[0044] The clustering results for each monitoring time are obtained by clustering the clustered samples at that time. The specific process is as follows:

[0045] For clustered samples at monitoring times, the OPTICS clustering algorithm is used to cluster the samples at each monitoring time, obtaining clustering results for each monitoring time. These clustering results include multiple subclasses. For example, for monitoring times... Monitoring time Clustering of the samples yields the monitoring time. The clustering results include multiple subclasses, namely... , ,..., , , For monitoring time The number of all subclasses included in the clustering result.

[0046] It should be noted that when clustering the clustered samples at the monitoring time, since the number of subclasses included in the clustering results is uncertain, this embodiment selects the OPTICS (Ordering points to identify the clustering structure) clustering algorithm to cluster the clustered samples at the monitoring time. The OPTICS clustering algorithm is a well-known technology and will not be described in detail here.

[0047] In this process, any monitoring location is taken as the target location. Based on the data distance between the target location and each subclass in the clustering results at the monitoring time, the subclass to which the target location belongs at the monitoring time and the membership degree between the target location and its subclass are obtained. The specific process is as follows:

[0048] (1) Based on the clustering results at the monitoring time, obtain the data distance between the target location and all subclasses; the calculation method for the data distance between the target location and any subclass is: the average of the Euclidean distances between the target location and the hydrogeological data of all monitoring locations in that subclass.

[0049] (2) Sort all subclasses in descending order of the data distance between the target location and each subclass.

[0050] (3) Calculate the distance between the target location and the data of each sorted subclass, and then... The preference of each subclass as the subordinate subclass of the target position; The formula for calculating the preference of a subclass as a member subclass of the target position is:

[0051] ;

[0052] In the formula, For the future The preference of each subclass as a member subclass of the target position. For the target position and the sorted front The variance of the data distance between each subclass. For the target position and the sorted front The variance of the data distance between each subclass.

[0053] It should be noted that when obtaining the subclass of the target location, when adding... After subclassing, a sudden change in the variance of the data distance between the target location and all subclasses indicates that... The subclass does not conform to the data characteristics of the target location, therefore, it is not suitable as a subordinate subclass of the target location.

[0054] (4) Among them, The range of values ​​is , To monitor the number of all subclasses contained in the clustering results at a given time, iterate through... The range of values, until the previous The optimality of each subclass as a member subclass of the target position Stop traversing, and move the previous... Each subclass serves as the subordinate subclass of the target location at the monitoring time. This is the preset first threshold.

[0055] Among them, the first threshold The specific value can be set according to the actual application scenario and requirements, and the first threshold The range of values ​​is Therefore, in this embodiment, the first threshold is... Set as , This indicates the number of types of all hydrogeological data at the monitoring location.

[0056] (5) The normalized result of the data distance between the target location and each of its subclasses at the monitoring time is taken as the membership degree of the target location and its subclasses at the monitoring time; wherein, the sum of the data distances between the target location and all its subclasses at the monitoring time is recorded as the sum value. Then the data distance and sum of the target location and each of its subcategories at the monitoring time will be calculated. The ratio is used as the normalized result of the data distance between the target location and each subordinate subclass at the monitoring time.

[0057] It should be noted that by clustering the clustered samples at each monitoring time, the clustering results at each monitoring time are obtained. These results reflect the similarity and differences of the hydrogeological data of all monitoring locations at that monitoring time. Based on the data distance between each monitoring location and each subclass in the clustering results at each monitoring time, the subclass to which each monitoring location belongs and the membership degree of each subclass are obtained. The subclass and membership degree provide the data characteristics of each monitoring location at different monitoring times. This comprehensive consideration of the temporal correlation of hydrogeological data improves the accuracy and stability of subsequent classification of all monitoring locations.

[0058] S3. Classify all monitoring locations according to their subclass at each monitoring time and their degree of membership in each subclass.

[0059] It should be noted that step S2 obtains the subclass of each monitoring location at the monitoring time and the degree of membership with each subclass, providing the temporal correlation of hydrogeological data for each monitoring location. In order to further capture the correlation and trend information of hydrogeological data over a long period of time and achieve the accuracy and stability of subsequent classification of all monitoring locations, this embodiment calculates the membership distance between every two monitoring locations based on the subclass of each monitoring location at each monitoring time and the degree of membership with each subclass.

[0060] Specifically, all monitoring locations are classified according to their subclass at each monitoring time and their degree of membership to each subclass. The specific process is as follows:

[0061] 1. Calculate the membership distance between every two monitoring locations based on the subclass to which each monitoring location belongs at each monitoring time and the degree of membership with each subclass.

[0062] It should be noted that the calculation of membership distance comprehensively considers the data characteristics of the monitoring location at different time points, and not only the data at a single time point, but also the long-term trend of the data. This can more accurately capture and measure the correlation and trend information of hydrogeological data from different monitoring locations over a long period of time, thereby improving the comparability of data in the classification results of all monitoring locations.

[0063] 2. Calculate the geographical distance between any two monitoring locations based on their geographical locations; the geographical distance between any two monitoring locations is equal to the Euclidean distance between their geographical locations.

[0064] 3. The sum of the membership distance and geographical distance between any two monitoring locations is used as the cluster distance between any two monitoring locations. Specifically, the membership distance between any two monitoring locations is normalized to obtain a normalized result; the geographical distance between any two monitoring locations is normalized to obtain a normalized result; the sum of the normalized result of the membership distance and the normalized result of the geographical distance between any two monitoring locations is used as the cluster distance between any two monitoring locations. The normalization method includes, but is not limited to, maximum and minimum value normalization.

[0065] It should be noted that combining membership distance and geographical distance comprehensively considers the data similarity and spatial distribution characteristics of monitoring locations over a long period of time. This not only improves the comparability of hydrological data of various clusters in the classification results, but also takes into account the geographical distribution of monitoring locations, thereby improving the accuracy and reliability of the classification results.

[0066] 4. Based on the clustering distance between every two monitoring locations, the OPTICS clustering algorithm is used to classify all monitoring locations, and the classification results of all monitoring locations are obtained. The classification results include multiple clusters.

[0067] The method for calculating the membership distance between any two monitoring locations is as follows:

[0068] (1) For any two monitoring locations, the monitoring time will be... The monitoring location with the fewest number of subclasses is denoted as the monitoring location. Record another monitoring location as the monitoring location. ; monitor location During monitoring time The subclasses are as follows: , For monitoring location During monitoring time The number of all subordinate subclasses; monitoring locations During monitoring time The subclasses are as follows: , For monitoring location During monitoring time The number of all subordinate subclasses; and .

[0069] (2) For monitoring locations During monitoring time The Subclasses Calculate monitoring location During monitoring time Each subclass and its subclass The data distance, where the data distance between two subclasses refers to the data distance between the center points of the two subclasses; if the monitoring location During monitoring time Among all the subordinate subclasses, there exists a subclass that is related to the subordinate subclass. Data distance less than If the subclass is a member of the class, then these will be related to the subclass. Data distance less than Subclasses of a class, as subclasses In the case of subclasses belonging to the same subclass, the subclass is... With monitoring location During monitoring time Optimal distance of the subclass The calculation formula is: , For monitoring location During monitoring time and its subclass membership degree For monitoring location During monitoring time and its subclass The sum of the membership degrees of all their subclasses. To take the absolute value; if the monitoring location During monitoring time Among all the subordinate subclasses, there is no subclass that is related to the subordinate subclass. Data distance less than Subclasses to which the monitoring location belongs, i.e. During monitoring time All subclasses and their subclasses The data distances are all greater than or equal to Then it belongs to a subclass In this case, if there are no subclasses, then the subclass is... With monitoring location During monitoring time Optimal distance of the subclass The calculation formula is: .

[0070] in, ; Indicates a subclass The maximum distance from the center point among all sample points. The second threshold is a preset threshold. The specific value can be set according to the actual application scenario and requirements, and the second threshold The range of values ​​is Therefore, in this embodiment, the second threshold is... Set to 1.2.

[0071] (3) Then the monitoring location and monitoring location During monitoring time Membership distance , For monitoring location During monitoring time The number of all subordinate subclasses, For monitoring location During monitoring time The Each subclass Subclass With monitoring location During monitoring time The optimal distance to the member subclass; the monitoring location and monitoring location The sum of the membership distances at all monitoring times is used as the monitoring location. and monitoring location The membership distance is denoted as . .

[0072] S4. For each cluster in the classification results, an edge computing device is set at the center of all monitoring locations in the cluster. This device is used to determine in real time whether there is a faulty sensor based on the real-time hydrogeological data of all monitoring locations in the cluster, and then issue a maintenance work order through the cloud platform.

[0073] The edge computing layer in the "cloud-edge-device" architecture consists of multiple edge computing devices, which are edge computing servers or gateways. Edge computing devices typically have certain computing and storage capabilities and can perform data cleaning, feature extraction, and anomaly detection, thereby reducing data transmission volume and improving data quality. Therefore, in this embodiment, the hydrogeological data collected by the sensor is preliminarily processed, and the processed hydrogeological data is transmitted to the cloud platform through wired or wireless networks.

[0074] Therefore, for each cluster in the classification results, the center location of all monitoring locations in the cluster is obtained based on the geographical location of all monitoring locations in the cluster, and an edge computing device is set up at the center location.

[0075] It should be noted that this embodiment uses stable data from the initial installation phase as clustering samples, combined with cluster analysis and comprehensive distance calculation, to gradually solve the problem of low data comparability among monitoring locations in conventional classification methods. This improves the comparability of hydrogeological data at monitoring locations, provides a scientific basis for the rational deployment of edge computing devices, and ensures the efficient operation and reliability of the hydrological environment monitoring system.

[0076] Furthermore, during the execution of hydrological and environmental monitoring tasks, for all monitoring locations included in each cluster in the classification results, real-time hydrogeological data of each monitoring location are collected by the installed sensors, and the real-time hydrogeological data of each monitoring location are normalized. The normalization process includes, but is not limited to, maximum and minimum value normalization. The maximum and minimum values ​​used here are the maximum and minimum values ​​of the hydrogeological data in the sample data.

[0077] Furthermore, for any edge computing device, based on the real-time hydrogeological data of all monitoring locations in the cluster corresponding to the edge computing device, it is determined in real time whether there are faulty sensors among the sensors installed at all monitoring locations in the cluster. This includes: constructing a box plot for the real-time hydrogeological data of all monitoring locations in the cluster corresponding to the edge computing device. In the obtained box plot, there is a line in the middle of the box, which represents the median of the real-time hydrogeological data of all monitoring locations. The upper and lower bases of the box are the upper and lower quartiles of the real-time hydrogeological data of all monitoring locations, respectively. This means that the box contains 50% of the real-time hydrogeological data, so the height of the box reflects the degree of fluctuation of the real-time hydrogeological data to a certain extent. Therefore, based on the upper and lower quartiles, upper and lower limits are set, and sensors at monitoring locations whose real-time hydrogeological data exceeds the upper or lower limit are recorded as faulty sensors.

[0078] Furthermore, the edge computing device transmits real-time hydrogeological data of all monitoring locations in its corresponding cluster, as well as the geographical locations of all faulty sensors, to the cloud platform.

[0079] In the "cloud-edge-device" architecture, the cloud platform is primarily responsible for storing, analyzing, and managing data transmitted from the edge computing layer. Additionally, by analyzing hydrogeological data collected from multiple monitoring locations by each edge computing device, it determines whether any sensors are faulty. This mainly includes: for any given sensor, if the sensor is continuously... Each time the sensor is judged to be faulty, it will be continuously... The first moment of each time point is taken as the time when the faulty sensor begins to malfunction, and a maintenance work order is issued to the sensor. The maintenance work order includes the geographical location of the faulty sensor and the time when the malfunction began, and the maintenance work order is assigned to the maintenance personnel with the nearest geographical location.

[0080] in, Preset quantity The specific values ​​can be set according to the actual application scenario and requirements, and the preset quantity is available. The range of values ​​is Therefore, this embodiment will preset the quantity. Set it to 6.

[0081] This invention also discloses an intelligent monitoring system for hydrogeological parameters, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent monitoring method for hydrogeological parameters according to the present invention is implemented.

[0082] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. An intelligent monitoring method for hydrogeological parameters, characterized in that, include: When carrying out hydrological and environmental monitoring tasks, multiple monitoring locations are set up and sensors are installed in the monitoring area to collect hydrogeological data at each monitoring location at each monitoring time. The hydrogeological data of all monitoring locations during the initial installation will be used as sample data to classify all monitoring locations. For each cluster in the classification results, an edge computing device will be set at the center of all monitoring locations in the cluster. This device will be used to determine in real time whether there are faulty sensors based on the real-time hydrogeological data of all monitoring locations in the cluster, and then issue maintenance work orders through the cloud platform. The process of classifying all monitoring locations is as follows: Hydrogeological data from all monitoring locations collected at each monitoring time in the sample data are used as clustering samples for each monitoring time; clustering is performed on these samples to obtain clustering results for each monitoring time; based on the data distance between each monitoring location and each subclass in the clustering results for each monitoring time, the subclass to which each monitoring location belongs at each monitoring time and its membership degree with each subclass is obtained, including: taking any monitoring location as the target location; sorting all subclasses in descending order of the data distance between the target location and each subclass; and calculating the distance between the target location and each sorted subclass. The preference of each subclass as a member subclass of the target position. The range of values ​​is , To monitor the number of all subclasses contained in the clustering results at a given time point; iterate through... The range of values, until the previous When the preference of a subclass as a member subclass of the target position is greater than a preset first threshold, the traversal stops, and the previous subclass is set to [a higher priority]. Each subclass is considered as a member subclass of the target location at the monitoring time; the normalized result of the data distance between the target location at the monitoring time and each member subclass is used as the membership degree of the target location between the target location and each member subclass at the monitoring time. Based on the subclass to which each monitoring location belongs at each monitoring time and the degree of membership with each subclass, calculate the membership distance between any two monitoring locations, including: for any two monitoring locations, at each monitoring time... The monitoring location with the fewest number of subclasses is denoted as the monitoring location. Record another monitoring location as the monitoring location. Monitoring location and monitoring location During monitoring time Membership distance , For monitoring location During monitoring time The number of all subordinate subclasses, For monitoring location During monitoring time The Each subclass Subclass With monitoring location During monitoring time The optimal distance to the member subclass; the monitoring location and monitoring location The sum of the membership distances at all monitoring times is used as the monitoring location. and monitoring location Membership distance; The sum of the membership distance and geographical distance between any two monitoring locations is used as the cluster distance between any two monitoring locations; all monitoring locations are classified according to the cluster distance between any two monitoring locations.

2. The intelligent monitoring method for hydrogeological parameters according to claim 1, characterized in that, In the clustered samples, each sample point consists of all hydrogeological data from each monitoring location; therefore, each sample point is... A point in 3D space, This indicates the number of types of all hydrogeological data at the monitoring location.

3. The intelligent monitoring method for hydrogeological parameters according to claim 1, characterized in that, The former The formula for calculating the preference of a subclass as a member subclass of the target position is: ； In the formula, For the future The preference of each subclass as a member subclass of the target position. For the target position and the sorted front The variance of the data distance between each subclass. For the target position and the sorted front The variance of the data distance between each subclass.

4. The intelligent monitoring method for hydrogeological parameters according to claim 1, characterized in that, The optimal distance is calculated as follows: Calculate monitoring location During monitoring time Each subclass and its subclass The data distance between two subclasses refers to the data distance between the center points of the two subclasses. If monitoring location During monitoring time Among all the subordinate subclasses, there exists a subclass that is related to the subordinate subclass. Data distance less than The subclass of the class will be related to the subclass of the class. Data distance less than Subclasses of a class, as subclasses The same subclass, and , For monitoring location During monitoring time and its subclass membership degree For monitoring location During monitoring time and its subclass The sum of the membership degrees of all their subclasses; otherwise, the membership degree of the subclasses. There are no subclasses, and ; ; Indicates a subclass The maximum distance from the center point among all sample points. This is the preset second threshold.

5. The intelligent monitoring method for hydrogeological parameters according to claim 1, characterized in that, The step of determining in real time whether there is a faulty sensor based on real-time hydrogeological data from all monitoring locations in the cluster includes: For any edge computing device, a box plot is constructed for the real-time hydrogeological data of all monitoring locations in the cluster corresponding to the edge computing device. The obtained box plot includes the upper quartile and the lower quartile. Based on the upper quartile and the lower quartile, an upper limit and a lower limit are set. Sensors at monitoring locations whose real-time hydrogeological data exceed the upper limit or the lower limit are recorded as faulty sensors. In addition, the edge computing device transmits real-time hydrogeological data of all monitoring locations in its corresponding cluster, as well as the geographical location of all faulty sensors, to the cloud platform.

6. The intelligent monitoring method for hydrogeological parameters according to claim 1, characterized in that, The process of issuing maintenance work orders through the cloud platform includes: On the cloud platform, for any given sensor, if the sensor is continuously... The sensor was identified as faulty at every moment. For the preset quantity, it will be continuous The first moment of each time point is taken as the time when the faulty sensor begins to malfunction, and a maintenance work order is issued to the sensor. The maintenance work order includes the geographical location of the faulty sensor and the time when the malfunction began, and the maintenance work order is assigned to the maintenance personnel with the nearest geographical location.

7. An intelligent monitoring system for hydrogeological parameters, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement an intelligent monitoring method for hydrogeological parameters according to any one of claims 1-6.

Images